SAJOG Vol 23, No 2 (2017) by HMPG

ISSN 2305-8862

September 2017 Vol. 23 No. 2 +

• CD56 natural killer cells in women with unexplained infertility • Women’s intention to request caesarean delivery in Tanzania • Factors affecting initiation of ART in pregnant women in Swaziland • Termination of pregnancy in Tunisia

SAJOG September 2017 Vol. 23 No. 2

THE SOUTH AFRICAN JOURNAL OF OBSTETRICS AND GYNAECOLOGY

Editor William Edridge Editorial Board SAJOG Alan Alperstein (Cape Town) Geoffrey Buga (Walter Sisulu) Hennie Cronje (Free State) Franco Guidozzi (Witwatersrand) Justus Hofmeyr (East London) Thinus Kruger (Stellenbosch) Gerhard Lindeque (Pretoria) Eddie Mhlanga (KwaZulu-Natal) Sam Monokoane (Limpopo) Jack Moodley (KwaZulu-Natal) Dan Ncayiyana (Durban) Hein Odendaal (Stellenbosch) Zephne van der Spuy (Cape Town) HEALTH & MEDICAL PUBLISHING GROUP (HMPG)

CONTENTS

CEO and Publisher Hannah Kikaya Email: hannahk@hmpg.co.za

Editorial

Executive Editor Bridget Farham

The continuing problem with misoprostol

W Edridge

Research 38 Does interval laparoscopic sterilisation influence the risk of lower genital tract infections and menstrual abnormalities?

G Kistan, J S Bagratee, M Panday

43 Perinatal outcomes in pregnant women presenting with preterm premature rupture of membranes at a regional hospital in KwaZulu-Natal Province, South Africa

G Msomi, M Naidoo, B Hira

48 The peritoneum at laparotomy: A survey of gynaecological practise among Nigerian gynaecologists

A A Bamigboye, F O Okonofua

51 An endometrial histomorphometric study of CD56+ natural killer cells in women with unexplained infertility

M Muller, G Kalmeier, P Eyal, A de Bruin, R Pool, C du Rant, A Stander, A van Schoor, E Nortje, P du Toit

56 Factors associated with women’s intention to request caesarean delivery in Dar es Salaam, Tanzania

C G Misaeli, B A Kamala, A H Mgaya, H L Kidanto

63 Factors associated with the lack of antiretroviral therapy initiation among eligible HIV-positive pregnant women in Swaziland

Managing Editors Naadia van der Bergh Claudia Naidu Technical Editor Christelle Cronje Kirsten Morreira Emma Buchanan Paula van der Bijl Production Manager Emma Jane Couzens DTP and Design Travis Arendse Clinton Griffin Chief Operating Officer Diane Smith | Tel. 012 481 2069 Email: dianes@hmpg.co.za Online Support Gertrude Fani Tel. 021 532 1281 | Cell. 072 463 2159 Email: publishing@hmpg.co.za ISSN 2305-8862 Journal website: www.sajog.org.za Use of editorial material is subject to the Creative Commons Attribution – Non-commercial Works Licence. https://creativecommons.org/licenses/ by-nc/4.0

C Chouraya, G Louwagie, B Nhlabatsi, M A Mahdi, B V Girdler-Brown

Review 69 Termination of pregnancy for fetal anomaly in a Tunisian population

E B Hamida, I Ayadi, A Bezzine, B Rabii, S B Hammouda, B Bouguerra, Z Marrakchi

Case report 71 Bucket-handle tear of posterior uterine cervical lip in a secondtrimester unscarred uterus after use of misoprostol: A first report of two cases

S R Singhal, S K Singhal

CPD Questions

Listed in DOAJ, AIM, AJOL, Excerpta Medica (EMBASE), EBSCO, Scopus, Sabinet, Emerging Sources Citation Index (Web of Science) Published by the Health and Medical Publishing Group, a subsidiary of the South African Medical Association HEAD OFFICE Block F, Castle Walk Corporate Park, Nossob Street, Erasmuskloof Ext. 3, Pretoria, 0181 Tel. 012 481 2069 EDITORIAL OFFICE Suite 11, Lonsdale Building, Lonsdale Way, Pinelands, 7405 Tel. 021 532 1281 | Cell. 072 463 2159 | E-mail. publishing@hmpg.co.za All letters and articles for publication must be submitted online at www.sajog.org.za ©Copyright: Health and Medical Publishing Group (Pty) Ltd.

This open-access article is distributed under Creative Commons licence CC-BY-NC 4.0.

EDITORIAL

The continuing problem with misoprostol An article in this edition describes two cases of serious genital tract injury that occurred with standard doses of misoprostol, commonly sold as Cytotec, when used to achieve evacuation of the uterus in the second trimester. Misoprostol is the only prostaglandin E1 analogue uterine stimulant that is currently commercially available in obstetrics and gynaecology; all other prostaglandin preparations, commercially known as Prostin, Prandin, Prepidil, and by many other names, are prostaglandin E2 analogues, and long preceded the use of misoprostol. Misoprostol was originally synthesised for use in peptic ulceration, but early in its genesis its use as a uterine stimulant began. It is not unusual for a drug to have a use different to the one for which it was originally devised; chlorpromazine, or Largactil, for example, was originally a premedicating agent for anaesthesia. Now, owing to a discovered additional benefit, it is predominantly a psychotropic agent for the treatment of psychotic delusions in schizophrenia. Misoprostol can be used in low doses to soften the cervix to allow instrumentation for suction evacuation of the uterus in termination or incomplete miscarriage, or to soften the cervix prior to instrumentation in, for example, hysteroscopy. But it is in its use as an abortifacient or as an agent for inducing labour that greater problems may occur. A further use is for contracting the atonic uterus in primary postpartum haemorrhage, for which it remains recommended primarily when other agents are not available and where medical supervision is limited. As an abortifacient in the first trimester, misoprostol may be used alone or in combination with mifepristone or methotrexate. The side-effects of misoprostol vary from the relatively minor to the very serious. It may cause shivering, or a mild pyrexia with repeated doses, and, particularly if taken orally, it may cause gastrointestinal upset, especially diarrhoea (in a dehydrated patient, this may not be a minor complication). But the greatest concern regarding misoprostol is an idiosyncratic, excessive uterotonic response at what are considered reasonable therapeutic doses. Idiosyncratic means particular to the individual â&#x20AC;&#x201C; that the response may far exceed the response in many others, and that it may far exceed the average response. This unusual response cannot be predicted. An individual given reasonable doses of misoprostol may, therefore, develop excessive uterine contractions leading to, as in the two cases reported here, tearing of the undilated cervix with excessive pressure from above. Tearing of the undilated cervix more usually occurs when the cervix is fibrosed from previous traumatic instrumentation and is therefore resistant to dilatation. In the two cases reported here, no such history existed (information in a history may, of course, be withheld). In induction of labour, misoprostol may be hazardous. Induction of labour is designed to mimic the natural evolution of contractions leading to dilation of the cervix and descent of the presenting part. At times, the induction agent alone is necessary to sustain labour and effect delivery, but at times, oxytocin, after an appropriate interval, is required. Overstimulation or hyperstimulation is well recognised with prostaglandin E2 analogues, as with E1 analogues, and for this reason

delivery systems for E2 analogues that do not completely dissolve and can be removed from the vagina were developed. With E2, as with E1 analogues, an excessive response cannot be predicted. The consequence of overstimulation can be grave: fetal distress may lead to the passage of meconium, with the attendant risk of aspiration. Acidosis may lead to fetal and neonatal asphyxia, and cerebral palsy or death. In very excessive responses, the uterus may rupture, resulting in fetal and even maternal death, or the undilated cervix may tear (as occurred in the cases reported here), resulting in life-threatening haemorrhage. Are any of these complications greater with misoprostol than with the previously developed E2 analogues? How can such complications be avoided? It is not known whether serious complications with misoprostol exceed those with E2 analogues. Many studies have looked at case series of misoprostol and the question remains unanswered. Sizeable comparative data are absent. And there is concern that harmful effects of uterine stimulants may go unreported â&#x20AC;&#x201C; such reports may reflect badly on institutions if, on review, preventive and corrective measures are deemed inadequate, and in an era of litigation, a culture of silence is inevitable. In induction of labour, monitoring should commence by cardiotocogram (CTG) once contractions are established. This is especially the case when tachysystole (six contractions or more in two 10-minute episodes) is suspected, or prolonged or coupled contractions are suspected that may prevent fetal recovery. A midwife cannot observe a patient undergoing induction of labour throughout the many hours it may take, and so all patients undergoing induction must be asked to inform supervising staff when regular contractions occur and assessment can begin. While baseline changes and late decelerations can be detected by intermittent auscultation, baseline variability cannot, and, for many, assessment by CTG would be an essential prerequisite for induction, and for the high-risk labour that follows. Teaching of correct CTG interpretation is essential. In the developing world, CTG machines may not be available. Induction of labour should be avoided in cases where uteroplacental insufficiency is suspected or present. In the developing world, sonar machines may not always be available, and clinical judgement of poor growth and oligohydramnios is unreliable. However, excessive uterine contractions may exceed the reserve even of the fetus in whom liquor levels, doppler investigations (where available) and growth are normal. Excessive response to misoprostol as an abortifacient may perhaps be limited, as in induction of labour, by withholding further doses once a response is established. The importance of adequate medical assessment cannot be overemphasised: assessment cannot prevent an excessive response, but patients being administered misoprostol in repeated doses should not be ignored in side-rooms or a corner of the ward, but observed and assessed by an enquiry into contractions and palpation. Misoprostol is popular because of its efficiency at achieving delivery/evacuation, and because it is inexpensive. This latter benefit may contribute to its distribution outside hospitals, clinics and pharmacies. Away from medical supervision, dose control may be

SAJOG â&#x20AC;˘ September 2017, Vol. 23, No. 2

EDITORIAL non-existent. Patients may also arrive in maternity and gynaecology admission areas seemingly undergoing spontaneous labour or miscarriage, and suffer the dire consequences of excessive doses of misoprostol, the administration of which is never confessed. It is possible that such patients may be given further doses when in hospital, thereby exaggerating the total dose given. The pharmacodynamics of possible previous and undeclared doses of the drug are not uniform across a population, and for a few, the effects of previously administered doses may persist. The inhibition of excessive misoprostol contractions by tocolytic agents, as with the inhibition of any uterine stimulant, may be effective. Sustained inhibition protocols and the safety of repeated tocolytic doses are less well known. Wherever misoprostol is used, rapid recourse to a safe operating theatre is an essential requirement. The absence of such can only magnify any complications that occur. Is misoprostol safe for all who use it? No, it is not. The risks for the few that react excessively to its stimulus are great. Vigilance and anticipation of idiosyncratic responses are essential. Thorough

reporting of complications may lead others to question its suitability, as has the National Institute for Health and Care Excellence (NICE)[1] in the UK, which has stated that its use in induction of labour should be confined to induction of labour or evacuation in the presence of intrauterine fetal death (as in the cases reported here, though they were in the second trimester), or in the context exclusively of formal clinical trials. While this opinion is not universally held, it was informed by more than a casual concern for potentially very serious, harmful effects. Will Edridge Editor 1. National Institute for Health and Care Excellence. Clinical Guidance: Inducing Labour. London: NICE, 2008. https://www.nice.org.uk/guidance/cg70 (accessed 30 August 2017).

S Afr J Obstet Gynaecol 2017;23(2):36-37. DOI:10.7196/SAJOG.2017.v23i2.1543

SAJOG â&#x20AC;˘ September 2017, Vol. 23, No. 2

This open-access article is distributed under Creative Commons licence CC-BY-NC 4.0.

RESEARCH

Does interval laparoscopic sterilisation influence the risk of lower genital tract infections and menstrual abnormalities? G Kistan,1 MB ChB, FCOG (SA), MMed; J S Bagratee,1 MB ChB, FCOG (SA), FRCOG (UK), MMed, PhD; M Panday,2 MB ChB, FCOG (SA) Department of Obstetrics and Gynaecology, Nelson R Mandela School of Medicine, College of Health Sciences, University of KwaZulu-Natal, Durban, South Africa 2 Department of Family Planning, King Dinuzulu Hospital Complex, Durban, KwaZulu-Natal, South Africa 1

Corresponding author: G Kistan (gaysheenkistan@gmail.com)

Background. Tubal sterilisation is a safe, accessible and effective contraceptive method. There is a paucity of data regarding the risk of genital tract infections and menstrual abnormalities post sterilisation in Durban, South Africa. Objectives. To evaluate the risk of lower genital tract infections (LGTIs) and menstrual abnormalities following interval laparoscopic sterilisation. Methods. A prospective cohort study of 225 women undergoing sterilisation between August 2012 and April 2013, with follow-up 1 year post procedure, was conducted at King Dinuzulu Hospital, Durban. Results. Following sterilisation, LGTIs were increased only in women with a history of infection pre sterilisation (odds ratio 6.7; 95% CI 2.2 - 20.9; p=0.002). There was no significant risk of HIV acquisition post sterilisation. In women who had not used contraception or used barrier methods pre sterilisation, we found no significant change in menstrual patterns post sterilisation. There was an increase in menstrual bleeding and dysmenorrhoea post sterilisation among previous combined oral contraceptive users. Among women with amenorrhoea on injectable contraception pre sterilisation, 73.8% reported return to regular menses and 26.2% reported abnormal uterine bleeding post sterilisation. Among injectable contraceptive users with regular menses pre sterilisation, 71.4% reported no change in menses post sterilisation and 28.6% reported abnormal uterine bleeding post sterilisation. Conclusion. In women undergoing interval tubal sterilisation, the risk of LGTIs was only increased in those women with a history. Menstrual abnormalities post sterilisation were more likely in women who used steroid contraception prior to sterilisation. S Afr J Obstet Gynaecol 2017;23(2):38-42. DOI:10.7196/SAJOG.2017.v23i2.1115

Lower genital tract infections (LGTIs) manifest as abnormal vaginal discharge, pruritus, cutaneous lesions or dysuria. The aetiology is either non-sexually transmitted infections such as candidiasis, or sexually transmitted infections (STIs) (trichomoniasis, bacterial vaginosis, oral or genital herpes, chlamydia, gonorrhoea and genital warts).[1] The prevalence of STIs in Durban, South Africa (SA) remains extremely high, with a disproportionate burden among women. STIs prevail in up to 13% of women, with an incidence rate of up to 20 per 100 women years.[2] The sequelae include facilitated HIV transmission and acquisition, pelvic inflammatory disease (PID), infertility and chronic pelvic pain.[2] Literature suggests that high-risk women are at an increased risk of STIs due to low rates of condom use post sterilisation.[3] In a study of African American women residing in low-income communities, sterilised women were less likely to report using condoms with their main partners than non-sterilised women.[3] Contrary to this finding, SangiHaghpeykar et al.[4] reported that women at high risk of contracting STIs (a history of STIs or multiple partners) are significantly more likely to plan condom use post sterilisation with insignificant risk of exposure to disease post procedure. Literature regarding the influence of sterilisation on the risk of HIV transmission and risk behaviours is scarce. Armstrong et al.[5] compared HIV risk behaviours of sterilised and non-sterilised women. They found that sterilised women were more sexually active

and tended not to use condoms, with only 12% of participants using condoms. No difference in HIV risk behaviour due to multiple sexual partners or prostitution was observed between the two groups, mainly due to monogamy post sterilisation.[5] Since 1951, it has been hypothesised that sterilisation might increase a woman’s risk of abnormal uterine bleeding (post tubal ligation syndrome – risk of heavy bleeding and intermenstrual bleeding).[6] Some authors found an increased risk of menstrual abnormalities up to 5 years post sterilisation. Others found that women reported shorter or irregular menstrual cycles, decreased bleeding and dysmenorrhoea.[7,8] Controlling for previous contraceptive use, further studies have found no direct influence of sterilisation on menstrual patterns.[9-11]

Methods

The study population included women who were referred to King Dinuzulu Hospital, Durban, SA for voluntary elective sterilisation from 1 August 2012 - 30 April 2013, and who returned for followup 1 year post procedure. Preceding the sterilisation procedure, a detailed history, gynaecological examination and an HIV test were performed on all women. Following the laparoscopic tubal sterilisation with Filshie clips, women were followed-up for 1 year post sterilisation, i.e. 1 August 2013 - 30 April 2014 on an outpatient basis. Following informed consent for participation in the study,

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH women were routinely educated about safe sexual practices and offered a full gynaecological consultation and an HIV test. Women who failed to return for a follow-up consultation were contacted telephonically to return for their follow up 12 - 18 months post procedure. These women were subsequently followed-up until ethical approval expired, i.e. 25 July 2014. Women who failed to return for follow-up were excluded from the study. Data were collected from interviews and hospital charts. Family planning registrars were trained on the use of the data sheet and the conduct of the interview. These interviews were conducted by registrars and the consultant in charge of the unit. Data were confidential and strictly for the purpose of the study. Information regarding LGTIs, sexual behaviours and menstrual patterns was gathered by interview and women’s subjective impressions. Genital tract infections (GTIs) were categorised into LGTIs (vaginal discharge, genital warts or genital ulcers) and upper genital tract infection, i.e. pelvic inflammatory disease (PID). Women who reported vaginal discharge were asked about the nature of the discharge and treatment sought to distinguish between vaginal candidiasis from STIs. Data were captured and analysed using SPSS version 19.0 (IBM Corp., USA). Normally distributed data were analysed using a paired samples Student’s t-test when comparing two groups. The Wilcoxon sign rank test was used on comparison of the pre- and post sterilisation differences in menstrual variables and LGTI variables. Pearson’s χ2 or Fisher’s exact tests were utilised to identify trends between categorical data variables as appropriate. A p-value <0.05 was considered as statistically significant. Permission was granted by King Dinuzulu Hospital to conduct this research study at their premises and have access to women’s files. Ethical clearance was obtained from the Biomedical Research Ethics Committee of the University of KwaZulu-Natal (ref. no. BE 139/13), where the study was registered.

Results

A total of 225 women returned at 1 year post sterilisation and were included in the study cohort.

Demographics The demographic characteristics of our cohort are shown in Table 1. The mean (SD) age of women was 35 (4.26) years. The racial profile varied; 60% were black (n=135), 35.6% indian (n=80), 1.8% white (n=4) and 2.7% were coloured (n=6).

Genital tract infections (GTIs) Pre sterilisation, 34 women (15.1%) reported a history of a GTI due to an STI (Table 2). Post sterilisation, 11 women (4.9%) reported LGTIs owing to STIs. Eleven women (4.9%) sought medical attention for a vaginal discharge post procedure. Ten women received syndromic treatment for STIs and 1 woman received treatment for vaginal candidiasis. Among the women with STIs, 10 women sought medical attention for a vaginal discharge and 1 woman for genital warts. None of these women reported a history of PID post sterilisation. Of these 11 women, 6 had previously sought medical attention for LGTIs owing to STIs pre-sterilisation (17.6%) and 5 women (2.6%) had no history of prior infection. There was an increase in LGTIs due to STIs post sterilisation in those women who had a history of STIs (OR 6.7; 95% CI 2.2 - 20.9; p=0.002) compared with women who had no history of infection pre sterilisation. In

women with no history of a LGTI, sterilisation conferred a lower risk of LGTIs (OR 0.8; 95% CI 0.7 - 0.9; p=0.001).

Sexual behaviour Most women had 1 sexual partner at the time of sterilisation (n=209, 92.9%), with 9 women (4%) reporting 1 - 3 partners and 7 women (3.1%) had no partners. Post sterilisation, 213 women (94.7%) reported having 1 partner, 8 women (3.6%) had no partner and 4 women (1.8%) reported having 1 - 3 partners. There was no significant change in the number of sexual partners post sterilisation (p=0.31). Of the cohort, 53.3% (n=120) reported having sexual intercourse more than once weekly, 41.8% (n=94) reported weekly or monthly sexual intercourse, while 4.9% (n=11) reported not being sexually active post procedure. A total of 132 women (58.7%) reported having the same sexual interest and pleasure compared with their experiences pre procedure, 27.1% (n=61) had more interest and pleasure, and 14.2% (n=32) experienced less interest and pleasure. Among the women with STIs post sterilisation, 6 (54.5%) women had a previous history of STIs. Of these 6 women, 4 (66.7%) were HIV-negative and 2 (33.3%) women were HIV-positive. Among the HIV-positive women, 1 woman (50%) reported an increase in the number of partners post procedure. Both women reported occasional and inconsistent condom use during sexual intercourse. Both women reported an increase in sexual activity post sterilisation with the same sexual interest and pleasure compared with their pre sterilisation experiences. The remaining 4 women, who were HIVnegative, all reported no condom use post sterilisation. Two of these 4 women (50%) reported an increase in the number of sexual partners post sterilisation. Three of the 4 women (75%) reported increased sexual activity and pleasure post sterilisation and 1 woman (25%) reported the same sexual interest and pleasure. Of the 5 women (45.5%) who had no prior history of STIs pre sterilisation, all were HIV-negative. Four women (80%) reported 1 partner pre and post sterilisation, and 1 (20%) reported 1 partner pre sterilisation and 2 partners post sterilisation. None of Table 1. Demographic characteristics of study cohort (N=225) Mean (SD) Age (years)

35 (4.26)

Partner’s age (years)

39 (5.55)

Parity (n)

3 (1)

Haemoglobin (g/dL)

12.4 (6.19)

Body mass index (kg/m2)

33.3 (7.42)

SD = standard deviation.

Table 2. Genital tract infections pre and post TL Pre TL, n (%)

Post TL, n (%)

34 (15.1)

11 (4.9)

LGTI: Vaginal discharge

30 (13.3)

10 (4.4)

Genital warts

2 (0.9)

1 (0.4)

UGTI: PID

2 (0.9)

GTIs

TL = tubal ligation; LGTI = lower genital tract infection; UGTI = upper genital tract infection; PID = pelvic inflammatory disease; GTI = genital tract infection.

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH these women reported condom use pre and post sterilisation. Four of these women reported more frequent sexual intercourse with the same interest and pleasure. One woman reported the same frequency, sexual interest and pleasure compared with her pre sterilisation experience.

The majority of HIV-positive women (66.7%) always used condoms during coitus post sterilisation (n=38). The remaining 21.1% of HIVpositive women (n=12) used condoms occasionally and 12.3% (n=7) did not use condoms post sterilisation.

HIV status

Menstrual patterns

A total of 169 women were HIV-negative (75%). Of the 56 women (25%) who were HIV-positive, 38 (67.9%) were on antiretroviral drugs (ARVs) and 18 (32.1%) were not on treatment owing to CD4 cell counts >350 cells/µL. Post sterilisation, 168 women tested HIVnegative. One woman, who had tested HIV-negative pre sterilisation, had tested positive post sterilisation. She reported an increase in sexual partners, non-use of condoms and increased sexual activity post sterilisation. Among the study cohort, there was no risk of HIV acquisition post sterilisation (relative risk 0.006; 95% CI 0.001 - 0.042, p=1). In women who were HIV-positive, the mean CD4 cell count pre procedure was 426 cells/μL and 601 cells/μL post procedure. There was a significant rise in CD4 cell counts (p=0.02), with 66.7% of HIV-positive women (38/57) being on ARVs.

Table 3 outlines menstrual patterns pre and post sterilisation controlling for previous contraceptive use. Women on barrier methods and those not on any contraception pre sterilisation had no statistically significant changes in their number of days of menstrual bleeding (p=1) or dysmenorrhoea (p=0.89). Among the women with amenorrhoea on injectable contraception pre sterilisation, 73.8% (n=79) reported return to regular menses, and 26.2% (n=28) reported abnormal uterine bleeding post sterilisation (prolonged bleeding, heavy bleeding and cycle irregularity). Among injectable contraceptive users with regular menses or spotting pre sterilisation, 71.4% (n=30) reported no change in menses post sterilisation and 28.6% (n=12) reported abnormal uterine bleeding 225

Contraception

200 175 149

150

Cases, n

The majority of women were on injectable contraception (DepoProvera, Nur-isterate) as outlined in Fig. 1 with the remaining on combined oral contraception (COC), barrier methods or no contraception. Of the cohort, 15.5% (n=35) of women used dual contraception. Pre sterilisation, 19.1% (n=43) women used condoms occasionally, 20% (n=45) always used condoms and 60.9% (n=137) women did not use condoms during sexual intercourse. Post procedure, 37.3% of women (n=84) used condoms either occasionally or always and 62.7% (n=141) did not use condoms. This increase in non-use of condoms did not reach statistical significance (p=0.23).

