Impact on perinatal mortality of missed opportunities to treat maternal syphilis in rural South Africa: baseline results from a clinic randomized controlled trial
correspondence Professor David Wilkinson, South Australian Centre for Rural and Remote Health, University of South Australia–Whyalla Campus, Nicolson Avenue, Whyalla Norrie SA 5608, Australia. E-mail: email@example.com
Summaryobjective To demonstrate the impact on perinatal mortality of inadequate treatment for maternal syphilis despite adequate screening.
method In 12 clinics providing antenatal care in Hlabisa, South Africa 1783 pregnant women were screened for syphilis at their first antenatal visit between June and October 1998. Pregnancy outcome was determined among those with syphilis.
results A total of 158 women were diagnosed with syphilis: prevalence 9% (95% CI 8–10%). Mean gestation at first antenatal visit was 24 weeks. Thirty women (19%) received no treatment and 96 (61%) received all three recommended doses of penicillin. Among those receiving at least one dose, mean delay to the first dose was 20 days. Among those fully treated mean delay to treatment completion was 34 days. Pregnancy outcome was known for 142 women (90%) and there were 17 perinatal deaths among 15 women (11%). Eleven of 43 women (26%) who received one or fewer doses of penicillin experienced a perinatal death whilst only four of 99 women (4%) who received two or more doses of penicillin did so (P = 0.0001). Protection from perinatal death increased with the number of doses of penicillin: linear modelling suggests that one dose reduced the risk by 41%, two doses by 65% and three doses by 79%, compared with no doses. A dose-specific, categorical model confirmed reduction in risk by 79% for all three doses.
conclusion Despite effective screening, many pregnant women with syphilis remain inadequately treated, resulting in avoidable perinatal mortality. Delays in starting and finishing treatment, as well as incomplete treatment occur. Near-patient syphilis testing in the antenatal clinic with early treatment could improve treatment of syphilis and reduce perinatal mortality, and a randomized trial to test this is underway.
Maternal syphilis in pregnancy continues to be an important and avoidable cause of pregnancy loss in developing countries (Guinness et al. 1988; Lindstrand et al. 1993; Pattinson et al. 1995). The risk of perinatal mortality among women with syphilis is at least twice that amongst those without (Bam et al. 1994). Syphilis prevalence among pregnant women ranges from 6 to 19% in different parts of Africa (Ratnam et al. 1982; Guinness et al. 1988; Lindstrand et al. 1993; Swingler & Van Coeverden de Groot 1993; Bam et al. 1994; Jenniskens et al. 1995; Aseffa et al. 1998), and syphilis is also re-emerging as a serious concern in the newly independent states of the former Soviet Union (Tichonova et al. 1997) and in disadvantaged parts of the United States (CDC 1998).
Syphilis is treatable and screening pregnant women at their first antenatal visit is recommended by the World Health Organization (WHO 1991). However, effective syphilis screening does not always ensure adequate treatment. Previous work in rural South Africa demonstrated that less than half of the identified cases of maternal syphilis in a mobile primary care clinic were fully treated (Wilkinson et al. 1997a). These findings are supported by other work documenting, for example, adequate treatment in only 20% of women diagnosed with syphilis during pregnancy (Ballot & Rothberg 1993; Bam et al. 1994), and previous work in rural South Africa has shown that testing for syphilis in the community antenatal clinics (near-patient testing) as opposed to the district hospital laboratory is accurate, cheap and associated with improved syphilis treatment (Wilkinson & Sach 1998). Here we report the baseline findings from a clinic randomized controlled trial designed to test the impact of near-patient testing for syphilis on perinatal mortality.
The study took place in Hlabisa health district KwaZulu/ Natal, South Africa. This district is home to 220 000 people and health facilities include a 450-bed hospital, 12 community clinics and two mobile clinic teams.
In 1997, 5368 women received antenatal care at the clinics in Hlabisa district. These women seem to be representative of pregnant women in the community, as a survey carried out in the district in 1996 estimated that 95% of women received antenatal care and that 83% delivered in a health facility (Wilkinson et al. 1997b). In 11 of the 12 clinics, as a part of routine antenatal care, all women have blood drawn at their first visit and are advised to return for the results of syphilis testing 2 weeks later. Blood taken in the clinic is sent to the district hospital laboratory where serum is tested for syphilis using a rapid plasma reagin (RPR) test. In the 12th antenatal clinic, sited at the district hospital, blood is processed the same way and RPR results are available the same day, although not all women wait to receive them. Women who test positive for syphilis in the antenatal clinics are counselled, treated and asked to bring their partners for treatment. Treatment consists of three weekly intramuscular doses of 2.4 mega-units of benzathine penicillin (Department of Health South Africa 1995).
