Maternal and congenital syphilis in Bolivia, 1996: prevalence and risk factors.

Maternal and congenital syphilis in Bolivia, 1996: prevalence and risk factors Karen L. Southwick,1 Stanley Blanco,3 Ana Santander,3 Miguel Estenssoro,4 Faustino Torrico,5 Guillermo Seoane,3 William Brady,3 Martha Fears,2 Joel Lewis,2 Victoria Pope,2 Jeannette Guarner,2 & William C. Levine2 Objectives The present study was carried out in seven maternity hospitals to determine the prevalence of maternal syphilis at the time of delivery and the associated risk factors, to conduct a pilot project of rapid syphilis testing in hospital laboratories, to assure the quality of syphilis testing, and to determine the rate of congenital syphilis in infants born to women with syphilis at the time of delivery — all of which would provide baseline data for a national prevention programme in Bolivia. Methods All women delivering either live-born or stillborn infants in the seven participating hospitals in and around La Paz, El Alto, and Cochabamba between June and November 1996 were eligible for enrolment in the study. Findings A total of 61 out of 1428 mothers (4.3%) of live-born infants and 11 out of 43 mothers (26%) of stillborn infants were found to have syphilis at delivery. Multivariate analysis showed that women with live-born infants who had less than secondary-level education, who did not watch television during the week before delivery (this was used as an indicator of socioeconomic status), who had a previous history of syphilis, or who had more than one partner during the pregnancy were at increased risk of syphilis. While 76% of the study population had received prenatal care, only 17% had syphilis testing carried out during the pregnancy; 91% of serum samples that were reactive to rapid plasma reagin (RPR) tests were also reactive to fluorescent treponemal antibody-absorption (FTA-ABS) testing. There was 96% agreement between the results from local hospital laboratories and national reference laboratories in their testing of RPR reactivity of serum samples. Congenital syphilis infection was confirmed by laboratory tests in 15% of 66 infants born to women with positive RPR and FTA-ABS testing. Conclusion These results indicate that a congenital syphilis prevention programme in Bolivia could substantially reduce adverse infant outcomes due to this disease. Keywords: syphilis, congenital, prevention and control; syphilis, congenital, epidemiology; fetal death; infant mortality; pregnancy complications; syphilis serodiagnosis; fluorescent treponemal antibody-absorption test; Bolivia. Mots cle´s: syphilis conge´nitale, lutte; syphilis conge´nitale, e´pide´miologie; syphilis, se´rodiagnostic; mort fœtale; mortalite´ du nourrisson; complications de la grossesse, diagnostic; test FTA-ABS; facteurs socio-e´conomiques; Bolivie. Palabras clave: sı´filis conge´nita, prevencio´n y control; sı´filis conge´nita, epidemologı´a; muerte fetal; mortalidad infantil; complicaciones infecciosas del embarazo, diagno´stico; serodiagno´stico de la sı´filis; test de absorcio´n del anticuerpo fluorescente de treponema; factores socioecono´micos; Bolivia. Bulletin of the World Health Organization, 2001, 79: 33–42. Voir page 40 le re´sume´ en franc¸ais. En la pa´gina 41 figura un resumen en espan˜ol. Introduction Much attention and considerable resources are being focused on limiting the vertical transmission of human immunodeficiency virus (HIV) infection from preg- nant women to their infants (1). At the same time, it is essential for public health programmes not to neglect the prevention of another congenital infection, namely congenital syphilis. This is a potentially devastating condition that can cause fetal death, prematurity, meningitis, vasculitis, bone and joint destruction, and multi-system disease (2). Studies in the pre-penicillin era demonstrated that a woman with primary or secondary syphilis had a 70% or greater chance of infecting her fetus (3). More recently it was reported that 30–60% of infected live-born infants show no signs of congenital syphilis at birth (4). Penicillin treatment is effective for both the pregnant woman and the fetus if given a sufficient time before birth (3). Congenital syphilis causes a large burden of disease in resource-poor settings (5). For example, a 1 US Centers for Disease Prevention, Atlanta, GA, USA. Current address: Center for Disease Prevention and Epidemiology, Oregon Health Division, 800 NE Oregon St., Suite 772, Portland, Oregon, 97232, USA (email: Karen.L.Southwick@state.or.us). Correspondence should be addressed to this author at the latter address. 2 US Centers for Disease Prevention, Atlanta, GA, USA 3 MotherCare, La Paz, Bolivia, and Arlington, VA, USA. 4 INLASA — Instituto Nacional de Laboratorios de Salud — ‘‘Nestor Morales Villazon’’, Miraflores, La Paz, Bolivia. 5 Universidad Mayor de San Simon, Cochabamba, Bolivia. Ref. No. 99-0371 33Bulletin of the World Health Organization, 2001, 79 (1) # World Health Organization 2001 study in Zambia reported that almost 1% of infants born in a major university hospital in Lusaka had signs of congenital syphilis at delivery (6). Also, a recent population-based study inMalawi showed that 26% of stillbirths, 11% of neonatal deaths, 5% of post-neonatal deaths, and 8% of infant deaths were attributable to active maternal syphilis infection (7). Maternal syphilis is also a significant problem in South America. Several studies during the 1990s in Central and South America showed reactive syphilis tests in pregnant women ranging from 1.7% in Panama to 11.5% in Recife, Brazil (8). Consequently, in 1995, the Pan American Health Organization (PAHO) began a campaign to reduce the rate of congenital syphilis in the Americas to 0.5 (or fewer) cases per 1000 live births by increasing the coverage of antenatal care, establishing routine syphilis serology testing during prenatal care and at delivery, and promoting diagnostic strategies to ensure rapid treatment of infected pregnant women (8). In Bolivia, neonatologists and hospital-based paediatricians have reported anecdotally that con- genital syphilis is a common problem, but few data are available. In 1994, the prevalence of symptomatic congenital