125 100 75 52

50 25 0

COC

Barrier

Nil

Contraceptive method

Fig. 1. Contraceptive use pre sterilisation. (IC = injectable contraceptive; COC = combined oral contraceptive).

Table 3. Menstrual patterns pre and post TL controlling for contraceptive use IC (N=149) Parameter

COC (N=52)

Barrier (N=11)

Nil (N=13)

Pre TL

Post TL

Pre TL

Post TL

Pre TL

Post TL

Pre TL

PostTL

Amenorrhoea

107 (71.8)

27 (18.1)

79 (53.0)

38 (73.1)

4 (7.7)

7 (63.6)

8 (61.5)

69 (46.2)

4-8

14 (9.4)

65 (43.6)

14 (26.9)

44 (84.6)

3 (27.3)

5 (38.5)

7 (53.8)

1 (0.7)

5 (3.4)

4 (7.7)

1 (9.1)

Amenorrhoea

107 (71.8)

Light

21 (14.1)

46 (30.9)

24 (46.2)

2 (3.9)

4 (36.4)

4 (30.8)

3 (23.1)

Average

18 (12.1)

76 (51.0)

28 (53.9)

27 (51.9)

5 (45.5)

4 (36.4)

6 (46.1)

7 (53.9)

Heavy

3 (2.0)

27 (18.1)

23 (44.2)

2 (18.2)

3 (27.3)

3 (23.1)

Amenorrhoea

107 (71.8)

Regular

32 (21.5)

130 (87.3)

52 (100)

47 (90.4)

10 (90.9)

13 (100)

Irregular

10 (6.7)

19 (12.8)

5 (9.6)

1 (9.1)

Nil

149 (100)

104 (69.8)

52 (100)

19 (36.5)

11 (100)

10 (76.9)

9 (9.2)

Yes, not requiring meds

39 (26.2)

28 (53.9)

3 (23.1)

Yes, requiring meds

6 (4.0)

5 (9.6)

1 (7.7)

Days of menses, n (%)

Flow of menses, n (%)

Cycle regularity, n (%)

Pain with menses, n (%)

TL = tubal ligation; IC = injectable contraceptive; COC = combined oral contraceptive.

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH post sterilisation (prolonged bleeding and heavy menstrual bleeding). Among previous injectable users, there was a significant increase in days of bleeding (p=0.03), amount of menstrual bleeding (p=0.01) and dysmenorrhoea post sterilisation (p=0.005). Fourteen women (9.4%) sought medical attention post sterilisation for heavy menses (n=4, 28.6%) and cycle irregularity (n=10, 71.4%). These women were assessed clinically for abnormalities and 1 woman (7.1%) was found to have leiomyomas and was referred to her base hospital for further treatment. The remaining women in whom no abnormalities were found were successfully treated medically. Among previous COC users, 37 women (71.2%) reported heavy or prolonged menses post sterilisation. Cycle irregularity was reported by 5 women (9.6%) and dysmenorrhoea by 33 women (63.5%). An increase in duration of menses (p=0.02), flow of menses (p=0.01) and dysmenorrhoea (p=0.001) was found in previous COC users post sterilisation. Five women (9.6%) sought medical attention post sterilisation for heavy menses and cycle irregularity. Clinically, no organic cause was found and these women were successfully treated.

Discussion

In SA, female sterilisation accounts for up to 14% of contraceptive use among sexually active women.[12] In our setting with a high prevalence of STIs and HIV,[2] there are concerns regarding the impact of sterilisation on the spread of STIs and HIV. We found the risk of LGTIs owing to STIs was not directly affected by interval tubal sterilisation but is rather a reflection of sexual behaviour prior to sterilisation. A history of a prior STI in women undergoing tubal sterilisation conferred a 6-fold risk of a repeat STI after sterilisation compared with women who had no history of infection pre sterilisation. In this subset of women, non-use of condoms post sterilisation, increase in sexual partners and sexual activity, interest and pleasure was found. There was no risk of HIV acquisition post sterilisation; however, an increase in STIs and HIV risk behaviour was found in women with a past history of STIs. Similarly, low rates of condom use in high-risk groups post sterilisation is supported by previous studies.[3] Women in our cohort were unlikely to use condoms if they were HIV-negative or reported just one sexual partner post sterilisation. Previous literature found that women with a past history of STIs are more likely to plan condom use post sterilisation,[4] which differs from our study findings. In women with a past history of STIs, interventions to increase condom use post sterilisation are needed to reduce both the risk of subsequent infections and HIV. Education of both women and partners and offering ample free supplies may alleviate this problem. Sweat et al.,[13] found condom social marketing (condom branding, media-based marketing campaign and targeted population-based campaigns) may double the odds of condom use. This may be an effective strategy in high-risk populations to advocate for and increase condom use. Reassuring findings from our study were that the majority of HIVpositive women used condoms post sterilisation (66.7%), which may reflect women’s use of the health system and counselling at HIV clinics which needs to be emphasised and extended to all health facilities for both HIV-positive and -negative women. In women with no prior history of STIs, we found no risk of infection post sterilisation. This was due to the older age of the women, monogamous relationships, no change in partners post sterilisation and no change in condom use post sterilisation. Furthermore, consistent with previous studies,[14] the majority of women reported no change in sexual interest and pleasure after interval tubal sterilisation (58%).

In women who were not on steroid contraception pre sterilisation, we found no change in menstrual patterns post sterilisation. This is supported by Wilcox et al.,[7] who found that menstrual abnormalities (dysmenorrhoea, heavy bleeding) are more common 5 years post sterilisation than within the first 2 years post procedure. Menstrual abnormalities in our study cohort were significant in women who previously used steroid contraception. There was an increase in the duration and amount of menstrual bleeding and dysmenorrhoea post sterilisation among previous COC users. Among previous injectable contraceptive users, abnormal uterine bleeding and dysmenorrhoea was found in our study cohort. Our findings suggest that menstrual changes post sterilisation are due to recent discontinuation of steroid contraception, rather than the sterilisation procedure itself. The literature further supports that sterilisation has no direct influence on menstrual patterns.[9-11] Accounting for prior contraceptive use, changes in menses post sterilisation have been shown to result in the return of pre-contraceptive use characteristics, including dysmenorrhoea, heavy menstrual flow and cycle irregularity.[10,11] Previously, it has been shown that women on injectable contraception have re-establishment of menstrual cycles up to 6 - 8 months following the last injection.[15] Women in our cohort subjectively reported changes in menses due to prolonged amenorrhoea (women on injectable contraception) and due to previous COC use.

Study limitations Selection bias was encountered in that only women who were willing and able to present themselves one year post sterilisation were included in the study cohort. A further limitation was that data pertaining to menstrual patterns and the presence of STIs were obtained from the women’s history and from subjective impressions.

Conclusions

The risk of LGTIs due to STIs in women undergoing tubal ligation was increased only in those women with a history of infection pre sterilisation. Women with a history of STIs need to be counselled and offered follow-up visits to allow for screening for infection post procedure. They need to be counselled on the correct and consistent use of condoms post sterilisation. In women who had used steroid contraception previously, menstrual abnormalities were more likely post sterilisation, which highlights the importance of counselling these women prior to sterilisation. Further research on the correlations between HIV, sterilisation and risk behaviour is needed. Large trials examining risk behaviours in HIVpositive and HIV-negative women post sterilisation is recommended.

Acknowledgements. We would like to acknowledge Ms Jesca Batidzirai, statistician at the University of KwaZulu-Natal, for her assistance with the statistical analysis of the data. We wish to express our gratitude to the management and staff of King Dinuzulu Hospital for permitting us to conduct our study at their institution and to the women from King Dinuzulu Hospital who participated in our study. Author contributions. GK designed the study with input from JB and MP. MP monitored data collection and GK collected data for analysis. GK, JB and MP analysed and interpreted the data and together prepared the manuscript. All authors contributed to and approve the manuscript. Funding. None. Conflicts of interest. None.

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH 1. Landers DV, Wiesenfeld HC, Heine RP, Krohn MA, Hillier SL. Predictive value of the clinical diagnosis of lower genital tract infection in women. Am J Obstet Gynecol 2004;190(4):1004-1008. https://doi.org/10.1016/j.ajog.2004.02.015 2. Naidoo S, Wand H, Abbai S, Ramjee G. High prevalence and incidence of sexually transmitted infections among women living in Kwazulu-Natal, South Africa. AIDS Res Ther 2014;11(31):2-7. https://doi.org/10.1186/1742-6405-11-31 3. Semaan S, Lauby J, Walls C. Condom use with main partners by sterilized and non-sterilized women. Wom Health 1997;25(2):65-85. https://doi.org/10.1300/J013v25n02_04 4. Sangi-Haghpeykar H, Poindexter A. Planned condom use among women undergoing tubal sterilisation. Sex Transm Dis 1998;25(7):335-341. https://doi.org/10.1097/00007435-19980800000002 5. Armstrong KA, Samost L, Tavris DR. HIV-risk behaviors of sterilized and nonsterilized women in drug-treatment programmes-Philadelphia, 1989-1991. MMWR 1992;41(9):149-152. 6. Peterson HB, Pollack AE, Warshaw S. Tubal Sterilization. Chapter 23. In: Rock JA, Jones HW, Te Linde RW, eds. Te Linde’s Operative Gynaecology. Philadelphia: Wolters Kluwer Health/Lippincott Williams & Wilkins, 2008:609-629. 7. Wilcox LS, Martinez-Schnell B, Peterson HB, Ware JH, Hughes JM. Menstrual function after tubal sterilization. Am J Epidemiol 1992;135(12):1368-1381. https://doi.org/10.1093/oxfordjournals.aje. a116248 8. Peterson HB, Jeng G, Folger S, Hillis S, Marchbanks P, Wilcox L. The risk of menstrual abnormalities after tubal sterilization. N Engl J Med 2000;343(23):1681-1687. https://doi.org/10.1056/ NEJM200012073432303

9. Harlow B, Missmer SA, Cramer DW, Barbieri RL. Does tubal sterilization influence the subsequent risk of menorrhagia or dysmenorrhea? Fertil Steril 2002;77(4):754-760. https://doi.org/10.1016/ S0015-0282(01)03253-8 10. Westhoff C. Tubal sterilization – safe and effective. N Engl J Med 2000;343:1724-1726. https://doi. org/10.1056/NEJM200012073432310 11. Rulin MC, Davidson AR, Philliber SG, Graves WL, Cushman LF. Long-term effect of tubal sterilization on menstrual indices and pelvic pain. Obstet Gynecol 1993;82(1):118-121. 12. South African Demographic and Health Survey. 2003. http://www.gov.za/documents/south-africandemographic-and-health-survey-sadhs (accessed 2 December 2015). 13. Sweat MD, Denison J, Kennedy C, Tedriwc V, O’Reilly K. Effects of condom social marketing on condom use in developing countries: A systematic review and meta-analysis, 1990 - 2010. Bull World Health Organ 2012;90(8):613-622. https://doi.org/10.2471/BLT.11.094268 14. Costello C, Hillis S, Marchbanks P, Jamieson D, Peterson H. The effect of interval tubal sterilization on sexual interest and pleasure. Obstet Gynaecol 2002;100(3):511-517. https://doi.org/10.1016/ S0029-7844(02)02042-2 15. Westhoff C. Depot-medroxyprogesterone acetate injection (Depo-Provera): A highly effective contraceptive option with proven long-term safety. Contraception 2003;68(2):75-87. https://doi. org/10.1016/S0010-7824(03)00136-7

SAJOG • September 2017, Vol. 23, No. 2

This open-access article is distributed under Creative Commons licence CC-BY-NC 4.0.

RESEARCH

Perinatal outcomes in pregnant women presenting with preterm premature rupture of membranes at a regional hospital in KwaZuluNatal Province, South Africa G Msomi,1 MB ChB; M Naidoo,2 MB ChB, MFamMed, FCFP, MSc (Sports Med), PhD; B Hira,1 MB BCh, FCOG Discipline of Obstetrics and Gynaecology, School of Clinical Medicine, University of KwaZulu-Natal, Durban, South Africa Discipline of Family Medicine, School of Nursing and Public Health, University of KwaZulu-Natal, Durban, South Africa

1 2

Corresponding author: M Naidoo (naidoom@ukzn.ac.za)

Background. Worldwide, the incidence of preterm premature rupture of membranes (PPROM) is between 1% and 4% of all pregnancies. Objectives. The primary objectives of this study were to describe and compare the perinatal outcomes of HIV-positive and HIV-negative women presenting with PPROM to a regional hospital in KwaZulu-Natal. Methods. This was a retrospective analytical cross-sectional study which reviewed files of pregnant women presenting with premature rupture of membranes at gestation between 28 and 36 completed weeks. These were identified from the labour ward birth register and from the neonatal ward admission book. Categorical and numerical variables pertaining to the method of confirmation of diagnoses, clinical profiles, modes of delivery, maternal outcomes and neonatal outcomes were considered. Results. A total of 87 files were analysed. Forty-six women (53%) were HIV-negative and 41 (47%) were HIV-positive. Fifty-two percent were in the gestational age <34 weeks. Fifty-nine percent (n=51) of women delivered vaginally and 31% (n=27) delivered by caesarean delivery (CD). There was also no statistical significance between the Apgar scores of the HIV-exposed and HIV-unexposed neonates, birth weights and modes of delivery. There was no statistical significance in sepsis rates, the need for ventilation and the duration of hospital stay between the two groups. The odds of developing neonatal jaundice (NNJ) in the HIV-positive group was 0.14 (95% confidence interval (CI) 0 - 0.93), which was statistically significant. There was no reported maternal or neonatal mortality and no maternal morbidity associated with PPROM in either groups. Conclusion. This study suggests that there are no immediate significant differences in neonatal and maternal outcomes in pregnancies complicated by PPROM between HIV-negative and HIV-positive women on ART except that of NNJ. S Afr J Obstet Gynaecol 2017;23(2):43-47. DOI:10.7196/SAJOG.2017.v23i2.1124

Preterm delivery and prematurity are significant contributors to perinatal morbidity and mortality, and their inclusion in the list of sustainable development goal epitomises this. Worldwide, the incidence of preterm delivery is between 1% and 4% of all pregnancies, with higher incidence notably in low-income countries.[1] This figure rises significantly if a patient has a history of premature rupture of membranes (PROM). It is estimated that 30 - 40% of cases of preterm delivery are preceded by PROM.[2-6] As a complication of pregnancy, PROM has potentially devastating perinatal and maternal outcomes, not to mention the management dilemma it poses to the obstetrician. It is therefore no surprise that much international research has been done on this subject, with varying and sometimes conflicting findings. The role of HIV as a confounding factor to preterm PROM (PPROM) has, however, not yet been adequately described. There are no clear distinctions in the management of PPROM in HIV-positive and HIV-negative groups. Studies have found an increased incidence of preterm deliveries, low birth weights and perinatal deaths as the main associated complications in the HIV-positive group.[7-11] Other disputed complications in this group include increased risk of intrapartum hypoxia and neonatal encephalopathy.[12-15] HIV and chorioamnionitis have mutually deleterious effects on PPROM, as HIV is implicated in the pathogenesis of

chorioamnionitis, and HIV-positive patients who develop PPROM are much more likely to develop chorioamnionitis.[16] The presence of chorioamnionitis, with or without rupture of membranes, can in turn increase the risk of vertical transmission of the virus from mother to the fetus.[16,17] Vertical transmission of the virus from mother to child has been researched in great depth, with encouraging results regarding the use of antiretroviral treatment (ART) in reducing transmission. The risk of vertical transmission varies based on the stages of pregnancy, with the highest risk occurring during parturition.[18] Increases in the duration of the rupture of membranes, prematurity and prolonged labour after rupture of membranes increase the risk of perinatal HIV transmission.[3,18] Some studies have reported an increased risk of preterm deliveries with the use of antiretroviral drugs (ARVs).[20,21] It is against this background that this study investigated the perinatal outcomes of HIV-positive, compared with HIV-negative women presenting with PPROM at the regional hospital in KwaZulu-Natal Province (KZN), South Africa.

Methods

We retrospectively analysed the files of pregnant women who presented with PPROM to a regional hospital in KZN from 1 January 2013 to 31 of December 2013. This hospital offers a regional level of

SAJOG â&#x20AC;˘ September 2017, Vol. 23, No. 2

RESEARCH care to the surrounding suburban population. It also serves as a referral point for a nearby district hospital, rural and urban primary healthcare clinics, and serves a population of ~400 000 people. Due to the limitations of our neonatal facilities, viability is restricted to gestation of 28 weeks or a birth weight of 1 000 g. Inclusion criteria were clinical files of women who presented or were referred with a history of having ruptured membranes, without exhibiting any signs of labour, and were between the 28th and 37th week of gestation. In the management of a fully developed fetus with PROM decision-making is generally less complicated. Conversely, the management of pregnancies of gestation less than 28 weeks is far more complex and inconsistent, thereby making it difficult to study retrospectively. Women with existing medical conditions such as hypertensive diseases that could influence perinatal outcomes were excluded. Further exclusion criteria were unknown gestational age, unknown HIV-status, congenital abnormalities and cases where corresponding neonatal files were missing. When reviewing the clinical files, rupture of membranes was confirmed by either visualisation of amniotic fluid in the vagina on speculum examination, a convincing history of drainage of amniotic fluid, plus reduced amniotic fluid index on ultrasound scan or by observing persistent amniotic fluid staining of sanitary pads. The corresponding neonatal charts were also reviewed. It is important to note that, regardless of the gestational age, this institution does not offer expectant management for women with PPROM. A data collection sheet was used to extract the required information from the medical files. Maternal and obstetric characteristics included demographic data, gestation, number of previous pregnancies and relevant obstetric history, HIV-status, rhesus factor and syphilis serology. Clinical data were assessed in respect of relevant clinical findings: laboratory investigations (haematology and microbiology), ultrasound scans, administration of antibiotics and corticosteroids, induction of labour and the eventual mode of delivery. The primary outcomes of this study were the perinatal and maternal morbidity and mortality occurring before discharge. Collated perinatal data included Apgar scores at 1 and 5 minutes, admission to the neonatal ward or intensive care unit, need for ventilation, development of any of the following conditions: hyaline membrane disease, hypoxic ischaemic encephalopathy, intraventricular haemorrhage, congenital or nosocomial infections, necrotising enterocolitis and death. Maternal morbidity included increased length of hospital stay (more than 12 hours following vaginal delivery and more than 72 hours following caesarean section (CS)). The variables on the data sheet were captured on a Microsoft Excel spreadsheet and imported into Stata version 13 (StataCorp LP, USA) for analysis. The HIV-positive and HIV-negative subgroups were then compared. The statistical analysis was performed using Pearson χ2 test with p<0.05 considered significant. We determined the mean and odds ratios (OR) for various outcomes between the two cohorts and their 95% confidence intervals (CI) were calculated. The study was approved by the Biomedical Ethics Committee at the University of KwaZulu-Natal (ref. no. BE 275/14), the provincial Department of Health and the facility manager.

Results

There was a total of 7 694 deliveries in 2013. Out of a total of 130 (0.2%) files of patients who potentially met the inclusion criteria

only 87 files were included in this study. Forty-three files were either missing or lacked the crucial information required for the study. Most women (n=64; 74%) fell in the age bracket of 18 - 34 years, with 13 (15%) women <18 years, and (n=10; 11%) >34 years of age. Forty-six women (53%) were HIV-negative and 41 (47%) were HIV-positive. In the HIV-positive cohort 5 (12%) patients had CD4 counts <200 cells/µL, 13 (32%) between 200 and 350 cells/µL, 20 (49%) >500 cells/µL, and 3 (7%) patients had unknown CD4 counts. All HIV-positive women were on a triple ART regimen. The clinical information of all reviewed patients is presented in Table 1. Regarding the patients' gestational age, 24% (n=21) were between 28 and 32 weeks, 28% (n=24) were between 32 and 34 weeks, while 48% (n=42) were between 34 and 36 weeks. There were no statistical Table 1. Comparison of clinical information of HIV-positive (n=41) and HIV-negative (n=46) participants HIVnegative, n (%)

HIVpositive, n (%)

p-value

28 - 32

9 (20)

12 (29)

0.21

33 - 34

13 (28)

11 (27)

0.53

35 - 36

24 (52)

18 (44)

0.29

<12

3 (7)

0.10

>12 and <24

10 (22)

8 (19)

0.51

>24

30 (65)

24 (59)

0.54

Unknown

6 (13)

6 (15)

Speculum

20 (43)

18 (44)

0.57

Pad inspection

5 (11)

9 (22)

0.13

History and low AFI

4 (9)

3 (7)

0.57

Digital exam

6 (13)

9 (22)

0.21

Unknown

11 (24)

2 (5)

0.01

Yes

5 (11)

9 (22)

0.13

35 (76)

29 (71)

0.37

Unknown

6 (13)

3 (7)

0.30

Yes

14 (30)

21 (51)

0.38

27 (59)

18 (44)

0.12

Unknown

5 (11)

2 (5)

0.27

NVD

30 (65)

28 (68)

0.47

16 (35)

13 (32)

0.47

Yes

11 (24)

11 (27)

0.54

14 (30)

13 (32)

0.43

Unknown

21 (46)

17 (41)

Gestation (weeks)

Time from PPROM to delivery (hours)

Confirmation of SROM

Tocolysis

Steroid administration

Mode of delivery

Induction of labour

SROM = spontaneous rupture of membranes; NVD = normal vaginal delivery; CD = caesarean delivery; AFI = amniotic fluid index; PPROM = preterm premature rupture of membranes.

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH differences pertaining to the clinical profile and management of the HIV-positive and HIV-negative women, except that women who were HIV-positive had steroids administered significantly more often. The mean number of antenatal visits in the HIV-negative group was 4.17 (95% CI 3.43 - 4.91) whereas in the HIV-positive group it was 4.31 (95% CI 3.79 - 4.84) and there was no statistically significant difference between the 2 groups. Only 2 patients had no antenatal visits and both were HIV-negative. The number of patients who booked before 20 weeks of gestation in the HIV-negative and HIV-positive groups were 15 (33%) and 16 (39%), respectively, and this was not statistically significant. All the booked patients in this study were rhesus-positive. Only one patient was HIV-positive and had a positive screening test for syphilis. In 38 (44%) women, the drainage of amniotic fluid was confirmed with the recommended use of a speculum. Ultrasound amniotic fluid index and regular pad inspection confirmed rupture of membranes in 7 (8%) and 14 (16%) women, respectively. Notably, 15 (17%) had the rupture of membranes confirmed by the observation of leakage of amniotic fluid during digital per vaginam examination. The common bacteriological risk factors for premature rupture of membranes could not be determined in both groups, as very few patients had specimens sent for analysis. One HIV-negative patient had a Bartholin’s abscess and one HIV-positive patient had a urinary tract infection. None of the women exhibited any other pregnancy-related infections. A total of 21 women (24%) had a latency period from rupture of membranes to delivery of <24 hours; 10 (22%) were in the HIV-negative Table 2. Comparison of neonatal outcomes of HIV-negative (N=46) and HIV-positive (N=41) women

group and 11 (27%) were in the HIV-positive group. The difference was not statistically significant. Three women delivered within 12 hours and all were in the HIV-positive group. The majority of women (62%) had a latency of >24 hours, 30 (65%) and 24 (56%) in the HIV-negative and HIV-positive groups, respectively. Women who had ruptured membranes for >12 hours and those who warranted suppression of labour to facilitate steroid lung maturity were routinely given prophylactic antibiotics. Fifty-seven (66%) women were given oral erythromycin and metronidazole in accordance with the hospital management protocol. Four women who were eligible for antibiotics were not prescribed any and 5 had antibiotics mentioned in the management plan, but no prescriptions were found in the files. Eight women were already in labour when the 12-hour period had elapsed. Antibiotic treatment was not prescribed in 9 patients. Fifteen women (16%) had their labour suppressed using nifedipine to allow for fetal lung maturity. Betamethasone was given to 35 of the 45 women eligible to receive steroids for fetal lung maturity. There were no statistically significant differences between the 2 groups with respect to tocolysis and the use of steroids. Twenty-two (25%) women had induced labour; 51 (59%) delivered vaginally and 27 (31%) delivered by caesarean delivery (CD). Indications for CD included fetal distress (n=7), breech presentation (n=5), anhydramnios (n=3), failed induction of labour (n=3), previous CD ×1 (n=4), previous CD ×2 (n=2), poor progress of labour (n=2), cord prolapse (n=1), high parity (n=1) and very low fetal weight (n=1). There were no statistically significant differences in the induction of labour and mode of delivery between the HIVnegative and HIV-positive groups.

