Syphilis screening during pregnancy helps prevent congenital syphilis. The harms associated with false positive (FP) screens and whether screening leads to correct treatments has not been well determined.
The population included mothers and infants from 75 056 pregnancies. Using laboratory-based criteria we classified initial positive syphilis screens as FP or true positive (TP) and calculated false discovery rates. For mothers and infants we determined treatments, clinical characteristics, and syphilis classifications.
There were 221 positive screens: 183 FP and 38 TP. The false discovery rate was 0.83 (95% confidence interval [CI], 0.78–0.88). False discovery rates were similar for traditional 0.83 [95% CI, 0.72–0.94] and reverse algorithms 0.83 (95% CI, 0.77–0.88), and for syphilis Immunoglobin (Ig) G 0.79 (95% CI, 0.71–0.86) and total 0.90 (95% CI, 0.82–0.97) assays. FP screens led to treatment in 2 women and 1 infant. Two high-risk women were not rescreened at delivery and were diagnosed after hospital discharge; 1 infant developed congenital syphilis. Among 15 TP women with new syphilis, the diagnosis was before the late third trimester in 14 (93%). In one-half of these women, there was concern for reinfection, treatment failure, inadequate treatment or follow-up care, or late treatment, and their infants did not achieve an optimal syphilis classification.
Syphilis screening identifies maternal syphilis, but limitations include FP screens, which occasionally lead to unnecessary treatment, inconsistent risk-based rescreening, and among TP mothers failure to optimize care to prevent birth of infants at higher risk for congenital syphilis.
Screening for syphilis during pregnancy is an important measure to prevent congenital syphilis. The likelihood of harm associated with false positive screens and of failure of true positive screens to result in correct treatment has not been well determined.
Syphilis screening identifies maternal syphilis, but limitations include false positive screens, which occasionally lead to unnecessary treatment, inconsistent risk-based rescreening, and among true positive mothers failure to optimize care to prevent birth of infants at higher risk for congenital syphilis.
Screening for syphilis is an important public health practice necessary but not sufficient to prevent the birth of infants at risk for congenital infection.1–4 Syphilis screening and treatment early in pregnancy helps prevent intrauterine infection, which can result in complications such as stillbirth, hydrops fetalis, or preterm birth. For women at high risk for syphilis and in populations at high risk, screening needs to be repeated early in the third trimester and again at delivery to detect new infection or reinfection. Based on maternal treatment and an evaluation for clinical or laboratory features of syphilis, infants are classified according to how likely they are to have congenital syphilis.1 These infants are assigned 1 of 4 congenital syphilis risk classifications. Treatment recommendations and duration vary according to the infant’s congenital syphilis risk classification and whether follow-up can be assured. For infants at higher risk of congenital syphilis, including those born to women treated <4 weeks before delivery, the preferred treatment is 10 days of parenteral penicillin.
The traditional syphilis screening approach starts with an initial nontreponemal test, such as the rapid plasma reagin (RPR) test or Venereal Disease Research Laboratory (VDRL) test, and if positive reflexes to a confirmatory treponemal test (Fig 1). An alternative reverse screening approach starts with a treponemal test, such as the syphilis Immunoglobin (Ig) G assay or a syphilis total IgM and IgG assay, and if positive reflexes to a nontreponemal test (Fig 2).5,6 A second treponemal test can be used to adjudicate discordant results.7,8 A single screening test is insufficient to make the diagnosis of syphilis, because additional testing is required to distinguish true positive (TP) and false positive (FP) results. In many populations, FP results are more common than TP results.
In a large population, we sought to quantify the frequency of noncompliance with recommended maternal and neonatal evaluation after an initial positive screen for syphilis.
Methods
Study Population
The population included mothers from 75 056 pregnancy episodes who received perinatal care in Northeast Ohio through the Cleveland Clinic’s health care system and their 77 410 infants delivered between June 1, 2014 and February 28, 2021. In Cuyahoga County, Ohio, in which the health care system is centered, the syphilis rate increased from 26.4 to 41.7 per 100 000 from 2016 to 2020.9 Recommended screening for all women was at the first prenatal visit and for high-risk women again in both the early third trimester and at delivery. Universal early third trimester rescreening was recommended July 1, 2019. The Cleveland Clinic’s institutional review board reviewed the study and determined it an exempt quality improvement project.
