Methods

Top Methods Results Discussion Author & Article Info References

Study Group

Routine examination of patients attending the sexually transmitted diseases clinics of the Baltimore City Health Department includes serologic screening for syphilis. Since January 1988, patients have also been offered voluntary HIV testing. Clinical staging of syphilis is based on results of history, physical findings, dark-field examination, serologic testing for syphilis, and, when possible, evaluation of recent sexual partners. Patients suspected of having syphilis are routinely evaluated for the presence of neurologic signs and symptoms that may prompt referral for lumbar puncture. Patients diagnosed with primary, secondary, or early latent syphilis (<1-year duration) are routinely treated with 2.4 million units of benzathine penicillin or, if they are allergic to penicillin, 14 days of doxycycline (100 mg orally twice daily) according to the treatment guidelines from the Centers for Disease Control and Prevention (CDC) [16].

Case Definition

Newly identified patients with syphilis are classified as having primary, secondary, early latent, late latent, or tertiary syphilis according to the CDC criteria [5]. Patients are diagnosed with primary syphilis if they have a genital ulcer; have typical spirochetes observed on dark-field microscopy of the lesion exudate; or if they have a genital ulcer with a newly documented, serologic test for syphilis (the reactive fluorescent treponemal antibody absorption [FTA-ABS] test). The FTA-ABS test was used in this definition because it may become positive before the rapid plasma reagin card test [17]. Secondary syphilis is diagnosed in patients with a rash, mucous patches, or condylomata lata with a reactive rapid plasma reagin card test and with a confirmatory FTA-ABS test. Early latent syphilis is diagnosed when the patient presents with no apparent clinical signs of syphilis infection but presents with a documented conversion of a previously negative, rapid plasma reagin card test and a FTA-ABS test within the past 12 months.

Patients in whom the duration of infection cannot be determined are considered to have latent syphilis of unknown duration. For this study, to diagnose a new infection in patients with a previous history of syphilis, a fourfold or greater increase in card test titers for rapid plasma reagin was required. Patients with latent syphilis who did not have documented serologic test conversion from nonreactive to reactive or patients with a history of syphilis who did not have a documented fourfold increase in titer were excluded from these analyses.

Patients diagnosed with HIV infection by serologic testing are routinely given counseling after the test and are offered on-site CD4 lymphocyte enumeration to help guide decisions about therapy [15].

Laboratory Testing

Serologic testing for syphilis was done using the rapid plasma reagin card test (Macro-Vue, Becton Dickinson, Maryland) and using the FTA-ABS test (Zeus Scientific Inc., Raritan, New Jersey) at the laboratories of the Baltimore City Health Department and the Maryland State Health Department according to the manufacturers' instructions. Sera reactive for rapid plasma reagin were titered to nonreactivity using twofold serial dilutions. Serologic testing for HIV infection was done as previously described using a commercially available enzyme-linked immunosorbent assay (ELISA) (ENI Corporation, Columbia, Maryland) at the disease control laboratory of the Baltimore City Health Department with confirmation of serum samples doubly reactive in the ELISA test by Western blot (Dupont, Wilmington, Delaware) [5, 15]. Serum samples were scored for Western blot reactivity according to Food and Drug Administration criteria. Western blot patterns meeting two or more alternative criteria (American Red Cross, Department of Defense, or U.S. Public Health Service) were also reported as positive. A single T-cell subset evaluation was done on site within 24 hours of collection using the Coulter Epics Profile II System (Coulter, Hialeah, Florida). Quality control was done daily according to the manufacturer's quality assurance protocol.

Therapeutic Response Criteria

Information was abstracted from any study patient who was treated for syphilis at our clinics and who had follow-up rapid plasma reagin titers reported to the disease registry of the Baltimore City Health Department. The registry comprises the identifying and demographic information, dates and provider-specific information about syphilis serologic tests (including nontreponemal titers), and treatment dates of any persons with syphilis who were diagnosed and reported to the health department within the city. The rate of titer change was determined using the time from syphilis treatment at the sexually transmitted diseases clinics to the date the rapid plasma reagin titer decreased sufficiently or failed to decrease sufficiently to meet CDC criteria for an adequate response or to the last recorded titer within 1 year of therapy [16].

