IJG-08145; No of Pages 5 International Journal of Gynecology and Obstetrics xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
International Journal of Gynecology and Obstetrics journal homepage: www.elsevier.com/locate/ijgo
CLINICAL ARTICLE
Gestational syphilis and stillbirth in Latin America and the Caribbean Lauren Arnesen ⁎, Gerardo Martínez, Luis Mainero, Suzanne Serruya, Pablo Durán Pan American Health Organization, Centro Latinoamericano de Perinatología, Salud de la Mujer y Reproductiva, Montevideo, Uruguay
a r t i c l e
i n f o
Article history: Received 10 April 2014 Received in revised form 3 September 2014 Accepted 3 November 2014 Keywords: Caribbean Latin America Mother-to-child transmission Stillbirth Syphilis
a b s t r a c t Objective: To measure the association between gestational syphilis and stillbirth in Latin America and the Caribbean. Methods: In a retrospective study, data on stillbirth and gestational syphilis extracted from the Sistema Informático Perinatal database were analyzed for deliveries in 11 countries between January 1, 2009, and December 31, 2012. Potential confounders were examined, and binary logistic regression analysis was performed to assess the association between gestational syphilis and stillbirth. Results: Among 368 151 deliveries, 3875 (1.1%) were by women with a positive syphilis test, and 1461 (0.4%) were stillbirths. Among the stillbirths, 29 (2.0%) were delivered by women with a positive syphilis test. After controlling for country, congenital anomalies, gestational age at labor, maternal age, and previous stillbirth, gestational syphilis was significantly associated with stillbirth (odds ratio 1.88, 95% confidence interval 1.25–2.83; P = 0.002). Conclusion: Gestational syphilis contributes to stillbirth in Latin America and the Caribbean. Interventions targeting gestational syphilis are highly cost-effective and should be implemented across the region. © 2014 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved.
1. Introduction WHO estimates that 1.36–2 million pregnancies are affected by syphilis every year [1–3], of which 50%–80% will result in adverse outcomes without proper treatment [2,4–6]. Prominent among these outcomes is stillbirth, which occurs in 25%–40% of cases [4–6]. Globally, syphilis is the main cause of more than 212 000 stillbirths annually [2,3,7], and, in areas with a high prevalence of syphilis, as many as half of all stillbirths can be attributed to this infection [5,8]. Worldwide, Latin America and the Caribbean (LAC) has the highest incidence of syphilis, and accounts for up to 25% of the 2 million annual cases of gestational syphilis [1,9]. The prevalence of gestational syphilis in LAC varies from 0.08% to 7.0% by country [2,10]. Every year, an estimated 100 000 stillbirths in the region are attributable to congenital syphilis, which is defined as a neonate delivered to a mother who was untreated or inadequately treated for syphilis during pregnancy, or a neonate with a positive syphilis test [1,10,11]. The Pan American Health Organization (PAHO) approved the Regional Elimination of Congenital Syphilis Plans of Action in 1995 and 2010, in which elimination is defined as an incidence of 0.5 or fewer cases of congenital syphilis, including stillbirths, per 1000 deliveries [1,9,10]. Elimination of congenital syphilis is possible with proper treatment—one dose of penicillin is almost 100% effective at preventing syphilis-associated adverse outcomes in pregnancy [2,5,7, 9,12]. Moreover, the cost-effectiveness of prenatal syphilis screening ⁎ Corresponding author at: Centro Latinoamericano de Perinatología, Salud de la Mujer y Reproductiva, Avenue Italia # Hospital de Clinicas, Piso 16, Montevideo 11100, Uruguay. Tel.: +598 2 487 2929. E-mail address: [email protected] (L. Arnesen).
and treatment has been shown repeatedly and is considered one of the most cost-effective interventions available [2,4,5,10,12]. Nevertheless, syphilis-attributable stillbirths are still occurring in LAC. There are few studies on the relationship between gestational syphilis and stillbirth in LAC. Two previous studies in the region showed that the risk of stillbirth increases with exposure to gestational syphilis: one [13] reported that women with a positive syphilis test in Jamaica have a 3.04-fold increased odds of fetal death, whereas the other [14] indicated that women with a positive syphilis test in LAC countries have a 2.41-fold higher risk of fetal death. Generally, research in LAC has focused on the prevalence of gestational syphilis and calls to action to test and treat pregnant women [14], rather than examining factors that influence the relationship between gestational syphilis and stillbirth, and only one study has approached the problem at the regional level [13,15–19]. Sistema Informático Perinatal (SIP) is a PAHO/WHO database, developed by the Latin American Center for Perinatology and Human Development (CLAP), which includes a set of standardized instruments designed for obstetric and neonatal health services [14]. Every entry in SIP corresponds to a birth record containing data on maternal demographics; family and obstetric history; prenatal visits, delivery, and postpartum details; and discharge information for both the mother and neonate. Information is prospectively recorded at the clinic level during prenatal, delivery, and postnatal care. In addition to recording clinical data, SIP facilitates the electronic storage of information for further processing and analysis for managerial, epidemiologic, and other purposes. The primary aim of the present study was to use SIP data for deliveries occurring from 2009 to 2012 to measure the association between gestational syphilis and stillbirth at the regional level, with adjustments
http://dx.doi.org/10.1016/j.ijgo.2014.09.017 0020-7292/© 2014 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved.
Please cite this article as: Arnesen L, et al, Gestational syphilis and stillbirth in Latin America and the Caribbean, Int J Gynecol Obstet (2014), http:// dx.doi.org/10.1016/j.ijgo.2014.09.017
2
L. Arnesen et al. / International Journal of Gynecology and Obstetrics xxx (2014) xxx–xxx
for potential confounding factors. A secondary aim was to identify highrisk groups through tests for confounders, mediators, and effect measure modifiers, to focus future gestational syphilis interventions both in LAC and globally. 2. Materials and methods In a retrospective study, SIP data on stillbirth and gestational syphilis were analyzed for deliveries recorded in LAC countries between January 1, 2009, and December 31, 2012. Because data were extracted via a computer-generated identification code that was not linked in any way to patients, the study was exempt from institutional review board approval and informed consent was not needed. Use and coverage of SIP varies within the LAC region and within countries. Data collection and entry are done at the institutional level, and data are reported to CLAP through the respective national ministry of health. Deliveries in 15 LAC countries that use SIP and share their data with CLAP were considered for the present study. Of these countries, four did not have syphilis data recorded for the time period of the study and so were excluded. The 11 countries included in our study were: Argentina, Bolivia, Colombia, Ecuador, Guyana, Haiti, Honduras, Nicaragua, Paraguay, El Salvador, and Uruguay. The countries were randomly labeled with letters to maintain confidentiality because the study aim was not to compare specific countries, but to carry out a regional analysis of the relationship between gestational syphilis and stillbirth. For the analysis, data were extracted from the SIP records on syphilis test results, delivery outcome, and potential confounders that might biologically influence the relationship between gestational syphilis and stillbirth (Box 1). Records that were not complete were excluded from the study. The definition of stillbirth varies by study [5,8,12,13,15]. In line with recent studies [5,12], for the present analysis, stillbirth was defined as the neonate being dead at delivery. Gestational syphilis was defined as a positive result on the mother’s most recent prenatal non-treponemal or treponemal syphilis test [19]. The results from at least one syphilis test for the mother during pregnancy were available for all included deliveries. Traditionally, syphilis screening has been carried out in laboratories with non-treponemal tests, the results of which can take a