Data Collection
Clinical data from our health care system’s internal perinatal outcomes database (STORK), which uniquely links mothers to their delivered children, was joined with syphilis laboratory data within a secure DataLab. Clinical data were managed using secure REDCap (Research Electronic Data Capture) tools.10,11 Chart reviews documented treatment, clinical characteristics, and syphilis stage.
Syphilis Screening Approaches
Syphilis screening assays were performed at the Cleveland Clinic’s main campus laboratory using consistent methodology. Traditional screening was performed using an RPR first assay until the phased introduction of reverse screening in 2014. For reverse screening, the laboratory used a syphilis IgG assay (BioPlexTM2200 syphilis IgG, Bio-Rad assay) with enhanced cutoffs (≤0.8 nonreactive, 0.9–5.9 weak reactive, and ≥ 6.0 reactive). In October 2019, reverse screening was introduced using a syphilis IgM and IgG assay (Bio-PlexTM2200 syphilis total IgM and IgG Bio-Rad) using the manufacturer's cutoffs (<0.8 nonreactive, 0.9–1.1 weak reactive, and ≥1.1 reactive).
Definition of True Positive and False Positive
Laboratory-based criteria were used to classify the first positive screen during the pregnancy as FP or TP. For traditional screening, a positive RPR was confirmed by a fluorescent treponemal antibody absorption test. For reverse screening, a positive or equivocal syphilis IgG or syphilis IgM and IgG was confirmed by an RPR, or if the RPR was negative by a positive Treponema pallidum antibody.7 For the traditional approach, a screen was defined as FP if the RPR was positive but not confirmed by either a florescent treponemal antibody absorption test or a T. pallidum antibody. For the reverse approach, the screen was judged FP if a positive or equivocal syphilis IgG or syphilis IgM and IgG assay was followed by both a negative RPR and a negative T. pallidum antibody. When no confirmatory test result was available and the initial screen could not be adjudicated, the patient was excluded from analysis.
TP women were classified as past diagnosis if diagnosed and treated for syphilis before pregnancy or as new diagnosis if diagnosed by screening with a new infection. Women treated for syphilis before the pregnancy were classified as new diagnosis if evidence of reinfection was documented on their initial screen. For women treated during pregnancy, the initial treatment was classified as adequate if the mother received stage appropriate therapy at least 4 weeks before delivery.1
Congenital Syphilis Stage
Infants born to TP mothers were assigned to 1 of 4 syphilis risk classifications according to Centers for Disease Control and Prevention (CDC) criteria: proven or highly probable congenital syphilis, possible congenital syphilis, congenital syphilis less likely, or congenital syphilis unlikely (Table 1).12
. | Definition . | Recommended Evaluation . |
---|---|---|
Confirmed proven or highly probable congenital syphilis | Any neonate with: • An abnormal physical examination that is consistent with congenital syphilis; • A serum quantitative nontreponemal serologic titer that is fourfold (or greater) higher than the mother’s titer at delivery (eg, maternal titer = 1:2, neonatal titer ≥1:8 or maternal titer = 1:8, neonatal titer ≥1:32); or • A positive darkfield test or PCR of placenta, cord, lesions, or body fluids or a positive silver stain of the placenta or cord. | CSF analysis for VDRL, cell count, and protein CBC and differential and platelet count Long-bone radiographs Other tests as clinically indicated (eg, chest radiograph, liver function tests, neuroimaging, ophthalmologic examination, and auditory brain stem response) |
Possible congenital syphilis | Any neonate who has a normal physical examination and a serum quantitative nontreponemal serologic titer equal to or less than fourfold of the maternal titer at delivery (eg, maternal titer = 1:8, neonatal titer ≤1:16) and 1 of the following: • The mother was not treated, was inadequately treated, or has no documentation of having received treatment. • The mother was treated with erythromycin or a regimen other than those recommended in these guidelines (ie, a nonpenicillin G regimen). • The mother received the recommended regimen but treatment was initiated <30 d before delivery. | CSF analysis for VDRL, cell count, and protein CBC, differential, and platelet count Long-bone radiographs |
