From the Johns Hopkins University School of Medicine, Baltimore, MD.
The authors report no conflict of interest.
Correspondence: Khalil G. Ghanem, MD, PhD, Johns Hopkins University School of Medicine, JHUBMC ID Division, 5200 Eastern Ave, MFL Center Tower Suite 378, Baltimore, MD 21224. E-mail: firstname.lastname@example.org.
Received for publication January 1, 2013, and accepted January 22, 2013.
The use of the reverse sequence syphilis algorithm has identified a population with serodiscordant test results: nonreactive nontreponemal tests and confirmed reactive treponemal tests. This is a heterogeneous population. Most individuals have treated syphilis resulting in a decline in nontreponemal antibody titers after therapy. Others have early infection (testing preceded the development of detectable antibodies or was affected by the prozone phenomenon), and subsequent nontreponemal tests are usually found to be reactive. Nontreponemal antibody titers may also decline over time even in the absence of therapy, so asymptomatic individuals with persistently nonreactive nontreponemal tests and no history of syphilis treatment may have late syphilis. In the antibiotic era, several important clinical and public health questions arise in individuals with serodiscordant test results without documented past syphilis treatment. What are the clinical risks of progression to tertiary syphilis? What are the public health risks of (a) sexual transmission and (b) vertical transmission from mother to infant?
In this issue of Sexually Transmitted Diseases, Peterman and colleagues1 try to address the question of vertical transmission of syphilis in women with persistently nonreactive nontreponemal tests. They reviewed public health reports over a 28-year period to determine if any cases of congenital syphilis could be identified. Their data suggest that vertical transmission in these women is very unlikely. Are their conclusions justified?
The probability of vertical transmission from an untreated mother depends on the duration of the infection in the mother and the stage of pregnancy (reviewed by Wicher and Wicher2). Transmission is 70% to 100% in primary and secondary syphilis, 40% in early latent syphilis, and 10% in late latent disease.3,4 In 1922, Williams5 reported that 50 of 109 infants born to 50 seronegative mothers (of which 46 had never received treatment for syphilis) had probable (n = 7) or definitive (n = 43) congenital syphilis (also reviewed by Moore6). Although the study suggests the possibility of vertical transmission among untreated seronegative women with syphilis, there are several issues worth highlighting. The study offered limited data on the persistence of seronegativity (thus, early infection could not be ruled out). Nontreponemal tests used in the modern antibiotic era are more sensitive than those used in the preantibiotic era. Finally, routine exposures to β-lactam antibiotics, tetracyclines, and macrolides for indications other than syphilis in the modern era further complicate historical comparisons. Modern reports of vertical transmission of syphilis in seronegative women, when not caused by seronegative early infection in the mother7,8 or the prozone phenomenon yielding false-negative nontreponemal test results,9 are often incompletely documented.10,11 As such, the current study by Peterman, despite its limitations, deserves attention.
The major strength of this study is the large number (n = 23,863) of case reports of congenital syphilis collected over a 28-year period. This study has several limitations that mostly pertain to the quality of the data collected from public health records and are discussed by the authors. A few are worth highlighting. The screening of pregnant women during the study period was mainly conducted using the traditional syphilis testing algorithm, which consists of screening for the presence of nontreponemal antibodies followed, if positive, by a confirmatory treponemal test. In the setting of a negative nontreponemal test result, only clinical suspicion would dictate whether a treponemal test was ordered. Given that congenital syphilis is rare and may not manifest at the time of delivery, clinicians with limited experience may miss the diagnosis when the nontreponemal test is nonreactive (a form of outcome identification bias). In addition, the missing data on public health case-report forms introduce another source for potential misclassification. For example, missing data made it difficult to definitively classify the outcome in an infant whose mother had persistently nonreactive nontreponemal tests and whose bone radiographs were suggestive of congenital syphilis. Although the Peterman study correctly concludes that vertical transmission from mothers with persistently nonreactive nontreponemal tests is a rare phenomenon (should it occur at all), how “rare” is “rare enough” when trying to formulate clinical and public health policies?