Maternal outcomes None of the mothers suffered any intrapartum or postpartum complications until discharge. Mothers who stayed in hospital for >3 days were kept as boarders awaiting completion of the intravenous antibiotic course for their babies and/or for adequate neonatal weight gain.

HIVnegative, n (%)*

HIVpositive, n (%)*

p-value

1 000 - 1 499

4 (9)

7 (17)

0.20

1 500 - 1 999

9 (20)

12 (29)

0.21

Neonatal outcomes

2 000 - 2 499

27 (59)

16 (39)

0.05

>2 500

6 (13)

6 (15)

0.30

Male

18 (39)

25 (61)

0.03

Female

28 (61)

16 (39)

0.03

8 - 10

45 (98)

39 (95)

0.64

1 (2)

2 (5)

0.64

<24

14 (30)

12 (15)

0.55

24 - 48

2 (4)

7 (8)

0.06

>48

8 (17)

11 (14)

0.21

2 (4)

2 (5)

0.64

22 (48)

21 (51)

0.46

3-7

19 (41)

13 (32)

0.40

5 (11)

7 (17)

0.30

In the baby-weight category, 43 (49%) babies weighed 2 000 - 2 500 g, 21 (24%) weighed 1 500 - 1 999 g, 12 (14%) weighed >2 500 g and 11 (13%) weighed 1 000 - 1 500 g. There were significantly more males in the HIV-positive group and more females in the HIVnegative group (p=0.03). Clinical information on neonatal outcomes is presented in Table 2. Overall, 54 babies were admitted to the nursery, 24 were HIV-negative and 30 were HIV-positive. Invasive ventilation was administered to 2 babies in each group. Twenty-six (30%) were discharged from the nursery within 24 hours, 9 (10%) were discharged between 24 and 48 hours and 19 (23%) were in the nursery for >48 hours. The odds of being admitted to the nursery for >48 hours in the HIV-positive group was 1.74 (95% CI 0.64 - 4.76) and the odds of staying in hospital for more than 7 days in the HIV-positive infants was 1.68 (95% CI 0.51 - 5.51). Both odds ratios were not statistically significant. There were no correlations between CD4 cell counts of the mothers and the length of their stays in either the nursery or the hospital. Twenty-three (26%) babies received intravenous antibiotics for suspected congenital sepsis on the basis of a raised white cell count. Twelve (26%) were in the HIV-negative group and 7 (17%) were HIV-positive. This finding was not statistically significant

Birth weight (grams)

Sex

Apgar at 5 min

Time spent in nursery (hours)

umber needing invasive N ventilation Length of hospital stay post delivery (days)

*Percentages are of births within each group.

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH (p=0.44). Routine neonatal septic screening by means of C reactive protein (CRP) and blood culture investigations were not helpful in confirming infection, except in 2 babies, where contamination was suspected. A total of 8 babies (9%) developed neonatal jaundice (NNJ). Of these, only 1 was HIV-negative; 3 were male and 5 were female. The odds of developing NNJ in the HIV-positive group was 0.14 (95% CI 0 - 0.93), which was statistically significant. The mean birth weight was 2 450 g, with a range of 1 200 - 2 900 g. Six women delivered vaginally, while 2 delivered by CD. No babies suffered neurological or gastrointestinal complications associated with PPROM before discharge. There was also no perinatal mortality in the analysed files.

Discussion

This retrospective study found no significant differences in the perinatal and maternal outcomes of HIV-positive and HIV-negative women with pregnancies complicated by PPROM. There were no differences in respect of respiratory complications, neonatal intensive care unit admission, neonatal infections and length of hospital stay post delivery. None of the babies developed any neurological or gastrointestinal complications until they were discharged and there was surprisingly no perinatal mortality. This study also concurs with previous studies performed in other African countries, comparing HIV-positive and HIV-negative cohorts, which found little or no differences between HIV-positive and HIV-negative women with respect to gestational age, birth weight, and Apgar scores.[22,23] We could find no explanation for the skewed distribution of the sexes among the 2 groups. The differences were statistically significant. This could be a statistical error due to the higher number of missing or disqualified files. Eight babies in this study developed NNJ and of these, only 1 was HIV-positive – this was a statistically significant finding. The higher incidence of NNJ among the babies of HIV-negative women is in keeping with the findings by Nakanga et al.[24] in a study undertaken in Malawi in 2015. They hypothesised that the efavirenz (EFV) in the prevention of mother-to-child transmission (PMTCT) of HIV acted as a fetal liver enzyme inducer that would aid the conjugation of bilirubin. All women in this study were on a triple ART regimen that included EFV. However, in contrast to this and other studies, the male sex was not a risk factor for neonatal jaundice in this study.[25,26] The seropositivity rate in this study was 47%, which was higher than the KZN provincial seropositivity rate of 40.1%, but this may be due to the fact that HIV-positive women are more prone to sepsis and hence PPROM.[27] Half of the women had CD4 counts of >350 cells/µL, while 12% had CD4 counts of <200 cells/µL . It is important to note that all the women were either on long-term or prophylactic ART, which consists of the fixed-dose combination of tenofovir (TDF), emtricitabine (FTC) and EFV as per national protocol. There was no association between low CD4 count and adverse outcomes, possibly because of the effect of the ART on suppressing the viral load.[28] The mean number of antenatal visits and mean gestational age at booking were similar between the HIV-positive and HIV-negative groups. There was no obvious association between PPROM and maternal age, poor antenatal clinic attendance or parity in this study. High-risk maternal age groups, teenagers and women >34 years of age accounted for 26% of the women in this study. Nulliparous women constituted 33% of women with PPROM, which was lower than in previous similar studies.[29,30]

Seventeen percent of the women had the rupture of membranes confirmed by digital per vaginam examination despite the leakage of fluid having been noted in the patient history. Standard protocol recommends the use of a speculum to prevent ascending infection.[31] We found no obvious explanation for this, however, disregard for protocol and the shortage of speculums were considered among the reasons for diagnosing PPROM by doing digital per vaginam examinations. In his study of 97 women with PROM in Chris Hani Baragwanath General Hospital (CHBGH), Iloanusi[30] found that 35% of the women had a digital examination prior to the onset of labour, which was double the percentage compared with our study. The most common microorganisms causing PPROM in this setting could not be determined due to the consistent omission of urogenital septic screening in nearly all of the women in this study. Iloanusi[30] also found the same transgression at CHBGH. The lack of microbiological investigations did not seem to have any negative impact on the outcomes, as sepsis was not a significant finding in this study. In accordance with hospital protocol based on scientific evidence nearly all women were given prophylactic antibiotics (erythromycin and metronidazole).[32-34] Babies delivered after prolonged rupture of membranes also received a course of intravenous antibiotics as standard neonatal protocol. In resource-constrained areas, it could be argued that routine urogenital septic screening of patients with PPROM is not essential, particularly if immediate delivery is anticipated and if antibiotics are used routinely. Even if clinical chorioamnionitis is present, Osmanagaoglu and Unal[35] found that neonatal outcomes in such patients were similar to those without clinical chorioamnionitis. Following completion of steroids, a choice between expectant management and delivery is made.[36] The protocol in our hospital setting recommends delivery of the neonate regardless of gestation because of concerns about the risk of infection and other complications associated with prolonged rupture of membranes. The fewer adverse outcomes in this study compared with others elsewhere seem to validate this school of thought. The rate of CD in this study was 33%, which was slightly higher than in similar studies by Iloanusi[30] at CHBGH and by Patil and Patil[29] in India, who reported CD rates of 25% and 27%, respectively. The rate of CD for failed induction of labour was 11.1% compared with 7.6% in the Indian study and 16% in the CHBGH study.

Study limitations Despite the limitations inherent to a retrospective study and the small sample size, this study provides a more objective comparison between HIV-positive and HIV-negative women with PPROM, given the homogeneity of the 2 groups in our study and the similarity in their management. Perhaps, more significantly, this study also reports some encouraging data on the benefit of the use of ART in all HIV-positive pregnant women regardless of their CD4 counts. A major limitation of this study was that a significant number of files were either missing or were excluded due to a deficiency of crucial information required for the study, but this may not have had a negative impact on the overall findings because files from HIV-positive and HIV-negative women stood equal chances of being excluded. Deviation from the standard protocol of screening for infections in a form of midstream urine and vaginal swabs was noted. The extent to which sepsis is a risk factor for the PPROM and

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH the most common pathogenic organism could not be established. The above two points are concerning in view of the rising litigation claims being made against the public and private sectors and the cost of settling such claims.[37] If the effect of medical litigation in obstetrics is to be minimised, it is absolutely imperative that clinicians and nurses ensure that evidence-based practice guidelines are followed, that medical records are comprehensively completed and that administrative staff ensure that records are safely stored.[35]

Conclusion

This study suggests that there are no immediate significant differences in neonatal and maternal outcomes in pregnancies complicated by PPROM between HIV-negative and HIV-positive women who are on prophylactic or long-term ART, except in the development of NNJ, where ART was protective. PPROM did not seem to significantly compound adverse perinatal outcomes associated with preterm delivery when the management strategy of immediate delivery after steroid fetal lung maturation, regardless of gestation, was employed. A larger and more powered study is recommended to confirm the findings of this study. Furthermore, intermediate and longterm follow up is suggested to determine the development of late neonatal and maternal complications in this setting. Acknowledgements. The authors would like to thank the administrative, nursing and medical staff who assisted with the study. Author contributions. GS was the principal investigator and MN was the main research supervisor while BH provided specialist input. All authors were responsible for the drafting of the publication. Funding. None. Conflicts of interest. None. 1. French JI, McGregor JA. The pathobiology of premature rupture of membranes. Semin Perinatol 1996;20(5):344-368. https://doi.org/10.1016/s0146-0005(96)80002-4 2. Taylor J, Garite TJ. Premature rupture of membranes before fetal viability. Am J Obstet Gynecol 1984;64:615-620. http://dx.doi.org/10.1016/0002-9378(90)91080-V 3. Smith G, Rafuse C, Anand N, et al. Prevalence, management, and outcomes of preterm prelabour rupture of the membranes of women in Canada. J Obstet Gynaecol Can 2005;27(6):547-553. https:// doi.org/10.1016/s1701-2163(16)30711-3 4. Maymon E, Chaim W, Sheiner E, Mazor M. A review of randomised clinical trials of antibiotic therapy in preterm premature rupture of membranes. Arch Gynecol Obstet 1998;261(4):173-181. https://doi.org/10.1007/s004040050218 5. McGregor JA, French JI, Lawellin D, Todd JK. Preterm birth and infection: Pathogenic possibilities. Am J Reprod Immunol Microbiol 1988;16:123-132. 6. Simhan HN, Canava TP. Preterm premature rupture of membranes: diagnosis, evaluation and management strategies. Br J Obstet Gynaecol 2005;112:32-37. 7. Temmerman M, Chomba EN, Ndinya-Achola, Plummer FA, Coppens M, Piot P. Maternal human immunodeficiency virus-1 infection and pregnancy outcome. Obstet Gynecol 1994;83:495-501. https://doi.org/10.1097/00006250-199404000-00002w 8. Kumar RM, Uduman SA, Khurrana, AK. Impact of maternal HIV-1 infection on perinatal outcome. Int J Gynecol Obstet 1995;49(2):137-143. https://doi.org/10.1016/0020-7292(95)02356-h 9. Markson LE, Turner BJ, Houchens R, Silverman NS, Cosler L, Takyi BK. Association of maternal HIV infection with low birth weight. J Acquir Immune Defic Syndr Hum Retrovirol 1996;13:227234. https://doi.org/10.1097/00042560-199611010-00004

10. Langston C, Lewis DE, Hammill HA, Popek EJ, Kozinetz CA, Kline MW. Excess intrauterine fetal demise associated with maternal human immunodeficiency virus infection. J Infect Dis 1995;172:1451-1460. https://doi.org/10.1093/infdis/172.6.1451 11. Goldberg RL, Thompson C. The infectious origins of stillbirths. Am J Obstet Gynecol 2003;189(3):861-873. https://doi.org/10.1067/s0002-9378(03)00470-8 12. Martin R, Boyer P. Incidence of premature birth and neonatal respiratory disease in infants of HIVpositive mothers. J Pediatr 1997;131:851-856. 13. Kennedy D, Fawcus S, Kroon M. The effect of maternal HIV status on perinatal outcome at Mowbray Maternity Hospital and referring midwife obstetric units, Cape Town. SA J Obstet Gynaecol 2012;18(1):6-10. doi:10.7196/sajog.417 14. Pattison RC, Hulsbergen MH, van Hoorick L. The effects of maternal HIV infection on maternal condition and perinatal death in southwest Tshwane. Facts Views Vis ObGyn 2010;2(4):227-231. 15. De Knijf A, Pattison RC. Confidential enquiries into quality of care of women in labour using hypoxic ischaemic encephalopathy as a marker. Facts Views Vis ObGyn 2010;2(4):219-225. 16. Temmerman M, Nyong’o AO, Bwayo J, Fransen K, Coppens M, Piot P. Risk factors for motherto-child transmission of human immunodeficiency virus-1 infection. AJOG 1995;172(2):700-705. https://doi.org/10.1016/0002-9378(95)90597-9 17. St Louis ME, Kamenga M, Browa C, et al. Risk of perinatal HIV-1 transmission according to maternal immunologic, virologic and placental factors. JAMA 1993;269(22):2853-2859. https://doi. org/10.1001/jama.1993.03500220039023 18. Kuhn L, Abrahams E, Matheson P, et al. Timing of maternal-infant HIV transmission: Association between intrapartum factors and early polymerase chain reaction results. AIDS 1997;11(4):429-435. https://doi.org/10.1097/00002030-199704000-00005 19. Garcia-Tejedor A, Perales A, Maiques V. Duration of rupture of membranes and extended labor are risk factors for HIV transmission. Int J Gynecol Obstet 2003;82(1):17-23. https://doi.org/10.1016/ s0020-7292(03)00123-1 20. Toumala RE, Shapiro DE, Mofenson LM, et al. Antiretroviral therapy during pregnancy and the risk of an adverse outcome. N Engl J Med 2002;346:1863-1870. https://doi.org/10.1056/nejmoa991159 21. Townsend CL, Cortina-Borja M, Peckham CS, Tookey PA. Antiretroviral therapy and premature delivery in diagnosed HIV infected women in the United Kingdom and Ireland. AIDS 2007;21(8):1019-1026. https://doi.org/10.1097/qad.0b013e328133884b 22. Braddick MR, Kreiss JK, Embree JE, et al. Impact of maternal HIV infection on obstetrical and early neonatal outcome. AIDS 1990;4(10):1001-1006. https://doi.org/10.1097/00002030-19901000000009 23. Lepage P, Dabis F, Hitimana DG, et al. Perinatal transmission of HIV-1: Lack of impact of maternal HIV infection on characteristics of livebirths and on neonatal mortality in Kigali, Rwanda. AIDS 1991;5(3):295-300. https://doi.org/10.1007/978-1-4615-2530-1_18 24. Nakanga W, Patel P, Panjwani S, Kennedy N, Kawaza K. Supra-treatment threshold neonatal jaundice: Incindence in HIV-exposed compared to non-exposed neonates at Queen Elizabeth Central Hospital in Blantyre, Malawi. Malawi Med J 2015;27(3):104-108. https://doi.org/10.4314/mmj.v27i3.7 25. Maisel MJ, Clifford K, Antle CE, Leib GR. Jaundice in the healthy newborn infant: A new approach to an old problem. Pediatrics 1988;81(4):505-511. 26. Najib KS, Saki F, Hemmati F. Incidence, risk factors and causes of sever neonatal hyperbilirubinemia in south of Iran (Fars province). Iran Red Crescent Med J 2013;15(3):260-263. https://doi. org/10.5812/ircmj.3337 27. National Department of Health, South Africa. The 2013 National Antenatal Sentinel HIV Prevalence Survey South Africa. Pretoria: NDoH, 2015:27. 28. Geetha S, Ravindra P, Rashmi M. A prospective study of obstetric and newborn outcome in a cohort of HIV-affected pregnant women. JEMDS 2015;4:514-524. https://doi.org/10.14260/jemds/2015/77 29. Patil S, Patil V. Maternal and foetal outcome in premature rupture of membranes. IOSR JDMS 2014;13(12):56-83. https://doi.org/10.9790/0853-131275683 30. Iloanusi NE. Evaluation of pregnant women admitted with prelabour rupture of membranes. June 2013, http://wiredspace.wits.ac.za/jspui/bitstream/10539/13746/3/Iloanusi%20MMed%20Final%20 Research%20Report%20Submission.pdf (accessed 4 June 2016). 31. Lewis DF, Major CA, Towers CV, et al. Effects of digital vaginal examination on latency period in preterm rupture of membranes. Obstet Gynecol 1992;80:630-634. https://doi.org/10.1016/00029378(91)91234-n 32. Kenyon S, Boulvain M, Neilson J. Antibiotics for preterm rupture of the membranes: A systematic review. Obstet Gynecol 2004;104(5 Part 1):1051-1057. https://doi.org/10.1097/01. aog.0000143268.36682.21 33. Mercer BM. Antibiotic Therapy for reduction of infant morbidity after preterm premature rupture of the membranes. JAMA 278(12):989-995. https://doi.org/10.1001/jama.1997.03550120049032 34. Osmanağaoğlu MA, Ünal S, Bozkaya H. Chorioamnionitis risk and neonatal outcome in preterm premature rupture of membranes. Arch Gynecol Obstet 2004;271(1):33-39. https://doi.org/10.1007/ s00404-004-0644-8 35. Cox SM, Leveno KJ. Intentional delivery versus expectant management in premature rupture on membranes at 30 - 34 weeks gestation. Obstet Gynaecol 1995;875:879. https://doi.org/10.1016/00297844(95)00303-9 36. Bateman C. Medical negligence pay-outs soar by 132% – subs follow. SAMJ 2011;101(4):216-218. https://wdoi.org/10.7196/samj.4881

SAJOG • September 2017, Vol. 23, No. 2

This open-access article is distributed under Creative Commons licence CC-BY-NC 4.0.

RESEARCH

The peritoneum at laparotomy: A survey of gynaecological practice among Nigerian gynaecologists A A Bamigboye,1 MB ChB, MMed, FCOG (SA), PhD; F O Okonofua,2 PhD, FAS aculty of Clinical Sciences, University of Medical Sciences, Ondo, Nigeria and School of Clinical Sciences, Department of Obstetrics and F Gynaecology, University of the Witwatersrand, Johannesburg, South Africa 2 Faculty of Clinical Sciences, University of Medical Sciences, Ondo, Nigeria 1

Corresponding author: A A Bamigboye (aabamig@gmail.com)

Background. The debate on the benefit of leaving the peritoneum unsutured is ongoing among gynaecologists. There is yet to be a final decision on this surgical step, especially considering the long-term complications of adhesion formation, subfertility and chronic pelvic pain. Objectives. The objective of the study is to determine the attitudes of Nigerian gynaecologists toward the peritoneum during laparotomy for both benign conditions and caesarean sections. Methods. The survey was conducted during the 50th International Conference of Gynaecologists hosted by the Society of Gynaecology and Obstetrics of Nigeria in October 2015. A questionnaire was distributed to attending participants across the board regarding peritoneal closure at laparotomy. A total of 125 respondents completed the survey. Data were entered into SSPA statistical software for analysis. Results. The attitude towards the peritoneum was not associated with the year of experience as a gynaecologist. The reasons for the decision not to close the peritoneum was mostly based on the little evidence there is as regards the benefits. Conclusion. The attitude towards the peritoneum during gynaecological procedures amongst Nigerian specialists appears to be in congruity with international practices where there is no clear-cut preference on the status of the peritoneum. However, the younger specialists gravitate towards non-closure. S Afr J Obstet Gynaecol 2017;23(2):48-50. DOI: 10.7196/SAJOG.2017.v23i2.1131

Surgeons in obstetrics and gynaecology have different approaches to the fate of the peritoneal membranes after they have been breached during laparotomy. Animal experiments had not shown any advantages in the wound strength when the peritoneum was sutured. If anything, there was an increase in adhesions formation when the peritoneum was surgically approximated.[1-4] In the last decade, in humans, evidence from earlier studies has shown short-term benefits with non-closure of the peritoneal surfaces[5] but recent trials have questioned the practice without looking into the long-term sequelae.[6-8] Many clinicians may have adopted a change in practice owing to extrapolations from animal studies and earlier clinical trials. However, a meta-analysis of clinical trials in a 2014 Cochrane review that included more recent reviews, questioned change in practice without looking at the long-term sequelae of non-closure. Notable advantages of non-closure of the peritoneum during caesarean section were shorter operative periods and shorter periods of hospitalisation, which effectively reduces the cost of operative procedures. The long-term complications of adhesion formation and chronic pelvic pain were not convincingly addressed in the analysed studies.[6] An American study has confirmed the advantage of parietal peritoneal closure over non-closure in the prevention of adhesions, while leaving the visceral peritoneum unsutured; there was a 3 - 5-fold reduction in adhesion formation taking into cognisance many cofounding variables such as ethnicity, medical status, and infectious morbidity.[7] Another trial showed decreased adhesion formation in the allocation group, where the parietal peritoneum and the rectus sheath were approximated.[8] Therefore the debate as to whether to close the peritoneal surfaces or not may be resurfacing.

In a previous population survey among registered South African obstetricians and gynaecologists, many reasons were provided regarding their practices toward the integrity of the peritoneum. The motivation for closing the peritoneum included anatomical restoration, prevention of wound dehiscence, reducing haemorrhage and adhesion formation.[4] This survey aimed to assess the attitudes towards and practices of peritoneal non-closure among a cohort of Nigerian obstetricians and gynaecologists attending the annual gynaecological conference marking the 50th anniversary of the society foundation in 2015.

Methods

A questionnaire was distributed among delegates attending the Society of Gynaecologists and Obstetrics of Nigeria 2015. The proforma, which was self-administered, asked about the baseline data and their attitude towards closure and non-closure of the peritoneum during caesarean section and laparotomy for benign conditions. Data were entered into Microsoft Excel and analysed using SPSS epidemiological software (IBM Corp., USA).