Congenital syphilis less likely | Any neonate who has a normal physical examination and a serum quantitative nontreponemal serologic titer equal or less than fourfold of the maternal titer at delivery (eg, maternal titer = 1:8, neonatal titer ≤1:16) and both of the following are true: • The mother was treated during pregnancy, treatment was appropriate for the infection stage, and the treatment regimen was initiated ≥30 d before delivery. • The mother has no evidence of reinfection or relapse. | No evaluation is recommended. |
Congenital syphilis unlikely | Any neonate who has a normal physical examination and a serum quantitative nontreponemal serologic titer fourfold of the maternal titer at delivery and both of the following are true: • The mother’s treatment was adequate before pregnancy. • The mother’s nontreponemal serologic titer remained low and stable (ie, serofast) before and during pregnancy and at delivery (eg, VDRL ≤1:2 or RPR ≤1:4). | No evaluation is recommended. |
. | Definition . | Recommended Evaluation . |
---|---|---|
Confirmed proven or highly probable congenital syphilis | Any neonate with: • An abnormal physical examination that is consistent with congenital syphilis; • A serum quantitative nontreponemal serologic titer that is fourfold (or greater) higher than the mother’s titer at delivery (eg, maternal titer = 1:2, neonatal titer ≥1:8 or maternal titer = 1:8, neonatal titer ≥1:32); or • A positive darkfield test or PCR of placenta, cord, lesions, or body fluids or a positive silver stain of the placenta or cord. | CSF analysis for VDRL, cell count, and protein CBC and differential and platelet count Long-bone radiographs Other tests as clinically indicated (eg, chest radiograph, liver function tests, neuroimaging, ophthalmologic examination, and auditory brain stem response) |
Possible congenital syphilis | Any neonate who has a normal physical examination and a serum quantitative nontreponemal serologic titer equal to or less than fourfold of the maternal titer at delivery (eg, maternal titer = 1:8, neonatal titer ≤1:16) and 1 of the following: • The mother was not treated, was inadequately treated, or has no documentation of having received treatment. • The mother was treated with erythromycin or a regimen other than those recommended in these guidelines (ie, a nonpenicillin G regimen). • The mother received the recommended regimen but treatment was initiated <30 d before delivery. | CSF analysis for VDRL, cell count, and protein CBC, differential, and platelet count Long-bone radiographs |
Congenital syphilis less likely | Any neonate who has a normal physical examination and a serum quantitative nontreponemal serologic titer equal or less than fourfold of the maternal titer at delivery (eg, maternal titer = 1:8, neonatal titer ≤1:16) and both of the following are true: • The mother was treated during pregnancy, treatment was appropriate for the infection stage, and the treatment regimen was initiated ≥30 d before delivery. • The mother has no evidence of reinfection or relapse. | No evaluation is recommended. |
Congenital syphilis unlikely | Any neonate who has a normal physical examination and a serum quantitative nontreponemal serologic titer fourfold of the maternal titer at delivery and both of the following are true: • The mother’s treatment was adequate before pregnancy. • The mother’s nontreponemal serologic titer remained low and stable (ie, serofast) before and during pregnancy and at delivery (eg, VDRL ≤1:2 or RPR ≤1:4). | No evaluation is recommended. |
The material presented in this table was developed by the Centers for Disease Control.12 Reproduction and use of this material does not imply an endorsement by the US Government, Department of Health and Human Services, or Centers for Disease Control and Prevention. CBC, complete blood count; PCR, polymerase chain reaction.
Statistical Analysis
Data were described using medians, interquartile ranges, and minimums and maximums for continuous variables, and counts and percentages for categorical variables. Groups were defined by syphilis screen results and were compared on demographic and clinical characteristics and outcomes by using nonparametric Wilcoxon rank or Kruskal-Wallis tests for continuous and ordinal characteristics and χ2 or Fisher’s exact tests for categorical characteristics. False discovery rates (the number of false positive tests divided by the total number of positive tests) and their 95% binomial confidence intervals (CI) were calculated. Sample sizes for individual variables reflect missing data. All analyses were performed on a complete-case basis. All tests were 2-tailed and performed at a significance level of 0.05. SAS 9.4 software was used for all analyses.