few weeks to be delivered in rural areas [20]. New point-of-care (POC) treponemal tests react to antibodies for Treponema pallidum, which usually remain in an individual’s system even after successful treatment and require only 20 minutes for detection [20]. Use of a treponemal or nontreponemal syphilis test, and what constituted a positive response for gestational syphilis, was determined at the institutional level. Among the SIP variables analyzed, “previous stillbirth” indicated that the mother had one or more previous stillbirths. Congenital anomalies were classified at the institutional level in SIP as “no,” “minor,” or “major,” and were analyzed as an ordinal variable. Both gestational age at labor and maternal age were analyzed as continuous variables, and then were grouped in accordance with international norms [21,22] for comparison. Statistical analyses were performed with SPSS version 17.0 (SPSS Inc, Chicago, IL, USA). After extraction of delivery outcome, gestational syphilis exposure, and all potential confounders from SIP, an unadjusted binomial logistic regression analysis was carried out for gestational syphilis and stillbirth. Each potential covariate was then analyzed separately with syphilis and delivery outcome in unique binary logistic regression models. Syphilis and all variables that were significantly associated with stillbirth when controlling for syphilis in their respective binary logistic regression model were then analyzed in a larger model, which also adjusted for syphilis exposure status. Covariates that were not significant in the larger model were excluded, resulting in a final model. All independent variables in the final model were tested for multicollinearity with linear regression by using dummy variables for all nominal variables. In the final model, confounding among covariates not thought to be in the causal pathway (all but gestational age at labor) was checked with Pearson χ2 tests of the covariate with syphilis and birth outcome separately, and the crude and Mantel-Haenszel odds ratios (ORs) were compared. Covariates that were not confounders, but also not in the causal pathway, were examined for statistical evidence of effect measure modification with calculations of the difference in the syphilis– stillbirth relationship across strata of the covariate with the beta for the interaction term (covariate*syphilis) in a binary logistic regression, in which syphilis, the covariate, and the interaction term are the independent variables. For the variable thought to be in the causal pathway—
Box 1 Covariates tested for significance in the relationship between gestational syphilis and stillbirth. Active smoker Maternal age Alcohol use Rubella vaccine Group B streptococcus carrier Tetanus vaccine Bacteriuria Birth weight Cervix check Chagas disease Chronic hypertension Civil status Companion at delivery Companion at labor Congenital defect Contraceptive failure Country Chorioamnionitis infection Diabetes Eclampsia Heart disease Kidney disease Pre-eclampsia
Urinary tract infection Drug use Education Ethnic origin Gestational age at labor Hemorrhage Family history of diabetes Maternal history of diabetes Family history of eclampsia Maternal history of eclampsia Maternal history of heart disease Family history of hypertension Maternal history of hypertension Maternal history of kidney disease Family history of preeclampsia Maternal history of preeclampsia History of violence toward mother HIV infection Hospitalized during pregnancy Intrauterine growth restriction Literate Multiple pregnancy Number of antenatal visits
Passive smoker Planned pregnancy Pregnancy-induced hypertension Presentation at labor Previous body mass index Previous pregnancies Previous live births Previous stillbirths Premature rupture of membranes Rupture of membranes b 37 wk Syphilis test treponemal Syphilis test non-treponemal Mother lives alone Three consecutive spontaneous abortions Toxoplasmosis Violence to mother Treated for syphilis Partner treated for syphilis Week of syphilis test nontreponemal Week of syphilis treatment Week of syphilis test treponemal
Please cite this article as: Arnesen L, et al, Gestational syphilis and stillbirth in Latin America and the Caribbean, Int J Gynecol Obstet (2014), http:// dx.doi.org/10.1016/j.ijgo.2014.09.017
L. Arnesen et al. / International Journal of Gynecology and Obstetrics xxx (2014) xxx–xxx
3
3. Results
All births in SIP from 2009–2012 (n=712 081) Missing syphilis and/or birth outcome (n=213 217) Have response for syphilis and birth outcome (n=480 864) Missing a variable/response for at least one covariate a (n=112 713) Have response/value for all covariates a in model (n=368 151)
Fig. 1. Flow diagram of the study population. Abbreviation: SIP, Sistema Informático Perinatal. a Country, gestational age at labor, maternal age, congenital anomaly, previous stillbirths.
gestational age at labor—the crude and Mantel-Haenszel ORs were examined to check for mediation. A P value of 0.05 was used as a cutoff for statistical significance. A difference of 10% or more was considered to be significant for the tests.
Of the 712 081 birth records contained in SIP during the study period, 368 151 (51.7%) met the inclusion criteria (Fig. 1). There were 3875 deliveries by women with a positive syphilis test and 1461 stillbirths in the study population. Of the stillbirths, 29 (2.0%) were delivered by women with a positive syphilis test, and 1432 (98.0%) were delivered to women with a negative syphilis test. Table 1 shows the distribution of each variable by both gestational syphilis status and delivery outcome. With the presence and increased severity of congenital anomalies, the proportion of cases with gestational syphilis and stillbirths increased. A total of 3830 (1.0%) of 365 172 deliveries without congenital anomalies, 21 (1.3%) of 1569 deliveries with minor anomalies, and 24 (1.7%) of 1410 deliveries with major anomalies were by women with a positive syphilis test. Additionally, 1279 (0.4%) deliveries without congenital anomalies, 23 (1.5%) with minor anomalies, and 159 (11.3%) with major anomalies were stillbirths. The mean gestational age at labor was 38.1 weeks (range 1.0–44.0) and mean maternal age was 24.9 years (range 10.0–55.0). Most mothers were aged between 20 and 34 years (Table 1). Overall, 318 (0.4%) of 87 191 deliveries to mothers younger than 20 years, 869 (0.4%) of 243 018 deliveries to mothers aged 20–34 years, and 274 (0.7%) of 37 942 deliveries to mothers aged 35 years or older were stillbirths. The proportion with a previous stillbirth was higher among women with a positive syphilis test than among women with a negative syphilis test, and was also higher among those who had a stillbirth in the study period than among those with a live birth (Table 1). The crude binary logistic regression model showed that gestational syphilis significantly increased the odds of stillbirth (P b 0.001) (Table 2). After controlling for country, congenital anomaly, previous stillbirth, gestational age at labor, and maternal age, gestational syphilis remained associated with stillbirth (P = 0.002) (Table 2). In the adjusted model, major congenital anomalies were significantly associated with stillbirth (P b 0.001), as was previous stillbirth
Table 1 Covariates by gestational syphilis status and delivery outcome.a Covariate
Syphilis Positive Negative Country A B C D E F G H I J K Congenital anomaly No Minor Major Previous stillbirth Yes No Gestational age at labor, wk b28 28–31 32–36 ≥37 Maternal age, y b20 20–34 ≥35
Gestational syphilis status
Delivery outcome
Total
Positive (n = 3875)
Negative (n = 364 276)
Stillbirth (n = 1461)
Live birth (n = 366 690)
– –
– –
29 (2.0) 1432 (98.0)
3846 (1.0) 362 844 (99.0)
3875 364 276
318 (8.2) 190 (4.9) 659 (17.0) 38 (1.0) 3 (0.1) 150 (3.9) 132 (3.4) 2 (b0.1) 0 630 (16.3) 1753 (45.2)
101 422 (27.8) 54 149 (14.9) 53 468 (14.7) 4366 (1.2) 296 (0.1) 11 337 (3.1) 11 645 (3.2) 49 (b0.1) 15 (b0.1) 17 102 (4.7) 110 427 (30.3)
500 (34.2) 246 (16.8) 231 (15.8) 78 (5.3) 0 27 (1.9) 166 (11.4) 1 (0.1) 0 44 (3.0) 168 (11.5)
101 240 (27.6) 54 093 (14.8) 53 896 (14.7) 4326 (1.2) 299 (0.1) 11 460 (3.1) 11 611 (3.2) 50 (b0.1) 15 (b0.1) 17 688 (4.8) 112 012 (30.5)
101 740 54 339 54 127 4404 299 11 487 11 777 51 15 17 732 112 180
3830 (98.8) 21 (0.6) 24 (0.6)
361 342 (99.2) 1548 (0.4) 1386 (0.4)
1279 (87.5) 23 (1.6) 159 (10.9)
363 893 (99.2) 1546 (0.4) 1251 (0.4)
365 172 1569 1410
193 (5.0) 3682 (95.0)
5188 (1.4) 359 088 (98.6)
77 (5.3) 1384 (94.7)
5304 (1.4) 361 386 (98.6)
5381 362 770
30 (0.8) 57 (1.5) 388 (10.0) 3400 (87.7)
1926 (0.5) 2871 (0.8) 27 658 (7.6) 331 821 (91.1)
263 (18.0) 171 (11.7) 378 (25.9) 649 (44.4)
1693 (0.5) 2757 (0.8) 27 668 (7.5) 334 572 (91.2)
1956 2928 28 046 335 221
788 (20.3) 2647 (68.3) 440 (11.4)
86 403 (23.7) 240 371 (66.0) 37 502 (10.3)
318 (21.7) 869 (59.5) 274 (18.8)
86 873 (23.7) 242 149 (66.0) 37 668 (10.3)
87 191 243 018 37 942
a