Chart Review

Patients at the sexually transmitted diseases clinics who were diagnosed with syphilis between January 1990 and November 1991 were identified retrospectively from syphilis case logs maintained by the sexually transmitted diseases control program of the Baltimore City Health Department. Patients with genital ulceration attributed to culture-proven herpes simplex virus or Haemophilus ducreyi were excluded. A total of 527 patients with diagnoses of early syphilis were identified from a review of the syphilis logs from the health department; of these patients, 174 (33%) were excluded because of insufficient evidence to permit classification as having early syphilis using study definitions. Of the 353 remaining patients with early syphilis, an additional 44 were excluded because they declined voluntary HIV testing. Patients accepting HIV testing did not differ statistically from patients not accepting HIV testing in terms of age, race, gender, history of syphilis, or stage of current syphilis diagnosis (data not shown). Clinic records of patients identified in this way were reviewed using a standardized data collection instrument that included demographic information, recent sexual history, sexually transmitted disease history, risk behaviors for HIV infection, HIV testing history, syphilis and HIV laboratory results, and physical examination data pertaining to the syphilis diagnosis.

Statistical Analyses

Data extracted from clinical records were double entered into D-Base data management files for analysis. Data analysis was done using version 6.04 of SAS-PC software (SAS Institute Inc., 1990). Hypotheses about categorical data were tested using chi-square and Fisher exact tests. For two-by-two tables, chi-squares were calculated by using the Yates correction for continuity. Risk differences and 95% confidence intervals (CIs) were calculated [17]. Comparisons between log-transformed card test titers for rapid plasma reagin were done using the Wilcoxon test.

Results

Top Methods Results Discussion Author & Article Info References

Of the 309 patients with syphilis who had sufficient data to meet study definitions of early syphilis and had HIV test results, 208 (67%) were men; 108 (35%) patients had primary syphilis, 116 (38%) had secondary syphilis, and 85 (27%) had early latent disease. The study group was composed almost entirely of young African-American residents of inner-city Baltimore. Forty-four percent acknowledged risk behaviors for HIV infection, such as a history of homosexuality or bisexuality, intravenous drug use, or a known sexual partner who was infected with HIV or at risk for HIV infection (Table 1). Seventy patients (23%) were seropositive for HIV. Previous documented syphilis was common (17.5% [54 of 309]) and was associated with HIV seropositivity. Thirty-nine percent (27 of 70) of patients with HIV infection had previously had syphilis (which was documented in clinic records) compared with 11% (27 of 239) of patients without HIV infection (P < 0.001). This association remained significant for men and women when analyzed separately (P < 0.001 and P = 0.04, respectively).

Early syphilis patients with concomitant HIV infection were more likely to present with secondary syphilis than were patients who were seronegative for HIV (Table 1). For patients with or without HIV infection, 26% (18 of 70) compared with 38% (90 of 239) presented with primary syphilis, 53% (37 of 70) compared with 33% (79 of 239) presented with secondary syphilis, and 21% (15 of 70) compared with 29% (70 of 239) presented with early latent syphilis (P = 0.01). When stratified by sex, this pattern was observed in men and women but remained statistically significant only for men (Table 1). Fifty-one percent of men with HIV infection presented with secondary syphilis compared with 26% of men without infection (a difference of 25%; 95% CI, 10.4% to 39.8%; P = 0.003). Similarly, for the 101 women, 62% seropositive for HIV presented with secondary syphilis compared with 46% of women seronegative for HIV (a difference of 16%; CI, –12.4% to 44.4%; P = 0.60).