Results

Of the 150 questionnaires handed out to delegates during two conference plenary sessions, 125 doctors responded. The average age of the surgeons who completed the survey was (mean (standard deviation (SD)) 44.9 (8.1) years, of whom 74.7% were male, 84.7% were specialist gynaecologists, and 15.2% were registrars. Most respondents (68.8%) had 0 - 10 yearsâ&#x20AC;&#x2122; experience as specialists and 74.6% were practising at referral hospitals (Table 1). The mean (SD)

SAJOG â&#x20AC;˘ September 2017, Vol. 23, No. 2

RESEARCH constipation, and infertility. There was no difference in morbidity.[9] A collaborative study published in 2016, CORONIS, observed ~8 100 women during a 3-year follow-up and did not find any difference in the occurrence of pelvic pain, dyspareunia, bowel obstruction, bladder surgery, infertility and ectopic pregnancy, whether the peritoneum was closed with sutures or not.[11] Perhaps a longer follow-up of the CORONIS study in future might give clinicians the final answer to the fate of the peritoneum. A paradigm shift in practice cannot be succinctly advocated until the possible chronic sequelae have been assessed in studies incorporating longer follow-up periods. Our findings were in agreement with those of a similar study[4] conducted among registered South African obstetricians and gynaecologists, which showed that there was no clear preference concerning peritoneal closure. The result of a similar survey was published in 2016 among US gyneaecologists, where it was concluded that variation in surgical practice including attitude towards the peritoneum varies amongst practising gynaecologists, with no particular trend of one approach over the other.[12]

number of caesarean sections performed by the surgeons was 3.25 (1.8) per week, while mean (SD) 2.02 (1.2) laparotomies were performed weekly for a benign condition. Forty-eight percent (48.3%) stated that they would not close the peritoneum at all, or close the parietal peritoneum only when the visceral peritoneum prevented closure of the rectus sheath. The remaining specialists stated that they would close either one or both of the peritoneal membranes (Table 2). There was no association between the specialistsâ&#x20AC;&#x2122; experience and their attitudes toward the closure of the peritoneum; 41.2% of specialists would not close the peritoneum in the 0 - 10 years group v. 41.9% in the 11 - 20 years group. There was no statistically significant association between attitude towards the peritoneum and status (specialist v. registrar) or level of practice (secondary- or tertiary-level) in both procedures of hysterectomy and caesarean sections.

Discussion

The attitudinal practice of Nigerian obstetricians and gynaecologists towards the closing of the peritoneum did not differ based on the number of yearsâ&#x20AC;&#x2122; experience as a specialist. The doctors tend to gravitate toward what was advised globally, i.e. to leave the peritoneum unsutured,[10] and might indicate the influence of continuous medical education in the practice of obstetrics and gynaecology. A 2003 Cochrane review documented the short-term benefits of non-closure of the peritoneum and the National Institute for Health and Clinical Excellence, UK (NICE) guideline may have popularised non-closure.[5,10] Since the long-term complications of adhesion formation, infertility, and chronic pain is not fully ascertained, we believe that the motivation to change their practices may be for financial reasons where, following non-closure of the peritoneum, there was a reduction in theatre time and reduction in the use of suture materials and hospital stay.[5] A study involving 144 women looked into the issue of peritoneal closure v. non-closure and assessed the long-term complications of dyspareunia, adhesions,

Study limitation The study sample included delegates at SOGON 2015 only and therefore was not nationally representative.

Conclusion

The trend amongst the gynaecologists present at the SOGON regarding whether to leave or close the peritoneum has no particular preference. A multicentre trial has been conducted, and if anything, a worsened outcome with peritoneal non-closure or otherwise has not been found. A longer follow-up will be needed before a final change in practice could be advocated. However, the need to continue with medical education cannot be overemphasised as recommendations might change with time.

Table 1. Attitude of gynaecologists towards peritoneal closure during laparotomy (N=124) Years of experience as a specialist

I always close both VP and PP, n (%)

I close only the VP, n (%)

I close only the PP, n (%)

I do not close the peritoneum at all, n (%)

I close only when the omentum is in the way, n (%)

0 - 10

12 (15.3)

21 (24.7)

10 (11.8)

35 (41.2)

6 (7.1)

11 - 20

10 (32.3)

4 (12.9)

13 (41.9)

4 (12.9)

21 - 30

2 (66.7)

1 (20.0)

>30

2 (66.7)

1 (33.0)

Total

26 (24.1)

11 (10.2)

48 (44.4)

11 (10.2)

VP = visceral peritoneum; PP = parietal peritoneum.

Table 2. Attitude of obstetricians towards peritoneal closure during caeserean section (N=124) Years of experience as a specialist

I always close both VP and PP, n (%)

I close only the VP, n (%)

I close only the PP, n (%)

I do not close the peritoneum at all, n (%)

I close only when the omentum is in the way, n (%)

0 - 10

14 (16.3)

23 (25.7)

6 (7.0)

35 (40.7)

8 (9.3)

11 - 20

10 (32.3)

4 (12.9)

1 (1.16)

4 (12.9)

21 - 30

4 (80.0)

12 (38.7)

>30

2 (66.7)

1 (33.3)

Total

28 (22.2)

7 (5.65)

47 (37.3)

12 (9.52)

VP = visceral peritoneum; PP = parietal peritoneum.

SAJOG â&#x20AC;˘ September 2017, Vol. 23, No. 2

RESEARCH Acknowledgements. We acknowledge the statistical analysis done by Mr W Oshuntokun of the Registry Department of the University of Medical Sciences, Ondo and Louismed Hospital, Lagos, for sponsoring the logistics. We equally mention Prof. FO Okonofua, Vice Chancellor and an Obstetrics and Gynecology researcher of the same university for his interest in the survey. Author contributions. AB conceived the idea, planned the research, distributed and administered it, and wrote the body of the research, FO read and distributed the questionnaire and perused the draft report. Funding. LouisMed Specialist Hospital, Lekki, Lagos, Nigeria. Conflicts of interest. None. 1. Parulkar BG, Supe AN, Vora IM, Mathur SK. Effect of experimental nonclosure of peritoneum on development of suture line adhesions and wound strength in dogs. Indian J Gastroenterol 1986;5(4):251.

4. Bamigboye AA, Buchman E, Hofmeyr GJ. Closure of peritoneum at laparotomy: A survey of gynecological practice. S Afr Med J 1999;89(3):332-335. 5. Bamigboye AA, Hofmeyr GJ. Closure versus non-closure of the peritoneum at caesarean section. Cochrane Database Syst Rev 2003;4:CD000163. https://doi.org/10.1002/14651858.CD000163. 6. Bamigboye AA, Hofmeyr GJ. Closure versus non-closure of the peritoneum at caesarean section: Short- and long-term outcomes. Cochrane Database Syst Rev 2014;8:CD000163. https://doi. org/10.1002/14651858.CD000163.pub2 . 7. Lyell DJ, Caughey AB, Hu E, Daniels K. (2005). Peritoneal closure at primary cesarean delivery and adhesions. Obstet Gynecol 2005;106(2):275-280. https://doi.org/10.1097/01. AOG.0000171120.81732.4c 8. Hamel KJ. Incidence of adhesions at repeat cesarean delivery. Am J Obstet Gynecol 2007;196(5):e31e32. https://doi.org/10.1016/j.ajog.2006.09.011 9. Roset E, Boulvain M, Irion O. Nonclosure of the peritoneum during caesarean section: Long-term follow-up of a randomised controlled trial. Eur J Obstet Gynaecol Reprod Biol 2002;108:40-44. https://doi.org/10.1016/S0301-2115(02)00366-4 10. National Institute for Health and Clinical Excellence (NICE). Caesarean Section Guideline CG13. London: NICE, 2004.

2. Kapur ML, Daneshwar A, Chopra P. Evaluation of peritoneal closure at laparotomy. Am J Surg 1979;137(5):650-652. https://doi.org/10.1016/0002-9610(79)90040-0

11. The CORONIS collaborative group. Caesarean section surgical techniques: 3 year follow-up of the CORONIS fractional, factorial, unmasked, randomised controlled trial. Lancet 2016;388(10039):6272. https://doi.org/10.1016/S0140-6736(16)00204-X

3. Kyzer S, Bayer I, Turani H, Chaimoff C. The influence of peritoneal closure on the formation of intraperitoneal adhesions: An experimental study. Int J Tissue React 1986;8(5):355-359.

12. Lyell DJ, Power M, Murtough K, Ness A, et al. Surgical techniques at cesarean delivery: A US survey. Surg J 2016;2(4):e119–e125. https://doi.org/10.1055/s-0036-1594247

SAJOG • September 2017, Vol. 23, No. 2

This open-access article is distributed under Creative Commons licence CC-BY-NC 4.0.

RESEARCH

An endometrial histomorphometric study of CD56+ natural killer cells in women with unexplained infertility M Muller,1 MB ChB, MMed (Path); G Kalmeier,2 MSc; P Eyal,1 MB ChB, MMed (Path); A de Bruin,3 MB ChB, MMed; R Pool,4 MB ChB, MMed (Haem); C du Rant,5 PhD; R Ehlers,6 PhD; A Stander,2 PhD; A van Schoor,7 PhD; E Nortje,2 MSc; P du Toit,2 PhD Department of Anatomical Pathology, Faculty of Health Sciences, University of Pretoria, South Africa Department of Physiology, Faculty of Health Sciences, University of Pretoria, South Africa 3 Endocrine and Reproductive Unit, Faculty of Health Sciences, University of Pretoria, South Africa 4 Department of Haematology, Faculty of Health Sciences, University of Pretoria, South Africa 5 Department of Immunology, Faculty of Health Sciences, University of Pretoria, South Africa 6 Department of Statistics, Faculty of Natural and Agricultural Sciences, University of Pretoria, South Africa 7 Department of Anatomy, Faculty of Health Sciences, University of Pretoria, South Africa 1 2

Corresponding author: M Muller (drmarissa.muller@gmail.com)

Background. The number of peripheral blood and endometrial natural killer cells varies greatly during implantation and the first trimester of pregnancy and is thought to play a role in the maintenance of a healthy pregnancy. However, the role of endometrial CD56+ natural killer (NK) cells as an immunological mechanism in unexplained infertility is yet unknown. Objectives. The study aimed to enumerate the concentrations of CD56+ NK cells in endometrial samples, and to statistically compare these numbers between fertile and infertile women. Methods. A histomorphometric analysis was conducted using haematoxylin and eosin staining and an immunohistochemical approach was used for quantifying cell numbers. Results. Fifty samples were collected in equal parts between a study group of infertile female subjects (mean (standard deviation) age 35 (4), range 26 - 42 years) and a control group of multiparous fertile individuals (mean (SD) age 43.4 (6.3), range 30 - 55). The mean number of CD56+ NK cells present at different depths for both the study and control groups did not differ significantly. Age and group (study or control) were not significantly related to the mean number of CD56+ NK cells. However, for the late secretory phase the mean number of CD56+ NK cells was significantly higher than for the early phase. Conclusion. Our findings could not identify a statistically significant correlation between the number of CD56+ NK cells and infertility. S Afr J Obstet Gynaecol 2017;23(2):51-55. DOI:10.7196/SAJOG.2017.v23i2.1145

The standard clinical definition for infertility, as defined in the International Committee for Monitoring Assisted Reproductive Technology (ICMART) and the World Health Organization (WHO) glossary (2009), is ‘a disease of the reproductive system defined by the failure to achieve a clinical pregnancy after 12 months or more of regular unprotected sexual intercourse’.[1] Terms often used synonymously with infertility include: • Reproductive failure – describing the inability to conceive or maintain pregnancy, due to either recurrent miscarriages, infertility or repeated implantation failure (in vitro fertilisation).[2] • Repeated implantation failure – described as the failure to achieve pregnancy following 2 - 6 in vitro fertilisation cycles.[3] • Recurrent miscarriages – described as the loss of ≥3 consecutive pregnancies before gestational age (20 - 22 weeks).[4]

Hull et al.[5] first described infertility as a failure to conceive over a period of 3 years of unprotected intercourse, while others report non-conception after 1 year of unprotected intercourse during the fertile phase as sufficient to diagnose infertility.[6,7] More recently, infertility has been described as the failure to conceive after 6 cycles of unprotected intercourse, regardless

of age, while the National Institute for Health and Clinical Excellence has rendered a definition that failure to conceive after 2 years signifies infertility.[8] In accordance, it is reported that 92% of the general population shows successful conception after 2 years (84% after 1 year).[9] Prognostic factors for higher cumulative pregnancy rates include: female age <30 years; previous pregnancy in the same relationship; and <2 years, infertility. The prognosis worsens when the duration of infertility exceeds 3 years and the female partner is older than 35 years. Pregnancy rates are reported to decline by 9% each year beyond the age of 30 years.[9] According to the American Society of Reproductive Medicine, standard infertility evaluations should include semen analysis, post-coital test and assessment of ovulation, hysterosalpingogram, and sometimes laparoscopy.[10] In agreement, the Practice Committee Bulletin on Unexplained Infertility suggests that basic evaluations should provide evidence of ovulation, adequate sperm production, and patency of the Fallopian tubes; and that unexplained infertility can only be diagnosed if these results are normal.[11] Unexplained infertility is unfortunately often misdiagnosed due to the lack of dedicated testing measures, while endometriosis, mild degrees of

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH tubal infertility, and poor ovarian reserve are considered the most frequent reasons for the misdiagnosis of unexplained infertility.[10]

CD56+ as diagnostic investigation During implantation and the maintenance of pregnancy, there exists a close interaction between the endocrine and immune system. Under influence of sex steroids, there is a marked increase of a unique population of endometrial natural killer (NK) cells and lymphocytes, derived predominantly from a subset of peripheral blood NK cells and subsequently recruited to the uterus. These immune cells are believed to promote placental and trophoblast growth, provide immunomodulation of maternal-fetal interaction and have a presumptive role in the maintenance of healthy pregnancy.[3] The normal, healthy endometrium contains a variety of haematopoietic cells, including granulocytes like CD56+ NK cells, and its composition varies depending on the stage in the menstrual cycle and menopausal status. CD56+ NK cells have a primary role in the innate immune responses against viruses and transformed cells, and is thought to possibly be a critical mechanism in preventing maternal rejection of the fetus. It is thought that such immunological mechanisms influence reproductive failure, as successful pregnancy involves maternal immunity adaptation to the semi-allogeneic developing embryo.[13,14] It is known that peripheral blood NK cells are phenotypically and functionally different from endometrial NK cells, with less than 10% resembling endometrial NK cells and differing in surface antigen expression.[15] Endometrial NK cells have little cytotoxic activity, but are a rich source of cytokines, particularly angiogenic ones, which are possibly involved in tissue remodelling during the formation of the placenta and the regulation of trophoblast invasion and angiogenesis.[14,16,17] Granulated lymphocytes (including CD56+ NK cells) are present in large numbers in pre-decidualised endometrial stroma in the mid- and late-secretory phases. There are two theories to explain these large numbers: â&#x20AC;˘ The recruitment of CD56+ NK cells from peripheral CD56+ NK cells which differentiate in the uterine microenvironment into the endometrial phenotype. â&#x20AC;˘ That endometrial CD56+ NK cells stem from the proliferation and differentiation of stem cells in the endometrium.[12]

An immunohistochemical analysis in the non-pregnant endometrium shows increases of endometrial CD56+ NK cells during the secretory phase (alleged time of implantation).[12] These numbers remain high during early pregnancy, constituting 70% of the T lymphocytes at the interface between maternal decidua and the invading trophoblast and are the most predominant leucocyte population during the time of implantation and early pregnancy.[3,4] In contrast, 30% of the endometrial T lymphocytes were CD56+, as opposed to the peripheral blood (5 - 15%). This percentage remains constant during the menstrual cycle in both the proliferative and the secretory phases.[3,4] Successful pregnancy outcomes have been reported after intravenous immunoglobulin G (IgG) therapy in patients with recurrent pregnancy losses. Intravenous IgG therapy is shown to down-regulate elevated circulating peripheral blood CD56+ NK cells, suggesting an association between increased CD56+ NK cell numbers and unexplained infertility.[18] Research on the association of increased peripheral or endometrial CD56+ NK cells and infertility, however, remains ambiguous.

Research problem and objective The role of endometrial CD56+ NK cells in the immunological mechanism of unexplained infertility is yet unknown. Quantification of endometrial CD56+ NK cells, or the comparative percentages of CD56+ NK cells in the circulating peripheral blood during the secretory phase of individuals with unexplained infertility have not been documented. It is also unknown whether these values differ in infertile and multiparous fertile individuals. Determining these values could help to determine the role played by endometrial CD56+ NK cells in the immunology of unexplained infertility, and the aim of the study was therefore to enumerate and statistically compare endometrial CD56+ NK cell numbers in both infertile and control patients. The study only evaluated endometrial CD56+ NK cell numbers; peripheral blood levels were reserved for future research. The study was approved by the Faculty of Health Sciences Research Ethics Committee, University of Pretoria, and the National Health Research Ethics Council (ref. no. 8/2013).

Methods

Volunteering women with a history of infertility were recruited from the Endocrine and Reproductive units of three Gauteng-based hospitals in South Africa: Steve Biko Academic Hospital, Kloof Hospital and Kalafong Hospital. All women had normal menstrual cycles, normal ovarian and pelvic ultrasound examinations, normal hysterosalpingographies, normal mid-luteal progesterone levels, and their partners showed normal semen analysis (Table 1). The control group consisted of gynaecology outpatients with no associations with infertility. Anonymous samples were used for this histomorphometric, cross-sectional analysis and all participants signed informed consent.

Endometrial biopsies Endometrial biopsies were performed by clinician gynaecologists as an outpatient procedure using an endometrial sampler during the secretory phase of the menstrual cycle (Z-sampler, Bioteque America Inc., USA). Specimens were placed in 10% buffered formalin and study numbers were allocated to each sample.

Haematoxylin and eosin staining The first evaluation step was done on haematoxylin and eosin (H&E)-stained slides under light microscopy to verify the histological representation of the endometrium, and to exclude inflammation, hyperplasia, atypia or evidence of malignancy. Samples were categorised into three phases: Early secretory phase: presence of tubular-shaped glands; subnuclear vacuoles in more than 50% of the epithelial cells, involving more than 50% of glands; presence of mitotic activity in the glandular cells (Fig. 1). Mid-secretory phase: presence of angulated glands; supranuclear vacuoles in the epithelial cells; no mitotic activity; increased glandular secretions and progressively increased stromal oedema. Presence of spiral arterioles and eosinophilic cytoplasm in the stromal cells. Late-secretory phase: presence of closely packed serrated-shaped glands; pre-decidua changes around spiral arterioles forming a compact layer; presence of mitotic activity in the predecidual cells; apoptotic activity within the glands; fibrin thrombi in the small vessels; extravasation of erythrocytes into the stroma (Fig. 2).[12]

SAJOG â&#x20AC;˘ September 2017, Vol. 23, No. 2

RESEARCH Table 1. The selection criteria for subject recruitment used by gynaecologists collecting the endometrial samples Inclusion criteria

Exclusion criteria

Childbearing age

Advanced age (>35 years)

Normal menstrual cycles

History of previous ovarian surgery

Normal ovarian examination on ultrasound

Polycystic ovarian syndrome

Normal transvaginal ultrasound

Hyperprolactinaemia

Normal hysterosalpingography

Hypothalamic causes of infertility secondary to abnormal weight changes

Normal sonohysterography/ diagnostic office hysteroscopy Normal pelvic examination on ultrasound

Fig. 1. Early secretory phase with subnuclear vacuoles.

Normal mid-luteal progesterone/ Luteinising hormone levels Normal semen analysis of partner

Thyroid disease

Endometriosis Reversible adhesive tubal disease Uterine leiomyoma Uterine polyps Asherman’s syndrome (uterine synechiae) Congenital uterine cavity anomalies

Additional inclusion criteria for control group

Male partner infertility

Multiparous (≥2 full-term pregnancies)

Fig. 2. Late secretory exhaustion and predecidual changes.

Immunohistochemistry A second step of evaluation was done through the immunohistochemistry of each sample. Two sections (4 µm) were cut off all the specimens with a standard microtome and dried overnight at 37˚C. Sections were manually dewaxed in xylene twice for 5 minutes and brought to distilled water before being treated with 0.38% EDTA and incubated for 30 minutes. The sections were washed in distilled water and treated with 3% H2O2 solution for 10 minutes, washed again and placed in buffer wash for 5 minutes. All sections were covered with primary mouse anti-human NK cell, CD56+, monoclonal antibodies (Dako, Denmark) at a dilution of 1:100 for 30 minutes. The slides were washed in a buffer and dried, covered with FLEX detection system (Agilent Technologies Inc., USA) sera and incubated for 30 minutes at room temperature. Samples were washed again in a buffer wash solution and incubated with 3,3-diaminobenzidine tetrahydrochloride twice for 3 minutes at room temperature. After being washed again in tap water, samples were counterstained in haematoxylin (2 min) and eosin blue (5 min). The slides were dehydrated and mounted and adequate positive controls (lymphnode tissue) were placed on each slide. This morphometric evaluation was performed using transmitted light microscopy and an eyepiece to a magnification of ×400. The

magnification of ×400 was defined as the high-powered field (HPF) used in this study. The number of positively staining cells in ten, randomly selected, non-overlapping HPFs were counted and labelled as level 1. The same procedure was repeated from the same subject’s endometrial sample in a second (4 µm deeper) section and was labelled as level 2. A paired t-test was used to compare the mean number of CD56+ NK cells of the two different tissue sections (level 1 and level 2) for both study and control groups. A linear model was fitted to test for a relationship between the mean number of CD56+ NK cells and the covariates group (study or control), age and phase (early, mid or late). For covariates that were significant, where applicable, post-hoc comparisons of least-squares means of CD56+ NK were performed with a Bonferroni correction. After omission of 3 samples deemed to fall outside the inclusion criteria, the 53 original samples were reduced to 25 samples in both the study group (mean (standard deviation (SD)) age of 35.0 (4.0), range 26 - 42 years) and control group (mean (SD) age 43.4 (6.3), range 30 - 55). None of the participants in the study group had a history of live birth and had been diagnosed as infertile. Infertility specialist teams of the respective gynaecology units were used. In the control group, the median (range) number of successful pregnancies was 3 (3 - 7).

Results

H&E results

The paired t-test performed showed that for both the study and control groups there was not a significant difference between

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH Table 2. Mean number of endometrial CD56+ NK cells per 10 HPFs in the three stages of the secretory phase of the total population CD56+ cells in level 1 and 2,

Age of participants,

Stage of secretory phase

mean (SD)

mean (SE)

Early

16.17 (5.22)

38.07 (1.73)

Mid

30.88 (4.75)

39.00 (1.27)

Late

37.16 (4.91)

40.41 (1.96)

NK = natural killer; HPF = high-powered field; SD = standard deviation; SE = standard error.

Table 3. Mean numbers of endometrial CD56 NK cells per HPF in both the study and control group* +

Group

Control

Study

N Obs

Variable

Median

Mean (SD)

Min

Max

l1_hpf1

17.00

30.16 (29.01)

1.00

120.00

l1_hpf2

15.00

23.16 (21.08)

3.00

100.00

l1_hpf3

20.00

27.68 (28.29)

2.00

124.00

l1_hpf4

20.00

23.72 (16.72)

2.00

62.00

l1_hpf5

20.00

25.08 (18.62)

6.00

78.00

l1_hpf6

19.00

24.52 (17.01)

5.00

72.00

l1_hpf7

20.00

25.32 (17.58)

3.00

59.00

l1_hpf8

20.00

26.60 (20.87)

6.00

77.00

l1_hpf9

16.00

24.36 (17.05)

2.00

64.00

l1_hpf10

20.00

24.28 (17.45)

0.00

60.00

AvgL1

18.70

25.49 (18.18)

4.80

72.10

l1_hpf1

24.00

32.16 (28.61)

1.00

102.00

l1_hpf2

15.00

29.52 (35.94)

0.00

163.00

l1_hpf3

20.00

29.20 (25.68)

0.00

96.00

l1_hpf4

17.00

29.48 (32.87)

0.00

110.00

l1_hpf5

22.00

36.80 (39.03)

1.00

151.00

l1_hpf6

20.00

32.88 (33.69)

1.00

139.00

l1_hpf7

21.00

26.08 (22.18)

0.00

92.00

l1_hpf8

23.00

31.76 (28.77)

2.00

115.00

l1_hpf9

23.00

34.76 (29.70)

0.00

104.00

l1_hpf10

22.00

27.04 (22.07)

0.00

83.00

AvgL1

21.40

30.97 (26.95)

2.00

112.60

NK = natural killer; HPF = high-powered field; SD = standard deviation. *The magnification of ×400 was defined as the high-powered field (HPF) used in this study.

the mean CD56+ NK cells measured at level 1 and level 2 with p=0.6636 and p=0.8836, respectively. The results for the two levels were combined in further analysis. In the linear model with mean CD56+ NK cells as dependant variable and group, age and phase as covariates, the results indicated a significant difference (p=0.0173) in the mean number of endometrial CD56+ NK cells/10 HPFs between the phases. In the post-hoc test least-square means of the mean CD56+ NK cells were compared for the different phases making a Bonferroni correction. There was a significant difference in the mean CD56+ NK cells between early and late secretory stages, (p=0.0165 after the Bonferroni correction) (Table 2). Controlling for group and phase, there was no significant relationship between age and mean CD56+ NK cells (p=0.2171). There was also no significant difference in the mean CD56+ NK cells for the 2 groups when keeping age and

phase constant (p=0.8476; study and control mean (SD) values were 28.77 (4.65) and 27.36 (4.63), respectively.

