The incidence of syphilis has increased dramatically during the last decade [13], as has coinfection with HIV [14, 15]. Users of intravenous drugs and "crack" cocaine, already at high risk for HIV infection, are also at substantial risk for syphilis [16-20]. Diagnosis and treatment of syphilis in HIV-infected intravenous drug users may be further complicated by the atypical syphilis serologic profiles known to occur in narcotic addicts [21, 22].

To assess the role of HIV infection on stage at presentation, serologic profiles, and response to therapy of patients with syphilis, we studied a cohort of HIV-seropositive and HIV-seronegative intravenous drug users attending a methadone maintenance treatment program in New York City, the site of an ongoing longitudinal study of HIV infection [23, 24].

Methods

Top Methods Results Discussion Author & Article Info References

Cohort Description

Data for this study were pooled from two sources. A longitudinal study of HIV infection among current and former intravenous drug users, initiated in mid 1985 in a long-term methadone maintenance program in New York City, has been previously described [23, 24]. In addition, a primary care clinic on the same site as the methadone program provided mandatory annual syphilis screening to all program patients, as well as ongoing primary medical care that included syphilis diagnosis and treatment [25]. All patients attending the program at any time between July 1985 and April 1991 were eligible for inclusion in this study.

Baseline Measurements

Baseline historical data concerning syphilis were obtained at enrollment, at 6- to 12-month intervals thereafter for all patients, and as indicated clinically when a diagnosis of syphilis was suspected by on-site medical providers. All patients underwent mandatory annual screening for syphilis (Roche Laboratories, Raritan, New Jersey) with a nontreponemal test (automated reagin test [ART, 1985 to 1986] or rapid plasma reagin [RPR, 1986 to 1991]). Reactive specimens were tested with a treponemal test (fluorescence treponemal antibody absorption test [FTA-Abs, 1985 to 1990] or microhemagglutination assay for antibodies to Treponema pallidum [MHA-TP, 1990 to 1991]). Nontreponemal test sera were not titrated beyond a 1:256 dilution. Patients with reactive serologic tests were further evaluated, staged, and treated if clinically indicated.

Serum was analyzed at baseline and at 6-month intervals for evidence of HIV infection by enzyme immunoassay (Abbott Laboratories, North Chicago, Illinois), with confirmation by Western blot. Additional serum was frozen for future assay, including serologic testing for syphilis.

Chart Review

We reviewed the methadone program medical records and research database since July 1985 of all patients who 1) reported a recent history of syphilis; 2) showed reactive nontreponemal and treponemal tests for syphilis; 3) received treatment for syphilis at the on-site primary care clinic; or 4) were identified as possible syphilis cases through a review of hospital records. Data abstracted from medical and research records of each patient included previous history and treatment of syphilis, physical findings, date of confirmed HIV serologic status, CD4+ T-lymphocyte counts, treatment regimen, and baseline and post-treatment serologic studies. All serologic data reported here were obtained from fresh sera; results from frozen sera were only used to prompt chart review.

Case Definition

All patients with known HIV serologic status having both reactive nontreponemal and treponemal test results on at least one specimen were considered to have syphilis. Patients with reactive nontreponemal tests and nonreactive treponemal test results were classified as having biologic false-positive results, regardless of the titer [26, 27]. Patients with syphilis who had nonreactive FTA-Abs or MHA-TP test results after treatment were classified as FTA-Abs or MHA-TP reverters. The first date on which a patient had reactive nontreponemal and treponemal test results prompting treatment was considered the date of diagnosis.

Patients with positive HIV serologic test results before or within 6 months of the diagnosis of syphilis were considered to be HIV seropositive; those with negative HIV test results after syphilis diagnosis were considered to be HIV seronegative. When applicable, HIV seroconversion dates were calculated as the midpoint between last negative and first positive HIV test result. Seroconverters were considered to be HIV seropositive if the assigned seroconversion occurred closer to the date of syphilis diagnosis than to the previous negative HIV test result. Patients with unknown HIV serologic status within 6 months of syphilis diagnosis were excluded from this study.

