Introduction
Congenital syphilis (CS), caused by Treponema pallidum, is transmitted from mother to foetus and can lead to severe adverse outcomes [Reference Trinh1]. Globally, CS remains a significant public health concern, not limited to low-income countries, with the World Health Organization aiming to reduce mother-to-infant transmission to fewer than 50 cases per 100,000 live births [2, Reference Moseley3]. In Japan, syphilis cases have increased since the early 2010s [Reference Takahashi4], driven by heterosexual transmission, leading to a rise in CS cases [5–8]. The calculated incidence rate of CS per 100,000 live births rose from below 1.0 before 2013 to 5.09 in 2023 [8]. Despite robust screening programmes, CS has been increasing in recent years in Japan, emphasizing the need for ongoing vigilance.
Early screening and treatment of pregnant women can prevent CS [Reference Alexander9, Reference Lin, Eder and Bean10], but approximately 70% of affected infants are asymptomatic at birth, and passively transferred maternal antibodies complicate early diagnosis and treatment [Reference Fang11]. Therefore, guidelines in Japan [Reference Ito12] and other countries [Reference Trinh1, Reference Arnold and Ford-Jones13, 14] recommend estimating the likelihood of CS based on maternal and infant non-treponemal antibody tests, clinical symptoms, and maternal syphilis treatment history [Reference Arnold and Ford-Jones13, 14]. However, the complexity of current guidelines often leads to overtreatment in an effort to avoid missed cases. More accurate diagnostic tools are needed, requiring a better understanding of each indicator’s diagnostic contribution.
To address these issues, we conducted the first nationwide survey to assess the aetiology of CS and analyse risk factors in both mothers and infants, aiming to enhance clinical guidelines [Reference Ito12].
Materials and methods
Study protocol and data collection
This nationwide retrospective cohort study included pregnant women with syphilis from 2015 to 2024 and followed the outcomes of their pregnancies. Recruitment was conducted by inviting member physicians of the Japanese Society for Pediatric Infectious Diseases between April and August 2024. Participating physicians provided anonymized clinical data from their institutions; patients did not self-register. Participants included pregnant women with a positive syphilis serologic test, regardless of disease activity, and their infants with a potential risk of CS. Data were collected on maternal age, social history (e.g., engagement with commercial sex work in the past or currently), syphilis stage, time of diagnosis, treatment details, and serological test results. Infant data included sex, perinatal information, clinical symptoms, and serological test results. After excluding those who did not consent, maternal and infant risk factors were analysed to identify contributors to diagnosis.
Tests for diagnosis of CS
Serological testing is the most common method for diagnosing syphilis and is divided into non-treponemal (non-specific) and treponemal (specific) tests. The rapid plasma reagin (RPR) test was used as a non-treponemal test, and either T. pallidum latex agglutination or T. pallidum hemagglutination as treponemal tests. The automated quantitative RPR assay is widely used as it eliminates variability between laboratory technicians, making it more objective and easier to monitor antibody titre trends [Reference Ito12, Reference Lee15, Reference Tsuboi16]. Although some facilities in our study still used the conventional manual card test, previous studies in the US and Japan revealed a strong correlation between the two methods, allowing both results to be treated as comparable [Reference Tsuboi16–Reference Arbefeville, Lynch and Ferrieri18]. Furthermore, the optimal decrement rate in RPR titre by the automated test for a 4-fold decrement by manual card test was reported to be 76.54% [Reference Tsuboi16]. Therefore, given its superior speed, lack of interpersonal variation in interpretation, and higher throughput compared to the manual test, the automated assay is recommended in Japan as an excellent alternative. A positive result is indicated by a value greater than 1.0 R.U.
To explore alternative serum biomarkers for diagnosing CS, in addition to the infant-to-mother RPR ratio, we utilized the results of absolute RPR, total IgM, and fluorescent treponemal antibody absorption (FTA-ABS) IgM fractions of infants, which had been performed as part of routine clinical practice; among these, the FTA-ABS IgM was not covered by the national health insurance in Japan. However, these tests were not uniformly performed in all cases, and their implementation depended on the policy of each facility and attending physician.
