Screening Veterans for Syphilis: Implementation of the Reverse Sequence Algorithm

Abstract

We evaluated the syphilis reverse sequence algorithm (RSA) in a Veteran Affairs facility, finding 5.5% reactive Treponema pallidum enzyme immunoassay (EIA) tests. In a subset of EIA+/VDRL-/TP-PA+ cases, 48% were previously treated. Of veterans with unknown/no prior therapy, only 45% had documentation of subsequent treatment, suggesting suboptimal interpretation of RSA results.

The incidence of syphilis in the United States has been increasing since 2000, making accurate diagnosis essential [1]. Traditional syphilis screening involves a nontreponemal antibody test such as the Venereal Disease Research Laboratory (VDRL), followed by confirmation of positive results with a treponemal antibody specific assay such as the Treponema pallidum enzyme immunoassay (TP-EIA). More recently, many laboratories across the country have switched to the reverse sequence algorithm (RSA) for syphilis screening, which allows for high throughput and objective results [2]. This algorithm starts with a screening treponemal test such as an EIA and, if positive, a reflexive nontreponemal test. The drawback of using an initial treponemal test is the inability to distinguish between treated versus untreated infection [3]. In addition, the RSA may engender a diagnostic dilemma when there is a discordant result (i.e., positive treponemal test followed by nonreactive nontreponemal test), for which a second treponemal assay is recommended.

The Centers for Disease Control and Prevention (CDC) offers guidance in cases of discordant results, but evolving guidelines may be challenging for nonspecialists, especially those with limited syphilis management experience. Most recent guidelines suggest that in cases of a reactive second treponemal test, treatment should be pursued if patients have not been treated previously. For patients with prior treatment, no additional therapy is recommended unless sexual history suggests reexposure [4]. In addition, the US Preventive Services Task Force (USPSTF) states that there is “limited evidence on the accuracy of screening” using the RSA and that additional studies are required to better understand its implications for screening purposes [5]. Thus, the objective of this study was to evaluate implementation of the RSA in a Veterans Affairs (VA) population to assess its impact and understand how providers interpret discordant results.

METHODS

The Durham Veterans Affairs Medical Center (DVAMC) is part of Veterans Integrated Service Network (VISN) 6 and performs syphilis testing for the DVAMC, as well as several VA medical centers within the VISN. DVAMC initiated the RSA in September 2009, using the automated treponemal EIA, BioPlex® 2200 Syphilis immunoglobulin G (IgG) Kit, (Bio-Rad, Hercules, California), which is a multiplex flow immunoassay for the detection of T. pallidum IgG antibodies. The BD VDRL Antigen (Becton, Dickinson and Company, Sparks, Maryland) was used as the reflexive nontreponemal antibody test. Discordant results were then tested with the Serodia Treponema pallidum particle agglutination assay (TP-PA) Test (Fujirebo, Malvern, Pennsylvania).

We performed a retrospective review of all veterans tested for syphilis within VISN 6 between September 1, 2009, and June 1, 2015. Positive screening was defined as a reactive IgG-EIA with reactive VDRL. Discordant results included reactive IgG-EIA tests followed by a nonreactive VDRL. For discordant cases, we assessed the proportion with reactive TP-PA results. We included only the first diagnostic test (i.e., the first VDRL performed concurrently to or following IgG-EIA testing). Patients with cancelled or inconclusive TP-PA tests were excluded.

Data were extracted using a systematic query of the electronic medical record (EMR), supplemented by manual record review for a randomly selected subset of veterans with discordant results for whom charts were available. Veterans seen within VISN 6, but at centers other than DVAMC, did not have readily available records and were not included in analyses of patient characteristics. Collected data elements included: age, race, sex, human immunodeficiency virus (HIV) status, sexual risk factor, prior syphilis infection, reason for syphilis testing, and management after discordant results. Chart review was performed systematically using the search function of the VA electronic health record. Each chart was searched for the following terms: “syphilis,” “VDRL,” “RPR,” and “TP-PA”; at a minimum, all records for the year following the discordant test were searched. For veterans with discordant results and no or unclear prior treatment for syphilis, state health department records were also searched for any history of prior syphilis treatment.

For veterans with reactive TP-PA versus non-reactive TP-PA, comparisons were made using Student’s t-test for normally distributed continuous variables, and χ² or Fisher exact tests as indicated for categorical variables. A P-value of <.05 was considered to be statistically significant. All analyses were performed using R version 3.2.2 (2015). This study was approved by the Institutional Review Board at the Durham VA Medical Center.

RESULTS

Syphilis screening with the RSA was performed for 11152 veterans from September 1, 2009, until June 1, 2015, resulting in 617/11152 (5.5%) reactive T. pallidum IgG-EIA results. Of these, 438/617 (71.0%) were followed by non-reactive VDRL testing, of which 378/438 (86.3%) were TP-PA (+) and 60 (13.7%) were TP-PA (−) (Figure 1).

Figure 1.

This figure is a flow chart depicting reverse sequence algorithm (RSA) testing results performed for Veterans Integrated Service Network (VISN) 6 between September 1, 2009 and June 1, 2015.

Through chart review, we compared a random subset of veterans with reactive TP-PA tests and available charts (n = 160), extracted via systematic query of the EMR, versus all of those with nonreactive TP-PA tests and available charts (n = 37). We found that TP-PA (+) patients were more often HIV-infected and had received previous treatment (Supplementary Table 1). There were also statistically significant differences between the 2 groups with respect to patients’ race and reason for testing.