Please cite this article as: Arnesen L, et al, Gestational syphilis and stillbirth in Latin America and the Caribbean, Int J Gynecol Obstet (2014), http:// dx.doi.org/10.1016/j.ijgo.2014.09.017
4
L. Arnesen et al. / International Journal of Gynecology and Obstetrics xxx (2014) xxx–xxx
Table 2 Binomial logistic regression models of variables associated with stillbirth. Variable Crude Syphilis Positive Negative Adjusted Syphilis Positive Negative Country A B C D E F G H I J Kb Congenital anomaly No Minor Major Previous stillbirth Yes No Gestational age at laborc Maternal aged
Stillbirthsa
Wald χ2
Odds ratio (95% CI)
P value
29/3875 (0.7) 1432/364 276 (0.4)
–
2.18 (1.88–2.51) 1
b0.001 –
29/3875 (0.7) 1432/364 276 (0.4)
9.25 –
1.88 (1.25–2.83) 1
0.002 –
500/101 740 (0.5) 246/54 339 (0.5) 231/54 127 (0.2) 78/4404 (1.8) 0/299 27/11 487 (0.2) 166/11 777 (1.4) 1/51 (2.0) 0/15 44/17 732 (0.2) 168/112 180 (0.1)
234.22 166.59 144.36 360.59 0.00 99.76 415.18 0.14 0.00 8.52 –
4.26 (3.54–5.13) 3.94 (3.20–4.85) 3.63 (2.94–4.48) 15.63 (11.77–20.76) 0.00 (0.00–0.00) 0.08 (0.05–0.14) 10.52 (8.39–13.20) 2.22 (0.03–145.37) 0.00 (0.00–0.00) 1.68 (1.19–2.42) 1
b0.001 b0.001 b0.001 b0.001 0.998 b0.001 b0.001 0.708 0.998 0.004 –
1279/365 172 (0.4) 23/1569 (1.5) 159/1410 (11.3)
– 39.48 1106.93
1 4.01 (2.60–6.18) 25.90 (21.38–31.37)
– b0.001 b0.001
77/5381 (1.4) 1384/362 770 (0.4) 1461/368 151 (0.4) 1461/368 151 (0.4)
67.81 – 2238.60 86.76
2.85 (2.22–3.65) 1 0.81 (0.80–0.82) 1.04 (1.03–1.05)
b0.001 – b0.001 b0.001
Abbreviation: CI, confidence interval. a Values are given as total number of stillbirths/number of deliveries (percentage) unless indicated otherwise. b Randomly selected as the reference to allow for comparison between countries. c As a continuous variable, in increments of 1 week. d As a continuous variable, in increments of 1 year.
(P b 0.001) (Table 2). Gestational age at labor was the only protective factor (Table 2). There was a varying risk of stillbirth by country (Table 2). There was no evidence of multicollinearity among the independent variables in the final model (variance inflation factor b 2). Country was found to be a confounder of the relationship between gestational syphilis and stillbirth (Mantel-Haenszel OR 1.6; 95% confidence interval [CI] 1.4–1.8), and gestational age at labor was a mediator (Mantel-Haenszel OR 1.9; 95% CI 1.5–2.3) (Fig. 2). Mother’s age, congenital anomalies, and previous stillbirths were effect measure modifiers (Fig. 2). For mother’s age, the OR without the interaction effect was 2.60 (95% CI 1.47–4.61); the OR with the interaction effect was 0.99 (95% CI 0.97–1.01). The OR without congenital anomalies as the interaction term was 1.48 (95% CI 1.01–2.17); with major congenital anomalies, the OR was 1.59 (95% CI 0.59–4.30). Without previous
Maternal Age
Congenital Anomaly
Gestational Syphilis
Previous Stillbirth(s)
Stillbirth Gestational Age at Labor
KEY Confounder Mediator Effect measure modifier
Country
Fig. 2. Directed acyclic graph of gestational syphilis, stillbirth and associated covariates.
stillbirths as an interaction term, the OR was 2.41 (95% CI 2.05–2.83); with previous stillbirths, the OR was 0.42 (95% CI 0.21–0.85).
4. Discussion In the present study, gestational syphilis was associated with increased odds of stillbirth in LAC. This relationship remained when controlling for country, gestational age at labor, maternal age, congenital anomalies, and previous stillbirths. The present study analyzed recent data, used a very large study population, and included 11 countries from across the region. The increased likelihood of stillbirth after exposure to gestational syphilis recorded in the present study has been documented previously [1,13–19]. Outside LAC, a study of all congenital syphilis cases reported in LAC in 1982 in Texas, USA [23], found that the risk of stillbirth was 6.5 times higher for those exposed to gestational syphilis. Although odds and risk cannot be compared to one other, these previous results support the present conclusion that gestational syphilis is a clear factor in the incidence of stillbirth. The 11 LAC countries included in the study were not randomly selected; only those using SIP and sharing data with CLAP were eligible. Therefore, the regional representation of the current relationship between gestational syphilis and stillbirth could be skewed toward the national situation in the study countries. Nevertheless, the results remain the most representative of the situation in LAC to date. Furthermore, the inclusion of the 11 countries in the region does not alter the weight of the finding that gestational syphilis is related to increased odds of stillbirth. Country of birth was found to be a confounder in the relationship between gestational syphilis and stillbirth. As compared with the reference country, the OR for stillbirth in the presence of gestational syphilis
Please cite this article as: Arnesen L, et al, Gestational syphilis and stillbirth in Latin America and the Caribbean, Int J Gynecol Obstet (2014), http:// dx.doi.org/10.1016/j.ijgo.2014.09.017
L. Arnesen et al. / International Journal of Gynecology and Obstetrics xxx (2014) xxx–xxx
ranged from 0.08 to 15.63. Large differences in the odds of stillbirth by country are probably a reflection of healthcare systems—specifically, access to, and quality of, care—and whether POC syphilis testing and treatment is available. The ability to test and treat women on the same day has been cited as a cost-effective and successful way to prevent adverse birth outcomes associated with untreated gestational syphilis [3,7,12, 19,20]. Future studies looking at health systems with POC tests would help health facilities to identify other factors contributing to untreated cases of gestational syphilis. Gestational age at labor was a mediator in the final model. It seems that gestational syphilis influences the prematurity of the delivery, which affects the likelihood of a live birth. The transplacental passage of Treponema pallidum has been cited as a risk factor for prematurity [2,3], and prematurity has been shown to result in an increased likelihood of stillbirth [24,25], highlighting the importance of testing and treating women as early as possible to prevent these adverse birth outcomes. Increasing maternal age, previous stillbirth, and presence of congenital anomalies were effect measure modifiers in the final model. In addition to other routine testing at the standard prenatal care visit, syphilis testing and treatment for older mothers (≥ 35 years), those with at least one previous stillbirth, and those who meet both criteria, should be emphasized and incorporated into interventions targeting these high-risk groups. The present study has some limitations. First, what constituted a positive test for gestational syphilis was determined at the institutional level. Although unlikely, some institutions could deviate from international standards when determining whether a non-treponemal or POC syphilis test is positive. If some institutions were using different criteria to define cases of syphilis, this effect would most probably be uniform across exposures, and thus would not significantly affect the observed relationship between gestational syphilis and stillbirth. Second, gestational syphilis was defined by the treponemal or nontreponemal test result closest to delivery. The decision of whether a syphilis test was positive or negative was made at the institutional level. Internationally, a reactive treponemal test confirmed by a nontreponemal test is considered confirmatory for syphilis [6,19]. For the few cases that received only the treponemal test, a positive result indicated the presence of Treponemal pallidum antibodies, and not whether treatment was successful or not, which might bias the results away from the null. Third, there are new variables in SIP that were not available for analysis in the present study. These variables should be incorporated in future analyses, including maternal and partner treatment for syphilis and postpartum syphilis testing, because they might indicate reinfection of a woman who tests negative or has been successfully treated for syphilis. The absence of these variables from the present model might bias the results toward the null, a possibility that was taken into account by using the syphilis test result closest to delivery. In summary, the most recent data from SIP have been analyzed with the intention of conveying the present relationship between gestational syphilis and stillbirth in LAC. The present analysis represents a picture of the current situation. As more countries begin to use the system, share their data, and increase syphilis testing coverage, future studies based on SIP will improve the understanding of gestational syphilis and stillbirth at the regional level.