The character and distribution of cutaneous manifestations of patients with secondary syphilis were compared in patients stratified by HIV serologic status. Among 116 patients diagnosed with secondary syphilis, 2 had dark-field-positive condylomata lata without documented rash and 114 had rashes (37 patients with HIV infection and 77 without HIV infection). Specific dermatologic descriptions were charted for 84 patients. Rashes were present in the classic palmar-plantar distribution in 42 patients (48% [16 of 33] of patients seropositive for HIV and 51% [26 of 51] of patients seronegative for HIV) and were present as generalized maculopapular rashes with or without pustular components but without palmar-plantar involvement in 35 patients (45% [15 of 33] of patients seropositive for HIV and 39% [20 of 51] of patients seronegative for HIV). Two patients (1 positive and 1 negative for HIV) had nummular lesions on the penile shaft, and 5 patients (1 seropositive and 4 seronegative for HIV) had scaling maculopapular rashes limited to the groin or extremities. No statistical differences were noted in the dermatologic characteristics of rashes among patients with or without HIV infection (P = 0.84). Most of the evaluable patients (91% [77 of 84]), regardless of HIV serologic status, presented with classic palmar-plantar or generalized maculopapular rashes of secondary syphilis.

To correct for the fact that homosexually active men present more often with secondary syphilis [18], syphilis stages at the times of diagnosis were analyzed separately in heterosexual men seropositive or seronegative for HIV. Heterosexual men seropositive for HIV were more likely to present with secondary syphilis (46% [16 of 35]) than were heterosexual men seronegative for HIV (25% [35 of 142]), yielding a difference of 21% (CI, 3.1% to 39.0%; P = 0.04). Six of 22 (27%) homosexual men seropositive for HIV presented with primary syphilis compared with 1 of 7 (14%) homosexual men seronegative for HIV. Almost two thirds of homosexually active men presented with secondary disease, regardless of HIV serologic status (59% [13 of 22] of homosexual men with HIV infection compared with 57% [4 of 7] of homosexual men without infection; P = 0.70).

Because intravenous drug use has been associated with poor care-seeking behavior that could result in delayed presentation for diagnosis and treatment [19, 20], we compared the clinical stages of syphilis among intravenous drug users separately. At the time of diagnosis, the stage of syphilis did not statistically differ among intravenous drug users who were seropositive for HIV compared with those who were seronegative for HIV (primary syphilis: 22% compared with 42%; secondary syphilis: 44% compared with 37%; early latent syphilis: 34% compared with 21%, respectively; P = 0.17). In contrast, patients with HIV infection who denied intravenous drug use did present more frequently in the secondary stage of syphilis compared with patients without HIV infection (59% compared with 33%, respectively; P = 0.005).

Previously treated syphilis could modify the clinical and serologic manifestations of re-infection [21-23]. Previously having syphilis (documented) was common among patients with HIV infection in this study, and a higher percentage of persons with HIV infection presented in the secondary stage of syphilis than did persons without HIV infection. Therefore, we also did analyses stratified by history of previously having syphilis. Among patients having their first episode of syphilis, 49% (21 of 43) of patients seropositive for HIV presented in the secondary stage compared with 34% (72 of 211) of patients seronegative for HIV (a difference of 15%; CI, 0% to 31%; P = 0.07). Among patients with a previous history of syphilis, 59% (16 of 27) of patients seropositive for HIV were diagnosed in the secondary stage of syphilis compared with 26% (7 of 27) of patients seronegative for HIV (a difference of 33%; CI, 8.5% to 58.2%; P = 0.003).

Among the 116 patients with secondary syphilis, a greater percentage of patients with HIV infection presented with a persistent chancre at the time of diagnosis than did patients without infection (32% [12 of 37] of patients seropositive for HIV compared with 16% [13 of 79] of patients seronegative for HIV; P = 0.09). After stratification by syphilis history, these differences were statistically different for patients with initial episodes of syphilis (43% [9 of 21] of patients seropositive for HIV compared with 15% [11 of 72] of patients seronegative for HIV; P = 0.01) but were not for patients with previous infections (19% [3 of 16] of patients seropositive for HIV compared with 29% [2 of 7] of patients seronegative for HIV; P = 0.62).