Staging, Treatment, and Response Criteria

Centers for Disease Control (CDC) criteria were used to stage syphilis cases and to classify treatment regimens and response to therapy [28, 29]. Dark-field microscopy was not available. Patients with a history of previous treatment for syphilis without evidence of new infection were excluded. Serologic results for HIV among patients enrolled in the longitudinal study were not routinely available to clinicians who diagnosed and treated syphilis. Data on sexual contacts regarding syphilis were unreliable, and therefore we could not differentiate relapse from reinfection with syphilis.

The rate of titer change was determined using the time from syphilis treatment to the date the nontreponemal titer decreased sufficiently or failed to decrease sufficiently to meet CDC criteria for an adequate response [29].

Indications for Lumbar Puncture

The presence of uveitis, neurologic signs or symptoms, or treatment failure in patients with syphilis prompted referral for lumbar puncture. Neither serum nontreponemal titer of 1:32 or greater nor HIV-seropositive status routinely prompted such referral [29]. A diagnosis of neurosyphilis was made if the cerebrospinal fluid (CSF) Venereal Disease Research Laboratory (VDRL) test was reactive.

Statistical Analyses

Proportions were compared using chi-square analysis or the Fisher exact test. The Student t-test was used for comparisons of means of normally distributed data. Comparisons between non-Gaussian distributions (for example, nontreponemal titers) were done using the Wilcoxon test. Odds ratios and 95% confidence intervals (CIs) were calculated using standard methods. Correlation between non-Gaussian distributions were done using the Spearman correlation coefficient. All P values were derived using two-tailed tests and were considered significant at P = 0.05. Power calculations were done by standard methods [30].

Results

Top Methods Results Discussion Author & Article Info References

Fifty patients attending the methadone maintenance treatment program from mid 1985 to April 1991 met the case definition for syphilis (Table 1).

No significant differences were seen in stage of syphilis at presentation by HIV serologic status (Table 2). Clinical findings in all cases were typical for the stage of syphilis, with no evidence of unusually fulminant manifestations of syphilis among HIV-seropositive patients.

Four HIV-seropositive patients with late latent syphilis received lumbar puncture. All four had negative CSF VDRL results. A fifth HIV-seropositive patient with an uncertain history of syphilis but with reactive serologic tests was hospitalized for fever and found to have a reactive CSF VDRL, consistent with asymptomatic neurosyphilis. Two additional patients, one with HIV infection, were diagnosed with syphilitic uveitis. Each had a reactive CSF VDRL, consistent with symptomatic neurosyphilis. No patient receiving lumbar puncture for the evaluation of neurosyphilis had an elevated CSF leukocyte count with a nonreactive CSF VDRL.

Nontreponemal Titers

Maximum nontreponemal titers were higher among HIV-seropositive (median, 1:128) than among HIV-seronegative (median, 1:32) patients with syphilis (P = 0.05) (Figure 1). This association was seen only among the 18 patients reporting a previous history of syphilis, even when maximum nontreponemal titers were corrected for the level of serofast titer preceding the episode of syphilis under investigation.

View larger version (21K): [in this window] [in a new window]   Figure 1. Distribution of maximum nontreponemal titers by human immunodeficiency virus (HIV) serologic status. No sera titrated beyond a 1:256 dilution. P = 0.05 for difference between maximum nontreponemal titers by HIV serologic status by Wilcoxon test.

The Mantel-Haenszel summary odds ratio for the association between HIV seropositivity and nontreponemal titer of 1:256 or greater, across all stages of syphilis, was 2.6 (95% CI, 1.3 to 3.9).

No correlation was seen between maximum nontreponemal titer and CD4+ T-lymphocyte count for HIV-seropositive patients (Spearman's rho = –0.07,P > 0.2).

Follow-up

Median length of follow-up from the date of first treatment for syphilis until the last serologic evaluation was 11.4 months (range, 1.6 to 55.6 months) for 47 patients, with no difference by HIV serologic status. One HIV-seronegative and one HIV-seropositive patient left the methadone program after treatment, and the exact date of treatment was unknown for one HIV-seronegative patient.