Final diagnostic definition of CS
Definitive diagnosis of CS is often difficult due to the presence of transferred maternal antibodies. The definition of confirmed CS varies depending on the country and organization [Reference Moseley3]. Infants were classified into four major categories (A–D) based on the likelihood of CS. Each category was further subdivided according to the diagnostic criteria applied (e.g., A1, A2; B1, B2; C1, C2), resulting in seven patterns in total (Table 1). CS cases were categorized as follows: definite diagnosis included A1 (direct detection of T. pallidum in lesions, body fluids, or nasal discharge) [14] and A2 (detection of T. pallidum-specific IgM antibodies) [19, Reference Herremans, Kortbeek and Notermans20]. In placental pathology, identification was achieved using T. pallidum immunohistochemistry, after observing findings such as villitis. Highly Probable Diagnosis was assigned to infants with a positive non-treponemal test for syphilis and any of the following: clinical manifestations consistent with CS, radiographic evidence of long bone abnormalities, or abnormal cerebrospinal fluid analysis. Abnormal cerebrospinal fluid findings were defined, considering neonatal reference ranges, as a cell count ≥20/mm3 and a protein level ≥ 119 mg/dL [Reference K21]. Excluded diagnoses encompassed C1 (untreated infants with negative RPR testing after 3 months of age) and C2 (infants with negative treponemal serology confirmed before the age of 18 months) [22, Reference Workowski23]. The category C2 also included children who were considered unlikely to have CS but were treated. Cases that did not fit into any of the aforementioned categories due to an unclear diagnosis were designated as D (Undiagnosed). Maternal treatment history influences the likelihood of CS but is not part of the diagnostic criterion and is therefore excluded.
Table 1. Final diagnostic definition of congenital syphilis
Abbreviations: CSF, cerebrospinal fluid; EIA, enzyme immunoassay; FTA-abs, fluorescent treponemal antibody-absorption; IgM, immunoglobulin M; PCR, polymerase chain reaction; RPR, rapid plasma reagin; VDRL, venereal disease research laboratory test.
In Japan, CS is also a notifiable disease under the national surveillance system, which has its own notification criteria (including: [i] markedly higher infant than maternal nontreponemal titres, [ii] persistently elevated infant titres beyond the expected decline of maternal antibodies, [iii] positive infant IgM against T. pallidum, [iv] clinical signs of early CS, and [v] signs of late CS) [24]. These criteria differ from the case definitions used in this study, which were specifically designed for research purposes.
Potential predictors for CS
We evaluated potential predictors of CS by comparing definite or highly probable diagnoses with excluded diagnoses. Univariate analysis was conducted using Fisher’s exact test for categorical variables and univariate logistic regression analysis for continuous variables. A multivariate logistic regression analysis was then performed to identify potential predictors of CS. Considering multicollinearity and the appropriate number of variables for the multivariable analysis, the following covariates were included in the multivariate models based on their potential associations in the univariate analysis: p-values less than 0.05 were considered statistically significant. We also compared the results of serological tests (maternal and infant RPR, and infant RPR/maternal RPR) as diagnostic indicators between CS and non-CS cases.
Statistical analysis
All statistical analyses were conducted using the SPSS Statistics 29.0 software (SPSS, Chicago, IL).
Ethical considerations
This study was approved by the Keio University Ethics Committee (Approval Number 20231204, recently revised in 2024), with a waiver of informed consent because the data were anonymized. To ensure participant autonomy, an opt-out option was provided.
Results
Pregnant women with syphilis
In all, 39 of the 42 (92.9%) facilities responded to the study’s questionnaire. A total of 230 pregnant women with syphilis during pregnancy were enrolled in this study (Figure 1). Table 2 presents maternal background, serological tests, syphilis stage, and treatment details.
Figure 1. Flowchart of study participants.
A total of 230 pregnant women with syphilis and their infants were included. No patients were excluded due to opting out of consent. Based on diagnostic criteria, 49 infants were classified as having a definite or highly probable diagnosis of congenital syphilis (Categories A1, A2, B1, B2), 73 were classified as excluded (Categories C1, C2), and 108 remained undiagnosed (Category D).
Table 2. Clinical characteristics of pregnant women with syphilis
Abbreviations: IgM, immunoglobulin M; IM, intramuscular; IV, intravenous; PO, per Os; RPR, rapid plasma reagin.
a A documented history of sexually transmitted infections encompasses past diagnoses of conditions such as syphilis, trichomoniasis, chlamydial or gonorrheal infections, and genital herpes.
b In Japan, oral amoxicillin is frequently used for the treatment of syphilis, particularly because intramuscular benzathine penicillin G, the global standard treatment, was unavailable in the country until 2022.
c Median (interquartile range).