Histomorphometric results From the histomorphometric analysis, each sample was evaluated as two separate slide sections (level 1 and 2) and the number of endometrial CD56+ cells/HPF for each level was counted (Table 3). The results from the analysis are given above; there was no statistically significant difference in the number of CD56+ NK cells between levels and groups.

Discussion and conclusion

This cross-sectional histomorphometric study was conducted to establish the potential role of endometrial CD56+ NK cells as an immunological mechanism in unexplained infertility. The study revealed no statistically significant associations between the numbers of CD56+ NK cells present in the endometrium compared with age, number of successful pregnancies, mean number of cells per HPF, or between accumulated cell numbers per 10 HPFs in either of the groups. However, the only positive statistical association in our results was the difference in number of CD56+ endometrial NK cells found between the early and late secretory phase (Table 2).The significance of this finding warrants further investigation, but in agreement, literature does show variation in endometrial NK cells numbers throughout the menstrual cycle, with a dramatic increase between days 6 - 7 after the luteinising hormone (LH) surge.[15] This is the putative time of implantation and the number of endometrial CD56+ NK cells have been shown to remain high during early pregnancy, encompassing 70% of the endometrial leucocytes in the first trimester before steadily declining and being absent at term.[15] Despite NK cells being the most predominant leukocyte population during the time of implantation and early pregnancy, the precise role of endometrial NK cells and their relative contribution of cytokine secretion or cytotoxicity in implantation and maintenance of a successful pregnancy remain unfounded.[3,4,13] Other research still shows associations with increased numbers of endometrial CD56+ NK cells and infertility, while women with reproductive failure are often treated with steroids, intravenous immunoglobulin, and tumour necrosis factor-alpha (TNF-α) blocking drugs, that aim to suppress CD56+ NK cells. However, a review of various forms of immunotherapies also did not show significant differences between treatment and control groups.[19] In essence, the hypothesis that endometrial NK cells might play an immunological role in unexplained infertility could not be supported or denied by our findings, but more conclusive evidence for a causative role for NK cells in unexplained infertility could be yielded if the testing of secretory-phase endometrial cell numbers could be standardised.[4,20] At present there is no agreed method for assessing endometrial NK cell numbers, and currently adopted methods vary greatly in the mean numbers of endometrial NK cells present, the percentage of endometrial NK cells versus that of stromal cells, or versus CD45+ cells.[6,15] On our part, follow-up studies of peripheral blood v. endometrial NK cell population numbers would help to further this investigation. Acknowledgements. None. Author contributions. MM, project leader; PE, supervisor; AdB, RP, co-supervisors; CdR, RE, GK, AS, EN, AvS, PdT, collaborators. Funding. Discovery Foundation Research Awards. Conflicts of interest. None.

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH 1. Zegers-Hochschild F, Adamson GD, De Mouzon J, et al. The International Committee for Monitoring Assisted Reproductive Technology (ICMART) and the World Health Organization (WHO) revised glossary on art terminology, 2009. Hum Reprod 2009;24(11):2683-2687. https:// doi.org/10.1093/humrep/dep343 2. Quaas A, Dokras A. Diagnosis and treatment of unexplained infertility. Rev Obstet Gynecol 2008;1(2):69-76. 3. Dosiou C, Giudice LC. Natural killer cells in pregnancy and recurrent pregnancy loss: endocrine and immunologic perspectives. Endocrine Rev 2005;26(1):44-62. https://doi.org/10.1210/er.20030021 4. Tang A, Alfirevic Z, Quenby S. Natural killer cells and pregnancy outcomes in women with recurrent miscarriage and infertility: A systematic review. Hum Reprod 2011;26(8):1971-1980. https://doi.org/10.1093/humrep/der164 5. Hull M, Glazener C, Kelly N, et al. Population study of causes, treatment, and outcome of infertility. Br Med J (Clin Res Ed) 1985;291(6510):1693-1697. https://doi.org/10.1136/bmj.291.6510.1693 6. Collins J, Crosignani P. Unexplained infertility: A review of diagnosis, prognosis, treatment efficacy and management. Int J Ostet Gynaecol 1992;39(4):267-275. https://doi.org/10.1016/00207292(92)90257-j 7. Evers J. Female subfertility. Lancet 2002;360(9327):151-159. https://doi.org/10.1016/s01406736(02)09417-5 8. Gnoth C, Godehardt D, Godehardt E, Frank-Hermann P, Freundl G. Time to pregnancy: Results of the German prospective study and impact on the management of infertility. Hum Reprod 2003;18(9):1959-1966. https://doi.org/10.1093/humrep/deg366 9. National Collaborating Centre for Women's and Children's Health. Clinical Guideline 11, fertility: Assessment and treatment for people with fertility problems, commissioned by the National Institute for Clinical Excellence (NICE). 2004. 10. American Fertility Society. Investigation of the Infertile Couple. Birmingham: American Fertility Society, 1992.

11. Practice Committee of the American Society for Reproductive Medicine. Effectiveness and treatment for unexplained infertility. Fertil Steril 2006;86(5):S111-S114. https://doi.org/10.1016/j. fertnstert.2006.07.1475 12. McCluggage W. Benign diseases of the endometrium. In: Kurman R, Ellenson L, Ronnett B, eds. Blaustein's Pathology of the Female Genital Tract. New York: Springer Science Business Media 2011:314-315. https://doi.org/10.1007/978-1-4419-0489-8 13. Yokoyama WM, Kim S. How do natural killer cells find self to achieve tolerance? Immunity 2006;24(3):249-257. https://doi.org/10.1016/j.immuni.2006.03.006 14. McCrath E, Ryan EJ, Lynch L, et al. Changes in endometrial natural killer cell expression of CD94, CD158a and CD158b are associated with infertility. Am J Reprod Immunol 2009;61(4):265-276. https://doi.org/10.1111/j.1600-0897.2009.00688.x 15. Tuckerman E, Laird SM, Prakash A, Li TC. Prognostic value of the measurement of uterine natural killer cells in the endometrium of women with recurrent miscarriage. Hum Reprod 2007;22(8): 2208-2213. https://doi.org/10.1093/humrep/dem141 16. Manaster I, Mandelboim O. The unique properties of human NK cells in the uterine mucosa. Placenta 2008;29:60-66. https://doi.org/10.1016/j.placenta.2007.10.006 17. Hanna J, Goldman-Wohl D, Hamani Y, et al. Decidual NK cells regulate key developmental processes at the human fetal-maternal interface. Nature Med 2006;12(9):1065-1074. https://doi. org/10.1038/nm1452 18. Clifford K, Flanagan AM, Regan L. Endometrial CD56+ natural killer cells in women with recurrent miscarriages: A histomorphometric study. Hum Reprod 1999;14(11):2727-2730. https://doi. org/10.1093/humrep/14.11.2727 19. Scott JR. Immunotherapy for recurrent miscarriage. The Cochrane Libr 2003;20(1). https://doi. org/10.1002/14651858.cd000112 20. Drury A, Nik H, van Oppernaaij R, Tang A, Turner M, Quenby S. Endometrial cell counts in recurrent miscarriage: A comparison of counting methods. Histopathology 2011;59(6):1156-1162.

SAJOG â&#x20AC;˘ September 2017, Vol. 23, No. 2

This open-access article is distributed under Creative Commons licence CC-BY-NC 4.0.

RESEARCH

Factors associated with women’s intention to request caesarean delivery in Dar es Salaam, Tanzania C G Misaeli,1 MD, MMed; B A Kamala,2,3 MD, MPH; A H Mgaya,2,4 MD, MMed, PhD; H L Kidanto,2,4 MD, MMed, PhD Department of Obstetrics and Gynaecology, Muhimbili University of Health and Allied Science, Dar es Salaam, Tanzania Department of Obstetrics and Gynaecology, Muhimbili National Hospital, Dar es Salaam, Tanzania 3 Department of Health Science, University of Stavanger, Stavanger, Norway 4 Department of Women’s and Children’s Health/International Maternal and Child Health, Uppsala University, Sweden 1

Corresponding author: B A Kamala (kamala8086@gmail.com)

Background. In the past decade, the rate of caesarean section (CS) has increased dramatically in many parts of the world. At Muhimbili National Hospital (MNH) there has been a dramatic rise in the caesarean section rate over the past decade. Objective. To determine the incidence of maternal request for CS and factors associated with intention to request caesarean section at the MNH antenatal clinic. Methods. We conducted a cross-sectional study from August to October 2014. A structured questionnaire gathered participants’ background and obstetric information, perceptions and opinions regarding a request for caesarean section, and the respective reasons for the request. Confidence intervals were calculated and a p-value <0.05 was considered significant. Results. The incidence of CS on maternal request was about 6%. The intention to request for CS in the index pregnancy was 8%. Higher-level education and formal-sector employment had higher odds for requesting CS (p=0.01 and p=0.05, respectively). Half of the participants agreed that maternal request for CS should be allowed; more private patients agreed that it could affect the doctor-patient relationship (p=0.02); more private patients agreed that request for CS was due to fear of losing a child (p=0.03). Previous history of CS was an independent predictor of maternal request for caesarean section (OR 1.7; 95% CI 1.7 - 15.4) and (OR 5.8; 95% CI 1.6 - 20.1), respectively. Conclusion. Maternal requests for CS exist at the national referral hospital in Tanzania. This was associated with factors other than women’s preferences, including perceived fear of child loss and events associated with previous CS. S Afr J Obstet Gynaecol 2017;23(2):56-62. DOI: 10.7196/SAJOG.2017.v23i2.1158

Caesarean section (CS) on maternal request implies patient choice for caesarean delivery, or CS on demand without maternal or fetal indications.[1] The concept of CS on maternal request is not well-defined as a clinical entity despite progress in establishing policy guidance for the procedure for CS on maternal request.[2] Healthcare providers may be uncertain on how to respond to maternal request for CS. In Tanzania, the rates of maternal requests for CS have not been published.[3] In 2014, the birth registry at the national referral hospital, Muhimbili National Hospital (MNH), recorded the highest rate of CS, at 51%, compared with other public hospitals. The high rate of CS at MNH was associated with low-risk Robson groups, doubtful CS indications, and increased likelihood of performing CS on request among private patients performed at that hospital compared with public patients.[4,5] According to the Tanzanian Ministry of Health guidelines, CSs are commonly performed for saving the lives of the mothers and newborns. A study at MNH showed that the total risk of ‘near-miss’ events associated with CS procedures was 3 - 7 per 1 000 operations.[6] The risk of CS has been closely associated with unsafe anaesthesia, poor preoperative preparations and delayed interventions due to limited resources.[7] In some settings, limited access to CS has been shown to contribute to severe maternal morbidity, such as postpartum haemorrhage, uterine rupture, puerperal sepsis, genital fistula and maternal death.[8–12] Limited access to CS also imposes a risk of intrapartum asphyxia with subsequent neonatal

neurological damage, and perinatal death.[13] Some documented literature reports that patients’ autonomy in healthcare includes maternal request for CS, even in the presence of the considerable risk associated with CS and inequity in maternal and newborn healthcare.[14,15] Women’s perceptions and their involvement in decision-making regarding CS have been used to draw conclusions relating to women’s requests for elective CS without medical grounds.[16] Most women at MNH receive counselling around the decision to perform CS during admission for labour or when abnormal labour is detected.[17] However, the antenatal care (ANC) clinic provides a better environment with reduced anxiety for making an informed choice.[18] The strategies necessary for reducing maternal and perinatal mortality include adequate antenatal care and appropriate caesarean intervention as part of Comprehensive Emergency Obstetric Care (CEoC).[19,20] At MNH, there are no clear guidelines relating to indications for CS. This has led to subjective indications for CS, including prior recurrent fetal loss, history of infertility and in vitro fertilisation, meconium-stained liquor, and non-assuring fetal heart traces.[17] Also, fear and blame among care providers in case of poor outcome, poorly conducted perinatal audits, maternal perceived fear of birth trauma, and loss of the baby during childbirth are possible reasons for performing unnecessary CS; this may include maternal requests for CS.[21] Interventions to reduce CS rates include fetal monitoring using fetal doppler, training on partography, correct

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH audit meetings, engaging mothers in the decision-making process of mode of delivery during antenatal care. Healthcare providers’ unawareness of women’s opinions of CS on maternal request hinders efforts to reduce unnecessary CS. This study aimed to determine the women’s perceptions and intentions for CS on maternal request and the factors associated with that intent.

Definition of variables

Methods

Ethical clearance

Study design and setting

CS on maternal request was defined as a caesarean delivery for a singleton pregnancy on maternal request after 37 completed weeks of gestation in the absence of any medical or obstetric indications. ‘Perception of CS on request’ referred to participants’ opinions on cesarean section on maternal request.

This cross-sectional study was conducted at MNH antenatal care clinic from August to October 2014. MNH is the largest referral hospital in Tanzania, located in the city of Dar es Salaam. Patients who attend ANC at MNH include referrals cases from public health facilities and patients who come directly from private facilities or from home as private patients under intramural private practice management (IPPM).

The Muhimbili University of Health and Allied Sciences Research Ethics and Publication committee reviewed the study proposal and granted ethical approval. Permission to conduct the study was obtained from the Executive Director of MNH (ref. no. HD/ MUH/T.130/2012). Informed consent was obtained voluntarily from the participants, who were informed about the objectives of the study, and were assured of confidentiality and that their names would not be used for the purpose of identification.

Sample size and sampling technique

Results

The sample size was calculated using OpenEpi version 3 (EpiData Association, Denmark), a software program for population surveys, with the assumption that the anticipated proportion for the desired outcome was 50%, with a precision of 5%, power of 80%, and a design effect of 1. The minimum required sample size was 384 cases. After the daily health education session, the investigator and research nurse informed all antenatal clinic attendees about the study. On a daily basis, ANC cards of all women who attended the clinic were collected and listed in order to create a sampling frame that was used to assess the eligibility of the study participants. Every fifth card was selected during the day’s clinic registration process and the card-holder was identified. Women with communication difficulties were excluded. Also, women with either two previous scars or one previous scar with a percieved recurrent indication for CS, such as cephalopelvic disproportion, were excluded. Out of 462 eligible participants, 440 agreed to participate in the study, yielding a response rate of 95%.

Data collection and research tools

Of the 440 participants who were interviewed, more than half (57%, n=250) were private patients, as shown in Table 1. Both public and private patients had a mean (SD) age of nearly 30 (5) years. There were more public patients in the category of those participants aged less than 25 years compared with other categories (p=0.07). Most of the patients (90%) had been married at least once in their lifetime and there was no significant difference in marital status between public and private patients. Private patients were more educated compared with public patients (61% v. 25% for tertiary education; p=0.01). Similarly, there was a lower proportion of public patients who were employed in the formal sector compared with the private patients (63.5% v. 28.3%; p<0.001). Table 1. Sociodemographic characteristics of private and public category of pregnant women in the study cohort

Characteristics

Total (N=440),

Private (n=249),

Public (n=191),

n (%)

84 (19.1)

41 (16.5)

43 (22.5)

p-value

Age (years)

Data were collected using a structured questionnaire adapted from previous surveys and comprising four sections.[22,20] The first section contained participants’ background information. The second section included fertility history, pregnancy history and mode of deliveries, the associated pregnancy outcome, and feedback on previous pregnancies. The third section contained questions about willingness to request CS and the fourth section contained 12 questions addressing the women’s perceptions of CS on maternal request. The questionnaire was translated from English into Swahili and pretested to ensure accuracy and appropriateness of the questions and responses; an obstetrician reviewed the document. Data were collected by the principal investigator and two trained research assistants.

Data analysis SPSS version 19 (IBM Corp., USA) was used to perform data entry and cleaning. Descriptive statistics for sociodemographic, past obstetric history, and previous delivery experiences and perceptions were calculated. Pearson’s χ2, Fisher’s exact test and t-tests were performed to compare the association of predictor variables with history and desire to request CS without medical grounds. Factors that were significant at p<0.05 were analysed by logistic regression.

≤ 25 26 - 30

172 (39.1)

98 (39.4)

74 (38.7)

31 - 34

94 (21.4)

63 (25.3)

31 (16.2)

≥35

90 (20.5)

47 (18.9)

43 (22.5)

29.8 (5.01)

29.9 (4.87)

29.7 (5.19)

Single

45 (10.3)

22 (8.8)

23 (12.0)

Ever married

395 (89.7)

227 (91.2)

168 (88.0)

Age (years), mean (SD)

0.07

Marital status 0.27

Education level Primary school or less

95 (21.6)

31 (12.4)

64 (33.5)

Secondary school

144 (32.7)

65 (26.2)

79 (41.4)

College/university

201 (45.7)

153 (61.4)

48 (25.1)

0.01

Occupation Employed

216 (49.1)

158 (63.5)

58 (28.5)

Petty trader

127 (28.9)

54 (21.7)

73 (39.2)

Other

97 (22.0)

37 (14.8)

60 (32.3)

SD = standard deviation.

SAJOG • September 2017, Vol. 23, No. 2

0.01

RESEARCH When assessing the mode of delivery in the last pregnancy, the rate of previous CS was 42%, of which 14% were elective and 28% were emergency CS (Table 2). The rate of previous stillbirths was as high as 15%, and higher among private patients (19%) compared with their public counterparts (12%). The proportions of live births, stillbirths and miscarriages were comparable between the private and public groups (p=0.35). Twelve percent of the studied group had a history of infertility, and 38% of the patients were primiparas. The mean (SD) age at first pregnancy was 25 (4.6) years, with private clients being significantly older than their public counterparts (t (435) = 4.58; p=0.001). Different background characteristics were analysed for an association with maternal request for CS in the previous pregnancy (Table 3). Patients with secondary school education or higher were more likely to request CS than those with primary education or no formal education (7.7% v. 1.4%; p=0.05). Similarly, patients who had been employed in the formal sector (9.2%) were more likely to request CS than informal traders (2.1%) and others (3.5%) (p=0.005). Other background characteristics had no significant association with maternal request for CS (all p≥0.17). The perception of maternal request for CS based on ‘fear of losing a child on normal delivery’ was more prevalent among private patients compared with their public counterparts (51% v. 38%; p=0.03) (Table 4). More private patients also agreed that requesting to deliver by caesarean section could affect the doctorpatient relationship, compared with their public counterparts (44% Table 2. Comparison of past obstetric history between private and public pregnant women in the study cohort

Obstetric history

Total (N=440),

Private (n=249),

Public (n=191),

n (%)

Discussion

This study revealed that the proportion of women who had a history of requesting CS was 6% and those intending to request CS was 8%. The main reasons for requesting CS were fear of losing a baby and a history of previous CS. Even though the proportion requesting CS seems small, it is highly likely to increase, based on the trend of increasing numbers of CS birth at MNH. In this hospital, CS births Table 3. Comparison of maternal characteristics with the history of CS on request among pregnant women in study cohort Ever requested to deliver by CS?

p-value

Mode of delivery of the last pregnancy*

Characteristic

NVD

145 (57.8)

71 (60.7)

74 (55.2)

Elective CS

35 (13.9)

14 (12.0)

21 (15.7)

Emergency CS

71 (28.3)

32 (27.3

39 (29.1)

Yes, n (%)*

No, n (%)*

Public

4 (3.2)

122 (96.8)

Private (IPPM)

11(7.5)

136 (92.5)

Primary or less

1 (1.4)

76 (98.6)

Secondary and above

14 (7.7)

182 (92.3)

<25

2 (8.3)

22 (91.7)

26 - 30

6 (5.8)

97 (94.2)

31 - 34

6 (5.9)

95 (94.1)

>35

1 (2.2)

44 (97.8)

Employed

11 (9.2)

108 (90.8)

Informal trader

2 (2.1)

94 (97.9)

Others

2 (3.4)

56 (96.6)

Primiparous

10 (6.9)

135 (93.1)

Multiparous

5 (3.9)

123 (96.1)

0.279

ge at first pregnancy A (years), mean (SD)

23.7 (4.0)

24 (4.2)

0.76

p-value

Type of clinic 0.61

0.18

Education level

Outcome of last pregnancy† Live birth

206 (69.8)

91 (65.4)

115 (73.7)

Macerated still birth

23 (7.8)

13 (9.4)

10 (6.4)

Fresh still birth

22 (7.5)

13 (9.4)

9 (5.8)

Miscarriage

44 (14.9)

22 (15.8)

22 (14.1)

Yes

54 (12.3)

25 (13.1)

29 (11.6)

386 (87.7)

166 (86.9)

220 (88.4)

Nulliparous

167 (38)

102 (41.0)

65 (34.0)

Primiparous

145 (33)

87 (34.9)

58 (30.4)

Multiparous

128 (29)

60 (24.1)

68 (35.6)

0.35

0.65

Parity

25.05 (4.6)

25.9 (4.4)

23.9 (4.6)

NVD = Normal vaginal delivery; CS = caesarean section; SD = standard deviation. *Excluded primigravida and miscarriages. †Excluded miscarriages.

0.05

Age (years)

History of infertility

Age at first pregnancy (years), mean (SD)

v. 32%; p=0.02). Other perspectives of women’s perceptions were not significantly different between the private and public patients. Intention to request to deliver by CS was reported by approximately 8% of the respondents. Previous history of CS was the only factor that was a significant predictor of the intention to request CS in the index pregnancy (p=0.001) (Table 5). Factors that were associated with reproductive history and previous delivery outcomes in the bivariate analysis were entered in the regression model (Table 6). The intention to request for CS was 6-fold more likely among patients who had a previous elective CS delivery than those who had vaginal delivery (adjusted odds ratio (AOR) 5.8; 95% CI 1.6 - 20.2). Furthermore, patients who had had a previous emergency CS were 5-fold more likely to request CS compared with those who delivered vaginally in their previous pregnancy (AOR 5.1; 95% CI 1.7 - 15.4). Study participants were less willing to request CS in other public low-referral point health facilities compared with private health facilities. Other factors related to intention to deliver by CS were not statistically significant.