Between June and October 1998 all women attending the 12 clinics completed a questionnaire documenting demographic details, location of residence and obstetric history. For women with syphilis, the dates of subsequent clinic attendances for results and treatment were recorded. At each visit women were asked if their midwife had informed them of the importance of partner notification, whether they had informed their partner, and whether their partner had been treated. Pregnancy outcome was determined from hospital and clinic records and by home visits if necessary. The University of Natal Medical School Ethics Committee gave ethical approval for this study.
Definitions used throughout the paper are as follows:
Complete syphilis treatment: three doses of penicillin at weekly intervals;
Adequate syphilis treatment: at least two doses of penicillin at weekly intervals (Wilkinson et al. 1997a; Wilkinson & Sach 1998);
Inadequate treatment: one dose or less of penicillin received;
Stillbirth (SB): born dead at or after 28 weeks gestation;
Early neonatal death (ENND): born alive, but died during first week of life;
Perinatal death: stillbirth or early neonatal death.
Syphilis tests were performed on serum separated from whole blood within 48 h of venesection, blood being stored at 4 °C prior to serum separation. Serum was qualitatively tested for syphilis infection by means of the RPR test (Macro-Vue RPR Card Test, Becton Dickinson, USA) and results were reported as negative or positive. Standard quality control measures were used during the study period. Although using the RPR test only will tend to overestimate the prevalence of early latent syphilis, in high prevalence settings the number of false positive results arising from this can be expected to be low.
The primary endpoint was perinatal outcome in the mother (we did not model outcome in individual babies, as we did not have data on how many live births were twin pregnancies). Generalized estimating equations (GEE) were used to take into account the potential design effect of the clinic cluster sampling, in order to study the association between doses of penicillin given and demographic, health seeking and health care variables.
To model the relative risk of perinatal mortality on the number of doses received, adjusted for confounding variables (age and gravidity), a Cox regression model with Breslow conditioning (Lee & Chia 1993) was used. The estimated standard errors from the model were adjusted by the design effect for perinatal outcome. The 95% confidence intervals reported are based on the adjusted standard errors and reflect the adjustment for cluster sampling. Since the data set is quite small, extensive modelling was not possible and smoothing was used to overcome this problem. We present the results of two models, a categorical model with a parameter for each does of penicillin used and a linear model, assuming a linear relationship across doses of penicillin.
Between June and October 1998, all 1783 women who received antenatal care were screened for syphilis. Of these, 158 tested RPR-positive and syphilis prevalence was estimated as 9% (95% CI: 8–10%).
The mean age of the 158 women with syphilis was 25 years (range 15–42 years) and their mean gravidity was 2.8 (range 1–11). Mean gestational age at their first antenatal visit was 24 weeks (range 8–44 weeks), and a history of a previous perinatal death was reported by eight (7%) of the 115 previously pregnant women.
Of the 158 women, 30 (19%) received no treatment for their syphilis, 20 (13%) received one penicillin dose, 12 (7%) received two doses and 96 (61%) received all three recommended doses. Thus, 50 (32%) women were considered inadequately treated (receiving less than two doses of penicillin). The mean number of doses received was 2.1.
Treatment delays and partner treatment
Among the women returning for treatment at least once, the mean delay from diagnosis to first dose of penicillin was 20 days. Among women returning at least twice mean delay from diagnosis to the second dose was 27 days, and among those who completed treatment, the mean time to completion was 34 days. Information on partner treatment was available for 70 of the 91 women who attended for three or more antenatal visits. In all, 78 (86%) reported having been given a contact card by a midwife. Of these, 70 (77%) reported informing their partner of the need for treatment but only 24 (26%) reported being sure the partner had received it.
Data on pregnancy outcomes were available for 142 (90%) women and place of delivery was known for 137; 113 (83%) delivered in a health care facility and the rest delivered at home. There were 17 perinatal deaths among 15 (11%) women; six women experienced stillbirths and nine experienced early neonatal deaths. Mean gestation at delivery in those with a perinatal death was 37 weeks.
Association between syphilis treatment and perinatal outcome
Of the 43 women who were inadequately treated for syphilis and among whom pregnancy outcome was known, 11 (26%) experienced a perinatal death compared to four among the 99 (4%) adequately treated women (P < 0.0001, Table 1).