The mean maternal age was 26.9 years, with a median of 26 years. In total, 43% of mothers (96/223) were unmarried, and 34.7% (60/173) had a history of sexually transmitted infections other than syphilis. A total of 31.8% (55/173) were engaged as commercial sex workers, and 30.0% (66/220) had not received antenatal checkups or had irregular checkups. The stages of syphilis were as follows: asymptomatic latent syphilis 85.9% (164/191), primary syphilis 4.2% (8/191), and secondary syphilis 9.9% (19/191). Amoxicillin with or without probenecid was used in 58.1% of cases (129/222), while 24.8% (55/222) received penicillin G, including intramuscular formulations approved for use in 2022.
Infants born to pregnant women with syphilis
A total of 230 infants were born to mothers with syphilis during the study period (Table 3). There were no cases of stillbirth or miscarriage. Among these infants, 127 were male and 101 were female; 170 were born via vaginal delivery and 58 via caesarean section. The average birth weight was 2,772 g, and 23.7% (54/228) of the infants were classified as low birth weight (< 2,500 g). Of these, 49 were diagnosed with definite or highly probable CS; 73 were determined not to have CS, and 108 had an unknown final diagnosis. Among the 49 definite or highly probable cases, 13 (26.5%) were diagnosed by placental histological identification of T. pallidum, 21 (42.9%) by FTA-ABS IgM, and 20 (40.8%) by a positive non-treponemal test result along with any of the following: physical abnormalities, long bone abnormalities, or abnormal cerebrospinal fluid findings.
Table 3. Clinical characteristics of infants born to mothers with syphilis
Abbreviations: CSF, cerebrospinal fluid; FTA-abs, fluorescent treponemal antibody-absorption; IgM, immunoglobulin M; PCR, polymerase chain reaction; RPR, rapid plasma reagin; VDRL, venereal disease research laboratory test.
a Including cases that meet higher-level definitions.
A total of 108 undiagnosed cases were excluded from the risk factor analysis, as most were ultimately lost to follow-up due to insufficient testing rather than insufficient surveillance. For example, 82.5% (33/40) had a negative infant RPR (<1.0 R.U.) at birth, and none of these cases underwent a repeat RPR test at 3 months. The antibiotic treatment rate was 56.1% (60/107), and the FTA-ABS IgM measurement rate, calculated as the number of patients measured divided by the total number of patients with known measurement status (i.e., whether measured or not), was 65.7% (36/105), and all were FTA-ABS IgM-negative. Furthermore, the total IgM measurement rate and infant RPR test rate were 62.0% (67/108) and 89.8% (97/108), respectively.
Risk factor analysis for CS
To evaluate the potential indicators of CS, we compared cases with a definite or highly probable diagnosis to those with an excluded diagnosis (Table 4). In the univariate analysis, being engaged as a commercial sex worker (odds ratio [OR]: 5.33, 95% confidence interval [CI]: 2.12–13,40, p < 0.001), failure to receive prenatal health checkups or irregular visits (OR: 3.91, 95%CI: 1.76–8.68, p < 0.001), preterm birth (OR: 10.29, 95%CI: 3.90–21.14, p < 0.001), and low birth weight (OR: 13.78, 95%CI: 5.40–35.16, p < 0.001) were significantly associated with CS. The authors also found that certain blood test results (low platelet count, high CRP, high total IgM, and maternal RPR >4) were more commonly observed in the definite or highly probable CS group. Furthermore, infants born to mothers who had not completed treatment more than 4 weeks before delivery had a significantly higher rate of CS (OR: 16.92, 95%CI: 6.39–44.82, p < 0.001).
Table 4. A univariate analysis of predictors of congenital syphilis
Abbreviations: OR, odds ratio; 95%CI, 95% confidence interval; IgM, immunoglobulin M; IM, intramuscular; IV, intravenous; PO, per Os; CRP, C-reactive protein; AST, aspartate aminotransferase; ALT, alanine aminotransferase; LD, lactate dehydrogenase.
a Median (interquartile range).
To avoid multicollinearity in the multivariate analysis, ‘failure to receive prenatal health checkups or irregular visits’, ‘unmarried at delivery’, and ‘maternal history as a commercial sex worker’ were excluded, as these three factors were strongly correlated with ‘maternal treatment not completed more than 4 weeks before delivery’, making it difficult to assess their independent effects. The authors also excluded some blood test results due to their non-specific nature. Based on these considerations, the following covariates were selected: maternal treatment not completed more than 4 weeks before delivery, total IgM > 20 mg/dL, preterm birth, and platelet count.