0.14

0.001

0.71

Occupation

0.05

Parity

CS = caesarean section; IPPM = intramural private practice management; SD = standard deviation. *Unless otherwise specified

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH Table 4. Comparison of different perceptions of CS on request among pregnant women in study cohort Perceptions of CS on request

Total (N=440), n (%)

Private (n=249), n (%)

Public (n=191), n (%)

202 (46.0)

111 (44.4)

91 (47.9)

p-value

CS on request should be allowed Disagree Not sure

31 (7.0)

19 (7.6)

12 (6.3)

Agree

207 (47.0)

120 (48.0)

87 (45.8)

90 (20.5)

52 (20.9)

38 (19.9)

0.77

Doctor has the right to overrule CS on maternal request Disagree Not sure

64 (14.5)

29 (11.6)

35 (18.3)

Agree

286 (65.0)

168 (67.5)

118 (61.8)

255 (58.0)

139 (55.8)

116 (60.7)

0.14

Women who request delivery by CS had history of infertility Disagree Not sure

130 (29.5)

75 (30.1)

55 (28.8)

Agree

55 (12.5)

35 (14.1)

20 (10.5)

0.44

Women who request delivery by CS due to previous miscarriage Disagree

205 (46.6)

111(44.4)

94 (49.5)

Not sure

88 (20)

49 (19.6)

39 (20.5)

Agree

147 (33.4)

90 (36.0)

57 (30.0)

0.39

Women who request delivery by CS are >35 years old Disagree

230 (52.3)

123(49.4)

107 (56.0)

Not sure

76 (17.3)

45 (18.1)

31 (16.2)

Agree

134 (30.5)

81 (32.5)

53 (27.7)

0.38

Women who request delivery by CS have psychological problems Disagree

182 (41.4)

100 (40.2)

82 (42.9)

Not sure

103 (23.4)

53 (21.3)

50 (26.2)

Agree

155 (35.2)

96 (38.6)

59 (30.9)

206 (46.8)

111 (44.6)

95 (49.7)

0.21

Women who request delivery by CS are financially well supported Disagree Not sure

55 (12.5)

28 (11.2)

27 (14.1)

Agree

179 (40.7)

110 (44.2)

69 (36.1)

0.21

Women who request delivery by CS have a fear of losing a child on normal delivery Disagree

167 (38)

84 (33.7)

83 (43.5)

Not sure

73 (16.6)

38 (15.3)

35 (18.3)

Agree

200 (45.5)

127 (51.0)

73 (38.2)

0.03*

Women who request delivery for CS have a fear of labour pain Disagree

54 (12.3)

25 (10.0)

29 (15.2)

Not sure

37 (8.4)

20 (8.0)

17 (8.9)

Agree

349 (79.3)

204 (81.9)

145 (75.9)

167 (38)

88 (35.2)

79 (41.6)

0.23

Requesting delivery by CS can affect doctor-patient relationship Disagree Not sure

100 (22.7)

50 (20)

50 (26.3)

Agree

173 (39.3)

112 (44.8)

61 (32.1)

130 (29.5)

78 (31.3)

52 (27.2)

0.02*0

Women who request delivery by CS have a need to undergo tubal ligation Disagree Not sure

209 (47.5)

110 (44.2)

99 (51.8)

Agree

101 (23)

61 (24.5)

40 (20.9)

124 (28.2)

63 (25.3)

61 (31.9)

0.28

Women who request delivery by CS will have less pelvic flow injury compared with normal delivery Disagree Not sure

157 (35.7)

96 (38.6)

61 (31.9)

Agree

159 (36.1)

90 (36.1)

69 (36.1)

CS = caesarean section. *Statistically significant.

0.22

RESEARCH increased from 22% in 2002 to 51% in 2014, and therefore there is a need to examine the factors that affect women’s decisions to elect for CS, based on their attitudes and experiences. Fear of losing a child was a predictor for maternal request for CS, as was reported in a study in Sweden in which 28% of respondents believed that requesting for CS was based on their concerns for the newborn.[23] Having a negative perception of childbirth based on the mother’s experience in a prior pregnancy and a history of obstetric complications could be the underlying Table 5. Factors associated with intention to undergo CS on maternal request* Intend to request delivery by CS Yes, n (%)

No, n (%)

Total, n (%)

NVD

4 (2.7)

141 (97.3)

145 (100)

Elective CS

6 (17.1)

29 (82.9)

35 (100)

Emergency CS

11 (15.5)

60 (84.5)

71 (100)

Factors

p-value

Mode of delivery of the last pregnancy

0.001†

Outcome of your last pregnancy Live birth

20 (9.7)

186 (90.3)

206 (100)

Macerated birth

1 (4.3)

22 (95.7)

23 (100)

Stillbirth

1 (4.5)

21 (95.5)

22 (100)

Miscarriage

3 (6.8)

41 (93.2)

44 (100)

0.687

Type of antenatal care clinic Public

14 (7.3)

177 (92.7)

191 (100)

IPPM (private)

20 (8.0)

229 (92.0)

249 (100)

Primary school or less

7 (7.4)

88 (92.6)

95 (100)

Secondary school

11 (7.6)

133 (92.4)

144 (100)

College/university

16 (8.0)

185 (92.0)

201 (100)

<21

6 (7.1)

78 (92.2)

84 (100)

26 - 30

10 (5.8)

162 (94.2)

172 (100)

31 - 34

13 (10.0)

117 (90.0)

130 (100)

>35

5 (9.3)

49 (90.7)

54 (100)

0.798

Education level

0.99

Age (years)

0.79

Occupation Employed

19 (8.8)

197 (91.2)

216 (100)

Informal trader

9 (7.1)

118 (92.9)

127 (100)

Other

6 (6.2)

91 (93.8)

97 (100)

0.25

Ever delivered by CS Yes

19 (16.7)

95 (83.3)

114 (100)

4 (2.5)

155 (97.5)

159 (100)

Yes

4 (20.0)

16 (80.0)

20 (100)

13 (14.6)

76 (85.4)

89 (100)

0.001†

Have you ever had an infertility problem 0.55

CS = caesarean section; NVD = normal vaginal delivery; IPPM = intramural private practice management. * Some numbers do not add up to the total due to missing values in some variables. † Statistically significant.

reason for expressing concern for the new baby and, subsequently, the maternal request for CS.[24] Prior local evidence has highlighted that mothers’ perceptions of maternal complications and risk, the chance of delivering a healthy baby following previous poor outcome, or an external influence such as seeking experience and advice from peers, are all contributing factors for CS on request.[17] At MNH and other health facilities in Tanzania, understaffing and limited health resources have led to the provision of substandard care; thus, upgrading CEoC could not only prevent adverse delivery outcomes, but might also contribute to reducing the number of unnecessary CSs based on maternal request. As shown in this study, previous history of CS was a significant factor associated with requesting CS, and previous emergency CS had a higher likelihood of maternal request for CS compared to previous elective CS. Similar findings have been reported in Canada, Sweden and Nigeria.[20,25,26] Furthermore, evidence has shown that fear of childbirth associated with obstetric complications is usually related to emergency rather than elective CS. Parturients and their newborns in health-resource-limited settings such as in Tanzania have a high risk of severe morbidity during birth due to abnomal labour, which could lead to avoidance of vaginal delivery as a result of a negative perception of childbirth.[17] The findings regarding the influence of advanced maternal age, psychological problems and previous pregnancy loss to intention to request CS were contrary to the findings of other studies.[27] Cultural differences could explain these differences. The participants’ unwillingness to request CS in another public referral point at a lower referral level, such as a regional hospital, compared with private health facility could imply women’s lack of trust/confidence and the provision of a relatively lower quality of service in these facilities compared with MNH.[6] This finding is supported by Okonkwo et al.,[20] who reported that the rate of maternal request for CS was higher in a tertiary hospital compared with secondary and primary health centres. Furthermore, almost half of the participants were of the opinion that CS on request should be allowed and that doctors had the right to overrule maternal request for CS. Nonetheless, participants felt that requesting CS might affect the doctor-patient relationship, which aligns with the findings of a previous qualitative study at MNH that highlighted maternal anxiety and poor client counselling during childbirth and thus found room for improvement.[17] Evidence relating to the potential benefits of elective CS compared with vaginal delivery has been inconsistent. While some literature supports the notion that elective CS is associated with a decreased risk of urinary incontinence, pelvic organ prolapse, anal sphincter damage, fecal incontinence and flexible timing for the mother,[1,28] others advocate vaginal delivery, considering the risk of adverse outcomes of CS, including haemorrhage, admission to the ICU, blood transfusion and hysterectomy, especially in low-resource settings.[29,30] As with any major surgical procedure, there are risks associated with CS, including complications of anaesthesia, excessive blood loss, breathing problems, infection, urinary tract injury, and injury to the baby.[6,7] In addition, recovery time and hospital stay following caesarean delivery are longer than following vaginal delivery and therefore CS is associated with increased cost of care for the individual, the family and the health system. Therefore, adequate information should be made available to the clients when considering delivery by CS when vaginal delivery is also possible. This study was conducted at the largest tertiary hospital in Tanzania, which has a rapidly increasing CS rate that represents

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH Table 6. Logistic regression on factors associated with intention of maternal request for CS Factor

COR

95% CI

AOR

95% CI

p-value

Mode of delivery of last pregnancy NVD

Elective CS

6.7

1.4 - 9.3

5.8

1 1.6 - 20.2

0.006

Emergency CS

8.1

2.6 - 10.1

5.1

1.7 - 15.4

0.004

Outcome of last pregnancy Live birth

Stillbirth

0.38

0.43 - 3.2

0.5

1 0.12 - 1.4

0.83

Miscarriage

0.11

0.14 - 1.2

0.24

0.21 - 3.1

0.37

0.35 - 4.4

1.25

0.47 - 3.39

0.65

Have you ever had an infertility problem No

Yes

1.25

CS = caesarean section; COR = crude odds ratio; AOR = adjusted odds ratio; NVD = normal vaginal delivery; CI = confidence interval.

a real problem in Tanzanian referral health facilities. The environment within a referral centre could also be the reason for the increase in the number of CS associated with lowrisk pregnancy.[4] The seemingly rare complications of CS, including subsequent placenta praevia with placenta acreta, and endometriosis should be addressed as possible adverse effects of CS during counselling. The reported numbers of maternal request for CS in this study might be biased by the fact that this study was conducted in the largest tertiary urban health facility, which received referred complicated pregnancies as well as private patients with low-risk pregnancies. Furthermore, the national CS rate of 6% in Tanzania is still below the 10% that was deemed acceptable based on WHO literature. However, CS rates should not be demand-driven but based on optimal indications. CS without maternal or fetal indications unnecessarily depletes the scarce resources available in a low-income country like Tanzania.

Conclusion

Maternal requests for CS do exist at the highest national referral hospital in Tanzania. The maternal requests for CS were associated with factors other than women’s preferences, including perceived fear of child loss following vaginal delivery and events associated with previous CS. In the absence of maternal or fetal indications for CS, a plan for vaginal delivery is a safe and appropriate option and should be recommended. Care providers have the responsibility of alleviating women’s perceived fear of childbirth by providing adequate counselling on their decision of mode of delivery to assist women in making an informed choice. Furthermore, good-quality intrapartum care encourages women to attempt vaginal birth. A deeper understanding of women and health care providers’ perceptions and attitudes towards maternal requests for CS through a qualitative approach is also recommended. Acknowledgements. We acknowledge the voluntary participation of women attending ANC clinic at MNH. We also appreciate the contribution of academic staff in the Department of Obstetrics and Gynaecology of MUHAS, and all physicians and nurses in the Department of Obstetrics and Gynaecology of MNH, who played an invaluable role in the success of this work. Author contributions. CM designed the study, carried out the collection and analyses of data, and drafted the first version of the manuscript. BK participated in data analysis, interpretation of data and development of

the manuscript. AM participated in the interpretation of data, and drafted the first version and final manuscript. HK contributed to the design of the study, data analysis, interpretation of the results and participated in the development of the first and final version of the manuscript. All authors read and approved the final manuscript. Funding. None. Conflicts of interest. None.

1. Viswanathan M, Visco AG, Hartmann K, et al. Cesarean delivery on maternal request. Evid Rep Technol Assess 2006;(133):1-138. 2. American College of Obstetricians and Gynecologists. ACOG committee opinion no. 559: Cesarean delivery on maternal request. Obstet Gynecol 2013;121(4):904-907. https://doi.org/10.1097/01. AOG.0000428647.67925.d3 3. National Bureau of Statistics (NBS), ICF Macro. Tanzania Demographic and Health Survey 2010. Dar es Salaam, Tanzania and Calverton, MA: NBS and ICF Macro, 2011. 4. Litorp H, Kidanto HL, Nyström L, Darj E, Essén B. Increasing caesarean section rates among lowrisk groups: A panel study classifying deliveries according to Robson at a university hospital in Tanzania. BMC Preg Childbirth 2013;13:107. https://doi.org/10.1186/1471-2393-13-107 5. Mdegela MH, Muganyizi PS, Pembe AB, Simba DO, Van Roosmalen J. How rational are indications for emergency caesarean section in a tertiary hospital in Tanzania? Tanzan J Health Res 2012;14(4):236-242. 6. Litorp H, Kidanto HL, Rööst M, Abeid M, Nyström L, Essén B. Maternal near-miss and death and their association with caesarean section complications: A cross-sectional study at a university hospital and a regional hospital in Tanzania. BMC Preg Childbirth 2014;14:244. https://doi. org/10.1186/1471-2393-14-244 7. Eriksson J, Baker T, Jörnvall H, Irestedt L, Mulungu M, Larsson E. Quality of anaesthesia for caesarean sections: A cross-sectional study of a university hospital in a low-income country. Trop Med Int Health 2015;20(10):1329-1336. https://doi.org/10.1111/tmi.12553 8. World Health Organization. WHO Guidelines for the Management of Postpartum Haemorrhage and Retained Placenta. Geneva: WHO, 2009. 9. Kongnyuy EJ, Mlava G, van den Broek N. A criterion based audit of the management of obstructed labour in Malawi. Arch Gynecol Obstet 2009;279(5):649-654. https://doi.org/10.1007/s00404-008-0786-1 10. Sorensen BL, Elsass P, Nielsen BB, Massawe S, Nyakina J, Rasch V. Substandard emergency obstetric care – a confidential enquiry into maternal deaths at a regional hospital in Tanzania. Trop Med Int Health 2010;5(8):894-900. https://doi.org/10.1111/j.1365-3156.2010.02554.x 11. Maaløe N, Sorensen BL, Onesmo R, Secher NJ, Bygbjerg IC. Prolonged labour as indication for emergency caesarean section: A quality assurance analysis by criterion-based audit at two Tanzanian rural hospitals. BJOG 2012;119(5):605-613. https://doi.org/10.1111/j.1471-0528.2012.03284.x 12. van Beekhuizen HJ, Unkels R, Mmuni NS, Kaiser M. Complications of obstructed labour: Pressure necrosis of neonatal scalp and vesicovaginal fistula. Lancet 2006;368(9542):1210. https://doi. org/10.1016/S0140-6736(06)69477-4 13. Kabakyenga JK, Östergren P-O, Turyakira E, Mukasa PK, Pettersson KO. Individual and health facility factors and the risk for obstructed labour and its adverse outcomes in south-western Uganda. BMC Preg Childbirth 2011;11(11):73. https://doi.org/10.1186/1471-2393-11-73 14. Nilstun T, Habiba M, Lingman G, et al. Cesarean delivery on maternal request: Can the ethical problem be solved by the principlist approach? BMC Med Ethics 2008;9(1):11. https://doi. org/10.1186/1472-6939-9-11 15. Signore C, Hemachandra A, Klebanoff M. Neonatal mortality and morbidity after elective cesarean delivery versus routine expectant management: A decision analysis. Semin Perinatol 2006;30(5):288-295. https://doi.org/10.1053/j.semperi.2006.07.010 16. Nama V, Wilcock F. Caesarean section on maternal request: Is justification necessary? Obstet Gynaecol 2011;13(4):263-269. https://doi.org/10.1576/toag.13.4.263.27693 17. Litorp H, Mgaya A, Kidanto HL, Johnsdotter S, Essén B. ‘What about the mother?’ Women’s and caregivers’ perspectives on caesarean birth in a low-resource setting with rising caesarean section rates. Midwifery 2015;31(7):713-720. https://doi.org/10.1016/j.midw.2015.03.008 18. Selinger H. Maternal request for caesarean section: An ethical consideration. J Med Ethics 2016;40(12):857-860. https://doi.org/10.1136/medethics-2013-101558

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH 19. Campbell OMR, Graham WJ, Lancet Maternal Survival Series steering group. Strategies for reducing maternal mortality: Getting on with what works. Lancet 2006;368(9543):1284-1299. https://doi.org/10.1016/S0140-6736(06)69381-1 20. Okonkwo NS, Ojengbede OA, Morhason-Bello IO, Adedokun BO. Maternal demand for cesarean section: Perception and willingness to request by Nigerian antenatal clients. Int J Wom Health 2012;4:141-148. https://doi.org/10.2147/IJWH.S10325 21. Litorp H, Mgaya A, Mbekenga CK, Kidanto HL, Johnsdotter S, Essén B. Fear, blame and transparency: Obstetric caregivers’ rationales for high caesarean section rates in a low-resource setting. Soc Sci Med 2015;143:232-240. https://doi.org/10.1016/j.socscimed.2015.09.003 22. Bettes BA, Coleman VH, Zinberg S, et al. Cesarean delivery on maternal request: Obstetriciangynecologists’ knowledge, perception, and practice patterns. Obstet Gynecol 2007;109(1):57-66. https://doi.org/10.1097/01.AOG.0000249608.11864.b6 23. Wiklund I, Edman G, Andolf E. Cesarean section on maternal request: Reasons for the request, selfestimated health, expectations, experience of birth and signs of depression among first-time mothers. Acta Obstet Gynecol Scand 2007;86(4):451-456. https://doi.org/10.1080/00016340701217913 24. Størksen HT, Garthus-Niegel S, Vangen S, Eberhard-Gran M. The impact of previous birth experiences on maternal fear of childbirth. Acta Obstet Gynecol Scand 2013;92(3):318-324. https:// doi.org/10.1111/aogs.12072

25. Nieminen K, Stephansson O, Ryding EL. Women’s fear of childbirth and preference for cesarean section – a cross-sectional study at various stages of pregnancy in Sweden. Acta Obstet Gynecol Scand 2009;88(7):807-813. https://doi.org/10.1080/00016340902998436 26. Kornelsen J, Hutton E, Munro S. Influences on decision making among primiparous women choosing elective caesarean section in the absence of medical indications: Findings from a qualitative investigation. J Obstet Gynaecol Canada 2010;32(10):962-969. https://doi.org/10.1016/ S1701-2163(16)34684-9 27. Hildingsson I, Nilsson C, Karlström A, Lundgren I. A longitudinal survey of childbirth-related fear and associated factors. J Obstet Gynecol Neonatal Nurs 2011;40(5):532-543. https://doi. org/10.1111/j.1552-6909.2011.01274.x 28. Cacciatore A, Giordano R, Romano M, La Rosa B, Fonti I. Putative protective effects of cesarean section on pelvic floor disorders. J Prenat Med 2010;4(1):1-4. 29. Shah A, Fawole B, M’Imunya JM, et al. Cesarean delivery outcomes from the WHO global survey on maternal and perinatal health in Africa. Int J Gynecol Obstet 2009;107(3):191-197. https://doi. org/10.1016/j.ijgo.2009.08.013 30. Lumbiganon P, Laopaiboon M, Gülmezoglu AM, et al. Method of delivery and pregnancy outcomes in Asia: The WHO global survey on maternal and perinatal health 2007-08. Lancet 2010;375(9713):490-499. https://doi.org/10.1016/S0140-6736(09)61870-5

SAJOG • September 2017, Vol. 23, No. 2

This open-access article is distributed under Creative Commons licence CC-BY-NC 4.0.

RESEARCH

Factors associated with the lack of antiretroviral therapy initiation among eligible HIV-positive pregnant women in Swaziland C Chouraya,1,2 MB ChB, MSc; G Louwagie,3 MMed, FCPHM, PhD; B Nhlabatsi,4 MPH; M A Mahdi,5 MPH; B V Girdler-Brown,2 MB ChB, MMed, FCPHM, FFPH Elizabeth Glaser Paediatric AIDS Foundation, Mbabane, Swaziland School of Health Systems and Public Health, University of Pretoria, Pretoria, South Africa 3 Department of Public Health Medicine, School of Health Systems and Public Health, University of Pretoria, Pretoria, South Africa 4 Sexual Reproductive Health Unit, Ministry of Health, Mbabane, Swaziland 5 Elizabeth Glaser Paediatric AIDS Foundation, Washington DC, US 1 2

Corresponding author: C Chouraya (cchouraya@pedaids.org)

Background. Antiretroviral therapy (ART) initiation is critical for the prevention of mother-to-child transmission (PMTCT) of HIV. Objectives. To quantify factors that were barriers or facilitators to the initiation of ART in pregnant HIV-infected women in Swaziland. Methods. We conducted a cross-sectional survey in HIV-infected women with at least one antenatal care (ANC) visit, who had delivered in maternity wards between April and August 2013 in Swaziland. Variables collected included intrapersonal, interpersonal and organisational factors. Logistic regression models were used to calculate univariate and adjusted multivariate measures of association between ART initiation and the independent variables. Results. Among the 163 pregnant women who were eligible for ART, 110 (67.5%) were initiated on ART by the time of delivery. The most commonly cited reason for not initiating ART (n=53) was women not being ready to initiate life-long treatment (24.5%). On multivariate logistic regression, favourable perceptions of the benefits of ART (adjusted odds ratio (AOR) 3.04; 95% CI 1.55 - 5.96) and presence of partner support (AOR 4.75; 95% CI 2.11 - 10.67) remained significantly and independently associated with ART initiation. Conclusion. ART initiation among ART-eligible pregnant women in Swaziland was independently associated with the presence of partner support and favourable perceptions of the benefits of ART. Stronger counselling and education for pregnant women and male involvement strategies need to be implemented as universal life-long ART for all HIV-infected pregnant women is implemented. S Afr J Obstet Gynaecol 2017;23(2):63-68. DOI:10.7196/SAJOG.2017.v23i2.1184

Prior to 2013, antiretroviral therapy (ART) was recommended by the World Health Organization (WHO) for HIV-infected pregnant women with CD4 cell counts of <350 cells/µL or those who were at WHO clinical stage III or IV Women who did not meet these criteria,were advised to use one of two options. Option A: antepartum zidovudine, with single-dose nevirapine during labour, and a 7-day zidovudine/lamivudine ‘tail’ alongside daily infant nevirapine during breastfeeding. Option B: ART during pregnancy and through cessation of breastfeeding to prevent mother-to-child HIV transmission (PMTCT.)[1] Many countries, including Swaziland, chose to implement option A; a CD4 cell count was performed on all women identified to have HIV infection to determine those who should be initiated on life-long ART for both maternal health and PMTCT. In Swaziland, 88% of facilities that provide antenatal care (ANC) services also provide PMTCT services.[2] In 2009, 73% of women attending ANC facilities were tested for HIV, and 88% of HIV-infected pregnant women received a complete course of PMTCT prophylaxis in 2009.[3] However, a review of the data in the Early Infant Diagnosis (EID) database in Swaziland at the time of the introduction of option A guidelines, showed that ~50% of HIV-infected women who were eligible for ART actually initiated treatment.[4] In Kenya, pregnancy has been identified as one of the risk factors for ART non-initiation.[5] A number of studies identified barriers to

ART initiation during pregnancy in southern and eastern Africa, including fear of knowing one’s HIV status, lack of male partners’ support, negative attitudes of healthcare workers, late presentation in pregnancy, advanced HIV disease, lack of finances, fear of stigma and discrimination, non-disclosure of HIV status and long waiting times.[6-11] New WHO guidelines now recommend initiation of life-long therapy in all HIV-infected pregnant and breastfeeding women. Obtaining a better understanding of the factors associated with the acceptance of ART will be critical to implement these recommendations effectively.[12] We conducted a study to quantify factors that were barriers or facilitators to the initiation of ART in HIV-infected women in Swaziland who were identified as eligible for ART under option A.

Methods

Study design and setting An analytical cross-sectional study was performed through administration of a pre-tested questionnaire to HIV-infected pregnant women after delivery, with a review of antenatal care cards, in all 11 public hospital and health centre maternity wards in Swaziland.

Participant selection Participants were selected consecutively from April to August 2013. The inclusion criteria for the study were: HIV-infected women who

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH were eligible for ART (i.e. CD4 cell counts of <350 cells/µL or WHO clinical stage III or IV) as documented in the patient ANC card; and the women should have attended at least one ANC.

Measurements Variables were collected using an interviewer-administered structured questionnaire and a record review of clinical information from the ANC cards. The socio-ecological model (SEM) was used to study the factors associated with ART initiation among eligible pregnant women. The model outlines multiple levels of influence for the actions of individuals that include intrapersonal, interpersonal, organisational, community and policy levels.[13,14] Intrapersonal variables collected included sociodemographic factors (age, parity, level of education, employment status); disease factors (CD4 cell count, clinical staging); and perceptions of ART (fear that ARV might harm the baby, lack of understanding of the need for ART, fear of side-effects from ART, and fear of commitment to life-long ART). Interpersonal variables collected were marital status, HIV disclosure, partner support, and being a member of a support group. The main organisational variable collected was pregnant women’s perceptions of healthcare workers’ attitudes toward them.