syphilis reported to the Bolivian Ministry of Health was 3.1 per 1000 live births (9). However, this may be an underestimate because a study during the early 1990s reported that 10.9% of 892 women who gave birth in the German Urquidi maternity hospital in Cochabamba had a reactive VDRL test (F. Torrico, unpublished data, 1993). At the time of the present study, the Bolivian Ministry of Health recommended routine syphilis testing during preg- nancy. However, funds and services for such testing were not provided, and the recommendation was not widely followed. The present study’s objectives were as follows: 1) to determine the prevalence of syphilis in women giving birth in Bolivian maternity hospitals and the associated risk factors; 2) to conduct a pilot project on rapid plasma reagin (RPR) testing at the time of delivery; 3) to assure the quality of syphilis serological testing; and 4) to determine the rate of congenital syphilis in infants born to women with maternal syphilis at delivery. The results of this investigation were used as a baseline to develop a national congenital syphilis prevention programme. Materials and methods All women delivering either live-born or stillborn infants in the seven participating hospitals in and around La Paz, El Alto, and Cochabamba between June andNovember 1996 were eligible for enrolment in this investigation. The study was approved by the Bolivian Ministry of Health and Social Welfare, the participating local institutions, the Office for Protec- tion fromResearchRisks of theNational Institutes of Health, and the institutional review board of the Centers for Disease Control and Prevention (CDC) in the USA. Each participating woman was required to give her written informed consent. A standardized antenatal and obstetrics form and a short question- naire on risk factors were filled out for each woman. Venous blood samples were obtained from the women, while from infants samples of blood were collected by venepuncture or from heel stick. A 2-cm length of the umbilical cord was taken from the portion closest to the infant at birth and immediately placed in 10% formalin. The women and their infants were given syphilis treatment before discharge if the mother’s serum was reactive to the RPR test carried out in the hospital laboratory at the time of delivery (10), and if there was no record in either the mother’s medical notes or antenatal card (a standard card given to every pregnant woman who receives antenatal care in Bolivia) that adequate penicillin treatment had previously been given — in accord with CDC treatment guidelines — during the pregnancy at least 30 days before the delivery (11) Treatment for syphilis, based on CDC treatment guidelines, was begun for both infected mothers and infants prior to their discharge from hospital (11). A full course of treatment at home was arranged for infants who did not require hospitalization and were discharged. If any women or infants had inadvertently been discharged before receiving treatment, a trained individual was sent to find them in their homes, explain the treatment, and administer penicillin. Before the study began, the individual hospitals were asked to develop their own strategies for partner notification and treatment. During the study, women with reactive RPR tests were asked to assist with the notification of their partner(s). Participation rates in all hospitals where women gave birth to either live-born or stillborn infants were determined by inspection of birth registers or hospitalization registers. Their clinical records were also reviewed. Case definitions Awoman was considered to have maternal syphilis if her serum at the time of delivery was reactive to the RPR test and fluorescent treponemal antibody- absorption (FTA–ABS) test, as carried out by the Bolivian reference laboratories, and if there was no written record that she had received adequate treatment for syphilis during the pregnancy and at least 30 days before delivery. An infant was considered to have laboratory- confirmed congenital syphilis if he or she was born to a woman with maternal syphilis and also if any one of the following was positive: direct fluorescent anti- body (DFA) test, immunohistochemistry, or IgM western blot assay. Laboratory tests Extensive laboratory preparation was undertaken to ensure the quality of the syphilis serological test results. The Bolivian reference laboratories were 34 Bulletin of the World Health Organization, 2001, 79 (1) Research selected based on their performance of analysing sera of known syphilis reactivity prepared by CDC. Before the study, a CDC-trained microbiologist visited each participating hospital laboratory and trained the staff on qualitative and quantitative RPR testing. During the study, the same microbiologist returned to each hospital to review the RPR testing techniques. Qualitative and quantitative RPR tests were performed onmaternal serum samples in the hospital laboratory using a standardized procedure (10), and the results were given to clinicians before the woman was discharged. The sera were sent from the hospitals to one of two Bolivian reference laboratories. Since one of the participating hospitals did not have a functioning laboratory, all the tests for that hospital were carried out by the Bolivian reference laboratory, which was next door to the hospital. The reference laboratories performed qualitative and quantitative RPR tests on all the maternal serum samples. FTA- ABS assay was performed on all RPR-reactive sera and also on 2% of non-reactive sera (10). Aminimally reactive FTA-ABS was designated as non-reactive. Individuals performing the analyses in the Bolivian reference laboratories were not provided with the hospital laboratory results. For establishing quality assurance, thematernal and infant serum samples and umbilical cord speci- mens from both the pilot and the actual study participants were sent to CDC in Atlanta, GA, USA where the RPR and FTA-ABS tests were performed on all RPR and/or FTA-ABS reactive maternal sera and on a systematic 10% sample of non-reactive sera from the Bolivian reference laboratories. In the few cases where the Bolivian reference laboratories had diagnosed a woman to