In the multivariable logistic regression analyses, maternal treatment not completed more than 4 weeks before delivery (OR: 7.20, 95% CI: 1.38–37.56, p = 0.02) and total IgM >20 mg/dL (OR: 65.31, 95% CI: 4.53–941.39, p = 0.002) were significantly associated with a higher risk of CS. (Table 5).
Table 5. A multivariate analysis of predictors of congenital syphilis
Abbreviations: OR, odds ratio; 95%CI, 95% confidence interval; IgM, immunoglobulin M.
Comparison of serological findings as diagnostic indicators (Table 6)
The authors compared maternal and infant serological findings between cases with a definite or highly probable diagnosis and those with an excluded diagnosis. In CS cases, 93.8% (45/48) had a positive RPR, while only 25% (18/72) of the excluded cases had a positive RPR. Among the confirmed cases, 34.3% (12/35) and 11.4% (4/35) had an infant-to-mother RPR ratio of ≥1 and ≥ 4, respectively. In contrast, 97.0% (65/67) and 100% (67/67) of the excluded cases did not reach a ratio of ≥1 and ≥ 4, respectively, suggesting that high ratios are much less common in non-CS cases. The number of cases evaluated using infant RPR ≥1 was 48, whereas only 35 cases could be assessed using the infant-to-mother RPR ratio because of the limited availability of maternal RPR data.
Table 6. Comparison of serological findings
Abbreviations: RPR, rapid plasma reagin.
Discussion
This study is the first nationwide survey of pregnant women with syphilis and their newborns, including those with CS, conducted in Japan. Incomplete maternal treatment more than 4 weeks before delivery and infant total IgM > 20 mg/dL were found to be significant risk factors for suspected CS. Infant RPR titre, rather than the infant-to-mother RPR ratio, appears to be a more useful diagnostic indicator. Also, this study identified 41 cases between 2015 and 2024 that met the national surveillance of diagnosed CS definition for CS [25], accounting for 23.7% of the 173 cases reported during the same period, as documented in the national surveillance data [24]. The fact that the majority of pregnant women with syphilis in the present study had latent syphilis and were asymptomatic reinforces the importance of screening tests.
Of the two identified risk factors, ‘incomplete maternal treatment more than 4 weeks before delivery’ has already been acknowledged as a risk factor in prior literature [Reference Thornton, Chaisson and Bleasdale26]. It has been reported that in seropositive mothers with non-treponemal antibodies, if antimicrobial treatment is completed more than 4 weeks prior to delivery, most of their infants become seronegative within 6 months after birth, even without receiving treatment [Reference Chang27].
The second risk factor, ‘infant total IgM > 20 mg/dL’, is newly recognized. In this study, total IgM >20 mg/dL at birth showed an OR of 65 for CS. For early diagnosis of CS, FTA-ABS IgM can be measured, as it detects treponemal antibodies that do not cross the placenta. However, its insufficient sensitivity has been a concern [Reference Scotti and Logan28–Reference Stoll30]. Moreover, the assay is subjective, relies on visual interpretation by technicians, is not covered by the national health insurance system, and is no longer offered by many diagnostic laboratories. In contrast, total IgM can be measured promptly and objectively at many facilities using automated assays, though its positivity is not specific to syphilis and may reflect other congenital infections [Reference Neu, Duchon and Zachariah31].
Among the examined potential risk factors, maternal history of sexually transmitted infections and marital status were not significantly associated with the outcome. In contrast, a history of sex industry employment and inadequate or no antenatal health checkups were significantly higher, with ORs of 5.33 and 3.91, respectively. Although these factors were excluded from multivariate analysis to avoid multicollinearity, they should still be considered important risk factors for CS. These two factors were closely related to and might be the underlying cause of maternal treatment not completed more than four weeks before delivery, which was also significantly associated with CS (Table 4). Prompt antimicrobial therapy and regular antenatal checkups were deemed most important in treating pregnant women with syphilis. Similar maternal risk factors, such as limited access to antenatal care and socio-economic vulnerability, have also been reported in studies from other high-income countries [Reference Alberto32]. In addition, low birth weight has consistently been described as an adverse outcome associated with CS in these settings [Reference Carlson33].
