Characteristics and Risk of Syphilis Diagnosis Among HIV-Infected Male Cohort

A Population-Based Study in Houston, Texas

Abstract

Background 

This population-based study assessed the characteristics, timing, and risk of syphilis diagnoses among HIV-infected males in Houston, Texas.

Methods 

A retrospective cohort of males newly diagnosed as having HIV between January 2000 and December 2002 was constructed using HIV surveillance data. These individuals were cross-referenced to sexually transmitted disease surveillance data to ascertain early syphilis diagnoses for the subsequent 10 years. Multivariable Cox regression was used to identify risk factors for syphilis diagnosis while controlling for the effects of covariates.

Results 

Approximately 6% of the HIV-infected male cohort received early syphilis diagnoses during a 10-year period. Of these comorbid individuals, 40.8% received an incident syphilis diagnosis 5 years or more after their HIV diagnosis. Men who have sex with men (MSM) transmission risk was associated with significantly increased hazard of having a syphilis diagnosis in multivariable analysis (adjusted hazard ratio [HR] of a syphilis diagnosis, 5.24; 95% confidence interval, 3.41–8.05). Compared with men who were older than 40 years at HIV diagnosis, those 13 to 19 years old were 4.06 (2.18–7.55) times more likely to obtain a syphilis diagnosis. The HRs of having an HIV-syphilis comorbidity decreased as age increased. Compared with whites, non-Hispanic African Americans had 1.59 (1.11–2.26) times increased risk of having a subsequent syphilis diagnosis. Risk-stratified HRs showed that MSM had an increased risk of contracting syphilis in all race/ethnicity and age groups.

Conclusions 

This study suggests that HIV-positive African Americans, youth, and MSM had increased risk of having a subsequent syphilis diagnosis. Targeting these groups with STI prevention messaging may be beneficial to reducing comorbidity.

In Brief

A population-based study in Houston, Texas, discovered that men’s risk of HIV-syphilis comorbidity decreased as age increased. Men who have sex with men had an increased risk of comorbidity in all race/ethnicity and age groups.

Recent studies have summarized sexually transmitted infection (STI) point prevalence among HIV-infected populations ¹ and identified risk factors associated with STI infection after HIV diagnosis.² The most commonly studied STI was syphilis, with a median prevalence of 9.5%.¹ Although syphilis infection is not associated with initial CD4 counts³ and does not seem to affect HIV disease progression,⁴ syphilis infection among HIV-infected patients is associated with an increased plasma HIV viral load and decreased CD4 cell count.^5–7 Routine screening of syphilis in HIV-infected persons is justified not only by its yield⁸ but also by its accuracy and efficiency to break the chain of syphilis transmission.⁹ After treatment of syphilis, HIV indicators such as CD4 cell count and HIV-RNA levels are both improved.⁷

HIV-infected individuals who are aware of their HIV status substantially reduced the prevalence of high-risk sexual behaviors.^6,10 However, research continues to show that these individuals, including men who have sex with men (MSM), still engage in high-risk behavior at some point after their HIV diagnosis.^2,11 In the post-highly active antiretroviral therapy era, almost one fifth of the HIV-infected MSM who were aware of their status reported having unprotected sex with casual partners without disclosure.¹² Meta-analysis suggests that people’s health beliefs about highly active antiretroviral therapy and viral load may promote unprotected sex.¹³ All of these factors possibly resulted in the resurgence of syphilis among MSM.^14,15

Compared with HIV-infected individuals, persons with both HIV and syphilis infections are more likely to be African American, MSM or those who report anonymous sex or sex with both sexes.¹⁶ Men who have sex with men have a 140-fold higher risk of diagnosis with HIV and syphilis compared with heterosexual men.¹⁷ Among HIV-infected MSM, factors associated with testing positive for STI include age, substance use, and years since HIV diagnosis.¹⁸ The growing literature supports the need to develop targeted interventions for HIV-infected MSM who are at risk for STI infection.^18,19

A research study conducted in 1994 in Houston showed that men were 3.5 times more likely than women to be diagnosed as having primary syphilis.²⁰ Young, African American men and women in Houston inner-city neighborhoods were disproportionately affected by syphilis.²⁰ However, to our knowledge, no study was conducted to track all HIV-infected males over a decade to investigate syphilis incidence and time elapsed to incident syphilis infection.