be non-reactive and CDC diagnosed the woman as having a reactive RPR and FTA-ABS test, the results were sent to the Bolivian investigators who then contacted the woman in order to start treatment. Serum samples from the infants of the women tested had the following analyses performed at CDC: qualitative and quantitative RPR tests, microhaemagglutination Treponema palla- dium (MHA-TP) test (10), IgM ELISA (12), and IgM western blot (13) assays. The individuals performing the analyses at CDC were not provided with the earlier test results. Because only a small percentage of the maternal sera was tested by CDC, all analyses of risk factors and outcomes were based on the test results from the Bolivian reference laboratories. From the group of infants described above, a subset was chosen for analysis of umbilical cord specimens at CDC— comprising all umbilical cords from the infants of women with presumptive maternal syphilis and a random sample of cords from infants born to women without syphilis. In addition, the formalin-fixed umbilical cords from the infants of six women with syphilis identified during the initial pilot study of 88 women were included and the DFA test (10) and immunohistochemistry (14) were performed on them. Statistical analysis Univariate analysis was carried out using the w2 test or Fisher’s exact test (if the expected cell values were less than 5). Median values of continuous variables were compared using nonparametric tests. Backwards stepwise logistic regression was used for multivariate analysis (SAS Institute, Cary, NC, USA). The chunk test was used to evaluate interaction terms (15). Results Characteristics of the participants The overall participation rate in all seven hospitals for women giving birth to live-born infants was 63% (1428/2264). Hospital-specific participation rates varied from 54% in one of the large referral maternity hospitals to 93% in one of the small hospitals. Information on antenatal history was available for 1368 (96%) of 1428 womenwith live-born infants; of these, 1043 (76%) received antenatal care, but only 227 (17%) had a record of having been tested for syphilis during pregnancy. A total of 210 womenwho gave birth to live-born infants reportedly had a negative VDRL test during pregnancy; 7 women had documentation of VDRL testing but no results were available, and 10 women had a record of a reactive VDRL test during pregnancy. Of the 325 women who did not receive antenatal care, 75 (23%) said they did not seek it because they were afraid of the doctor and medical services, 73 (22%) thought that antenatal care was not important, 65 (20%) said they lacked money to pay for such care, 60 (18%) said they did not live close enough to the antenatal care services, 47 (14%) reported that they could not leave their work, 24 (7%) said they had no one to watch over their children, and 14 (4%) said that they had no transport. The women could indicate more than one response. During the recruitment period, four hospitals reported the delivery of stillborn infants. The three hospitals with the fewest deliveries reported no stillbirths during the study period. Of a total of 95 women with stillborn infants, 43 (45%) were enrolled in the study. The median gestational age of the stillborn infants was 37.5 weeks (range, 24– 42 weeks), and the median birth weight was 2200 g (range, 400–3950 g). The median birth weight and median gestational age of stillborn infants were less than those of live-born infants (2200 g vs 3240 g, P = 0.003; 37.5 weeks vs 40 weeks, P = 0.006). The characteristics of the mothers of live-born and stillborn infants were similar with respect to their income, the urban/rural location of their homes, the number of previous pregnancies and births, and the number and educational level of their partners during pregnancy (Table 1 and Table 2). However, the mothers of stillborn infants were more likely to be older, to speak an indigenous language rather than Spanish at home, to have less education, and to have maternal syphilis at the time of delivery. 35Bulletin of the World Health Organization, 2001, 79 (1) Maternal and congenital syphilis in Bolivia, 1996 Non-participating women The characteristics of the participating and non- participatingwomenwith live-born infants are shown in Table 3. Clinical information was available for 44 (85%) of the 52 mothers of stillborn infants who did not participate in the study. There were no significant differences in the proportions with low birth weight (<2500 g) and prematurity (<37 weeks) between the stillborn infants whose mothers did or did not participate in the study. Laboratory results Repeat tests performed for quality assurance gen- erally showed a high level of agreement between the local hospital results and the national reference laboratory results. Of the 847 samples analysed by both a hospital laboratory and the Bolivian reference laboratory, there was 96% agreement for the qualitative RPR results. One small hospital laboratory accounted for 44% of the discrepant results. Of the 37 samples analysed by hospital laboratories and by CDC, there was 97% agreement on the qualitative RPR results. Qualitative RPR testing was carried out on the sera from 166 women by both a Bolivian reference laboratory and CDC, with 96% agreement (159/166). FTA-ABS testing was carried out on the sera from 78 women by both a Bolivian reference laboratory and the CDC laboratory, with 94% agreement. Confirmatory testing of the hospital labora- tories’ RPR results by the Bolivian reference laboratories showed that only a small proportion was false-positive. Of 80 specimens that were RPR reactive at the Bolivian reference laboratory, 73 (91%) had reactive FTA-ABS results. Of the 73 specimens with a reactive RPR and FTA-ABS, 62% had a titre of =1:8. The median titre of these 73 specimens was 1:16 (range, 1:1 to 1:256), compared with a median titre of 1:2 (range 1:1 to 1:16) for the 7 specimens that were RPR reactive and FTA-ABS non-reactive. The umbilical cords from 66 infants born to women with presumptive syphilis at delivery were analysed at CDC; 57 of these infants were born alive, 6 were stillborn, and the status of 3 at birth was not known. Of the 66 infants born