5
Conflict of interest The authors have no conflicts of interest. References [1] Pan American Health Organization, Latin American Center for Perinatology, Women and Reproductive Health, United Nations Children’s Fund, UNAIDS. 2010 situation analysis: Elimination of mother-to-child transmission of HIV and congenital syphilis in the Americas. http://www.paho.org/hq/index.php?option=com_docman&task= doc_download&gid=15893&Itemid. Published 2011. Accessed October 29, 2014. [2] Walker DG, Walker GJ. Forgotten but not gone: the continuing scourge of congenital syphilis. Lancet Infect Dis 2002;2(7):432–6. [3] Newman L, Kamb M, Hawkes S, Gomez G, Say L, Seuc A, et al. Global estimates of syphilis in pregnancy and associated adverse outcomes: analysis of multinational antenatal surveillance data. PLoS Med 2013;10(2):e1001396. [4] Schmid G. Economic and programmatic aspects of congenital syphilis prevention. Bull World Health Organ 2004;82(6):402–9. [5] Goldenberg RL, McClure EM, Saleem S, Reddy UM. Infection-related stillbirths. Lancet 2010;375(9724):1482–90. [6] Schmid GP, Stoner BP, Hawkes S, Broutet N. The need and plan for global elimination of congenital syphilis. Sex Transm Dis 2007;34(7 Suppl.):S5–S10. [7] Mabey DC, Sollis KA, Kelly HA, Benzaken AS, Bitarakwate E, Changalucha J, et al. Point-of-care tests to strengthen health systems and save newborn lives: the case of syphilis. PLoS Med 2012;9(6):e1001233. [8] Goldenberg RL, Thompson C. The infectious origins of stillbirth. Am J Obstet Gynecol 2003;189(3):861–73. [9] Pan American Health Organization, World Health Organization, Latin American Center for Perinatology, Women and Reproductive Health. Strategy and plan of action for elimination of mother-to-child transmission of HIV and congenital syphilis: regional monitoring strategy. http://www.paho.org/hq/index.php?option=com_ docman&task=doc_view&gid=24379&Itemid=. Published 2013. Accessed October 29, 2014. [10] Pan American Health Organization. Epidemiological profiles of neglected diseases and other infections related to poverty in Latin America and the Caribbean. http://www2.paho.org/hq/dmdocuments/2009/nds-epi-profiles.pdf. Published September 2009. Accessed October 29, 2014. [11] Centers for Disease Control and Prevention (CDC). Congenital syphilis – United States, 2003–2008. MMWR Morb Mortal Wkly Rep 2010;59(14):413–7. [12] Di Mario S, Say L, Lincetto O. Risk factors for stillbirth in developing countries: a systematic review of the literature. Sex Transm Dis 2007;34(7 Suppl.):S11–21. [13] Greenwood R, Foster-Williams K, Ashley D, Keeling J, Golding J. The epidemiology of antepartum fetal death in Jamaica. Paediatr Perinat Epidemiol 1994;8(Suppl. 1): 98–109. [14] Conde-Aguadelo A, Belizán JM, Díaz-Rossello JL. Epidemiology of fetal death in Latin America. Acta Obstet Gynecol Scand 2000;79(5):371–8. [15] Southwick KL, Blanco S, Santander A, Estenssoro M, Torrico F, Seoane G, et al. Maternal and congenital syphilis in Bolivia, 1996: prevalence and risk factors. Bull World Health Organ 2001;79(1):33–42. [16] Fitzgerald DW, Behets FM, Lucet C, Roberfroid D. Prevalence, burden, and control of syphilis in Haiti’s rural Artibonite region. Int J Infect Dis 1998;2(3):127–31. [17] Ali Z. Resurgence of congenital syphilis in Trinidad. J Trop Pediatr 1990;36(3):104–8. [18] Jacquier N, Dos Santos L, Deschutter JD, Duarte B, Rodriguez Fermepin M, Martinelli M, et al. Syphilis in adolescent mothers in the city of Posadas, Province of Misiones (in Spanish). Medicina (B Aires) 1999;59(5 Pt 1):437–45. [19] Cruz AR, Castrillón MA, Minotta AY, Rubiano LC, Castaño MC, Salazar JC. Gestational and congenital syphilis epidemic in the Colombian Pacific Coast. Sex Transm Dis 2013;40(10):813–8. [20] Tinajeros F, Grossman D, Richmond K, Steele M, Garcia SG, Zegarra L, et al. Diagnostic accuracy of a point-of-care syphilis test when used among pregnant women in Bolivia. Sex Transm Infect 2006;82(Suppl. 5):v17–21. [21] Alexander JM, Sheffield JS, Sanchez PJ, Mayfield J, Wendel Jr GD. Efficacy of treatment for syphilis in pregnancy. Obstet Gynecol 1999;93(1):5–8. [22] Centers for Disease Control and Prevention (CDC). Epidemic of congenital syphilis— Baltimore, 1996–1997. MMWR Morb Mortal Wkly Rep 1998;47(42):904–7. [23] Mascola L, Pelosi R, Blount JH, Binkin NJ, Alexander CE, Cates Jr W. Congenital syphilis. Why is it still occurring? JAMA 1984;252(13):1719–22. [24] Räisänen S, Gissler M, Saari J, Kramer M, Heinonen S. Contribution of risk factors to extremely, very and moderately preterm births - register-based analysis of 1,390,742 singleton births. PLoS One 2013;8(4):e60660. [25] Baroutis G, Mousiolis A, Mesogitis S, Costalos C, Antsaklis A. Preterm birth trends in Greece, 1980–2008: a rising concern. Acta Obstet Gynecol Scand 2013;92(5):575–82.