Results
False Positive and True Positive Maternal Syphilis Results by Screening Algorithm and Gestational Age
Positive initial syphilis screens were documented for 221 women, 183 FP, and 38 TP (Table 2). The FP rate was 243.8 cases per 100 000 pregnancies, and the TP rate was 50.6 cases per 100 000. Of the TP women, 21 (55%) had a past diagnosis (28.0 cases per 100 000) and 17 (45%) had a new diagnosis (22.6 cases per 100 000). The false discovery rate, which is based on the number of FP tests divided by the total number of positive tests, was 0.83 [95% CI, 0.78–0.88]. False discovery rates did not differ overall by screening algorithm and were similar for traditional 0.83 [95% CI, 0.72–0.94] and reverse algorithms 0.83 [95% CI, 0.77–0.88], and for syphilis IgG 0.79 [95% CI, 0.71–0.86] and IgM and IgG 0.90 [95% CI, 0.82–0.97] assays. The gestational age at the first positive screening for FP and TP women did not differ (Supplemental Table 5).
Syphilis Screening Algorithm Used at Diagnosis . | Total Screened . | False Positive N (%) . | True Positive N (%) . | False Discovery Rate . | P . |
---|---|---|---|---|---|
All screens | 221 | 183 (82.8) | 38 (17.2) | 0.83 (0.78–0.88) | .16a |
Traditional algorithm (RPR first) | 46 | 38 (83) | 8 (17) | 0.83 (0.72–0.94) | .97b |
Reverse algorithms | 175 | 145 (83) | 30 (17) | 0.83 (0.77–0.88) | |
Syphilis IgG | 107 | 84 (79) | 23 (22) | 0.79 (0.71–0.86) | .055c |
Syphilis IgM/IgG | 68 | 61 (90) | 7 (10) | 0.90 (0.82–0.97) |
Syphilis Screening Algorithm Used at Diagnosis . | Total Screened . | False Positive N (%) . | True Positive N (%) . | False Discovery Rate . | P . |
---|---|---|---|---|---|
All screens | 221 | 183 (82.8) | 38 (17.2) | 0.83 (0.78–0.88) | .16a |
Traditional algorithm (RPR first) | 46 | 38 (83) | 8 (17) | 0.83 (0.72–0.94) | .97b |
Reverse algorithms | 175 | 145 (83) | 30 (17) | 0.83 (0.77–0.88) | |
Syphilis IgG | 107 | 84 (79) | 23 (22) | 0.79 (0.71–0.86) | .055c |
Syphilis IgM/IgG | 68 | 61 (90) | 7 (10) | 0.90 (0.82–0.97) |
Statistics presented as N (row %) and 95% confidence interval (P5–P95). P values: Pearson’s χ2 test.
Comparing traditional screen, syphilis IgG, and syphilis IgM and IgG.
Comparing traditional screen to reverse algorithms.
Comparing syphilis IgG to syphilis IgM and IgG.
Characteristics of Women With False Positive and True Positive Syphilis Tests
Women with FP and TP screens differed according to several characteristics (Table 3). Women with TP screens were more likely to be Black and to use marijuana. Likewise, women with a FP screen were more likely to be White and to have no history of drug use. Nonsyphilis sexually transmitted infections were rare among women with FP screens (5%) but occurred more commonly (33%) among women with TP screens (P < .001). Among women with TP screens, there were higher rates of primary and reactivated genital herpes simplex virus, Chlamydia trachomatis, and Trichomonas vaginalis. Mean birth weight for infants born to mothers with TP screens was lower than for infants born to mothers with FP screens (3035 g vs 3283 g; P = .05), although gestational age at delivery was similar.