As a diagnostic indicator, the infant RPR titre was found to be more informative than the well-established infant-to-mother RPR ratio in this study [Reference Rawstron and Bromberg34–Reference Oloya36]. Although the specificity was lower, the sensitivity was much higher. In addition, maternal information, including RPR data, is not always available, may be obtained from a different facility or test method, and the ratio cannot be calculated when the maternal RPR is negative. In fact, in this study, 30% of confirmed cases (14/49) had no available ratio data because the maternal RPR value was unknown. Of course, the traditional infant-to-mother RPR ratio may also be a useful indicator, as a significantly higher ratio may indicate a stronger likelihood of vertical transmission. Also, the ‘Proven or Highly Probable’ criterion has been defined as an infant RPR that is at least fourfold higher than the maternal RPR by the multiple dilution method, or 1.5 to 2.0 times higher by the automated method [Reference Ito12]. Typically, infant RPR titres are one to two times lower than maternal titres [Reference Rawstron and Bromberg34, Reference Stoll37]. This may be due to unknown reaction inhibitors in neonatal blood, limited placental transfer of maternal RPR IgM fractions (especially from early infection), or lower antibody levels in preterm infants due to incomplete placental transfer. However, since an elevated infant-to-mother RPR ratio inherently implies infant RPR positivity (≥1.0 R.U.), infant RPR positivity was considered more valuable for diagnosis.
This study has some limitations. First, as indicated by the finding that most infants with a negative RPR at birth were lost to follow-up, nearly half of the cases lacked sufficient diagnostic evaluation. Furthermore, it is recommended that infants with a negative infant RPR who do not receive treatment have their RPR retested at 3 months of age, and it is necessary to promote adherence to this recommendation by following a flowchart [Reference Ito12, 14]. Second, the risk factors were compared within the group of cases considered to have a potential for CS, rather than being compared with the general population of pregnant women. Third, if the mother contracts syphilis after her initial screening, syphilis-related antibody testing is not performed, potentially leading to missed cases of CS. However, universal syphilis screening for all neonates is not currently considered economically feasible. Fourth, more than half of the mothers were treated with amoxicillin with or without probenecid. This was because, while the international standard treatment is intramuscular benzathine penicillin G, it was not approved in Japan until 2022, necessitating the use of amoxicillin as an alternative. There is a report indicating that treatment with ampicillin or amoxicillin resulted in an overall CS occurrence rate of 14% and 33% in cases of late syphilis (including latent syphilis of unknown duration), respectively. However, no cases of CS were observed among mothers with early syphilis [Reference Nishijima38]. Nevertheless, with the recent introduction of intramuscular benzathine penicillin G, it is expected to become the standard treatment. Fifth, as this study was based on a survey of paediatricians, information on stillbirths and miscarriages was not available. However, although deceased individuals from infectious diseases are included in the criteria for notifiable disease surveillance in Japan [39], the number of reported deaths due to CS remains extremely low [40]. With the increasing incidence of syphilis in Japan, there is growing concern about a potential rise in the number of stillbirths, highlighting the need for high-precision surveillance.
Conclusion
While uncertainties regarding maternal treatment during pregnancy remain, incorporating ‘maternal treatment not completed more than 4 weeks before delivery’ as a key criterion, together with the addition of total IgM levels, could facilitate the establishment of a more practical and effective treatment flow. Moreover, positive infant RPR, rather than the infant-to-mother RPR ratio, is likely to be more beneficial for evaluating CS.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0950268825100629.
Data availability statement
The data presented are available upon reasonable request from the corresponding author after publication.
Acknowledgements
We would like to express our sincere gratitude to the following institutions for their endorsement of this study and their invaluable cooperation in patient enrolment. A list of participating institutions is provided in the supplementary material (Supplementary Table 1).
Author contribution
Conceptualization: H.S., N.T., K.I., Y.I., M.M., H.M., T.Y., M.S.; Investigation: H.S., N.T., K.I., Y.I., M.M., H.M., M.S.; Formal analysis: H.S., M.F., T.F., M.S.; Data curation: H.S., M.S.; Supervision: Y.I., M.M., H.M., T.Y., M.S.; Validation: M.F., N.T., K.I., T.F.; Methodology: M.F., T.F.; Writing - review and editing: H.S., M.F., M.S.; Writing - original draft: H.S., M.F., M.S.
Competing interests
The authors declare no competing interests exist.
Funding statement
This research was supported by the Ministry of Health, Labour and Welfare of Japan (Grant number 24HA2011).
Open access