Sample size, data collection, management and analysis The planned sample size was calculated to be 380, with α set at 0.05 and 80% power to detect a 15% difference in ART initiation between women who had disclosed their status and those who had not, assuming that the ART initiation was 50% in those who did not disclose their status.

Data collection was performed by trained midwives at the identified maternity units. Training for the midwives consisted of a one-day training session onsite involving presentations and role-play. Data were entered using EpiData 3.1 software (EpiData Association, Denmark). Double data entry was done using EpiData 3.1. Data were exported to Stata version 11 (StataCorp LP, USA) for analysis. Women’s characteristics were described using percentages and proportions for categorical data and means, medians, standard deviation, ranges and interquartile ranges for continuous data. The proportion of eligible women initiating ART was also computed. Logistic regression was used to calculate crude and adjusted measures of association between the outcome and the independent variables (intrapersonal, interpersonal and organisational variables) which were measured as odds ratios with χ2 test (univariate analyses) or z-test (multivariate analyses) p-values. The Cronbach alpha statistic was calculated for each construct variable, i.e. variables that consisted of scores derived from responses to a number of questionnaire items. Construct variable scores were included as predictor variables if their Cronbach alpha statistics were 0.70 or higher. Additional potential predictor variables were included in the model if there was information from other studies that the variable was associated with ART initiation and also if the χ2 test p-value was ≤0.25 in univariate analysis as recommended by Hosmer and Lemeshow.[15] Stepwise backward hierarchical elimination was used to produce a parsimonious model; the likelihood ratio test, with an alpha value of 0.05, was used to decide on the elimination of statistically non-significant variables. Post regression analysis consisted of estimation of the receiver operating characteristic (ROC) area under the curve[16] and the Hosmer-Lemeshow goodness-of-fit test.[15]

Table 1. Demographic characteristics of the participants Total (N=163) Characteristic

n (%)

Median (IQR)

ART initiated (N=110) n (%)

Median (IQR)

ART not initiated (N=53) n (%)

Median (IQR)

Age (years)

26 (23 - 30)

26 (23 - 31)

25 (22 - 29)

Parity

2 (1 - 3)

Level of education Some primary

43 (26.4)

28 (25.5)

15 (28.3)

Completed primary

10 (6.1)

7 (6.4)

3 (5.7)

Some secondary

46 (28.2)

32 (29.1)

14 (26.4)

Completed secondary

52 (31.9)

34 (30.9)

18 (34.0)

Tertiary

4 (2.5)

2 (1.8)

2 (3.8)

Never attended school

8 (4.9)

7 (6.4)

1 (1.9)

Professional

4 (2.5)

3 (2.7)

1 (1.9)

Semi-skilled

35 (21.5)

21 (19.1)

14 (26.4)

Self-employed

8 (5.0)

4 (3.6)

4 (7.5)

Not employed

116 (71.2)

82 (74.5)

34 (64.2)

Pentecostal

34 (20.9)

21 (19.1)

13 (24.5)

Protestant

67 (41.1)

48 (43.6)

19 (35.8)

Zionist

41 (25.2)

27 (24.5)

14 (26.4)

Other

19 (11.6)

13 (11.9)

6 (11.4)

Not stated

2 (1.2)

1 (0.9)

1 (1.9)

Employment status

Religion

ART = antiretroviral therapy; IQR = interquartile range. *Fisher's exact test for categorical variables and Mann Whitney (Wilcoxon rank-sum) test for continuous variables.

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH

Ethical considerations

Clinical characteristics

All patients were asked to provide written informed consent before inclusion into the study. This study was approved by the University of Pretoria Ethics Committee (ref. no. 217/2011) and by the Swaziland Ministry of Health’s Scientific and Ethics Committee (ref. no. MH/599C).

The median baseline CD4 cell count for participants was 248 cells/µL and was slightly higher in women who initiated ART compared with those who did not; however, the difference was not statistically significant (p=0.6413) (Table 2). Excluding women who were already known HIV-infected at the time of conception, the median gestational age at HIV diagnosis was 19 weeks; women who did not initiate ART were diagnosed with HIV later in pregnancy compared with women who initiated ART. More than 90% of the women were at WHO clinical stage I or II. The majority of clients (88%) were seen at

Results

Participants enrolled in the study Of the 1 371 HIV-infected pregnant women delivering at the study sites, 6 had no ANC cards, 431 were already on ART before the current pregnancy, 549 were not eligible for ART, 204 had no documented CD4 cell count results, 16 decided not to participate in the study, and 2 were missed by the data collectors (Fig. 1). Therefore, 163 women were enrolled in the study.

Participants enrolled for study (N=163) No ANC cards, no ANC visits (n=6) N=1 365

ART initiation

Already on ART before current pregnancy (n=431)

Among the 163 pregnant women who were eligible for ART, only 110 (67.5%) had been initiated on ART by the time of delivery. N=934

Sociodemographic characteristics The demographic characteristics of the participants are summarised in Table 1. The median age was 26 years. The women who did not initiate ART were slightly younger (median age of 25 years) than those who initiated ART (median age of 26 years) but this difference was not statistically significant (p=0.2253). Only 4.9% of the women never attended school. The majority of the women (71.2%) were not employed. A higher proportion of women who initiated ART were unemployed compared with the women who did not initiate ART. The median parity for the women was 2 and this was not different between women who initiated ART and those who did not (p=0.2735). The most common religion among the women was Protestant followed by Zionists (a Christian religious group that generally does not believe in the efficacy of Western medicines).[17]

No CD4 results (n=204) N=730 No CD4 results (n=204) N=181 18 not enrolled into study: 2 missed by data collectors; 16 declined to participate Participants enrolled for study (N=163)

Fig. 1. Flow chart for patients enrolled into study. (ANC = antenatal care; ART = antiretroviral therapy.)

Table 2. Clinical characteristics of the women (N=163) Total Median (IQR)

ART initiated n/N (%)

Median (IQR)

ART not initiated n/N (%)

Median (IQR)

Characteristic

n/N (%)

CD4 cell count (cells/µL)

n/a

248 (190 300)

n/a

247 (179 300.5)

n/a

254 (199 300)

0.6413

Known positive at entry, n/N (%)

38/161 (23.6)

n/a

25/108 (23.1)

n/a

13/53 (24.5)

n/a

GA at HIV diagnosis (weeks)

n/a

19 (7 - 23)

n/a

19 (7.5 - 22)

n/a

21 (7 - 24)

0.2302

n/a

81/110 (73.6)

n/a

36/53 (67.9)

n/a

p-value*

WHO clinical stage, n/N (%) I

117/163 (71.8)

35/163 (21.5)

23/110 (20.9)

12/53 (22.6)

III

10/163 (6.1)

5/110 (4.5)

5/53 (9.4)

1/163 (0.6)

1/110 (0.9)

0/53 (0)

Client referred for ART, n/N (%) Yes

15/163 (9.2)

4/163 (2.5)

n/a

0/110 (0)

11/110 (10)

n/a

4/53 (7.5)

n/a

ART available at site

144/163 (88.3)

99/110 (90)

45/53 (84.9)

n/a

ART = antiretroviral therapy; IQR = interquartile range; GA = gestational age; n/a = not applicable. *Fisher's exact test for categorical variables and Mann-Whitney (Wilcoxon rank-sum) test for continuous variables.

SAJOG • September 2017, Vol. 23, No. 2

4/53 (7.5)

n/a

0.3460

RESEARCH Table 3. Women’s perceptions towards ART* Total

ART initiated

ART not initiated

Median (IQR)

Cronbach’s alpha

p-value†

Perceptions on benefits of ART

163

4.0 (3.6 - 4.6)

110

4.2 (3.8 - 4.8)

3.8 (3.6 - 4.2)

0.71

0.0001

Perceptions on harms/difficulties with ART

162

3.9 (3.4 - 4.4)

109

4.0 (3.6 - 4.6)

3.6 (3.0 - 4.0)

0.63

0.0011

Perceptions on need for ART

162

4.5 (4.0 - 5.0)

109

4.5 (4.0 - 5.0)

4.0 (3.8 - 4.8)

0.61

0.0155

Fear of stigma and discrimination

162

4.0 (3.8 - 4.8)

109

4.0 (3.8 - 4.8)

4.0 (3.3 - 4.5)

0.54

0.0280

ART = antiretroviral therapy; IQR = interquartile range. *Maximum score = 5 (positive perception). Scores for harms of ART, stigma and need for ART were reversed so that values closer to 5 indicate less harm and stigma; and greater need for ART. † Mann-Whitney (Wilcoxon rank-sum) test.

Table 4. Interpersonal factors Total (N=163), n (%)

ART initiated (N=110), n (%)

ART not initiated (N=53), n (%)

p-value*

Yes

150 (92)

107 (97.3)

43 (81.1)

<0.0001

13 (8)

3 (2.7)

10 (18.9)

Yes

10 (6.1%)

9 (8.2)

1 (1.9)

153 (93.9)

101 (91.8)

52 (98.1)

Yes

120 (73.6)

84 (76.4)

36 (67.9)

12 (7.4)

7 (6.4)

5 (9.4)

Unknown

31 (19)

19 (17.3)

12 (22.6)

Yes

91 (55.8)

65 (59.1)

26 (49.1)

72 (44.2)

45 (40.9)

27 (50.1)

Yes

117 (84.2)

94 (91.3)

23 (63.9)

22 (15.8)

9 (8.7)

13 (36.1)

Characteristic Disclosed HIV status (N=163)

Member of support group (N=163) 0.117

Religion supports ART (N=163)

ART, 18.9% had not disclosed their status compared with 2.7% among those who initiated ART (p<0.0001). A higher proportion of women who initiated ART were members of a support group compared with women who did not initiate ART (8.2% v. 1.9%), but this difference was not statistically significant. More than 40% of participants did not know the HIV status of their partner and this proportion was higher among women who did not initiate ART compared with those who initiated ART, although the difference was not statistically significant. Of the 139 women who consulted their partners on ART initiation, 84.2% said that their partners supported the decision to start ART. This proportion was lower (63.9%) among women who did not initiate ART compared with women who initiated (91.3%; p<0.0001). Interpersonal factors are summarised in Table 4.

Organisational factors 0.252

Knows HIV status of partner (N=163) 0.227

Partner supportive of ART (N=139) <0.0001

Women who did not initiate ART reported a less favourable attitude from healthcare workers compared with those who did initiate ART (median score for healthcare workers’ (HCWs’) attitude towards pregnant women of 3.5 v. 3.61, p=0.0045) (Table 5).

Reasons for not initiating ART The most common reason cited for not initiating ART (n=53) despite being eligible for ART, was that they were not yet ready to commit to life-long treatment (24.5%). There were, however, numerous other reasons (Table 6).

Univariate analysis

ANC healthcare facilities where ART services were available. There was no difference in ART initiation between women at ANC facilities with ART services (n=144) and those without ART services (n=15; 69% v. 73%, respectively; p=0.346).

On univariate analysis, factors that were associated with ART initiation were: perceptions of the benefits of ART (OR 3.49; p<0.001), perceptions on the harms of/difficulties with antiretrovirals (OR 2.14; p=0.002), perceptions on the need for ART (OR 1.94; p=0.011), fear of stigma and discrimination (OR 1.70; p=0.011), HIV status disclosure (OR 8.29; p=0.002), partner support for ART (OR 7.66; p<0.001).

Perceptions towards ART

Multivariate logistic regression

Perceptions towards initiating ART were estimated by the scores obtained for the construct variables; the median scores were statistically significantly different between those initiated on ART and those not initiated for all the domains. The median scores and MannWhitney (Wilcoxon rank-sum) test p-values are presented in Table 3.

The following variables were included in the initial regression model: employment status; perceptions on the benefits of ART; HIV disclosure status; member of a support group; partner support for ART; and knowledge of partner status (Table 7). The other variables were not included, either because of a p-value of >0.25 or Cronbach’s alpha <0.70, or because of many missing variables. The final model had two variables as independent predictors of ART initiation among pregnant women: partner support and perceptions of the benefits of ARVs.

* Fisher's exact test.

Interpersonal factors More than 90% of the participants had disclosed their HIV status to at least one individual. However, among those who did not initiate

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH Table 5. Organisational factors Total

ART initiated

ART not initiated

Characteristic

Median (IQR)

Cronbach’s alpha

p-value†

HCWs’ attitudes towards pregnant women*

108

3.57 (3.16 – 3.86)

3.61 (3.27 – 3.88)

3.5 (2.98 – 3.73)

0.86

0.045

ART = antiretroviral therapy; HCW = healthcare worker; IQR = interquartile range. *Expected score = 4 (excellent attitude). † Mann-Whitney (Wilcoxon rank-sum) test.

Table 6. Reasons for not initiating ART (N=53)* n (%) Not ready

13 (24.5)

Delayed by clinic procedures

9 (17.0)

Partner/family refused

7 (13.2)

Limited ANC visits

7 (13.2)

Not offered ART by nurses

7 (13.2)

Lack of food/finances

4 (7.6)

Fear of ART

3 (5.7)

Felt healthy

2 (3.8)

Work challenges

1 (1.9)

ART = antiretroviral therapy; ANC = antenatal care. *Respondents were asked to select one reason for not initiating ART.

Table 7. Multivariate logistic regression results for the final model (N=158)* Variable

Univariate OR

AOR

95% CI

p-value

Perceptions of benefits of ART

3.49

3.04

1.55 - 5.96

0.001

Partner support for ART

7.66

4.75

2.11 - 10.67 <0.001

OR = odds ratio; CI = confidence interval; AOR = adjusted odds ratio; p-value = double-sided z-test p-value; ART = antiretroviral therapy. *Area under the receiver operating characteristic (ROC) curve = 0.77; Hosmer-Lemeshow goodness-of-fit test p-values for 8, 10 and 12 groupings were 0.47, 0.54 and 0.71, respectively.

Discussion

In this study we found that only 67.5% of HIV-infected pregnant women who met the eligibility criteria for ART, actually initiated ART before they delivered. This rate was comparable with those reported globally.[18-23] The sociodemographic, HIV disease (WHO clinical stage or CD4 cell count), and gestational age at HIV diagnosis were not associated with ART initiation in our study. Data on the importance of these factors in ART acceptance have been questionable, with conflicting results between studies.[10,11,24-31] Partner support was strongly and independently associated with ART initiation among eligible pregnant women in our study, with the odds of initiating ART 4.8-fold higher in women whose partners supported ART initiation compared with those women whose partners did not support ART initiation (p<0.001). Our study contributes to the body of literature that shows that men’s support influences women’s decisions to accept ART and PMTCT interventions.[32-38] Disclosure of HIV status has been associated with the acceptance of ART for PMTCT in studies in South Africa and Zimbabwe.[26,33,39]

In univariate analysis, the odds of ART initiation among women who had disclosed their HIV status to their respective partners were higher than those of women who had not disclosed. However, in multivariate analysis, disclosure of HIV status was not associated with ART uptake when controlling for employment status, presence of male support, being a member of a support group, perceptions of the benefits of ARVs, and knowledge of partners’ HIV status. In a study in Zimbabwe, women who were members of an HIV support group were more than twice as likely to access HIV care and treatment as those who did not belong to any group.[11] In our study, being a member of a support group was not significantly associated with ART initiation, but only 6% of women were members of a support group during pregnancy. This might be attributed to the fact that pregnancy is short and women might not have the time to join a support group before they deliver. Owing to small numbers, we lacked sufficient power to draw a correlation between ART initiation and support group memberships. The odds of initiating ART among eligible pregnant women in Swaziland increased threefold for every one unit score increase in the positive perceptions of women regarding the benefits of ART (p=0.001). This was consistent with other studies which showed that women may refuse ARVs for PMTCT because they doubted their efficacy or they feared that ARVs could harm the unborn child.[40-43] In our study, both healthcare workers' attitudes towards pregnant women and women’s perceptions of access to ART were associated with ART initiation in univariate, but not multivariate, analyses. Other studies have reported that negative interactions between healthcare staff and pregnant women, negative staff attitudes and poor access to ART services can serve as barriers to the uptake of and adherence to PMTCT services by pregnant women.[6,10,40,42,44,45]

Study limitations Our study had some limitations. It only focused on women who were delivering at the healthcare facilities and this might have introduced some bias as women delivering at healthcare facilities might be motivated and committed individuals who are also likely to initiate ART. The interviews were conducted by healthcare workers and this might introduce bias especially on questions asking about healthcare workers’ attitude towards pregnant women. However, efforts were made through training and supervision of healthcare workers to minimise this. Finally, the cross-sectional nature of the study posed some difficulty in defining the temporal relationships

Conclusion

The most commonly cited reason given by women who did not initiate ART was that they felt they were not ready for a life-long commitment to ART. In multivariate analysis, positive perceptions of the benefits of ART and partner support were significantly associated with ART initiation. These results suggest that universal

SAJOG • September 2017, Vol. 23, No. 2

RESEARCH ART in HIV-infected women should be accompanied by improved counselling and education programmes. Strategies to implement male involvement and support could be to the advantage of both the women and PMTCT. Acknowledgements. We would like to thank all of the Swaziland Regional PMTCT coordinators: Phindile Msibi (Shiselweni Region), Clara Mujaya (Lubombo Region), Nkosinathi Maziya (Manzini Region) and Christabel Mahlambi (Hhohho Region) for the great work they did in coordinating the study and ensuring timely and quality data collection. We also acknowledge all of the midwives at the eleven maternity facilities in Swaziland for helping with data collection. And of course, all of the women who agreed to participate in this study. Author contributions. CC designed the study, wrote the protocol, supervised study implementation, conducted data analysis and wrote the manuscript. BN and MAM helped in designing and supervising study implementation. GL helped in study design, protocol writing and article review. BVG helped with data analysis and article editing. All authors read and approved the final manuscript. Funding. School of Health Systems and Public Health, University of Pretoria, Pretoria, South Africa Conflicts of interest. None.

1. World Health Organization. Antiretroviral drugs for treating pregnant women and preventing HIV infection in infants. Geneva: WHO, 2010. http://www.who.int/hiv/pub/mtct/antiretroviral2010/en (accessed 20 May 2011). 2. Ministry of Health, Swaziland. 2010 Service Availability Mapping. Mbabane: MoH, 2010. 3. World Health Organization, UNAIDS and UNICEF. Towards Universal Access: Scaling Up Priority HIV/AIDS Interventions in the Health Sector, 2010 Progress Report. 2010. http://data.unaids.org/ pub/Report/2010/swaziland_2010_country_progress_report_en.pdf 4. Adler MR, Sundaram M, Lukhele B, Vandelanotte J, Chouraya C, Mwanyumba F. Using Swaziland’s national early infant diagnosis database to inform PMTCT program interventions. Proceedings of the XVIII International AIDS Conference. Vienna, 2010. Abstract Number: THPE0416. 5. Karcher H, Omondi A, Odera J, Kunz A, Harms G. Risk factors for treatment denial and loss to follow-up in an antiretroviral treatment cohort in Kenya. Trop Med Int Health 2007;12(5):687-694. https://doi.org/10.1111/j.1365-3156.2007.01830.x 6. Kebaabetswe PM. Barriers to participation in the prevention of mother-to-child HIV transmission program in Gaborone, Botswana a qualitative approach. AIDS Care 2007;19(3):355-360. https://doi. org/10.1080/09540120600942407 7. Fitzgerald FC, Bekker LG, Kaplan R, Myer L, Lawn SD, Wood R. Mother-to-child transmission of HIV in a community-based antiretroviral clinic in South Africa. S Afr Med J 2010;100(12):827-831. https://doi.org/10.7196/samj.4045 8. Coetzee B, Kagee A, Vermeulen N. Structural barriers to adherence to antiretroviral therapy in a resource-constrained setting: the perspectives of health care providers. AIDS Care 2011;23(2):146151. https://doi.org/10.1080/09540121.2010.498874 9. Stinson K, Myer L. Barriers to initiating antiretroviral therapy during pregnancy: A qualitative study of women attending services in Cape Town, South Africa. Afr J AIDS Res 2012;11(1):65-73. https:// doi.org/10.2989/16085906.2012.671263 10. Duff P, Kipp W, Wild TC, Rubaale T, Okech-Ojony J. Barriers to accessing highly active antiretroviral therapy by HIV-positive women attending an antenatal clinic in a regional hospital in western Uganda.J Int AIDS Soc 2010;13(1):37. https://doi.org/10.1186/1758-2652-13-37 11. Muchedzi A, Chandisarewa W, Keatinge J, et al. Factors associated with access to HIV care and treatment in a prevention of mother to child transmission programme in urban Zimbabwe. J Int AIDS Soc 2010;13(1):38. https://doi.org/10.1186/1758-2652-13-38 12. World Health Organization. Consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV infection – recommendations for a public health approach. Geneva: WHO, 2016. http://www.who.int/hiv/pub/arv/arv-2016/en (accessed 10 March 2017). 13. McLeroy KR, Bibeau D, Steckler A, Glanz K. An ecological perspective on health promotion programs. Health Educ Q 1988;15(4):351-377. https://doi.org/10.1177/109019818801500401 14. Robinson T. Applying the socio-ecological model to improving fruit and vegetable intake among low-income African Americans. J Community Health 2008;33(6):395-406. https://doi.org/10.1007/ s10900-008-9109-5 15. Hosmer DW, Lemeshow S. Applied Logistic Regression. New Jersey: John Wiley and Sons Inc., 2004. 16. Hilbe JM. Logistic Regression Models. Florida; Chapman & Hall/CRC; 2009. 17. Encyclopedia Britannica. Zionist Church. http://www.britannica.com/EBchecked/topic/657514/ Zionist-church (accessed 7 November 2013).