Treatment and Response

Response to treatment could be assessed adequately in 43 of the 50 patients treated (26 HIV-seropositive and 17 HIV-seronegative patients). Seven patients did not have follow-up at the intervals necessary to assess serologic response to treatment.

The 43 evaluable patients received various treatments (Table 3). Twenty-one received the standard CDC-recommended regimen of penicillin appropriate to their stage of syphilis. These included 12 HIV-seropositive patients (1 with secondary syphilis, 9 with late latent syphilis, 1 with symptomatic neurosyphilis, and 1 with asymptomatic neurosyphilis) and 9 HIV-seronegative patients (1 with early latent syphilis, 7 with late latent syphilis, and 1 with symptomatic neurosyphilis). All responded appropriately (see Table 3). The HIV-seropositive patient with symptomatic neurosyphilis showed a decrease in CSF pleocytosis and in CSF VDRL titer after treatment. The HIV-seronegative patients with neurosyphilis did not undergo post-treatment lumbar puncture but responded clinically and with decreased titers of serum nontreponemal tests.

Twelve HIV-seropositive patients (two with primary, three with secondary, five with early latent, and two with late latent syphilis) and three HIV-seronegative patients (one with primary and two with secondary syphilis) were treated with higher doses of penicillin than recommended; all responded appropriately. Of these, 12 received 2 to 4 doses of 2.4 million units of benzathine penicillin for early (primary, secondary, or early latent) syphilis. Four patients with late latent syphilis (one HIV-seropositive and three HIV-seronegative patients) were treated with a lower dose of penicillin than recommended, but all responded adequately.

Three patients with late latent syphilis were treated with nonpenicillin regimens. One HIV-seropositive and one HIV-seronegative patient responded appropriately to 4 weeks of doxycycline and erythromycin, respectively. The third, an HIV-seronegative patient with late latent syphilis, had questionable compliance with and failed to respond to 2 weeks of tetracycline but was successfully retreated 1 year later with 4 weeks of doxycycline. This patient did not receive lumbar puncture.

No difference by HIV serologic status was noted in the rate of decrease of nontreponemal titer after treatment across all stages of syphilis. The expected trend toward slower post-treatment rates of decrease in more advanced stages of syphilis was observed in both HIV-seropositive and HIV-seronegative patients (Table 4). No consistent effect of history of syphilis on rate of titer decrease was observed.

Immune Status of Treated Syphilis Cases

The CD4⁺ T-lymphocyte counts taken close to the date of first antibiotic treatment for syphilis (median, 29 days) were available for 24 of the 26 HIV-seropositive patients evaluated for response to treatment. The median CD4⁺ T-lymphocyte count at the time of first treatment was 401/mm³ (range, 2 to 1230/mm³). The CD4⁺ T-lymphocyte count was less than 200/mm³ in 6 patients, between 200 and 500/mm³ in 7 patients, and more than 500/mm³ in 11 patients.

FTA-Abs and MHA-TP Reversion

Of the 43 patients with evaluable responses to treatment, 7 had reversion of initially reactive treponemal tests (FTA-Abs or MHA-TP) to negative (4 of 26 HIV-seropositive and 3 of 17 HIV-seronegative patients; P > 0.2). Patients with primary (n = 1), early latent (n = 2), and late latent (n = 4) syphilis had treponemal test reversion, occurring before adequate nontreponemal titer response in three cases (two with early latent and one with late latent syphilis). All five cases with follow-up treponemal tests showed persistent reversion, although two had intervening positive tests and one had relapse or reinfection before again showing reversion. Patients with treponemal test reversion had a trend toward lower peak nontreponemal titers (median, 1:32) than did patients with persistent FTA-Abs or MHA-TP reactivity (median, 1:96; P = 0.12). Treponemal test reversion was unrelated to HIV serologic status, stage of syphilis, CD4⁺ T-lymphocyte count, or previous history of syphilis.