Please cite this article as: Arnesen L, et al, Gestational syphilis and stillbirth in Latin America and the Caribbean, Int J Gynecol Obstet (2014), http:// dx.doi.org/10.1016/j.ijgo.2014.09.017
Contents lists available at ScienceDirect
International Journal of Gynecology and Obstetrics journal homepage: www.elsevier.com/locate/ijgo
CLINICAL ARTICLE
Gestational syphilis and stillbirth in Latin America and the Caribbean Lauren Arnesen ⁎, Gerardo Martínez, Luis Mainero, Suzanne Serruya, Pablo Durán Pan American Health Organization, Centro Latinoamericano de Perinatología, Salud de la Mujer y Reproductiva, Montevideo, Uruguay
a r t i c l e
i n f o
Article history: Received 10 April 2014 Received in revised form 3 September 2014 Accepted 3 November 2014 Keywords: Caribbean Latin America Mother-to-child transmission Stillbirth Syphilis
a b s t r a c t Objective: To measure the association between gestational syphilis and stillbirth in Latin America and the Caribbean. Methods: In a retrospective study, data on stillbirth and gestational syphilis extracted from the Sistema Informático Perinatal database were analyzed for deliveries in 11 countries between January 1, 2009, and December 31, 2012. Potential confounders were examined, and binary logistic regression analysis was performed to assess the association between gestational syphilis and stillbirth. Results: Among 368 151 deliveries, 3875 (1.1%) were by women with a positive syphilis test, and 1461 (0.4%) were stillbirths. Among the stillbirths, 29 (2.0%) were delivered by women with a positive syphilis test. After controlling for country, congenital anomalies, gestational age at labor, maternal age, and previous stillbirth, gestational syphilis was significantly associated with stillbirth (odds ratio 1.88, 95% confidence interval 1.25–2.83; P = 0.002). Conclusion: Gestational syphilis contributes to stillbirth in Latin America and the Caribbean. Interventions targeting gestational syphilis are highly cost-effective and should be implemented across the region. © 2014 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved.
1. Introduction WHO estimates that 1.36–2 million pregnancies are affected by syphilis every year [1–3], of which 50%–80% will result in adverse outcomes without proper treatment [2,4–6]. Prominent among these outcomes is stillbirth, which occurs in 25%–40% of cases [4–6]. Globally, syphilis is the main cause of more than 212 000 stillbirths annually [2,3,7], and, in areas with a high prevalence of syphilis, as many as half of all stillbirths can be attributed to this infection [5,8]. Worldwide, Latin America and the Caribbean (LAC) has the highest incidence of syphilis, and accounts for up to 25% of the 2 million annual cases of gestational syphilis [1,9]. The prevalence of gestational syphilis in LAC varies from 0.08% to 7.0% by country [2,10]. Every year, an estimated 100 000 stillbirths in the region are attributable to congenital syphilis, which is defined as a neonate delivered to a mother who was untreated or inadequately treated for syphilis during pregnancy, or a neonate with a positive syphilis test [1,10,11]. The Pan American Health Organization (PAHO) approved the Regional Elimination of Congenital Syphilis Plans of Action in 1995 and 2010, in which elimination is defined as an incidence of 0.5 or fewer cases of congenital syphilis, including stillbirths, per 1000 deliveries [1,9,10]. Elimination of congenital syphilis is possible with proper treatment—one dose of penicillin is almost 100% effective at preventing syphilis-associated adverse outcomes in pregnancy [2,5,7, 9,12]. Moreover, the cost-effectiveness of prenatal syphilis screening ⁎ Corresponding author at: Centro Latinoamericano de Perinatología, Salud de la Mujer y Reproductiva, Avenue Italia # Hospital de Clinicas, Piso 16, Montevideo 11100, Uruguay. Tel.: +598 2 487 2929. E-mail address: [email protected] (L. Arnesen).
and treatment has been shown repeatedly and is considered one of the most cost-effective interventions available [2,4,5,10,12]. Nevertheless, syphilis-attributable stillbirths are still occurring in LAC. There are few studies on the relationship between gestational syphilis and stillbirth in LAC. Two previous studies in the region showed that the risk of stillbirth increases with exposure to gestational syphilis: one [13] reported that women with a positive syphilis test in Jamaica have a 3.04-fold increased odds of fetal death, whereas the other [14] indicated that women with a positive syphilis test in LAC countries have a 2.41-fold higher risk of fetal death. Generally, research in LAC has focused on the prevalence of gestational syphilis and calls to action to test and treat pregnant women [14], rather than examining factors that influence the relationship between gestational syphilis and stillbirth, and only one study has approached the problem at the regional level [13,15–19]. Sistema Informático Perinatal (SIP) is a PAHO/WHO database, developed by the Latin American Center for Perinatology and Human Development (CLAP), which includes a set of standardized instruments designed for obstetric and neonatal health services [14]. Every entry in SIP corresponds to a birth record containing data on maternal demographics; family and obstetric history; prenatal visits, delivery, and postpartum details; and discharge information for both the mother and neonate. Information is prospectively recorded at the clinic level during prenatal, delivery, and postnatal care. In addition to recording clinical data, SIP facilitates the electronic storage of information for further processing and analysis for managerial, epidemiologic, and other purposes. The primary aim of the present study was to use SIP data for deliveries occurring from 2009 to 2012 to measure the association between gestational syphilis and stillbirth at the regional level, with adjustments
http://dx.doi.org/10.1016/j.ijgo.2014.09.017 0020-7292/© 2014 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved.
Please cite this article as: Arnesen L, et al, Gestational syphilis and stillbirth in Latin America and the Caribbean, Int J Gynecol Obstet (2014), http:// dx.doi.org/10.1016/j.ijgo.2014.09.017
2
L. Arnesen et al. / International Journal of Gynecology and Obstetrics xxx (2014) xxx–xxx
for potential confounding factors. A secondary aim was to identify highrisk groups through tests for confounders, mediators, and effect measure modifiers, to focus future gestational syphilis interventions both in LAC and globally. 2. Materials and methods In a retrospective study, SIP data on stillbirth and gestational syphilis were analyzed for deliveries recorded in LAC countries between January 1, 2009, and December 31, 2012. Because data were extracted via a computer-generated identification code that was not linked in any way to patients, the study was exempt from institutional review board approval and informed consent was not needed. Use and coverage of SIP varies within the LAC region and within countries. Data collection and entry are done at the institutional level, and data are reported to CLAP through the respective national ministry of health. Deliveries in 15 LAC countries that use SIP and share their data with CLAP were considered for the present study. Of these countries, four did not have syphilis data recorded for the time period of the study and so were excluded. The 11 countries included in our study were: Argentina, Bolivia, Colombia, Ecuador, Guyana, Haiti, Honduras, Nicaragua, Paraguay, El Salvador, and Uruguay. The countries were randomly labeled with letters to maintain confidentiality because the study aim was not to compare specific countries, but to carry out a regional analysis of the relationship between gestational syphilis and stillbirth. For the analysis, data were extracted from the SIP records on syphilis test results, delivery outcome, and potential confounders that might biologically influence the relationship between gestational syphilis and stillbirth (Box 1). Records that were not complete were excluded from the study. The definition of stillbirth varies by study [5,8,12,13,15]. In line with recent studies [5,12], for the present analysis, stillbirth was defined as the neonate being dead at delivery. Gestational syphilis was defined as a positive result on the mother’s most recent prenatal non-treponemal or treponemal syphilis test [19]. The results from at least one syphilis test for the mother during pregnancy were available for all included deliveries. Traditionally, syphilis screening has been carried out in laboratories with non-treponemal tests, the results of which can take a