This population-based study aims to describe the time elapsed between HIV and syphilis diagnoses using surveillance data in Houston, Texas. This method tracks populations and captures incident syphilis infections over a long period. In addition, this study describes the sociodemographic risk factors associated with increased susceptibility to early syphilis morbidity among HIV-infected males in Houston. The risk factors we examined include race/ethnicity, HIV transmission risk category (MSM/non-MSM), age at diagnosis, country of birth, sector of care (private/public), timing of HIV diagnosis (late/non-late HIV diagnosis), and year of HIV diagnosis. Identifying the time interval between HIV and syphilis diagnoses can help STI prevention programs locate the optimal timing of intervention. Because syphilis diagnoses occurred in a heterogeneous HIV-infected cohort, we hypothesized that risk of having a syphilis diagnosis may be different by sociodemographic and HIV transmission risk categories. To test this hypothesis, we examined the risk-stratified hazard ratios (HRs), adjusted survival functions, and cumulative hazard curves.

MATERIALS AND METHODS

Enhanced HIV/AIDS Reporting System

The Enhanced HIV/AIDS Reporting System (eHARS) is an HIV/AIDS surveillance system developed by the Centers for Disease Control and Prevention (CDC). The HDHHS uses eHARS to systematically collect data on HIV/AIDS in Houston/Harris County. Each year between 2008 and 2012, the Houston HIV/AIDS Surveillance Program received approximately 107,000 laboratory and provider reports resulting in 3500 investigations and 1275 newly reported cases of HIV infection. In 2011, a total of 95% of the expected number of cases were reported within 6 months of diagnosis.

Sexually Transmitted Disease Management Information System

The Sexually Transmitted Disease Management Information System, or STD*MIS, is an application created by the CDC and provided to state and local health departments. It serves as a repository for sexually transmitted disease (STD) surveillance data. Specifically, the HDHHS uses STD*MIS to house data from health care providers, disease intervention specialists, and laboratories on STD testing, morbidity, and treatment in Houston/Harris County. The HDHHS also actively uses STD*MIS for case management of public health follow-up activities, including partner services.

Definitions

A comorbid case was defined as an HIV case with a primary, secondary, or early latent syphilis diagnosis in the subsequent 10 years after the HIV diagnosis. The comparison group was defined as all remaining HIV cases diagnosed in 2000 to 2002 without any subsequent syphilis diagnosis in the 10-year follow-up period.

Sector of care was defined by the type of facility where a patient was first diagnosed as having HIV. Patients were grouped into the public care sector if they were diagnosed in the city’s sexually transmitted disease clinics, family planning clinics, tuberculosis clinics, any correctional facility, Health Resources and Services Administration–funded clinics, or County hospitals and clinics, which included county hospitals. In addition, other reporting sources include death certificates and reports from medical examiners or from any other facilities operated by federal, state, or local governments. Patients were categorized into the private care sector if they were diagnosed in private hospitals, primary care physicians’ offices, community-based clinics, or HIV counseling and testing sites operated by nongovernmental agencies. Patients first diagnosed as having HIV through out-of-state reports were grouped into “other’” sector of care. We used sector of care as a proxy measure of risk reduction messaging between facilities operated by government and other entities.

Late diagnosis was defined as an AIDS diagnosis within 12 months of an HIV diagnosis. We used late diagnosis as a proxy for access to care, as well as an indicator of disease severity at the time of HIV diagnosis.

Data Linkage

To ascertain HIV and syphilis comorbidity, we cross-referenced the eHARS and STD*MIS data sets by using a matching software called Link-King. It is a browser-based SAS application²¹ that allows for manual review and random sample validation. An STD*MIS data set was matched to an eHARS data set by first name, last name, date of birth, and social security number.

Study Population

Houston, with a population of 2.1 million, is the most populous city in Texas and the fourth most populous city in the United States. Among US metropolitan areas, the Houston–Sugar Land–Baytown Metropolitan Statistical Area ranked seventh in the prevalence of diagnosed HIV infection at the end of year 2010.²² African Americans had the highest HIV incidence rate (0.087%) among all race/ethnicity groups in Houston/Harris County. Men who had sex with men were reported in 61.1% of all new HIV infections. Similarly, the highest rates of primary and secondary syphilis infection in 2012 were among African American males (56.5 per 100,000). Among all males, the highest rates were in those 20 to 24 years of age (63.0 per 100,000) and those 25 to 29 years of age (42.9 per 100,000). Men who have sex with men transmission risk was reported in 62.9% of all interviewed primary and secondary syphilis cases.²³

This study is classified under public health program evaluation for improving HIV prevention messaging and does not require institutional review board approval. The study population consisted of 3170 males diagnosed as having HIV (regardless of AIDS diagnosis) between 2000 and 2002 in Houston/Harris County. Individuals had to be at least 13 years of age at the time of their HIV diagnosis for inclusion. We excluded 3 individuals who had an HIV diagnosis and a syphilis diagnosis on the same date because this study’s focus was the timing of syphilis diagnosis after a reported HIV diagnosis. We also excluded the 2 persons whose death dates coincided with their HIV diagnosis dates. Owing to the high burden of both syphilis and HIV among this population in Houston and the continued high proportion of cases attributed to MSM risk, males were highlighted as the study’s focus. We excluded females and categorized males based on HIV transmission risk factors into MSM and non-MSM groups.