To assess whether observed changes in clinical manifestations of syphilis were associated with HIV-related defects in cellular immunity, we evaluated the clinical presentations of early syphilis in the 52 patients seropositive for HIV for whom CD4 lymphocyte concentrations were available (Table 2). When patients were stratified according to CD4 lymphocyte concentrations of greater than or less than 500 cells/µL, we found that patients with less than 500 CD4 cells/µL were somewhat more likely to present with secondary syphilis than were patients with more than 500 CD4 cells/µL (68% [17 of 25] with < 500 cells/µL compared with 37% [10 of 27] with 500 cells/µL; P = 0.08). Interestingly, although only 4 patients with HIV infection in this study had less than 200 CD4 cells/µL, they all presented with secondary syphilis (P = 0.38).

Finally, to evaluate therapeutic response, we compared the rate of titer decline among patients with or without HIV infection as described by Gourevitch and colleagues [23] (Table 3). Card test titers for rapid plasma reagin that were reported to the Baltimore City Health Department after documented treatment in our clinics were abstracted from the disease registry of the health department. Therapy was either 2.4 million units of benzathine penicillin intramuscularly or doxycycline 100 mg orally twice a day for 14 days [16]. Repeat rapid plasma reagin titers were available for 106 of 309 patients (40% [28 of 70] of patients with HIV infection compared with 33% [78 of 239] of patients without infection; P = 0.32). Among patients with follow-up serologic tests permitting evaluation, no statistical difference by HIV serologic status was noted in the rate of decrease of card test titers for rapid plasma reagin after treatment across all stages of syphilis.

Among patients with sufficient data for analysis, 16 failed to respond appropriately (18% [5 of 28] of patients seropositive for HIV compared with 14% [11 of 78] of patients seronegative for HIV; P = 0.76). Of these 16 patients, 15 had been treated with one regimen of 2.4 million units intramuscular benzathine penicillin, and 1 patient seronegative for HIV with early latent syphilis was given the alternate regimen of 100 mg of doxycycline twice daily for 14 days. Only 1 patient, a man seronegative for HIV, who failed to respond to benzathine penicillin therapy returned to the sexually transmitted diseases clinics for follow-up evaluation. He had no neurologic signs or symptoms and a normal cerebrospinal fluid evaluation with a negative cerebrospinal fluid VDRL test. Unfortunately, follow-up information for all the other patients who failed to respond were collected at other provider sites, and, consequently, data about reinfection and neurosyphilis were not available.

Discussion

Top Methods Results Discussion Author & Article Info References

Although several epidemiologic studies have shown that syphilis [1-4], whether diagnosed by history or serologic testing, is an important risk marker for prevalent HIV infection, the effect of concomitant HIV infection on the clinical and laboratory manifestations of syphilis has been difficult to quantify. Several factors contribute to the difficulties in evaluating syphilis and HIV interactions. The natural history of syphilis is characterized by progression through a series of clinically defined stages [14]. Evaluating whether the natural history of the disease is modified by coexistent HIV infection is difficult because the clinical findings that define these stages are variable and the stages may blend into one another. In addition, despite the dramatic increases in numbers of patients with early syphilis in the United States during the late 1980s [24], few settings see sufficient numbers of patients with syphilis who do or do not have HIV infection to permit analyses controlling for potential confounders such as sex, sexual preference, drug use, or a history of previous syphilis. Nonetheless, because considerable overlap exists in factors associated with risk for syphilis and HIV infection (for example, homosexual activity and intravenous drug use), consideration of these variables is important for evaluation of how the clinical expression of syphilis might be altered in patients with HIV infection.

This study attempted to evaluate the clinical presentations of early syphilis in patients with or without HIV infection presenting to a sexually transmitted disease clinic for care. Experienced clinicians used well-defined CDC criteria for diagnosis of early syphilis. Using these criteria, we found that persons with HIV infection who had syphilis were more likely to present with secondary syphilis than patients without HIV infection, and among these patients with HIV infection who had secondary syphilis, coexistent chancres were more likely to be present. These differences persisted even after adjustment for the potential confounders listed above.