Among the 160 patients with EIA+/VDRL-/TP-PA+, 26 (16.3%) had documentation by a provider of not having received prior treatment (Supplementary Table 2). For an additional 60 (37.5%) of these veterans, treatment history was unclear; that is, there was absence of documentation (provider did not ask or did not document asking), or there was indeed documentation but to the effect that the patient or family did not know whether the patient had previously been treated. When health department records were searched, we confirmed no record of prior treatment for 83 (96.5%) of these 86 veterans with no/unclear prior syphilis history. However, 3 (3.5%) of these 86 patients had in fact received prior treatment, one of whom was treated again after discordant syphilis testing results.

Of the 83 veterans without prior treatment (per combination of VA and health department records), 37 (44.6%) received subsequent treatment, but 46 (55.4%) had no treatment documented after RSA results. Of 74 veterans who had received prior treatment per VA records, 10 went on to receive treatment again, 8 of whom had new symptoms. Among cases in which Infectious Diseases (ID) providers were not involved by default (e.g., HIV-infected patients), 27–33% received an ID consultation (Supplementary Table 3).

DISCUSSION

We investigated implementation of the RSA in a VA system, one of the first to adopt this process, in order to assess its clinical impact among the veteran patient population. We found that 5.5% of veterans screened by the RSA had reactive T. pallidum IgG-EIA results, and most had nonreactive VDRL testing or discordant results. Interestingly, the vast majority of cases with nonreactive VDRL results had reactive TP-PA testing, representing true infection either currently or at some time in the past. These high rates are consistent with prior reports that found a high burden of sexually transmitted infections (STIs) among military personnel [6]; indeed, the veteran population consists largely of groups disproportionately affected by STIs, particularly syphilis (e.g., men, racial and ethnic minorities, and those with psychiatric disease) [1, 7, 8]. Also consistent with prior studies is our finding that patients with reactive TP-PA results, as compared to those with nonreactive TP-PA results, were more often HIV-infected and previously treated [9].

The RSA has been studied in low prevalence settings (1–3% positive screening rate) [9–11], resulting in an increase in syphilis diagnoses, although with a high rate of false positive results [11]. Our “false positive” rate (EIA+/VDRL-/TP-PA-) rate was relatively low at 13.7%, consistent with data from high-prevalence populations [2]. To our knowledge, there has been limited evaluation of RSA performance in populations with high syphilis prevalence; in one study, investigators reported a 73% rate of previously treated infections and concluded that the RSA may not be suitable in their population [12]. Our rate of previous treatment among “true positives” (EIA+/VDRL-/TP-PA+) was not as high (48%), indicating that our VA population should likely benefit from the RSA for increased identification of untreated syphilis infections.

However, this rate of prior treatment is somewhat difficult to interpret, given that approximately a third of veterans had unclear history or lack of documentation regarding previous diagnosis and treatment. Additionally, we found that in veterans with EIA+/VDRL-/TP-PA+ results and no or unknown history of prior syphilis therapy, only a minority subsequently received documented treatment. These findings highlight the uncertainty that providers face in interpreting discordant RSA results, often not realizing the need to obtain a thorough medical history in this setting or in what cases treatment is indicated. Furthermore, although requesting health department records of syphilis diagnosis/ treatment is generally a helpful way to supplement medical history and guide treatment decisions, providers who are not ID specialists may not be aware of these existing health department databases.

Thorough history taking and adequate record checking is critical to avoid unnecessary treatment. In our study, a few veterans were treated for syphilis after discordant results were obtained despite having received treatment years earlier. These veterans could have avoided unnecessary treatment had VA and/or health department records been searched at the time of discordant testing. Having systems in place for providers to easily access records may help guide therapy for patients who need it and also eliminate treatment courses for cases in which it is unnecessary. Targeting treatment to only those who need it is ever more important in the current setting of benzathine penicillin shortages.

Our study was limited as a retrospective analysis subject to the shortcomings of chart review. We analyzed only a subset of charts with discordant serologic testing, which may have biased our results; however, this subset was randomly selected. Although we searched charts thoroughly for terms indicative of syphilis documentation, we may have missed cases in which providers actually discussed treatment with their patients if documentation was incomplete. Although we were able to view laboratory results for all veterans in our VISN, we only had readily available charts for review among veterans receiving care at the DVAMC, and thus our results may not be generalizable. Finally, we caution that our conclusions are not based on absolute occurrence of prior syphilis treatment but rather on prior results in the VA system or what the provider documented in the chart. We attempted to verify these findings by searching health department records and found that in the vast majority of cases, records were in fact consistent.

Our study brings to light important implications of RSA implementation, in that providers continue to face management dilemmas when used in populations that have a high prevalence of past syphilis and/or low documentation rate of prior therapy. In many cases of EIA+/VDRL-/TP-PA+ results and unclear prior syphilis treatment, the patient received a letter stating, “your results are consistent with previous infection,” and no subsequent investigation into treatment history or antibiotic therapy was pursued. In other cases, there was no acknowledgement in the chart regarding the discordant results or management plan. In addition, the ID team was only consulted in less than one-third of cases. Lack of ID involvement together with changing RSA guidelines and unclear syphilis histories may have contributed to confusion regarding appropriate syphilis management. These considerations should be anticipated prior to RSA implementation, including a recommended ID consultation when necessary for patients with discordant results.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Financial support. This work was supported by the Interdisciplinary Research Training Program in AIDS training grant [NIH/NIAID 5T32AI007392-25] to M. E. C.

Potential conflicts of interest. M. E. C. reports receiving royalties from UpToDate. Otherwise, the authors have no conflicts to report. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.