18. UNAIDS. Together we will end AIDS.2012. [cited 10 November 2013]. Available from: http://www. unicef.org/aids/files/aids__togetherwewillendaids_en.pdf (accessed 10 November 2013). 19. The Inter Agency Task Team on the Prevention and Treatment of HIV Infection In Pregnant Women, Mother and Children. Key Data Points for Priority Countries http://www.emtct-iatt.org/ priority-countries (accessed 10 November 2013). 20. Killam WP, Tambatamba BC, Chintu N, et al. Antiretroviral therapy in antenatal care to increase treatment initiation in HIV-infected pregnant women: A stepped-wedge evaluation. AIDS 2010;24(1):85-91. https://doi.org/10.1097/qad.0b013e32833298be 21. Stinson K, Boulle A, Coetzee D, Abrams EJ, Myer L. Initiation of highly active antiretroviral therapy among pregnant women in Cape Town, South Africa. Trop Med Int Health 2010;15(7):825-832. https://doi.org/10.1111/j.1365-3156.2010.02538.x 22. Chen JY, Ogwu AC, Svab P, et al. Antiretroviral treatment initiation among HIV-infected pregnant women with low CD4(+) cell counts in Gaborone, Botswana. J Acquir Immune Defic Syndr 2010;54(1):102-106. https://doi.org/10.1097/qai.0b013e3181c080bf 23. Weigel R, Hosseinipour MC, Feldacker C, et al. Ensuring HIV-infected pregnant women start antiretroviral treatment: an operational cohort study from Lilongwe, Malawi. Trop Med Int Health 2012;17(6):751-759. https://doi.org/10.1111/j.1365-3156.2012.02980.x 24. Barigye H, Levin J, Maher D, et al. Operational evaluation of a service for prevention of motherto-child transmission of HIV in rural Uganda: Barriers to uptake of single-dose nevirapine and the role of birth reporting. Trop Med Int Health 2010;15(10):1163-1171. https://doi.org/10.1111/j.13653156.2010.02609.x 25. Karcher H, Kunz A, Poggensee G, Mbezi P, Mugenyi K, Harms G. Outcome of different nevirapine administration strategies in preventing mother-to-child transmission (PMTCT) programs in Tanzania and Uganda. Med Gen Med 2006;8(2):12. https://doi.org/10.1186/1758-2652-8-2-12 26. Kuonza LR, Tshuma CD, Shambira GN, Tshimanga M. Non-adherence to the single dose nevirapine regimen for the prevention of mother-to-child transmission of HIV in Bindura town, Zimbabwe: A cross-sectional analytic study. BMC Public Health 2010;10(1):218. https://doi.org/10.1186/14712458-10-218 27. Feldacker C, Johnson D, Hosseinipour M, Phiri S, Tweya H. Who starts? Factors associated with starting antiretroviral therapy among eligible patienst in two public HIV clinics in Lilongwe, Malawi. PLoS One 2012;7(11):e50871. https://doi.org/10.1371/journal.pone.0050871 28. Kirsten I, Sewangi J, Kunz A, et al. Adherence to combination prophylaxis for prevention of motherto-child transmission of HIV in Tanzania. PLoS One 2011;6(6):e21020. https://doi.org/10.1371/ journal.pone.0021020 29. Kinuthia J, Kiarie JN, Farquhar C, et al. Uptake of prevention of mother to child transmission interventions in Kenya: Health systems are more influential than stigma. J Int AIDS Soc 2011;14(1):61. https://doi.org/10.1186/1758-2652-14-61 30. Albrecht S, Semrau K, Kasonde P, et al. Predictors of nonadherence to single-dose nevirapine therapy for the prevention of mother to child HIV transmission. J Acquir Immune Defic Syndr 2006;41(1):114-118. https://doi.org/10.1097/01.qai.0000179425.27036.d7 31. Stringer JSA, Sinkala M, Stout JP, et al. Comparison of two strategies for administering nevirapine to prevent perinatal HIV transmission in high-prevalence, resource-poor settings. J Acquir Immune Defic Syndr 2003;32(5):506-513. https://doi.org/10.1097/00126334-200304150-00007 32. Msuya SE, Mbizvo EM, Hussain A, Uriyo J, Sam NE, Stray-Pedersen B. Low male partner participation in antenatal HIV counselling and testing in northern Tanzania: Implications for preventive programs. AIDS Care 2008;20(6):700-709. https://doi.org/10.1080/09540120701687059 33. Peltzer K, Mosala T, Dana P, Fomundam H. Follow-up survey of women who have undergone a prevention of mother-to-child transmission program in a resource-poor setting in South Africa. J Assoc Nurses AIDS Care 2008;19(6):450-460. https://doi.org/10.1016/j.jana.2008.05.006 34. Kalembo FW, Zgambo M, Mulaga AN, Yukai D, Ahmed NI. Association between male partner involvement and the uptake of prevention of mother to child transmission of HIV (PMTCT) interventions in Mwanza district, Malawi. PLoS One 2013;8(6):e66517. https://doi.org/10.1371/ journal.pone.0066517 35. Farquhar C, Kiarie JN, Richardson BA, et al. Antenatal couple counseling increases uptake of interventions to prevent HIV-1 transmission. J Acquir Immune Defic Syndr 2004;37(5):1620-1626. https://doi.org/10.1097/00126334-200412150-00016 36. Kasenga F, Hurtig AK, Emmelin M. HIV-positive women’s experiences of a PMTCT programme in rural Malawi. Midwifery 2010;26(1):27-37. https://doi.org/10.1016/j.midw.2008.04.007 37. Theilgaard ZP, Katzenstein TL, Chiduo MG, et al. Addressing the fear and consequences of stigmatization – a necessary step towards making HAART accessible to women in Tanzania: A qualitative study. AIDS Res Ther 2011;8(1):28. https://doi.org/10.1186/1742-6405-8-28 38. O’Gorman DA, Nyirenda LJ, Theobald SJ. Prevention of mother-to-child transmission of HIV infection: views and perceptions about swallowing nevirapine in rural Lilongwe, Malawi. BMC Public Health 2010;10(1):354. https://doi.org/10.1186/1471-2458-10-354 39. Peltzer K, Sikwane E, Majaja M. Factors associated with short-course antiretroviral prophylaxis (dual therapy) adherence for PMTCT in Nkangala district, South Africa. Acta Paediatr 2011;100(9):1253-1257. 40. Kiarie JN, Kreiss JK, Richardson BA, John-Stewart GC. Compliance with antiretroviral regimens to prevent perinatal HIV-1 transmission in Kenya. AIDS 2003;17(1):65-71. 41. Mepham S, Zondi Z, Mbuyazi A, Mkhwanazi N, Newell ML. Challenges in PMTCT antiretroviral adherence in northern KwaZulu-Natal, South Africa. AIDS Care 2011;23(6):741-747. 42. Painter TM, Diaby KL, Matia DM, et al. Women’s reasons for not participating in follow up visits before starting short course antiretroviral prophylaxis for prevention of mother to child transmission of HIV: Qualitative interview study. Br Med J 2004;329(7465):543-546. 43. Levy JM. Women’s expectations of treatment and care after an antenatal HIV diagnosis in Lilongwe, Malawi. Reprod Health Matters 2009;17(33):152-161. 44. Varga C, Brookes H. Factors influencing teen mothers’ enrollment and participation in prevention of mother-to-child HIV transmission services in Limpopo Province, South Africa. Qual Health Res 2008;18(6):786-802. 45. Chinkonde JR, Sundby J, Martinson F. The prevention of mother-to-child HIV transmission programme in Lilongwe, Malawi: Why do so many women drop out? Reprod Health Matters 2009;17(33):143-151.

SAJOG • September 2017, Vol. 23, No. 2

This open-access article is distributed under Creative Commons licence CC-BY-NC 4.0.

REVIEW

Termination of pregnancy for fetal anomaly in a Tunisian population E B Hamida,1 MD; I Ayadi,1 MD; A Bezzine,1 MD; B Rabii,1 MD; S B Hammouda,2 MD; B Bouguerra,2 MD; Z Marrakchi,1 MD 1 2

Department of Neonatology, Charles Nicolle Hospital, Faculty of Medicine of Tunis, Tunis El Manar University, Tunisia Department of Obstetrics and Gynaecology, Charles Nicolle Hospital, Faculty of Medicine of Tunis, Tunis El Manar University, Tunis, Tunisia

Corresponding author: E B Hamida (emira@ben-hamida.com)

Background. Progress in prenatal diagnosis and prenatal screening in the past 2 decades has facilitated improved diagnostic methods for fetal abnormalities. Objectives. To assess the indications for late termination of pregnancy, gestational ages and to discuss ethical and legal considerations. Methods. A retrospective study compiling and analysing data from justified medical terminations of pregnancies, reported between January 2001 and June 2012 at Charles Nicolle Hospital. Congenital infectious embryofetopathies were excluded. Results. Over this period, there were 77 cases of clinically justified medical terminations of pregnancies – a rate of 1.9 per 1 000 total births. The mean gestational age was 27 weeks. In 33 cases, the medical abortion was performed beyond 26 weeks (42.8%). Fetal malformations were predominantly of the central nervous system. Conclusion. Medical terminations of pregnancy for fetal indications are relatively rare. However, they raise ethical, medical and legal issues, due in part to a late prenatal diagnosis. Such decisions should be taken through a multidisciplinary committee. S Afr J Obstet Gynaecol 2017;23(2):69-70. DOI:10.7196/SAJOG.2017.v23i2.1159

Advances in prenatal diagnosis allow the screening of various fetal malformations during pregnancy. In the absence of therapeutic options, the possibility of termination of pregnancy (TOP) for fetal anomaly is more and more likely to be considered. This raises many religious, social, and ethical questions.[1-3] Since 1973, Tunisian law has permitted TOP under two circumstances: voluntary and elective. Abortion is permissible prior to 12 weeks of gestation in cases of unwanted pregnancy. Pregnancy may be terminated at any period of gestation if there is a serious maternal health problem, including psychological conditions and/or fetal conditions with a high risk of severe disability or fatality. Approval for elective TOP can be given by one physician who declares that there is a maternal or fetal reason justifying the TOP. There is no upper gestation limit to late TOP stated in Tunisian legislation.[4,5] The aim of our study was to assess the indications of late TOP, gestational ages and to discuss ethical and legal considerations.

Results

Over the 11-year period there were 77 cases of TOP for fetal anomalies among 39 943 total births, giving a rate of 1.9 per 1 000 births. Forty-four cases (62.3%) were stillbirths. Indications for TOPs owing to fetal anomalies are summarised in Fig. 1. Central nervous system defects represented 53% of all indications, followed by chromosomal anomalies (21%). Anencephaly and spina bifida were responsible for 33.7% of all TOPs. Chromosomal anomalies included 10 cases of trisomy 21 (62.5%), 2 cases of trisomy 18 (12.5%), 3 cases of trisomy 13 (18.7%) and 1 case of a deletion of the short arm of chromosome 18. All fetal anomalies were suspected by antenatal ultrasound, except for a case of β-thalassaemia major. The first antenatal ultrasound was performed in the first trimester in 37 mothers (48%), in the second trimester in 26 mothers (33.7%) and in the third trimester in 14 mothers (18.1%).

Methods

This was a retrospective study conducted between January 2001 and June 2012 at the Department of Neonatology at Charles Nicolle Hospital (CNH) of Tunis, Tunisia, which is a referral centre for high-risk pregnancies. There are 3 000 - 4 000 births every year at CNH from an average of 170 000 annual births in Tunisia. Inclusion criteria were all births from a TOP at 22 or more weeks of gestation or weighing ≥500 g when gestational age was unknown. Births from medical terminations of pregnancy owing to congenital infectious embryofoetopathies were excluded. Data were obtained from the clinical records of the mother and the neonate. The data were recorded and analysed using Epi Info 6.04d for DOS (Centers for Disease Control and Prevention, USA). Descriptive parameters are presented as mean (SD) and frequencies are presented as percentages.

Central nervous system Genitourinary system

Skeletal system 53

Cardiovascular system Chromosomal diseases

Other 6

Fig. 1. Indications for termination of pregnancy (%).

SAJOG • September 2017, Vol. 23, No. 2

REVIEW Fetal anomalies were detected at a mean (SD) gestational age of 24 (3.2) weeks. No cases were suspected in the first trimester, 58 cases (75.3%) were recognised between 18 and 26 weeks, and 19 cases (24.6%) between 26 and 36 weeks. The mean (SD) gestational age at termination was 27 (3.5), (range 22 - 36) weeks. In 33 cases (42.8%), TOP was performed beyond 26 weeks and 15/33 (45.4%) were born alive. Among the infants born alive beyond 26 weeks, 10/15 (66.6%) died within the first hours after birth, 2 died between 2 and 7 days, one at day 17, and 2 at day 23. The 2 latter patients died after discharge: 1 neonate, born at 26 weeks, had Down’s syndrome, and the other, born at 30 weeks, had hydrocephaly. No neonatal intensive care was offered to the newborns.

Paediatricians may encounter legal problems if they decide not to resuscitate the newborn. In our study, TOP was performed beyond 26 weeks in more than 45% of cases. Two infants were even discharged before dying a few days later. Thus, before deciding on the means of terminating the pregnancy, it is important to define whether the fetus will be born alive, in which case feticide would have to be discussed. Without feticide, an infant may be born alive; however, the majority of obstetricians refuse to practise it for religious reasons. In such situations, TOP should be undertaken after careful discussion between attending obstetric and neonatal staff, and the woman and her family, with all parties agreeing to a written care plan before the termination takes place.

Discussion

Conclusion

Our study demonstrated a majority of central nervous system defects as indications of TOP for fetal anomaly. This pattern was consistent with previous studies.[1] Chromosomal abnormalities were the second most common indication for TOP, and trisomy 21 was the most common chromosomal abnormality as found by other studies. In our practice there is a lack of consensus around which abnormalities are severe enough to indicate termination, and up to what gestational age TOP is acceptable. The law allows TOP for fetal abnormality after the first trimester if the abnormality is ‘serious’, but there is a wide variation in the views of practitioners. Some obstetricians regard a cleft palate as a major abnormality, while some believe that TOP should be considered for only lethal conditions such as anencephaly and trisomy 18. Late TOP for achondroplasia, hydrocephaly or Down’s syndrome could not be justified since they are viable abnormalities, but not lethal.[3] If the anomaly is not lethal, the paediatrician may be confronted by a difficult dilemma if the fetus is born alive and requires resuscitation or intensive care. In Tunisia, viability is considered common beyond 26 weeks. From there, the fetus is regarded as a patient who is entitled to the legal protection available to any other child.

Medical TOPs for fetal indications are relatively rare. However, they raise ethical, medical and legal issues, due in part to a late prenatal diagnosis. Such decisions should be taken through a multidisciplinary committee containing at least an obstetrician, neonatologist and a geneticist. Acknowledgement. All authors. Author contributions: All authors contributed equally to the work presented in this paper. Funding. None. Conflicts of interest. None.

1. Aslan H, Yildirim G, Ongutb C, Ceylan Y. Termination of pregnancy for fetal anomaly. Int J Gynecol Obstet 2007;99:221-224. https//doi.org/10.1016/j.ijgo.2007.05.047 2. Gitsels-van der Wal JT, Manniën J, Ghaly MM, Verhoeven PS, Hutton EK, Reinders HS. The role of religion in decision-making on antenatal screening of congenital anomalies: A qualitative study amongst Muslim Turkish origin immigrants. Midwifery 2014;30:297-302. https://doi. org/10.1016/j.midw.2013.04.001 3. Statham H, Solomou W, Green J. Late termination of pregnancy: Law, policy and decision-making in four English fetal medicine units. BJOG 2006;113:1402-1411. https://doi.org/10.1111/j.14710528.2006.01144.x 4. Décret-Loi n° 73-2 du 26 Septembre 1973 portant modification de l'article 214 du code pénal. http://legislation-securite.tn/node (accessed 21 August 2017). 5. Blum J, Hajri S, Chelli H, Ben Mansour F, Gueddana N, Winikoff B. The medical abortion experiences of married and unmarried women in Tunis, Tunisia. Contraception 2004;69:63-69. https://doi.org/10.1016/j.contraception.2003.08.019

SAJOG • September 2017, Vol. 23, No. 2

This open-access article is distributed under Creative Commons licence CC-BY-NC 4.0.

CASE REPORT

Bucket-handle tear of posterior uterine cervical lip in a second-trimester unscarred uterus after use of misoprostol: A first report of two cases S R Singhal,1 MD; S K Singhal,2 MD 1 2

Department of Obstetrics and Gynaecology, Pandit Bhagwat Dayal Sharma Post Graduate Institute of Medical Sciences, Rohtak, Haryana, India Department of Anaesthesiology, Pandit Bhagwat Dayal Sharma Post Graduate Institute of Medical Sciences, Rohtak, Haryana, India

Corresponding author: S R Singhal (savita06@gmail.com)

A bucket-handle tear is a laceration of either the anterior lip or the posterior lip of the cervix so that it hangs like the handle of a bucket. These cervical injuries are more common in term deliveries and are associated with various risks factors, including cervical cerclage, induction of labour, young maternal age, assisted vaginal delivery, prostaglandin use and precipitate labour. We came across two patients who experienced a bucket handle tear of the posterior lip of the cervix in second-trimester abortions due to prostaglandin E1 use. Such a case has not been reported in the literature. In this case study, we report two such cases. S Afr J Obstet Gynaecol 2017;23(2):71-72. DOI:10.7196/SAJOG.2017.v23i2.1177

Cervical injuries are not uncommon during labour and abortion. Reported cervical injuries include cervical lacerations, lateral cervical tears, bucket-handle tears and annular detachment of the cervix.[1] A bucket-handle tear is a laceration of the anterior or the posterior lip of the cervix so that it hangs like the handle of a bucket. Although these cervical injuries are more common in term deliveries, alternative risk factors associated with cervical injuries are cervical cerclage, induction of labour, young maternal age, vacuum-assisted vaginal delivery, prostaglandin use and precipitate labour.[2,3] We examined two patients who each had a bucket-handle tear of the posterior lip of cervix in their secondtrimester abortions due to the use of prostaglandin E1. This has not been reported in literature.

Case 2

A 23-year-old primigravida presented at 18 weeks' gestation for medical termination of pregnancy for fetal hydrocephalus and meningomyelocele. She was given 200 µg of mifepristone vaginally and after 48 hours, 400 µg of misoprostrol was administered vaginally at 4-hourly intervals. After four doses of misoprostol, the cervix was still one-finger dilated and 50% effaced. Two more

Case 1

A 26-year-old G4P1A2 (G = no. of pregnancies; P = no. of delivered pregnancies; A = no. of abortions) at 19 weeks of pregnancy was admitted with a complaint of abdominal pain and 2 days of bleeding per vaginam. She had had one full-term vaginal delivery without any complications 3 years previously, and two spontaneous miscarriages at 2 months gestation, which had not been followed by evacuation. In this case, on examination the uterus was 22 weeks in size, tense, tender and fetal parts were not palpable. Ultrasound revealed a dead fetus of 18 weeks' gestation with a retroplacental clot of 10 × 9 cm. Augmentation of abortion was undertaken with 200 µg of misoprostol administered vaginally at 4-hour intervals. The patient aborted a 300 g abortus with 300 mL of blood clots present posterior to the placenta after four doses of misoprostol. The patient bled excessively after the abortion. On per vaginam and bimanual examination the uterus was well contracted. The cervix was examined in detail. The cervical os was found to be only 2 cm dilated and there was a 4 cm tear in the posterior cervical lip, which was bleeding, suggesting that the patient had aborted through the tear (Fig. 1). The defect was repaired with chromic catgut sutures. The patient had an uneventful recovery following the abortion.

Fig.1. Transverse cervical tear on posterior lip of cervix.

SAJOG • September 2017, Vol. 23, No. 2

CASE REPORT doses were given and the patient aborted a 250 g, dead, congenitally malformed abortus after six doses of misoprostol. The placenta was expelled completely. After the abortion, the patient had excessive bleeding. A per vaginam examination was done, which revealed that the cervical os was ~2 cm dilated and there was a 3 cm cervical tear on the posterior lip of the cervix, which was bleeding and through which the fetus had possibly aborted. The tear was stitched with chromic catgut sutures.

uterine rupture â&#x20AC;&#x201C; another serious complication of the unyielding cervix. There are clear-cut guidelines for managing a rigid cervix in term pregnancies, but this aspect of cervical stenosis is not taken into account at lower gestational ages. For mid-trimester abortions, further doses of prostaglandins should be withheld if the cervix is not dilating in spite of good uterine contractions to avoid unusual cervical tears.

Discussion

Conclusion

The cause of a bucket-handle tear may be an unyielding cervix due to fibrosis or stenosis, leading to pressure being transferred either to the anterior or posterior lip of the cervix. An added factor may be hypertonus caused by prostaglandins.[4] Cervical lacerations are common at term pregnancy. In the first case, the patient had a history of two abortions and one delivery. This led to fibrosis of the cervix, causing pressure on the posterior cervix which yielded to the uterine contractions, leading to the bucket-handle tear. In the second case, the patient was primigravida and the cervix was only one finger dilated after four doses of misoprostol despite good uterine contractions. The cervix might have been congenitally stenosed, causing it not to yield to uterine contractions, or there may have been a history of unconfessed early pregnancy losses. In both patients, the cervical os was 2 cm dilated after the abortion, favouring the aetiology of a rigid cervix. The patients most likely aborted through the tear in the posterior lip of the cervix. The patients were fortunate that contractions did not eventually lead to

In second-trimester abortions induced with prostaglandins, if the cervix is not yielding despite good uterine contractions, there is an option to withhold further doses to avoid cervical injuries. Acknowledgement. None. Author contributions. Both authors jointly carried out the study, compiled, analysed and prepared the manuscript. Funding. None. Conflict of interest. None.

1. Parikh R, Brotzman S, Anasti JN. Cervical laceration: Some surprising facts. Am J Obstet Gynecol 2007;196(5):e17-e18. https://doi.org/10.1016/j.ajog.2006.11.043 2. Melamed N, Ben-Haroush A, Chen R, Kaplan B, Yogev Y. Intrapartum cervical lacerations: Characteristics, risk factors, and effect on subsequent pregnancies. Am J Obstet Gynecol 2009;200(4):388.e1-4. https://doi.org/10.1016/j.ajog.2008.10.034 3. Landy HJ, Laughon SK, Bailit JL, Kominiarek MA, Gonzalez-Quintero VH, Ramirez M. Characteristics associated with severe perineal and cervical lacerations during vaginal delivery. Obstet Gynecol 2011;117(3):627-635. https://doi.org/10.1097/aog.0b013e31820afaf2 4. Premi HK, Sood R, Gupta T, Sood A. Bucket handle tear of cervix following prostaglandin instillation. The Antiseptic 1998;95(6):191.

SAJOG â&#x20AC;˘ September 2017, Vol. 23, No. 2

CPD

True (A) or false (B): Interval laparoscopic sterilisation 1. Monogamy post sterilisation may diminish the risk of sexually transmitted infections. Perinatal outcomes in preterm prelabour rupture of membranes (PPROM) 2. Worldwide, the incidence of PPROM is between 5% and 10% of all pregnancies. 3. In the study included in this journal, there were no differences in perinatal outcomes between HIV-positive and HIV-negative individuals. 4. A number of African studies have demonstrated no difference in gestation at delivery, birthweight and Apgar scores between HIV-positive and HIV-negative individuals. 5. Currently, standard fixed-dose combination antiretroviral therapy in South Africa consists of tenofovir, nevirapine and efavirenz.

11. A ccording to the American Society of Reproductive Medicine, routine evaluation of the infertile couple should include a post-coital test. 12. The prevalence of unexplained infertility remains debatable owing to the lack of dedicated testing measures. 13. The study included in this journal showed a difference in natural killer cells at different times of the menstrual cycle, but no difference between controls and those with unexplained infertility. Requesting caesarean section in Dar es Salaam, Tanzania 14. A previous study at a tertiary referral hospital in Tanzania reported that the incidence of ‘near-miss’ events associated with caesarean section was 3 - 7 per 1 000 procedures. 15. The previously recorded caesarean section rate at the national referral centre in Tanzania was 25%.

Peritoneum at laparotomy 6. The survey described here was performed at the 50th anniversary conference of the Society of Gynaecology and Obstetrics of Nigeria. 7. A meta-analysis of clinical trials in a 2014 Cochrane review questioned change in practice without looking at the longterm sequelae of non-closure.

Termination of pregnancy in Tunisia 16. Since 1973, termination of pregnancy has been permissible in Tunisia for unwanted pregnancy prior to 12 weeks of gestation. 17. In Tunisia, termination of pregnancy may be performed at any gestation due to serious maternal conditions, including psychological, and for fetal anomalies producing severe disability or incurable disease.

Endometrial natural killer cells in unexplained infertility 8. The World Health Organization (WHO) currently defines infertility as the failure to conceive after 18 months of regular sexual intercourse. 9. Repeated implantation failure is defined as failure to achieve a pregnancy after 2 - 6 in vitro fertilisation (IVF) cycles. 10. The National Institute for Health and Care Excellence (NICE) in the UK has defined infertility as a failure to conceive after 3 years of regular sexual intercourse.

Bucket-handle tear of the posterior cervix 19. Misoprostol is a prostaglandin E2 analogue. Lack of antiretroviral therapy (ART) initiation in HIVpositive patients in Swaziland 19. The most common cited reason for not initiating ART (24.5%) was not being ready to start lifelong therapy. 20. Partner support was independently associated with initiation of ART.

The CPD programme for SAJOG is administered by Medical Practice Consulting: Effective in 2014, the CPD programme for SAJOG will be administered by Medical Practice Consulting: CPD questionnaires must be completed online at www.mpconsulting.co.za CPD questionnaires must be completed online at www.mpconsulting.co.za A maximum ofwill 3 CEUs will be awarded per correctly completed test. A maximum of 5 CEUs be awarded per correctly completed test. Accreditation MDB015/173/02/2017(Clinical) Accreditation number:number: MDB015/178/02/2016(Clinical)

SAJOG • September 2017, Vol. 23, No. 2