to women with syphilis, 10 (15%) could be confirmed as having congenital syphilis by immunohistochemistry or DFA testing of the umbilical cord, or by IgMwestern blot assay of the infants’ sera. Of these 10 children, 8 were born alive and 2 were stillborn. Comparison of women with and without maternal syphilis A total of 61 (4.3%) out of 1428 mothers of live-born infants in the seven different hospitals were diagnosed with maternal syphilis (median, 4.5%; range, 2.7–6.4%; P = 0.3); there were no statistically significant differences in syphilis prevalence between the hospital sites. Physical signs were neither sensitive nor specific for maternal syphilis; upon being Table 1. Characteristics of women with an RPR test obtained at the time of delivery who had live-born or stillborn infants in seven Bolivian maternity hospitals, 1996 Characteristica No. of women No. of P-value with live- women with born infants stillborn infants Location of home Urban 1049 (79)b 23 (72) Rural 284 (21) 9 (28) 0.5 Age (years) < 20 286 (21) 7 (22) 20–29 746 (55) 9 (28) 530 333 (24) 16 (50) 0.002 Preferred language at home Spanish 675 (49) 12 (35) Aymara 291 (21) 15 (44) Quechua 414 (30) 7 (21) 0.006 Educational level None 99 (7) 7 (23) Primary 554 (41) 13 (43) Secondary 635 (47) 10 (33) University 52 (4) 0 (0) 0.007 Received antenatal care Yes 1043 (76) 20 (61) No 325 (24) 13 (39) 0.06 Maternal syphilis at delivery Yes 61 (4) 11 (26) No 1367 (96) 32 (74) 0.001 History of syphilis Yes 26 (2) 0 (0) No 1309 (98) 34 (100) 1.0 Watched television in the previous week Yes 1098 (80) 26 (77) No 283 (20) 8 (23) 0.8 Listened to the radio during the previous week Yes 1031 (75) 22 (65) 0.3 No 350 (25) 12 (35) Income less than the median monthly household (350 Bolivianos = US$ 70) Yes 595 (49) 16 (50) No 621 (51) 16 (50) 1.0 No. of partners during pregnancy None 200 (17) 7 (25) One 968 (83) 21 (75) More than one 3 (<1) 0 (0) 0.5 Had a steady partner at the time of delivery Yes 1141 (82) 27 (82) 0.9 No 244 (18) 6 (18) Educational level of steady partner Less than secondary 497 (44) 17 (63) 0.1 Secondary or more 524 (47) 7 (26) Don’t know 106 (9) 3 (11) a Individuals with unknown values were excluded. b Figures in parentheses are percentages for each category. 36 Bulletin of the World Health Organization, 2001, 79 (1) Research examined, only 3 (5%) out of 61 women with syphilis had physical signs indicative of syphilis (rash, condyloma latum, mucous patches on the tongue, or a genital ulcer), compared with 39 (3%) out of 1367 women without syphilis (relative risk (RR) = 1.6; 95% confidence interval (CI) = 0.5–5.0). Of the 10 women with live-born infants and documented reactive VDRL tests during pregnancy, only 4 had a record of treatment for syphilis in their antenatal files. A total of 49 of 61 (80%) women with live-born infants who were diagnosed with syphilis at the time of delivery had a record of receiving adequate treatment: 38 of the women with live-born infants had received their treatment in the hospital at the time of delivery. Of the 11 women with syphilis who gave birth to stillborn infants, 10 (91%) had received treatment, 8 of them in the hospital at the time of delivery. Of the 210 women with a non- reactive VDRL test during pregnancy recorded on their antenatal files, 6 had reactive RPR and FTA results at the time of delivery. Bivariate analysis showed that women with live-born infants were at increased risk of maternal syphilis if they spoke an indigenous language rather than Spanish at home, had no antenatal care, had less than a secondary school education, had a history of syphilis, had a monthly household income less than the median (US$ 70), had more than one partner during the pregnancy or had a partner with less than a secondary school education (Table 4). Several of these variables were highly correlated. For example, comparedwithwomenwhowatched television in the week before delivery, those who did not watch television were less likely to have received antenatal care (66% vs 79%; P <0.001), to have an income less than the median of the study population (63% vs 45%; P <0.001), or to have had a secondary education (72% vs 43%; P <0.001). In a multivariate analysis, four variables were independently associated with maternal syphilis (Table 5). Interaction terms did not contribute significantly to the model. Characteristics of live-born infants Women with syphilis who gave birth to live-born infants were more likely than womenwithout syphilis to have a low-birth-weight infant (16% vs 7%, OR = 2.6, 95% CI = 1.2–4.8). However, there was no significant difference between the proportions of premature live-born infants born to women with and without syphilis (5%vs 4%,OR=1.2, 95%CI=0.3– 4.3). A total of 23 (1.6%) out of 1428 live-born infants had physical signs that could be compatible with syphilis (sero-sanguineous nasal discharge, skin eruptions, hepatosplenomegaly, or jaundice); 5% of the live-born infants of women with syphilis had physical signs at birth, while only 1.5% of live-born infants of women without syphilis had such physical signs (RR = 3.2, 95% CI = 1.1–9.4). Women with signs compatible with early syphilis (genital ulcers, body rash, condyloma lata, or mucous patches on the tongue) were more likely to have live-born infants with physical signs compatible with syphilis than women who did not have such signs (7% vs 1%, RR = 5.0, 95% CI = 1.5–16.0). Adequate treatment for congenital syphilis was received by 50 infants (82%) and 12 (20%) of 61 of the live-born infants had all their treatment at home. Maternal and infant risk factors for laboratory- confirmed congenital syphilis Significant differences existed between the women with syphilis who had infants with and without confirmed congenital syphilis. Women with syphilis at delivery and with physical signs compatible with syphilis were more likely to give birth to an infant with laboratory-confirmed congenital syphilis than those who did not have physical signs (67% vs 11%, OR = 16.7, 95% CI = 1.0–555.6; P= 0.05). Of the 7 women whose infants had confirmed congenital syphilis and for whom physical examination results were available, Table 2. Summary characteristics of women with an RPR test obtained