Discussion

Top Methods Results Discussion Author & Article Info References

We did not observe unusually severe manifestations or differences in stage at presentation of syphilis among HIV-seropositive patients. Although such differences have been reported [1-4, 31], their incidence is unclear due to the absence of prospective studies [32]. Our data suggest that severe cases may be the exception and that syphilis presentation usually does not differ by HIV serologic status.

Nontreponemal titers were higher in HIV-seropositive than in HIV-seronegative patients with a previous history of syphilis. This association was not confounded by residual serofast titers from earlier episodes of syphilis and might have been even stronger had titration to end point dilution been done. A recent report found higher nontreponemal titers in HIV-seropositive than in HIV-seronegative patients with secondary syphilis but, in contrast to our findings, only among patients with no history of syphilis [33]. Although nontreponemal titers are typically higher in early syphilis than in late disease [34], we found that higher titers among HIV-seropositive patients were not confounded by a greater proportion of early syphilis diagnoses among these patients. Some investigators attribute this apparent association between HIV infection and high nontreponemal titers to HIV-related B-cell dysfunction [35]. Although we did not address this issue directly, we found no association between CD4⁺ T-lymphocyte count and level of nontreponemal titer among HIV-seropositive patients with syphilis.

We observed loss of reactivity of treponemal tests (FTA-Abs, MHA-TP) with equal frequency among HIV-seropositive and HIV-seronegative patients with syphilis, unrelated to the stage of syphilis or to the degree of immunosuppression. It is commonly accepted that reactive treponemal tests remain so for life; only in primary syphilis is post-treatment treponemal test reversion expected, and then only at a low (approximately 10%) rate after 1 year of follow-up [36, 37]. Yet, 4 of 28 patients with late latent syphilis showed reversion of treponemal tests after treatment, regardless of HIV serologic status. Additional recent findings suggest that treponemal test reversion may not be confined to early syphilis. Twelve percent of HIV-seropositive patients with late latent syphilis studied by Haas and colleagues [9] showed seroreversion. Reports that treponemal test reversion is more common among patients with HIV infection [9, 38] were not supported by our study, which we believe provides the first prospectively obtained data on this subject. Treponemal test reversion was weakly associated with low nontreponemal titer, in keeping with the results of Haas and colleagues [9], suggesting that strength of the initial immune response is related to the likelihood of persistent treponemal antibody titers.

No significant difference was seen in clinical or serologic response to standard penicillin treatment by HIV serologic status, nor between standard and higher-dose regimens among HIV-seropositive patients. Although this study was not designed as a treatment trial, these findings are important for several reasons. Despite the growing number of case reports and case series describing syphilis treatment failure in HIV-infected patients, no prospective data were available by which to assess the frequency of such failures. The validity of our data on serologic response to treatment is supported by our observation, described by others, that the rate of titer decline decreases as the stage of syphilis advances (see Table 4) [36, 37]. In addition, recent reports have found little effect of HIV infection on the response of serologic test results for syphilis to penicillin treatment [39, 40]. Therefore, any difference in response to conventional penicillin therapy that may exist between HIV-seropositive and HIV-seronegative patients is likely to be small and to be observed infrequently.

Although no differences in rates of relapse or reinfection, taken together, were noted by HIV serologic status, we were unable to distinguish between these types of infection.

Several cautions in the interpretation of these data are warranted. The number of patients studied was relatively small. With one HIV-seronegative treatment failure, 11 HIV-seropositive failures would have been necessary to detect an effect of HIV infection on treatment response with 80% power at the = 0.05 level. The lack of any observed effect of HIV infection on treatment response suggests, however, that a greater sample size would not have altered our conclusions. Larger studies are necessary to provide a more definitive answer to questions regarding treatment adequacy. We do not have data on CSF findings for most of our patients. No cases of clinical neurosyphilis evolved after treatment of HIV-infected patients, and serum nontreponemal tests indicated appropriate responses to treatment. Studies of pre- and post-treatment CSF data from HIV-seropositive and HIV-seronegative patients with syphilis obtained through a controlled trial are needed to determine definitively the appropriate treatment of and the role of lumbar puncture in managing HIV-infected patients with syphilis.