The reverse sequence algorithm, as well as the identification of persons with persistently nonreactive nontreponemal tests and no clear prior syphilis treatment, has challenged (and frustrated) clinicians and public health practitioners. The management of these patients requires additional clinical and public health resources that were not needed when the traditional syphilis screening algorithm was used because the latter method would not have identified these individuals. In an era of limited resources, this issue takes on added urgency. Is willful ignorance of this population acceptable? The answer to that question depends on the population and on our tolerance for risk. Studies from the preantibiotic era demonstrate a low risk of sexual transmission in persons with late syphilis (irrespective of the reactivity of their nontreponemal tests).6 As such, public health officials may be willing to tolerate a small amount of risk by deciding that partner services to track partners of serodiscordant persons are not a priority when resources are extremely limited. Similarly, the risk of vertical transmission (and, consequently, treatment of the mother and possible additional testing/treatment in the newborn) in persistently seronegative pregnant women is very low—but is it low enough to withhold all interventions in this population, even when resources are limited? Given the vulnerability of this population, most clinicians would be highly risk averse so that even a few well-described case reports of confirmed congenital syphilis in infants born to mothers with persistently nonreactive nontreponemal tests and reactive confirmatory treponemal tests would favor continued aggressive interventions in pregnant women. Peterman’s study, given its limitations, does not entirely rule out the possibility of vertical transmission. As such, it would be very difficult to recommend willful ignorance of this population based solely on their findings.
What is needed to move forward? Improved diagnostics that better reflect disease activity would be most welcome. Short of these, we need more data. It would be difficult to convince institutional review boards to allow randomization of pregnant women with no clear treatment history for syphilis and persistently negative nontreponemal test results to the observational arm of a trial. Consequently, the best we can hope for are well-described case series or case reports to better define the possibility of congenital syphilis in this population. Broader uptake of the reverse sequence algorithm is a double-edged sword: it may lead to enhanced detection of congenital syphilis cases (e.g., a woman presents at the time of delivery having had no prenatal care and the detection of treponemal antibodies would alert clinicians to the possibility of underlying syphilis), but it will likely increase the probability that many more women will require partner services and treatment (overtreatment?) for syphilis during their pregnancy after the detection of treponemal antibodies. To date, there are no well-documented published cases of congenital syphilis born to mothers with persistently negative nontreponemal serology results. If such cases fail to surface, we may enthusiastically reengage in willful (and blissful) ignorance.
1. Peterman TA, Newman DR, Davis D, et al.. Do women with persistently negative nontreponemal test transmit syphilis during pregnancy? Sex Transm Dis 2013; 40: 311–315.
2. Wicher V, Wicher K. Pathogenesis of maternal-fetal syphilis revisited. Clin Infect Dis 2001; 33: 354–363.
3. Fiumara NJ. Congenital syphilis in Massachusetts. N Engl J Med 1951; 245: 634–640.
4. Sanchez PJ, Wendel GD Jr, Grimprel E, et al.. Evaluation of molecular methodologies and rabbit infectivity testing for the diagnosis of congenital syphilis and neonatal central nervous system invasion by Treponema pallidum
. J Infect Dis 1993; 167: 148–157.
5. Williams JW. Influence of the treatment of syphilitic pregnant women on the incidence of congenital syphilis. Bull Johns Hopkins Hosp 1922; 33: 383.
6. Moore JE. The Modern Treatment of Syphilis. Springfield, IL: Charles C. Thomas, 1943.
7. Dorfman DH, Glaser JH. Congenital syphilis presenting in infants after the newborn period. N Engl J Med 1990; 323: 1299–1302.
8. Sanchez PJ, Wendel GD, Norgard MV. Congenital syphilis associated with negative results of maternal serologic tests at delivery. Am J Dis Child 1991; 145: 967–969.
9. Berkowitz KM, Stampf K, Baxi L, et al.. False negative screening tests for syphilis in pregnant women. N Engl J Med 1990; 322: 270–271.
10. Fiumara NJ, Hill W. Congenital syphilis in the absence of positive serology in the mother: A third indication for the Treponema pallidum
immobilization test. JAMA 1957; 163: 1037–1039.
11. Guinness LF, Sibandze S, McGrath E, et al.. Influence of antenatal screening on perinatal mortality caused by syphilis in Swaziland. Genitourin Med 1988; 64: 294–297.