at the time of delivery who had live-born or stillborn infants in seven Bolivian maternity hospitals, 1996 Characteristic Women with live- Women with P-value born infants stillborn infants Median Interquartile Median Interquartile range range Age (years) 23 20–29 29 21–34 0.07 Monthly income (US$) 70 50–100 65 30–85 0.92 No. of previous 1 0–3 3 0–5 0.11 pregnancies No. of previous births 1 0–3 2 0–4 0.19 Table 3. Characteristics of women who did and did not participate in RPR testing at the time of delivery, and of their live-born infants in seven Bolivian maternity hospitals, 1996 Characteristic Participating Non-participating P-value women women Median Interquartile Median Interquartile range range Age at delivery (years) 23 20–29 24 20–29 0.2 No. of pregnancies 1 0–3 2 1–4 0.001 Gestational age of 40 39–40 39 37–40 0.001 infant (weeks) Birthweightofinfant(g) 3240 2940–3540 3180 2810–3450 0.005 % who received 76 57 0.001 antenatal care % who gave birth to 7 12 0.001 a low-birth-weight infant % who gave birth to 16 0.001 a premature infant 37Bulletin of the World Health Organization, 2001, 79 (1) Maternal and congenital syphilis in Bolivia, 1996 two had condyloma lata at delivery. The women with syphilis whose infants had confirmed congenital syphilis had higher geometric mean RPR titres than the women with syphilis whose infants did not have confirmed congenital syphilis (1:32 vs 1:12; P = 0.001). They were also less likely to have a stable partner at the time of delivery (10% vs 44%,OR= 0.1, 95% CI = 0.0–0.9). There were no significant differences between the two groups with respect to the preferred language at home (Spanish vs indig- enous), educational level less than secondary, antena- tal care attendance, rural/urban residence, and age. There were also differences between the infants with and without laboratory-confirmed congenital syphilis born to women with syphilis at delivery. Infants with confirmed congenital syphilis were more likely to be low birth weight (<2500 g) than infants of women with syphilis without laboratory-confirmed congenital syphilis (78% vs 14%, OR = 22, 95% CI = 3.1–197.3) and to be premature (<37 weeks gestation) (50% vs 4%, OR = 24, 95% CI = 2.5–287.7). None of the six live-born infants with laboratory-confirmed congenital syphi- lis, for whom physical examination data were available at birth, had physical signs indicative of syphilis (hepatosplenomegaly, rash, sero-sangui- neous nasal discharge, or jaundice). The two stillborn infants with confirmed congenital syphilis were born macerated. Discussion Innovative approaches for the diagnosis and treat- ment of maternal syphilis, such as the on-site testing and treatment plan carried out in this study, have been documented in several resource-poor settings. In Zambia, for example, an on-site syphilis screening and treatment programme reduced by two-thirds the adverse outcomes attributable to syphilis (16). A Kenyan programme, also with an on-site testing and treatment strategy, screened a total of 13 131 women and documented a 6.5% prevalence of RPR test reactivity. The investigators calculated that 413 cases of congenital syphilis were prevented at a cost of US$ 50 per case prevented (17). Specific cost- effectiveness studies have not been carried out in Bolivia, but other reports have described the cost- effectiveness of antenatal syphilis testing pro- grammes even in areas with low syphilis seropreva- lence (18,19). The implementation of these strategies is feasible in countries that have limited medical and laboratory facilities, such as Bolivia. A broad-based maternal syphilis prevention programme, using on-site testing and treatment, would improve maternal and infant health in Bolivia. The present study showed a high prevalence of maternal syphilis, 4% in women giving birth to live- born infants and 26% inwomenwith stillborn infants in several Bolivian hospitals. Having either a known history of syphilis or more than one partner during pregnancy was a strong risk factor for maternal syphilis at delivery. However, these characteristics were uncommon and present in fewer than 50% of women with syphilis at delivery. Therefore, if only these characteristics were used to screen women, more than one half of the cases of maternal syphilis would be missed. The other two risk factors for maternal syphilis at delivery, i.e. lack of secondary education and not Table 4. Comparison between Bolivian women giving birth to live-born infants who had and did not have maternal syphilis at the time of delivery, 1996 Characteristica No. of women No. with Relative risk/ maternal 95% confidence syphilis at delivery intervalb Location of home Urban 1049 42 (4.0)c 0.7/0.4–1.3 Rural 284 16 (5.6) Age (years) < 20 286 12 (4.2) 1.0/0.5–1.8 5 20 1079 47 (4.4) Preferred language at home Indigenous 705 41 (5.8) 2.1/1.2–3.5 Spanish 675 19 (2.8) Less than secondary education Yes 653 46 (7.0) 3.7/2.0–6.8 No 687 13 (1.9) Received antenatal care Yes 1043 38 (3.6) 0.5/0.3–0.9 No 325 22 (6.8) History of syphilis Yes 26 4 (15.0) 3.6/1.4–9.2 No 1309 56 (4.3) Watched television during the previous week Yes 1098 37 (3.4) 0.4/0.3–0.7 No 283 23 (8.1) No. of partners during pregnancy More than one 3 1 (33.3) 7.5/1.5–38.7 None or one 1168 51 (4.4) Had steady partner at time of delivery Yes 1141 47 (4.1) 0.8/0.4–1.4 No 244 13 (5.3) Steady partner had less than secondary education Yes 497 28 (5.6) 2.7/1.4–5.4 No 524 11 (2.1) Income less than median weekly household income (US$ 70): Yes 595 36 (6.1) 2.0/1.2–3.4 No 621 19 (3.1) a Individuals with unknown values excluded. b Values shown are for the reference group. c Figures in parentheses are percentages. 