These are the first data from a large population to suggest that patients with syphilis who had coexistent HIV infection may have a modified clinical course of disease. In persons with HIV infection, both humoral and cell-mediated immunity are compromised [25, 26]. Qualitative immunologic changes in persons infected with HIV may lead to an accelerated appearance of secondary lesions, to delayed healing of primary lesions, or to both. Marra and colleagues [27] recently found delayed healing of experimental syphilitic chancres in simian immunodeficiency virus-infected macaques, supporting the thesis that delayed healing may also have contributed to persistence of chancres in patients with secondary syphilis who were also infected with HIV and seen in our clinic.

We did not have sufficient data to determine the effect of immunosuppression on the clinical course of syphilis, although we did observe a tendency for patients with HIV infection who had more advanced disease (as evidenced by lower CD4 lymphocyte concentrations) to present with secondary disease. Although only four patients in the study had less than 200 CD4 cells/µL, it is noteworthy that all presented with disseminated syphilis.

As suggested by Gourevitch and colleagues [23], who evaluated syphilis in drug users followed at drug treatment centers, we found no statistical difference in serologic response to standard therapy in patients with syphilis who were seropositive for HIV. Unfortunately, unlike patients followed in drug treatment centers, patients attending clinics for sexually transmitted diseases have relatively little incentive to return for follow-up, and only 34% (106 of 309) had serologic data permitting evaluation of response to treatment. Although no statistical difference was noted in the percentage of treatment failures by HIV serologic status, only 1 of the 16 patients who failed therapy by serologic response had documented re-evaluation for neurorelapse versus reinfection. Further prospective data in studies designed to evaluate this important issue are warranted to address persisting questions about the frequency of treatment failure and neurosyphilis prevalence in patients seropositive for HIV.

Our retrospective study had several limitations. The diagnosis of secondary syphilis was often made on the basis of an observed skin rash as well as other clinical and laboratory findings. Consequently, if patients with primary or latent syphilis and skin rashes attributed to other causes were seen at the clinic, they may have been misclassified as having secondary syphilis. We believe this unlikely because trained clinicians using careful diagnostic approaches sought to differentiate skin rashes from other processes. In reviewing available records of patients with secondary syphilis, 91% (77 of 84) were diagnosed on the basis of typical clinical findings such as palmar-plantar rashes or diffuse maculopapular or pustular rashes (or both). Further, syphilis stage was determined before knowledge of the HIV serologic status, thereby decreasing the likelihood that knowledge of HIV status modified patient classifications. We believe that most, if not all, patients described as having secondary syphilis were correctly classified.

Another potential limitation related to the retrospective nature of our study relates to the time at which these patients acquired syphilis. Many of the patients described gave a history of previously having syphilis, raising the question of whether some patients presented with relapse of previously treated syphilis rather than with re-infection. The fourfold increase in the card test titer for rapid plasma reagin used to define an active syphilis infection among patients with a previous history of treated syphilis can be seen with re-infection or relapse [28, 29]. Because most of our patients were sexually active with one or more new partners within the month before diagnosis, and almost 20% presented with known contact to a patient with documented syphilis, we believe that most of our patients were presenting with re-infection. Only a prospective study, however, would definitively answer this question.

We believe that these data help describe the interaction of syphilis and HIV. Our observations of a prolonged duration of chancres in patients with HIV infection who have infectious syphilis may help to explain how syphilis contributes to risk for HIV transmission. Because other investigators have shown that HIV can be recovered from the base of a genital ulcer in persons with HIV infection [30], a prolonged duration of genital ulcers might in turn contribute to increased efficiency of HIV transmission. Future studies are needed to elucidate the complete clinical and pathologic spectrum of the interaction of syphilis and HIV. Such knowledge will have important implications in more rapid and accurate diagnosis of syphilis and in the control of both infections.