Factor . | False Positive (N = 183) . | True Positive (N = 38) . | P . |
---|---|---|---|
Mothers | |||
Maternal age in years at positive screen | 30 (26–34) | 28 (24–32) | .11* |
Gravidity | 2 (1–3) | 2 (1–4) | .36* |
Parity | 1 (0–2) | 1 (0–2) | .64* |
Maternal race | .01*** | ||
White | 99 (54) | 12 (32) | |
Hispanic | 17 (9) | 3 (8) | |
Black | 40 (22) | 21 (55) | |
Multiracial | 11 (6) | 1 (3) | |
Asian | 10 (5) | 1 (3) | |
Indian | 1 (0.5) | 0 (0) | |
Unknown | 4 (2) | 0 (0) | |
Other | 1 (0.5) | 0 (0) | |
Substance use | |||
Tobacco | 7 (4) | 3 (8) | .38*** |
Alcohol | 2 (1) | 0 (0) | .99*** |
Marijuana | 8 (4) | 8 (21) | .002*** |
Cocaine | 2 (1) | 1 (3) | .43*** |
No drug use | 167 (91) | 28 (74) | .005** |
STI during pregnancy other than syphilis | |||
Chlamydia trachomatis | 4 (2) | 4 (11) | .03*** |
Neisseria gonorrheae | 0 (0) | 1 (3) | .17*** |
Trichomonas vaginalis | 0 (0) | 2 (5) | .03*** |
HIV | 0 (0) | 0 (0) | |
Herpes simplex virus primary | 0 (0) | 2 (5) | .03*** |
Herpes simplex virus reactivation | 1 (0.5) | 4 (11) | .003*** |
Hepatitis C | 1 (0.5) | 0 (0) | .99*** |
Human papillomavirusa | 3 (2) | 2 (5) | .21*** |
Hepatitis B | 1 (0.5) | 1 (3) | .31*** |
No STI | 173 (95) | 22 (66) | <.001*** |
Infants | |||
Gestational age at delivery in days | 274 (265–280) | 273 (267–276) | .36* |
Birth wt in gramsb | 3283 (2920–3584) | 3035 (2670–3440) | .05* |
Factor . | False Positive (N = 183) . | True Positive (N = 38) . | P . |
---|---|---|---|
Mothers | |||
Maternal age in years at positive screen | 30 (26–34) | 28 (24–32) | .11* |
Gravidity | 2 (1–3) | 2 (1–4) | .36* |
Parity | 1 (0–2) | 1 (0–2) | .64* |
Maternal race | .01*** | ||
White | 99 (54) | 12 (32) | |
Hispanic | 17 (9) | 3 (8) | |
Black | 40 (22) | 21 (55) | |
Multiracial | 11 (6) | 1 (3) | |
Asian | 10 (5) | 1 (3) | |
Indian | 1 (0.5) | 0 (0) | |
Unknown | 4 (2) | 0 (0) | |
Other | 1 (0.5) | 0 (0) | |
Substance use | |||
Tobacco | 7 (4) | 3 (8) | .38*** |
Alcohol | 2 (1) | 0 (0) | .99*** |
Marijuana | 8 (4) | 8 (21) | .002*** |
Cocaine | 2 (1) | 1 (3) | .43*** |
No drug use | 167 (91) | 28 (74) | .005** |
STI during pregnancy other than syphilis | |||
Chlamydia trachomatis | 4 (2) | 4 (11) | .03*** |
Neisseria gonorrheae | 0 (0) | 1 (3) | .17*** |
Trichomonas vaginalis | 0 (0) | 2 (5) | .03*** |
HIV | 0 (0) | 0 (0) | |
Herpes simplex virus primary | 0 (0) | 2 (5) | .03*** |
Herpes simplex virus reactivation | 1 (0.5) | 4 (11) | .003*** |
Hepatitis C | 1 (0.5) | 0 (0) | .99*** |
Human papillomavirusa | 3 (2) | 2 (5) | .21*** |
Hepatitis B | 1 (0.5) | 1 (3) | .31*** |
No STI | 173 (95) | 22 (66) | <.001*** |
Infants | |||
Gestational age at delivery in days | 274 (265–280) | 273 (267–276) | .36* |
Birth wt in gramsb | 3283 (2920–3584) | 3035 (2670–3440) | .05* |
Statistics presented as median (IQR), N (column %). STI, sexually transmitted infection.
P values:
Wilcoxon rank test,
Pearson’s χ2 test,
Fisher’s exact test.
Atypical squamous cells of undetermined significance and low-grade squamous intraepithelial lesions.