38 Bulletin of the World Health Organization, 2001, 79 (1) Research having a television to watch during the week before delivery (as an indicator of socioeconomic status), were highly prevalent in the study population; this is also true for the country as a whole. According to the 1994 National Health and Demographic Survey of Bolivia, 63% of Bolivian women aged 15–44 years had less than a secondary education and 33% of Bolivian women did not watch television at least once weekly (20). However, the prevalence of maternal syphilis was still substantial even in women without these risk factors. Antenatal care as currently practised in Bolivia appears to have been minimally effective in reducing the risk of maternal syphilis in women who give birth in hospitals. Although the majority of the participat- ing mothers received antenatal care, only a minority had documented proof that a syphilis test had been performed during the pregnancy. More women may have had undocumented screening and treatment during pregnancy. However, discussions with hospi- tal physicians suggest that pregnant women are not routinely screened. These physicians reported that they know these women ought to be screened for syphilis during pregnancy, but that the obstacle for most women is the cost. At the time of the study, pregnant women had to pay for both syphilis testing and treatment. Antenatal care providers usually do not recommend a syphilis test if they believe that the woman has no risk behaviour for sexually transmitted diseases (Ilanes, personal communication, 1996). This study demonstrates that a syphilis screen- ing programme for pregnant women in Bolivia is feasible. The high rate of agreement between the RPR results from the hospital laboratories and those of both the Bolivian reference laboratories and CDC showed that accurate RPR tests can be carried out on-site with minimal training of local staff in decentralized locations with basic facilities. The study also showed that the percentage of pregnant women with false-positive RPR results was small. Clinicians can be confident that a pregnant woman with a reactive RPR has a high probability of having syphilis. The low rate of false-positive RPR results alleviates the need for routine confirmatory testing, which is prohibitively expensive for most women and not available in most settings. The investigation’s laboratory data together with information collected from pre-existing antena- tal records suggest that some women may serocon- vert from a negative to a reactive syphilis test between an initial test during antenatal care and testing at the time of delivery. However, we cannot assess the quality of VDRL testing during pregnancy because these were done independently of our study; it is therefore not possible to state for certain that these women actually had seroconversions during preg- nancy. At present, we cannot make any specific recommendations about testing each woman twice for syphilis during pregnancy in Bolivia. It is nevertheless essential to ensure treatment of all asymptomatic infants of RPR-reactive mothers at birth. Our study showed that six live-born infants with laboratory-confirmed congenital syphilis were asymptomatic at birth. Therefore, all newborn infants of women with reactive syphilis tests, including asymptomatic infants, require immediate penicillin therapy. This study had several limitations: it only included women who gave birth in hospitals, which is estimated to be approximately 60%of Bolivian women who give birth (20). Furthermore, the participation rate of mothers was low in some of the hospitals. It is likely that the prevalence of syphilis in pregnant women found by this study underestimated the actual prevalence in such women attending these hospitals. Compared with women who participated in the investigation, the women who gave birth in the study hospitals and who did not participate were more likely to have a low-birth-weight infant and a small-for- gestational-age infant. It is known that syphilis is a cause of premature birth (2, 3), and the hospital staff may not have had the time to enrol women and collect samples during emergency or urgent deliveries. Furthermore, the women who participated may have been at lower risk for syphilis than the general Bolivian population of women of child-bearing age. According to the Bolivian NationalDemographic andHealth Survey in 1994 (20), 53% of pregnant Bolivian women received antenatal care, compared with 76% of the study participants, and 37% of Bolivian women of child-bearing age had at least a secondary school education, compared with 51% of the study population. The study probably also underestimated the prevalence of congenital syphilis in infants born to women with syphilis. The sensitivity of IgM tests for treponemal antibody in asymptomatic congenital syphilis or in babies infected late in pregnancy is Table 5. Multivariate analysis of risk factors associated with syphilis at the time of delivery in Bolivian women with live- born infants, 1996 Characteristica Total number No. with Odds ratio/ syphilis 95% confidence intervalb Less than secondary education Yes 653 (49)c 46 (7.0) 3.1/1.6–6.2 No 687 (51) 13 (1.9) Watched television during past week Yes 1098 (80) 37 (3.4) 0.5/0.3–0.9 No 283 (20) 23 (8.1) History of syphilis Yes 26 (<1) 4 (15.0) 6.7/2.1–21.7 No 1309 (99.9) 56 (4.3) No. of partners during the pregnancy More than one 3 (<1) 1 (33.3) 28.5/2.4–339.4 One or none 1168 (99) 51 (4.3) a Unknown values excluded. b Values shown are for the reference group. c Figures in parentheses are percentages. 39Bulletin of the World Health Organization, 2001, 79 (1) Maternal and congenital syphilis in Bolivia, 1996 unknown, and not all of the babies born to women with syphilis had serum available for testing. Also, the sensitivity and specificity of the immunohistochem- istry results were unknown. The immunohistochem- istry results were reported as positive only if whole spirochaetes were visualized. Four of the umbilical cords from infants of women with syphilis showed antigenic material that could have been treponemal fragments. While the fragments were considered highly suspicious for congenital syphilis, their exact significance was not known. Based on these results, we recommend the following measures for the control and prevention of maternal syphilis in Bolivia, as well as in other countries with substantial rates of this disease: – coverage and utilization of antenatal care by pregnant women must be increased; all women should receive free antenatal syphilis testing (providing free or subsidized syphilis testing and treatment is essential to remove cost barriers for pregnant women with limited means — it is far preferable to test for syphilis early during antenatal care