Data Analysis

HIV/STD surveillance data were used to conduct a retrospective cohort study of HIV and syphilis comorbidity. The dependent variable was HIV and syphilis comorbidity, and the independent variables were race/ethnicity, HIV transmission risk category (MSM vs. non-MSM), age at HIV diagnosis, country of birth, HIV diagnosis year, and sector of care. We used STD*MIS data through December 31, 2011, to evaluate the risk of developing syphilis among persons with HIV. If there were multiple diagnoses of syphilis in the 10-year follow-up, the second date of syphilis diagnosis was excluded from analysis. The χ² test was used to compare patient demographics and characteristics to examine associations between MSM status and the other covariates. For the survival analysis, patients were censored at date of death as reported in eHARS. The univariate analyses were performed by using the log-rank test. Cox proportional hazards regression with censoring for syphilis diagnosis was performed to adjust for covariate effects. The final regression model to determine the effect of potential risk factors on developing syphilis was constructed by aforementioned epidemiologically relevant variables, and the cumulative hazard function of syphilis diagnosis was estimated after fitting the Cox regression model at mean values of covariates. SAS version 9.3 (SAS Institute, Cary, NC) was used for data linkage and cleaning. Statistical analyses were performed by use of Stata SE version 11.0 (Stata Corp, College Station, TX). A 2-sided P value less than 0.05 was defined to be statistically significant.

RESULTS

A total of 3170 males were newly diagnosed as having HIV between 2000 and 2002 in Houston/Harris County, and their data were collected by the HDHHS HIV/STD surveillance program in the subsequent 10 years. Similar to the local HIV epidemic, 45% were African American, and the median age at HIV diagnosis was 36 years (range, 14–84 years). In these 3 years, the number of males diagnosed as having HIV was approximately 1000 each year. The most reported HIV transmission risk factor among males was MSM (51%), followed by high-risk heterosexual contact (14%), injection drug use (9%), and MSM/injection drug use (7%). Of these, 201 (6.3%) received early syphilis diagnoses during the 10-year period. When categorized by time from HIV diagnosis to subsequent syphilis diagnosis, we found that 32.8% (n = 66) received a syphilis diagnosis less than 2.5 years after their HIV diagnosis. Another 26.4% (n = 53) received a syphilis diagnosis within 2.5 to 5 years, whereas 31.8% (n = 64) received a diagnosis within 5 to 7.5 years and 9.0% (n = 18) within 7.5 to 10 years. Therefore, of the 201 individuals identified with a comorbidity in this period, 40.8% (n = 82) received a syphilis diagnosis 5 years or more after their HIV diagnosis. The median follow-up for survival analysis was 10 years.

Among HIV-infected individuals, the characteristics of MSM and non-MSM were significantly different by age (P < 0.001), race and/or ethnicity (P < 0.001), type of facility where they were diagnosed as having HIV (P < 0.001), and country of birth (P = 0.01; Table 1). HIV-infected MSM were younger than non-MSM, and most HIV-infected MSM were aged 30 to 39 years (42%). Non-Hispanic whites made up most of HIV-infected MSM (39%), whereas African Americans made up most of non-MSM (59%). Year of HIV diagnosis and timing of HIV diagnosis were not significantly associated with MSM status.

TABLE 1: Characteristics by HIV Transmission Risk Factor (N = 3170) in 2000 to 2002

We conducted both univariate and multivariable analyses to determine the association between HIV-syphilis comorbidity and covariates. Among HIV-infected males, MSM transmission risk was associated with significantly increased hazards of having a syphilis diagnosis in multivariable analysis (adjusted HR of a syphilis diagnosis, 5.24; 95% confidence interval, 3.41–8.05; Table 2). Compared with men who were older than 40 years at HIV diagnosis, those aged 13 to 19 years were 4.06 (2.18–7.55) times more likely to obtain a syphilis diagnosis. The HRs of having an HIV-syphilis comorbidity decreased as age increased. Compared with non-Hispanic whites, non-Hispanic African Americans had 1.59 (1.11–2.26) times increased risk of having a subsequent syphilis diagnosis. Year of HIV diagnosis remained associated with comorbidity in both univariate and multivariable analyses. Late diagnosis became a nonsignificant factor after adjusting for other covariates. In multivariable analysis, country of birth and type of facility at HIV diagnosis were nonsignificant. Therefore, MSM transmission risk factor, age, race and/or ethnicity, and year of HIV diagnosis are associated with HIV-syphilis comorbidity (P < 0.05 for each; Table 2).