Wt was not available for one infant.
Clinical characteristics were similar for women treated for syphilis before the pregnancy and those with a new diagnosis (Supplemental Table 6). Infants born to TP mothers with past and new syphilis diagnoses had similar gestational ages and birth weights.
Influence of False Positive Maternal Syphilis Test on Maternal or Infant Treatment
To assess for harm from screening, we determined if FP mothers or their infants received evaluations or treatment of syphilis. Although most did not receive unnecessary evaluations or treatments, 2 mothers and 1 of their infants were evaluated and treated (Fig 3). Maternal screening was by the traditional approach in 1 and by the syphilis IgG assay in the other. Four women with FP screens during this pregnancy had previously received treatment of syphilis that likely was unnecessary.
Influence of the Timing of True Positive Maternal Syphilis Tests on Maternal Treatment and Infant Congenital Syphilis Status
To determine if timing of maternal syphilis diagnosis and treatment influenced infant congenital syphilis status, and consequently the need for infant evaluation and treatment, we reviewed the charts of women with TP screens (Fig 4). All 21 infants born to 21 women with a past syphilis diagnosis were classified as congenital syphilis unlikely. Three of their infants received a single dose of intramuscular benzathine penicillin G and 2 because the mothers were at high-risk for recent reinfection. The 17 women with newly diagnosed syphilis were staged as having primary syphilis (n = 5, 29%), early latent syphilis (n = 2, 12%), and syphilis of unknown duration (n = 10, 59%). Screening during pregnancy identified 15 of these women. Two high-risk patients were not rescreened at delivery and were diagnosed with syphilis after hospital discharge. One of their infants developed classic features of congenital syphilis. A diagnosis of possible congenital syphilis was made in the other 2 weeks postpartum after maternal screening by syphilis IgM and IgG assay, with appropriate confirmatory testing, suggested acute infection. We have previously described these patients in a case report.13
Among the 15 women with new syphilis diagnosed during pregnancy (Supplemental Table 7), 93% (n = 14) were identified during or before the early third trimester. Seven of their babies were classified as having congenital syphilis less likely. The reason for a higher congenital syphilis classification status in the other 7 infants included maternal treatment less than 4 weeks before delivery (n = 1), syphilis stage inappropriate initial maternal treatment (n = 1), concern for maternal reinfection or relapse (n = 4), and inadequate follow-up care without repeat maternal or infant RPRs (n = 1).
Syphilis Screening at Delivery in True Positive and False Positive Mothers and Their Infants
Repeat syphilis screening at delivery was inconsistent among 21 women with past syphilis (Table 4). Only 57% (n = 12) had an RPR at delivery. Among 15 women with new syphilis, 87% (n = 13) had an RPR at delivery.
. | Infant’s Congenital Syphilis Risk Classification (N = Number of infants) . | Mother’s Syphilis Screen Result (N = Number of Mothers) . | |||||
---|---|---|---|---|---|---|---|
. | Proven or Highly Probable Congenital Syphilis (N = 1) . | Possible Congenital Syphilis (N = 9) . | Congenital Syphilis Less Likely (N = 7) . | Congenital Syphilis Unlikely (N = 21) . | True Positive New Diagnosis (N = 17) . | True Positive Past Diagnosis (N = 21) . | False Positive (N = 183) . |
Maternal RPR at delivery | 0 | 7 (78) | 6 (86) | 12 (57) | 13 (77) | 12 (57) | 27 (15) |
Infant RPR at birth | 0 | 7 (78) | 6 (86) | 9 (43) | 13 (77) | 9 (43) | 1 (0.5) |
Infant CSF VDRL at birth | 0 | 6 (67) | 1 (14) | 0 | 7 (41) | 0 | 1 (0.5) |
Infant long-bone radiograph at birth | 0 | 6 (67) | 1 (14) | 0 | 7 (41) | 0 | 0 |
. | Infant’s Congenital Syphilis Risk Classification (N = Number of infants) . | Mother’s Syphilis Screen Result (N = Number of Mothers) . | |||||
---|---|---|---|---|---|---|---|
. | Proven or Highly Probable Congenital Syphilis (N = 1) . | Possible Congenital Syphilis (N = 9) . | Congenital Syphilis Less Likely (N = 7) . | Congenital Syphilis Unlikely (N = 21) . | True Positive New Diagnosis (N = 17) . | True Positive Past Diagnosis (N = 21) . | False Positive (N = 183) . |
Maternal RPR at delivery | 0 | 7 (78) | 6 (86) | 12 (57) | 13 (77) | 12 (57) | 27 (15) |
Infant RPR at birth | 0 | 7 (78) | 6 (86) | 9 (43) | 13 (77) | 9 (43) | 1 (0.5) |
Infant CSF VDRL at birth | 0 | 6 (67) | 1 (14) | 0 | 7 (41) | 0 | 1 (0.5) |
Infant long-bone radiograph at birth | 0 | 6 (67) | 1 (14) | 0 | 7 (41) | 0 | 0 |
Statistics presented as N (column %).