few weeks to be delivered in rural areas [20]. New point-of-care (POC) treponemal tests react to antibodies for Treponema pallidum, which usually remain in an individual’s system even after successful treatment and require only 20 minutes for detection [20]. Use of a treponemal or nontreponemal syphilis test, and what constituted a positive response for gestational syphilis, was determined at the institutional level. Among the SIP variables analyzed, “previous stillbirth” indicated that the mother had one or more previous stillbirths. Congenital anomalies were classified at the institutional level in SIP as “no,” “minor,” or “major,” and were analyzed as an ordinal variable. Both gestational age at labor and maternal age were analyzed as continuous variables, and then were grouped in accordance with international norms [21,22] for comparison. Statistical analyses were performed with SPSS version 17.0 (SPSS Inc, Chicago, IL, USA). After extraction of delivery outcome, gestational syphilis exposure, and all potential confounders from SIP, an unadjusted binomial logistic regression analysis was carried out for gestational syphilis and stillbirth. Each potential covariate was then analyzed separately with syphilis and delivery outcome in unique binary logistic regression models. Syphilis and all variables that were significantly associated with stillbirth when controlling for syphilis in their respective binary logistic regression model were then analyzed in a larger model, which also adjusted for syphilis exposure status. Covariates that were not significant in the larger model were excluded, resulting in a final model. All independent variables in the final model were tested for multicollinearity with linear regression by using dummy variables for all nominal variables. In the final model, confounding among covariates not thought to be in the causal pathway (all but gestational age at labor) was checked with Pearson χ2 tests of the covariate with syphilis and birth outcome separately, and the crude and Mantel-Haenszel odds ratios (ORs) were compared. Covariates that were not confounders, but also not in the causal pathway, were examined for statistical evidence of effect measure modification with calculations of the difference in the syphilis– stillbirth relationship across strata of the covariate with the beta for the interaction term (covariate*syphilis) in a binary logistic regression, in which syphilis, the covariate, and the interaction term are the independent variables. For the variable thought to be in the causal pathway—
Box 1 Covariates tested for significance in the relationship between gestational syphilis and stillbirth. Active smoker Maternal age Alcohol use Rubella vaccine Group B streptococcus carrier Tetanus vaccine Bacteriuria Birth weight Cervix check Chagas disease Chronic hypertension Civil status Companion at delivery Companion at labor Congenital defect Contraceptive failure Country Chorioamnionitis infection Diabetes Eclampsia Heart disease Kidney disease Pre-eclampsia
Urinary tract infection Drug use Education Ethnic origin Gestational age at labor Hemorrhage Family history of diabetes Maternal history of diabetes Family history of eclampsia Maternal history of eclampsia Maternal history of heart disease Family history of hypertension Maternal history of hypertension Maternal history of kidney disease Family history of preeclampsia Maternal history of preeclampsia History of violence toward mother HIV infection Hospitalized during pregnancy Intrauterine growth restriction Literate Multiple pregnancy Number of antenatal visits
Passive smoker Planned pregnancy Pregnancy-induced hypertension Presentation at labor Previous body mass index Previous pregnancies Previous live births Previous stillbirths Premature rupture of membranes Rupture of membranes b 37 wk Syphilis test treponemal Syphilis test non-treponemal Mother lives alone Three consecutive spontaneous abortions Toxoplasmosis Violence to mother Treated for syphilis Partner treated for syphilis Week of syphilis test nontreponemal Week of syphilis treatment Week of syphilis test treponemal
Please cite this article as: Arnesen L, et al, Gestational syphilis and stillbirth in Latin America and the Caribbean, Int J Gynecol Obstet (2014), http:// dx.doi.org/10.1016/j.ijgo.2014.09.017
L. Arnesen et al. / International Journal of Gynecology and Obstetrics xxx (2014) xxx–xxx
3
3. Results
All births in SIP from 2009–2012 (n=712 081) Missing syphilis and/or birth outcome (n=213 217) Have response for syphilis and birth outcome (n=480 864) Missing a variable/response for at least one covariate a (n=112 713) Have response/value for all covariates a in model (n=368 151)
Fig. 1. Flow diagram of the study population. Abbreviation: SIP, Sistema Informático Perinatal. a Country, gestational age at labor, maternal age, congenital anomaly, previous stillbirths.
gestational age at labor—the crude and Mantel-Haenszel ORs were examined to check for mediation. A P value of 0.05 was used as a cutoff for statistical significance. A difference of 10% or more was considered to be significant for the tests.
Of the 712 081 birth records contained in SIP during the study period, 368 151 (51.7%) met the inclusion criteria (Fig. 1). There were 3875 deliveries by women with a positive syphilis test and 1461 stillbirths in the study population. Of the stillbirths, 29 (2.0%) were delivered by women with a positive syphilis test, and 1432 (98.0%) were delivered to women with a negative syphilis test. Table 1 shows the distribution of each variable by both gestational syphilis status and delivery outcome. With the presence and increased severity of congenital anomalies, the proportion of cases with gestational syphilis and stillbirths increased. A total of 3830 (1.0%) of 365 172 deliveries without congenital anomalies, 21 (1.3%) of 1569 deliveries with minor anomalies, and 24 (1.7%) of 1410 deliveries with major anomalies were by women with a positive syphilis test. Additionally, 1279 (0.4%) deliveries without congenital anomalies, 23 (1.5%) with minor anomalies, and 159 (11.3%) with major anomalies were stillbirths. The mean gestational age at labor was 38.1 weeks (range 1.0–44.0) and mean maternal age was 24.9 years (range 10.0–55.0). Most mothers were aged between 20 and 34 years (Table 1). Overall, 318 (0.4%) of 87 191 deliveries to mothers younger than 20 years, 869 (0.4%) of 243 018 deliveries to mothers aged 20–34 years, and 274 (0.7%) of 37 942 deliveries to mothers aged 35 years or older were stillbirths. The proportion with a previous stillbirth was higher among women with a positive syphilis test than among women with a negative syphilis test, and was also higher among those who had a stillbirth in the study period than among those with a live birth (Table 1). The crude binary logistic regression model showed that gestational syphilis significantly increased the odds of stillbirth (P b 0.001) (Table 2). After controlling for country, congenital anomaly, previous stillbirth, gestational age at labor, and maternal age, gestational syphilis remained associated with stillbirth (P = 0.002) (Table 2). In the adjusted model, major congenital anomalies were significantly associated with stillbirth (P b 0.001), as was previous stillbirth
Table 1 Covariates by gestational syphilis status and delivery outcome.a Covariate
Syphilis Positive Negative Country A B C D E F G H I J K Congenital anomaly No Minor Major Previous stillbirth Yes No Gestational age at labor, wk b28 28–31 32–36 ≥37 Maternal age, y b20 20–34 ≥35
Gestational syphilis status
Delivery outcome
Total
Positive (n = 3875)
Negative (n = 364 276)
Stillbirth (n = 1461)
Live birth (n = 366 690)
– –
– –
29 (2.0) 1432 (98.0)
3846 (1.0) 362 844 (99.0)
3875 364 276
318 (8.2) 190 (4.9) 659 (17.0) 38 (1.0) 3 (0.1) 150 (3.9) 132 (3.4) 2 (b0.1) 0 630 (16.3) 1753 (45.2)
101 422 (27.8) 54 149 (14.9) 53 468 (14.7) 4366 (1.2) 296 (0.1) 11 337 (3.1) 11 645 (3.2) 49 (b0.1) 15 (b0.1) 17 102 (4.7) 110 427 (30.3)
500 (34.2) 246 (16.8) 231 (15.8) 78 (5.3) 0 27 (1.9) 166 (11.4) 1 (0.1) 0 44 (3.0) 168 (11.5)
101 240 (27.6) 54 093 (14.8) 53 896 (14.7) 4326 (1.2) 299 (0.1) 11 460 (3.1) 11 611 (3.2) 50 (b0.1) 15 (b0.1) 17 688 (4.8) 112 012 (30.5)
101 740 54 339 54 127 4404 299 11 487 11 777 51 15 17 732 112 180
3830 (98.8) 21 (0.6) 24 (0.6)
361 342 (99.2) 1548 (0.4) 1386 (0.4)
1279 (87.5) 23 (1.6) 159 (10.9)
363 893 (99.2) 1546 (0.4) 1251 (0.4)
365 172 1569 1410
193 (5.0) 3682 (95.0)
5188 (1.4) 359 088 (98.6)
77 (5.3) 1384 (94.7)
5304 (1.4) 361 386 (98.6)
5381 362 770
30 (0.8) 57 (1.5) 388 (10.0) 3400 (87.7)
1926 (0.5) 2871 (0.8) 27 658 (7.6) 331 821 (91.1)
263 (18.0) 171 (11.7) 378 (25.9) 649 (44.4)
1693 (0.5) 2757 (0.8) 27 668 (7.5) 334 572 (91.2)
1956 2928 28 046 335 221
788 (20.3) 2647 (68.3) 440 (11.4)
86 403 (23.7) 240 371 (66.0) 37 502 (10.3)
318 (21.7) 869 (59.5) 274 (18.8)
86 873 (23.7) 242 149 (66.0) 37 668 (10.3)
87 191 243 018 37 942
a