Comparison of women inadequately and adequately treated for syphilis, adjusted for clustering
The risk of death was reduced with increasing numbers of doses of penicillin received (Table 2). Adjusted for possible confounding variables and for clustering and assuming a linear trend for dose, the estimated risk reduction for one dose of penicillin was 41% (95% CI 2–64%), for two doses it was 65% (42–79%) and for three doses, 79% (66–88%). Without adjustment for confounding variables, point estimates were 45, 69 and 83%, respectively. The estimated risk reduction for three doses of penicillin using this model was consistent with the estimate derived from a more general dose-specific model (Table 3).
Association between number of penicillin doses received and perinatal outcome
Comparison of the two perinatal mortality models used: model 1 (categorical model with a parameter for each dose of penicillin received) and model 2 (assumption of linear function across doses)
Our data demonstrate that despite successful screening syphilis-attributable perinatal mortality continues to occur among pregnant women attending these clinics. This is because of delays in starting and completing treatment, and because of incomplete courses of treatment. Near-patient testing for syphilis, done in the antenatal clinic by the attending midwife, with immediate therapy for those infected, offers an opportunity to improve pregnancy outcome.
All pregnant women in this typical rural health district who attended for antenatal care and who were diagnosed with syphilis were included in this study, and pregnancy outcome was known for 90%. A possible weakness of the study is the exclusive use of the RPR test to diagnose syphilis. However, the sensitivity and specificity of the test are high (Delport 1993) and in high prevalence settings few false positive results will occur. We chose to use the RPR test alone as this reflects real practice in such settings, and hence our findings may be applicable more broadly.
Poor management of maternal syphilis is not unique to Hlabisa (Ali 1990; Ballott & Rothberg 1993; Rutgers 1993; Swingler & Van Coeveren De Groot 1993; Bam et al. 1994; Fitzgerald et al. 1998). While antenatal syphilis screening may be done well, the proportion of women adequately treated is often unacceptably low (Delport 1993; Swingler & Van Coeveren De Groot 1993; Bam et al. 1994), resulting in substantial morbidity and mortality (Ali 1990; Swingler & Van Coeveren De Groot 1993; Pattinson et al. 1995; Fonn 1996). The challenge therefore is to increase the likelihood of receiving adequate treatment once syphilis has been diagnosed. Decentralizing the syphilis screening process and implementing near-patient RPR testing provides an opportunity to start treatment immediately. This would reduce the time taken to start treatment and should increase the mean number of doses received per woman (Wilkinson & Sach 1998). Furthermore, with an immediate diagnosis made and the first dose given, both midwife and patient can focus on strategies to ensure the second dose is given 1 week later. Our data presented here and previously (Wilkinson & Sach 1998) suggests that two doses of penicillin provide good protection against perinatal death. However, every effort should be made to give all three recommended doses.
Important ethical issues are raised by trials such as this. Previous work and these baseline data demonstrate high levels of inadequate therapy. First, this study was approved by the local Ethics Committee of the University of Natal Medical School and the local health service, conducted by the South African Medical Research Council, and funded by the Wellcome Trust. All these organizations have the protection of research subjects as their highest priority. The syphilis status of all women routinely attending the clinics in Hlabisa was documented from health service records, as were their subsequent clinic attendance patterns and pregnancy outcome. We carefully documented routine service activity and outcome. All women with syphilis were treated at delivery if found not to have been treated during pregnancy. It was not feasible to trace women who did not return for results at the recommended interval as the population is highly mobile, but all women were given clear written and verbal instruction on when to return for results.
Near-patient testing, using RPR rotators, has already been shown to be both feasible and economic in Kenya (Jenniskens et al. 1995) and sufficiently sensitive for use in antenatal clinics (Delport & Van den Berg 1998). However, its impact on perinatal mortality has not been demonstrated. To justify the expense of new equipment and the effort required to re-train busy staff it is important to have unbiased cost-effectiveness data available. For this reason we are now conducting a clinic randomized controlled trail to compare the current standard of care (testing serum for syphilis in the district hospital laboratory) with near-patient testing in the antenatal clinics, with perinatal mortality as the primary outcome. Our data also indicate the need to improve the proportion of partners treated and to reduce the gestation at which pregnant women attend the clinic for their first visit, thereby increasing the time available to successfully complete treatment.
The study was funded by the Wellcome Trust and the South African Medical Research Council, through the Africa Centre for Population Studies and Reproductive Health.