Among 38 infants born to TP mothers, RPRs were performed at delivery in 58% (n = 21). Stratified by past diagnosis (n = 21) or a new diagnosis (n = 15), RPRs were performed at birth in 43% (n = 9) and 87% (n = 13), respectively. Neither of 2 infants born to mothers diagnosed postpartum had RPRs at birth. By infant syphilis classification, the percent of infants with an RPR at birth were as follows: 0% (0 of 1) proven or highly probable congenital syphilis, 78% (7 of 9) possible congenital syphilis, 86% (6 of 7) congenital syphilis less likely, and 43% (9 of 21) congenital syphilis unlikely.
Among the women with FP screens, 27 of 183 (15%) had an RPR at delivery. One infant had both a serum RPR and a cerebrospinal fluid VDRL at birth.
Infant Treatment and Follow-Up
The 9 infants categorized as having possible congenital syphilis had normal physical exams at birth. Among the 8 recognized at birth, 6 had positive RPRs. All 6 were followed with serial RPRs, but 1 infant with a positive RPR at 2 months was lost to follow-up. Among 7 infants classified as congenital syphilis less likely, 2 of 3 with positive RPRs received a dose of intramuscular benzathine penicillin G. The 2 treated children had repeat RPR’s at 2 month; 1 was nonreactive. Of 16 infants classified as having possible congenital syphilis or congenital syphilis less likely, only 1 had a documented treponemal test at 18 months.
Discussion
In a cohort of over 75 000 pregnant women and their delivered infants, we explored the limitations of syphilis screening by examining the influence positive initial syphilis screens had on maternal and infant treatments. FP initial screens occurred commonly with both traditional and reverse sequence algorithms and on rare occasions resulted in unnecessary maternal and infant evaluations and treatments. Women with a TP result frequently did not receive the optimal follow-up care, management, or social support required to prevent the birth of infants at higher risk of congenital syphilis. In this regard, failure to achieve an optimal congenital syphilis classification was a proxy for reinfection or inadequate maternal treatment. Screening during pregnancy identified women with a new syphilis diagnosis during or before the early third trimester in 93% of cases. Despite this, one-half of their infants were classified at higher risk for congenital syphilis for various maternal indications, including concern for maternal reinfection or relapse, inappropriate initial maternal treatment, initial treatment not completed in a timely manner, or inadequate follow-up care without repeat maternal RPRs. In addition, failure to implement risk-based rescreening at delivery resulted in 2 missed maternal infections and a case of congenital syphilis.
Few studies have examined the potential harms associated with screening pregnant women for syphilis.14–16 At least 6 studies have reported that FP screening tests occur commonly when pregnant women are screened for syphilis.7,8,17–20 Potential harms associated with screening include time and expense associated with screening and FP results that lead to stress, incorrect labeling, and further diagnostic workup.16,21 In our population, 83% of initially positive screening tests were adjudicated by using laboratory-based criteria to be FP, with 243.8 FP results per 100 000 pregnancies. Overall, we did not observe a difference in false discovery rates for women screened using the traditional algorithm compared to reverse algorithms. Rarely, FP syphilis screens contributed to unnecessary treatments during pregnancy, as 2 women and 1 infant received evaluations and treatments because of concern for syphilis. In addition, 4 women were treated for syphilis before their pregnancy, yet were classified in our analysis as FP.