Please cite this article as: Arnesen L, et al, Gestational syphilis and stillbirth in Latin America and the Caribbean, Int J Gynecol Obstet (2014), http:// dx.doi.org/10.1016/j.ijgo.2014.09.017
4
L. Arnesen et al. / International Journal of Gynecology and Obstetrics xxx (2014) xxx–xxx
Table 2 Binomial logistic regression models of variables associated with stillbirth. Variable Crude Syphilis Positive Negative Adjusted Syphilis Positive Negative Country A B C D E F G H I J Kb Congenital anomaly No Minor Major Previous stillbirth Yes No Gestational age at laborc Maternal aged
Stillbirthsa
Wald χ2
Odds ratio (95% CI)
P value
29/3875 (0.7) 1432/364 276 (0.4)
–
2.18 (1.88–2.51) 1
b0.001 –
29/3875 (0.7) 1432/364 276 (0.4)
9.25 –
1.88 (1.25–2.83) 1
0.002 –
500/101 740 (0.5) 246/54 339 (0.5) 231/54 127 (0.2) 78/4404 (1.8) 0/299 27/11 487 (0.2) 166/11 777 (1.4) 1/51 (2.0) 0/15 44/17 732 (0.2) 168/112 180 (0.1)
234.22 166.59 144.36 360.59 0.00 99.76 415.18 0.14 0.00 8.52 –
4.26 (3.54–5.13) 3.94 (3.20–4.85) 3.63 (2.94–4.48) 15.63 (11.77–20.76) 0.00 (0.00–0.00) 0.08 (0.05–0.14) 10.52 (8.39–13.20) 2.22 (0.03–145.37) 0.00 (0.00–0.00) 1.68 (1.19–2.42) 1
b0.001 b0.001 b0.001 b0.001 0.998 b0.001 b0.001 0.708 0.998 0.004 –
1279/365 172 (0.4) 23/1569 (1.5) 159/1410 (11.3)
– 39.48 1106.93
1 4.01 (2.60–6.18) 25.90 (21.38–31.37)
– b0.001 b0.001
77/5381 (1.4) 1384/362 770 (0.4) 1461/368 151 (0.4) 1461/368 151 (0.4)
67.81 – 2238.60 86.76
2.85 (2.22–3.65) 1 0.81 (0.80–0.82) 1.04 (1.03–1.05)
b0.001 – b0.001 b0.001
Abbreviation: CI, confidence interval. a Values are given as total number of stillbirths/number of deliveries (percentage) unless indicated otherwise. b Randomly selected as the reference to allow for comparison between countries. c As a continuous variable, in increments of 1 week. d As a continuous variable, in increments of 1 year.
(P b 0.001) (Table 2). Gestational age at labor was the only protective factor (Table 2). There was a varying risk of stillbirth by country (Table 2). There was no evidence of multicollinearity among the independent variables in the final model (variance inflation factor b 2). Country was found to be a confounder of the relationship between gestational syphilis and stillbirth (Mantel-Haenszel OR 1.6; 95% confidence interval [CI] 1.4–1.8), and gestational age at labor was a mediator (Mantel-Haenszel OR 1.9; 95% CI 1.5–2.3) (Fig. 2). Mother’s age, congenital anomalies, and previous stillbirths were effect measure modifiers (Fig. 2). For mother’s age, the OR without the interaction effect was 2.60 (95% CI 1.47–4.61); the OR with the interaction effect was 0.99 (95% CI 0.97–1.01). The OR without congenital anomalies as the interaction term was 1.48 (95% CI 1.01–2.17); with major congenital anomalies, the OR was 1.59 (95% CI 0.59–4.30). Without previous
Maternal Age
Congenital Anomaly
Gestational Syphilis
Previous Stillbirth(s)
Stillbirth Gestational Age at Labor
KEY Confounder Mediator Effect measure modifier
Country
Fig. 2. Directed acyclic graph of gestational syphilis, stillbirth and associated covariates.
stillbirths as an interaction term, the OR was 2.41 (95% CI 2.05–2.83); with previous stillbirths, the OR was 0.42 (95% CI 0.21–0.85).
4. Discussion In the present study, gestational syphilis was associated with increased odds of stillbirth in LAC. This relationship remained when controlling for country, gestational age at labor, maternal age, congenital anomalies, and previous stillbirths. The present study analyzed recent data, used a very large study population, and included 11 countries from across the region. The increased likelihood of stillbirth after exposure to gestational syphilis recorded in the present study has been documented previously [1,13–19]. Outside LAC, a study of all congenital syphilis cases reported in LAC in 1982 in Texas, USA [23], found that the risk of stillbirth was 6.5 times higher for those exposed to gestational syphilis. Although odds and risk cannot be compared to one other, these previous results support the present conclusion that gestational syphilis is a clear factor in the incidence of stillbirth. The 11 LAC countries included in the study were not randomly selected; only those using SIP and sharing data with CLAP were eligible. Therefore, the regional representation of the current relationship between gestational syphilis and stillbirth could be skewed toward the national situation in the study countries. Nevertheless, the results remain the most representative of the situation in LAC to date. Furthermore, the inclusion of the 11 countries in the region does not alter the weight of the finding that gestational syphilis is related to increased odds of stillbirth. Country of birth was found to be a confounder in the relationship between gestational syphilis and stillbirth. As compared with the reference country, the OR for stillbirth in the presence of gestational syphilis
Please cite this article as: Arnesen L, et al, Gestational syphilis and stillbirth in Latin America and the Caribbean, Int J Gynecol Obstet (2014), http:// dx.doi.org/10.1016/j.ijgo.2014.09.017
L. Arnesen et al. / International Journal of Gynecology and Obstetrics xxx (2014) xxx–xxx
ranged from 0.08 to 15.63. Large differences in the odds of stillbirth by country are probably a reflection of healthcare systems—specifically, access to, and quality of, care—and whether POC syphilis testing and treatment is available. The ability to test and treat women on the same day has been cited as a cost-effective and successful way to prevent adverse birth outcomes associated with untreated gestational syphilis [3,7,12, 19,20]. Future studies looking at health systems with POC tests would help health facilities to identify other factors contributing to untreated cases of gestational syphilis. Gestational age at labor was a mediator in the final model. It seems that gestational syphilis influences the prematurity of the delivery, which affects the likelihood of a live birth. The transplacental passage of Treponema pallidum has been cited as a risk factor for prematurity [2,3], and prematurity has been shown to result in an increased likelihood of stillbirth [24,25], highlighting the importance of testing and treating women as early as possible to prevent these adverse birth outcomes. Increasing maternal age, previous stillbirth, and presence of congenital anomalies were effect measure modifiers in the final model. In addition to other routine testing at the standard prenatal care visit, syphilis testing and treatment for older mothers (≥ 35 years), those with at least one previous stillbirth, and those who meet both criteria, should be emphasized and incorporated into interventions targeting these high-risk groups. The present study has some limitations. First, what constituted a positive test for gestational syphilis was determined at the institutional level. Although unlikely, some institutions could deviate from international standards when determining whether a non-treponemal or POC syphilis test is positive. If some institutions were using different criteria to define cases of syphilis, this effect would most probably be uniform across exposures, and thus would not significantly affect the observed relationship between gestational syphilis and stillbirth. Second, gestational syphilis was defined by the treponemal or nontreponemal test result closest to delivery. The decision of whether a syphilis test was positive or negative was made at the institutional level. Internationally, a reactive treponemal test confirmed by a nontreponemal test is considered confirmatory for syphilis [6,19]. For the few cases that received only the treponemal test, a positive result indicated the presence of Treponemal pallidum antibodies, and not whether treatment was successful or not, which might bias the results away from the null. Third, there are new variables in SIP that were not available for analysis in the present study. These variables should be incorporated in future analyses, including maternal and partner treatment for syphilis and postpartum syphilis testing, because they might indicate reinfection of a woman who tests negative or has been successfully treated for syphilis. The absence of these variables from the present model might bias the results toward the null, a possibility that was taken into account by using the syphilis test result closest to delivery. In summary, the most recent data from SIP have been analyzed with the intention of conveying the present relationship between gestational syphilis and stillbirth in LAC. The present analysis represents a picture of the current situation. As more countries begin to use the system, share their data, and increase syphilis testing coverage, future studies based on SIP will improve the understanding of gestational syphilis and stillbirth at the regional level.