than to test for syphilis only at the time of delivery); – antenatal care providers and maternity hospital staff members in Bolivia must be taught that syphilis is common and that, in such settings, confirmatory serological testing is not necessary for treatment based on a presumptive diagnosis; – on-site testing is essential so that women can be tested and receive the results during the same visit to ensure prompt treatment, otherwise women may be lost to follow-up; – any pregnant woman who has a reactive RPR test result without documentation of adequate peni- cillin therapy earlier during the pregnancy (accord- ing to Ministry of Health and Social Welfare guidelines) should be given immediate penicillin treatment; – any infant born to a woman with syphilis diagnosed at the time of delivery should be treated with penicillin, regardless of the presence or absence of physical symptoms; – culturally appropriate partner notification techni- ques, which are sensitive to domestic relation- ships, need to be developed; – a functional reporting system should be estab- lished to notify the Ministry of Health and Social Welfare of all cases of maternal syphilis and congenital syphilis in order to improve the monitoring of syphilis cases and trends. As a result of this study, a national congenital syphilis prevention programme providing universal on-site testing of pregnant and parturient women was developed. In December 1998, a ministerial order was passed which decreed that all pregnant women in Bolivia can receive free basic antenatal care, including routine syphilis testing during antenatal care and again at the time of delivery, as well as free penicillin treatment for the mother, her partner, and the infant. Specific cost-effectiveness studies have not been carried out in Bolivia, but the total cost of each syphilis test is US$ 0.70 and is provided through municipal funds. Beginning in July 1999, routine free antenatal syphilis testing has been implemented in a number of Bolivian clinics and hospitals. n Acknowledgements This study was made possible through support provided by John Snow Inc / MotherCare Project and the Office of Health and Nutrition in the Bureau for Global Programs, Field Support and Research, USAgency for International Development (USAID), under the terms of contract number HRN-Q-00- 3039-00. The opinions expressed in this article are those of the authors and do not necessarily reflect the views of USAID or John Snow Inc. We are grateful to the following individuals and organizations for their collaboration: Dr Maria Luisa Melgar (Secretaria Nacional de Salud of Bolivia) for her support of the study; Ing. Alberto Saavedra and Eliana B Trujillo for their careful and complete data entry; the Community andChildHealthOrganization for donating penicillin; Freddy Tinajeros for his excellent teaching of the RPR technique; Kathleen Hutchins, Centers for Disease Control and Preven- tion (CDC) for developing the data entry screens; Raymond Ransom (CDC) for computer support; Dr Akbar Zaidi (CDC) for statistical support; Jeanine Bartlett (CDC) for processing the umbilical cord specimens; Robert George (CDC) for running the IgMwestern blot tests on infant sera; Patricia Daunas (MotherCare, Washington) and Silvia Garvizu, Liliana Medinicelli, and Blanca Gumucio (Mother- Care, Bolivia) for their administrative support; Jeanne McDermott (MotherCare, Washington) for her expert advice; the staff in Hospital de la Mujer, Hospital los Andes, Hospital ‘‘Solomon Klein’’ of Quintanilla, District Hospital Quillacollo, District Hospital Capinota and Hospital Mejico of Sacaba; Dr Michael St Louis andDrDouglas Fleming (CDC) for careful review of the manuscript; and George O’Hanlon for his support. We thank Elsevier Science Inc. and the American Academy of Pediatrics for donating sub- scriptions to the journals, Obstetrics and Gynecology and Pediatrics, to the hospitals participating in the study. Re´sume´ Syphilis maternelle et conge´nitale en Bolivie, 1996 : pre´valence et facteurs de risque Objectif La pre´sente e´tude a e´te´ re´alise´e dans sept maternite´s en vue de de´terminer la pre´valence de la syphilis maternelle au moment de l’accouchement et les facteurs de risque associe´s, d’entreprendre un projet pilote de de´pistage rapide de la syphilis dans les laboratoires hospitaliers, d’assurer la qualite´ des tests 40 Bulletin of the World Health Organization, 2001, 79 (1) Research se´rologiques et de de´terminer le taux de syphilis conge´nitale chez les nourrissons dont la me`re e´tait atteinte de syphilis au moment de l’accouchement, tous e´le´ments qui constitueront les donne´es de base d’un programme national de pre´vention en Bolivie. Me´thodes Toutes les femmes ayant mis au monde un enfant vivant ou mort-ne´ dans les sept hoˆpitaux participants a` La Paz, El Alto et Cochabamba, ou a` proximite´, entre juin et novembre 1996 re´pondaient aux crite`res de recrutement de l’e´tude. Re´sultats Au total, 61 me`res sur 1428 (4,3 %) ayant donne´ naissance a` un enfant vivant et 11 me`res sur 43 (26 %) ayant eu un enfant mort-ne´ e´taient atteintes de syphilis au moment de l’accouchement. Une analyse multivarie´e a montre´ que parmi les femmes ayant donne´ naissance a` un enfant vivant, celles qui n’avaient suivi qu’une scolarite´ primaire et qui n’avaient pas regarde´ la te´le´vision pendant la semaine pre´ce´dant l’accouchement (facteur utilise´ comme indicateur du niveau socio- e´conomique) et celles qui avaient des ante´ce´dents de syphilis ou qui avaient eu plus d’un partenaire pendant la grossesse avaient un risque plus e´leve´ de syphilis. Alors que 76 % des femmes participant a` l’e´tude avaient rec¸u des soins pre´natals, seules 17 % avaient fait l’objet d’un de´pistage se´rologique de la syphilis pendant la grossesse ; 91 % des e´chantillons de se´rum positifs pour le test RPR (test rapide a` la re´agine) e´taient e´galement positifs pour le test FTA-ABS (de´tection des anticorps spe´cifiques de tre´pone`me par immunofluorescence indirecte). Les re´sultats des tests RPR effectue´s sur les e´chantillons de se´rum dans les laboratoires hospitaliers et les laboratoires nationaux de re´fe´rence e´taient concordants a` 96 %. La pre´sence d’une