Factors unrelated to the timing of screening contributed to one-half of the infants born to mothers with new syphilis diagnosed during pregnancy being classified at higher risk for congenital syphilis, resulting in the need for more extensive infant evaluations and prolonged treatment in hospital. Our data indicate that maternal treatment failure is multifactorial and underscores the need to optimize maternal care and the importance of providing pregnant women with newly diagnosed syphilis the necessary education and social support to ensure that they not only receive appropriate initial treatment, but that they adhere to an intensive follow-up program, which focuses on preventing reinfection.22,23 It is recommended that pregnant women receive treatment appropriate for the stage of their disease and have serologic follow-up testing to gauge treatment response.1,24 Pregnant women who are seropositive for syphilis should be considered infected unless an adequate treatment history is documented and sequential serologic titers have declined appropriately after past therapy. Women treated before pregnancy who have not achieved a low stable or negative RPR titer should be retreated during pregnancy; otherwise, the infant will be assigned a higher syphilis risk classification. At a minimum, repeat screening should be performed during the early third trimester and again at delivery. For women at high-risk for reinfection and for communities or populations in which the prevalence of syphilis is high, monthly quantitative titers may be useful.
In our cohort, we observed inconsistent implementation of risk-based rescreening at delivery, and this resulted in 2 missed cases of maternal syphilis. Clinical differences between the women with TP and FP syphilis screens suggest that the definition of high-risk should include women with a history of sexually transmitted infections or illicit substance use. Given the reemergence of syphilis in many populations and the inconsistency with which risk-based rescreening guidelines are followed, a strong argument can be made for adopting universal rescreening protocols during pregnancy.13
Both CDC and American Academy of Pediatrics guidelines recommend that no infant should be discharged from the hospital without determination of the mother’s serologic status for syphilis.2,12 Moreover, all infants born to seropositive mothers require a careful physical examination and nontreponemal testing. Indeed, each of the 4 CDC infant congenital syphilis risk classifications recommends comparison of the maternal and infant RPR at delivery. Although these guidelines were followed for most infants born to mothers with reactive nontreponemal tests, compliance was inconsistent for babies born to women previously treated for syphilis. Failure to rescreen women with treated syphilis (a very high-risk population), and their infants, might result in failure to detect reinfection or treatment failure.
Our study has several limitations. Although we used a unique database that allowed us to capture maternal infant pairs from a large health care system, some infants did not receive follow-up care within our health care system. In this regard, it is possible that we missed some postnatally diagnosed cases of congenital syphilis. In addition, we pooled data for 3 different screening algorithms. We did not explore the influence of coinfections, such as other sexually transmitted infections, on FP serology tests or whether women with FP nontreponemal or treponemal assays during pregnancy have unique clinical characteristics, because this would be best performed using a control population without syphilis. Risk factors for FP nontreponemal and treponemal screens have been summarized by others.25,26
Conclusion
In a large population of pregnant women screened for syphilis, FP screens occasionally lead to unnecessary treatments in both mothers and neonates. Among women with new syphilis, the diagnosis was often early enough during pregnancy to allow time for treatment. However, in one-half there was concern for reinfection, treatment failure, inadequate treatment or follow-up care, or late treatment. In addition, inconsistent rescreening of high-risk mothers at delivery resulted in missed cases of syphilis and congenital infection. Among women with syphilis, especially among women treated before pregnancy, there was inconsistent RPR testing at delivery of both mothers and neonates. In addition to optimizing the care and management of women with positive syphilis screens, our results suggest a need for improvement in the screening of women at high-risk for syphilis, including those women with a history of treated syphilis.
Drs O’Connor and Foster conceptualized and designed the study, validated data extraction tools, collected data, carried out initial analyses, and drafted the manuscript; Mr Burke developed data extraction tools, collected data, and reviewed and revised the manuscript; Ms Worley provided statistical support and reviewed and revised the manuscripts; Drs Kadkhoda and Goje contributed to the study design and critically reviewed the manuscript for important intellectual content; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/peds.2022-057927.
FUNDING: No external funding.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no conflicts of interest relevant to this article to disclose.
Comments