5
Conflict of interest The authors have no conflicts of interest. References [1] Pan American Health Organization, Latin American Center for Perinatology, Women and Reproductive Health, United Nations Children’s Fund, UNAIDS. 2010 situation analysis: Elimination of mother-to-child transmission of HIV and congenital syphilis in the Americas. http://www.paho.org/hq/index.php?option=com_docman&task= doc_download&gid=15893&Itemid. Published 2011. Accessed October 29, 2014. [2] Walker DG, Walker GJ. Forgotten but not gone: the continuing scourge of congenital syphilis. Lancet Infect Dis 2002;2(7):432–6. [3] Newman L, Kamb M, Hawkes S, Gomez G, Say L, Seuc A, et al. Global estimates of syphilis in pregnancy and associated adverse outcomes: analysis of multinational antenatal surveillance data. PLoS Med 2013;10(2):e1001396. [4] Schmid G. Economic and programmatic aspects of congenital syphilis prevention. Bull World Health Organ 2004;82(6):402–9. [5] Goldenberg RL, McClure EM, Saleem S, Reddy UM. Infection-related stillbirths. Lancet 2010;375(9724):1482–90. [6] Schmid GP, Stoner BP, Hawkes S, Broutet N. The need and plan for global elimination of congenital syphilis. Sex Transm Dis 2007;34(7 Suppl.):S5–S10. [7] Mabey DC, Sollis KA, Kelly HA, Benzaken AS, Bitarakwate E, Changalucha J, et al. Point-of-care tests to strengthen health systems and save newborn lives: the case of syphilis. PLoS Med 2012;9(6):e1001233. [8] Goldenberg RL, Thompson C. The infectious origins of stillbirth. Am J Obstet Gynecol 2003;189(3):861–73. [9] Pan American Health Organization, World Health Organization, Latin American Center for Perinatology, Women and Reproductive Health. Strategy and plan of action for elimination of mother-to-child transmission of HIV and congenital syphilis: regional monitoring strategy. http://www.paho.org/hq/index.php?option=com_ docman&task=doc_view&gid=24379&Itemid=. Published 2013. Accessed October 29, 2014. [10] Pan American Health Organization. Epidemiological profiles of neglected diseases and other infections related to poverty in Latin America and the Caribbean. http://www2.paho.org/hq/dmdocuments/2009/nds-epi-profiles.pdf. Published September 2009. Accessed October 29, 2014. [11] Centers for Disease Control and Prevention (CDC). Congenital syphilis – United States, 2003–2008. MMWR Morb Mortal Wkly Rep 2010;59(14):413–7. [12] Di Mario S, Say L, Lincetto O. Risk factors for stillbirth in developing countries: a systematic review of the literature. Sex Transm Dis 2007;34(7 Suppl.):S11–21. [13] Greenwood R, Foster-Williams K, Ashley D, Keeling J, Golding J. The epidemiology of antepartum fetal death in Jamaica. Paediatr Perinat Epidemiol 1994;8(Suppl. 1): 98–109. [14] Conde-Aguadelo A, Belizán JM, Díaz-Rossello JL. Epidemiology of fetal death in Latin America. Acta Obstet Gynecol Scand 2000;79(5):371–8. [15] Southwick KL, Blanco S, Santander A, Estenssoro M, Torrico F, Seoane G, et al. Maternal and congenital syphilis in Bolivia, 1996: prevalence and risk factors. Bull World Health Organ 2001;79(1):33–42. [16] Fitzgerald DW, Behets FM, Lucet C, Roberfroid D. Prevalence, burden, and control of syphilis in Haiti’s rural Artibonite region. Int J Infect Dis 1998;2(3):127–31. [17] Ali Z. Resurgence of congenital syphilis in Trinidad. J Trop Pediatr 1990;36(3):104–8. [18] Jacquier N, Dos Santos L, Deschutter JD, Duarte B, Rodriguez Fermepin M, Martinelli M, et al. Syphilis in adolescent mothers in the city of Posadas, Province of Misiones (in Spanish). Medicina (B Aires) 1999;59(5 Pt 1):437–45. [19] Cruz AR, Castrillón MA, Minotta AY, Rubiano LC, Castaño MC, Salazar JC. Gestational and congenital syphilis epidemic in the Colombian Pacific Coast. Sex Transm Dis 2013;40(10):813–8. [20] Tinajeros F, Grossman D, Richmond K, Steele M, Garcia SG, Zegarra L, et al. Diagnostic accuracy of a point-of-care syphilis test when used among pregnant women in Bolivia. Sex Transm Infect 2006;82(Suppl. 5):v17–21. [21] Alexander JM, Sheffield JS, Sanchez PJ, Mayfield J, Wendel Jr GD. Efficacy of treatment for syphilis in pregnancy. Obstet Gynecol 1999;93(1):5–8. [22] Centers for Disease Control and Prevention (CDC). Epidemic of congenital syphilis— Baltimore, 1996–1997. MMWR Morb Mortal Wkly Rep 1998;47(42):904–7. [23] Mascola L, Pelosi R, Blount JH, Binkin NJ, Alexander CE, Cates Jr W. Congenital syphilis. Why is it still occurring? JAMA 1984;252(13):1719–22. [24] Räisänen S, Gissler M, Saari J, Kramer M, Heinonen S. Contribution of risk factors to extremely, very and moderately preterm births - register-based analysis of 1,390,742 singleton births. PLoS One 2013;8(4):e60660. [25] Baroutis G, Mousiolis A, Mesogitis S, Costalos C, Antsaklis A. Preterm birth trends in Greece, 1980–2008: a rising concern. Acta Obstet Gynecol Scand 2013;92(5):575–82.
Please cite this article as: Arnesen L, et al, Gestational syphilis and stillbirth in Latin America and the Caribbean, Int J Gynecol Obstet (2014), http:// dx.doi.org/10.1016/j.ijgo.2014.09.017