syphilis conge´nitale a e´te´ confirme´e par les tests de laboratoire chez 15 % des 66 nourrissons dont la me`re avait un RPR et un FTA-ABS positifs. Conclusion Ces re´sultats indiquent qu’un programme de pre´vention de la syphilis conge´nitale en Bolivie pourrait re´duire sensiblement les graves conse´quences de cette maladie chez le nourrisson. Resumen Sı´filis materna y conge´nita en Bolivia, 1996: prevalencia y factores de riesgo Objetivo El presente estudio se llevo´ a cabo en siete maternidades con miras a determinar la prevalencia de sı´filis materna en el momento del parto y los factores de riesgo asociados, evaluar un proyecto piloto de prueba ra´pida para la sı´filis en los laboratorios de hospitales, asegurar la calidad de las pruebas de la sı´filis, y determinar la tasa de sı´filis conge´nita entre los lactantes nacidos de mujeres con sı´filis, todo lo cual habı´a de proporcionar datos basales para un programa nacional de prevencio´n de la enfermedad en Bolivia. Me´todos Se decidio´ utilizar como muestra de partida para el estudio a todas las mujeres que dieran a luz (incluidos mortinatos) en los siete hospitales participan- tes, situados en o cerca de La Paz, El Alto y Cochabamba, entre junio y noviembre de 1996. Resultados Se detecto´ sı´filis en el momento del parto en 61 de 1428 madres (4,3%) de nin˜os nacidos vivos, y en 11 de 43 madres (26%) de mortinatos. El ana´lisis multifactorial revelo´ que, entre las primeras, presentaban ma´s riesgo de sı´filis aquellas que no habı´an llegado a recibir educacio´n secundaria, que no habı´an dispuesto de televisio´n durante la semana anterior al parto (indicador de la situacio´n socioecono´mica) y que tenı´an antecedentes de sı´filis o se habı´an relacionado con ma´s de una pareja durante el embarazo. Si bien un 76% de la poblacio´n estudiada habı´a recibido atencio´n prenatal, so´lo un 17% se habı´a sometido a la prueba de la sı´filis durante el embarazo; el 91% de las muestras de suero que dieron positivo en la RPR (prueba ra´pida de reaginas) tambie´n fueron positivas en la prueba FTA-ABS (deteccio´n por fluorescencia de anticuerpos anti- treponema absorbidos). Se registro´ una concordancia del 96% entre los resultados de los laboratorios de los hospitales locales y los laboratorios de referencia nacionales en lo que respecta a la prueba RPR con muestras se´ricas. La sı´filis conge´nita se vio confirmada por los ana´lisis de laboratorio en un 15% de los 66 nin˜os nacidos de mujeres con resultados positivos en las pruebas RPR y FTA-ABS. Conclusio´n Estos resultados indican que la aplicacio´n de un programa de prevencio´n de la sı´filis conge´nita en Bolivia podrı´a reducir sustancialmente los prono´sticos negativos por esa causa entre los lactantes. References 1. Guay LA et al. Intrapartum and neonatal single-dose nevirapine compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda: HIVNET 012 randomised trial. Lancet, 1999, 354: 795–802. 2. Schulz KF et al. Congenital syphilis. In: Holmes KK et al., eds. Sexually transmitted diseases, 2nd edition. New York, McGraw Hill, 1990. 3. Ingraham NR. The value of penicillin alone in the prevention and treatment of congenital syphilis. Acta Dermatovenereology, 1951, 31 (suppl. 24): 60–88. 4. Mascola L et al. Congenital syphilis revisited. American Journal of Diseases of Children, 1985, 228: 855–858. 5. Schulz KF, Cates W Jr, O’Mara PR. Pregnancy loss, infant death, and suffering: legacy of syphilis and gonorrhea in Africa. Genitourinary Medicine, 1987, 63: 320–325. 6. Hira SK et al. Congenital syphilis in Lusaka. II. Incidence at birth and potential risk among hospital delivered infants. East African Medical Journal, 1982, 59: 306–310. 7. McDermott J et al. Syphilis-associated perinatal and infant mortality in rural Malawi. Bulletin of the World Health Organization, 1993, 71: 773–780. 8. Plan of action for the elimination of congenital syphilis. Washington, DC, Pan American Health Organization, 1995 (unpublished document PAHO/WHO/CE116/14). 41Bulletin of the World Health Organization, 2001, 79 (1) Maternal and congenital syphilis in Bolivia, 1996 9. Secretaria Nacional de Salud. Plan a Mediano Plazo (PMPII) Para la Vigilancia y Prevencion del VIH/SIDA: Bolivia 1995–1997 [Medium-term plan (PMPII) for vigilance and prevention of HIV/ AIDS: Bolivia 1995–1997]. La Paz, 1995 (in Spanish). 10. Larsen SA et al, eds. A manual of tests for syphilis, 9th edition. Washington, DC, American Public Health Association, 1998. 11. Centers for Disease Control and Prevention. 1998 sexually transmitted diseases treatment guidelines. Morbidity and Mor- tality Weekly Report, 1998, 47, No. RR-1. 12. Stoll BJ et al. Clinical and serologic evaluation of neonates for congenital syphilis: a continuing diagnostic dilemma. Journal of Infectious Diseases, 1993, 167: 1093–1099. 13. Sanchez PJ et al. Molecular analysis of the fetal IgM response to Treponema palladium antigens: implications for improved serodiagnosis of congenital syphilis. Journal of Infectious Diseases, 1989, 159: 508–517. 14. Guarner J et al. Congenital syphilis in a newborn: an immunopathologic study. Modern Pathology, 1999, 12: 82–87. 15. Kleinbaum DG. Logistic regression: a self-learning text. New York, Springer-Verlag, 1994. 16. Hira SK et al. Syphilis intervention in pregnancy: Zambian demonstration project. Genitourinary Medicine, 1990, 66: 159–164. 17. Jenniskens F et al. Syphilis control in pregnancy: decentraliza- tion of screening facilities to primary care level, a demonstration project in Nairobi, Kenya. International Journal of Gynecology and Obstetrics, 1995, 48 (suppl): S121–S128. 18. Stray-Pedersen B. Economic evaluation of maternal screening to prevent congenital syphilis. Sexually Transmitted Diseases, 1983, 10: 167–172. 19. Garland S, Kelly N. Is antenatal screening for syphilis worthwhile? Medical Journal of Australia, 1989, 151: 368–372. 20. Secretaria Nacional de Planificacion, Instituto Nacional de Estadistica and MacroInternational Inc. Encuesta Nacional de Demografia y Salud [National Health and Demographic Survey]. La Paz, Bolivian Ministry of Sustainable Development and the Environment, 1994 (in Spanish). 42 Bulletin of the World Health Organization, 2001, 79 (1) Research