TABLE 2: Univariate and Multivariable Analyses for Factors Associated With HIV and Syphilis Comorbidity (N = 3170)

We further evaluated risk-stratified HRs for HIV-syphilis comorbidity associated with MSM status (Table 3), after adjusting for aforementioned covariates. We stratified the HIV-infected males by race/ethnicity and age groups. We found that compared with non-MSM HIV-infected males, MSM had an increased risk of contracting syphilis in all race/ethnicity and age groups. We constructed 2 age groups (younger vs. older), stratified by both age and race/ethnicity groups, and calculated HRs for each stratum. Among both younger and older African Americans and Hispanic cohorts, MSM had an increased risk of comorbidity in comparison with non-MSM. Owing to small sample size, the comparison could not be made among younger non-Hispanic white males. Among older non-Hispanic white males, there was no significant increase of risk of comorbidity in MSM compared with non-MSM (Table 3).

TABLE 3: Risk-Stratified HRs for Subsequent Syphilis Diagnosis Associated With MSM

After adjusting for all covariates, we plotted the cumulative hazard curves for all significant factors shown in the multivariable analysis (Table 2). Demonstrated differences were observed by MSM and non-MSM (Fig. 1A), as well as by age groups (Fig. 1B). These increases can be translated into an 8% increase of hazard estimates among MSM (2%) at the end of the 10-year follow-up period compared with non-MSM (10%; Fig. 1A). The probability of HIV-syphilis comorbidity at the end of the 10-year follow-up period decreased as age increased. This result is consistent with the multivariable analysis results presented in Table 2.

FIGURE 1: Estimated cumulative hazards by transmission risk and age.

DISCUSSION

This study identified the time elapsed between HIV and syphilis diagnoses and compared the characteristics and risk of syphilis diagnosis among HIV-infected males across different groups. We provided evidence to support our hypothesis that the risk of having a subsequent syphilis diagnosis is different by sociodemographic and HIV transmission risk categories. Our study showed that non-Hispanic whites made up most of HIV-infected MSM, whereas African Americans made up most of non-MSM. However, this finding is complicated by the potential underreporting of MSM behaviors by some subpopulations. Two studies, one using 2001–2006 National Health and Nutrition Examination Survey (NHANES) data and another using New York City’s community health survey, found that the prevalence of MSM was highest among non-Hispanic white men and lowest among non-Hispanic black men.^17,24 Evidence suggests that same-sex behaviors among non-Hispanic African American males might be underreported because of historical racial discrimination and social marginalization. Racial prejudice can contribute to African American males’ hesitancy to report same-sex behavior, a potential source of additional discrimination, and they may encounter isolation within their own community because of internal prejudices against MSM activities.^25,26 After controlling for covariates, HIV-infected MSM were at higher risk for receiving a syphilis diagnosis compared with other HIV-infected males. Further research may examine the role that frequency of health care visits contributes to volume of syphilis screening and diagnosis because MSM are significantly more likely to report a medical provider visit in the past year in comparison with heterosexual men.¹⁷

When we considered broader groupings of age at HIV diagnosis among MSM versus non-MSM, both younger (13–29 years) and older MSM (≥30 years) were strongly associated with a syphilis diagnosis in the subsequent 10 years. Overall, the HRs of having an HIV-syphilis comorbidity increased as age decreased. This finding was evident among MSM, in which younger MSM had higher HRs than older MSM (HRs of 5.95 and 4.47, respectively). Among all race/ethnicity and age stratum, younger African American MSM were more likely to experience comorbidity, which was consistent with a previous study conducted in North Carolina.¹⁶

Strengths

We used surveillance data, not self-reported diagnoses, to ascertain syphilis diagnoses among an HIV-infected cohort for 10 years. This study captures a larger sample size and limits recall bias often inherent to studies using self-reported diagnoses. A similar approach was used by Sena et al.¹⁶ in North Carolina but focused solely on those 18 to 30 years of age. The HDHHS HIV/STD surveillance program collects detailed information on demographic and risk factors for all individuals diagnosed as having HIV and/or STD in Houston/Harris County. Using time as a factor, we were able to generate HRs and estimate the risks of HIV-syphilis comorbidity among males by different sociodemographic groups.

Limitations

Although syphilis diagnoses occurring the same day as an HIV diagnosis were eliminated from the study population, incubation periods of syphilis were not accounted for. Therefore, it is possible the syphilis infection may have actually occurred before HIV acquisition. The results presented are reflective of timing of diagnoses, not acquisition, because infections often go undiagnosed and unreported for a period of time. Because local surveillance data were used, this study only reflects HIV and syphilis infections that are reported to the HDHHS. We are unable to account for those diagnosed as having HIV in Houston/Harris County that had a subsequent syphilis diagnosis reported outside the HDHHS’ jurisdiction.

Public Health Implications

Studies have established the bidirectional relationship between HIV and syphilis acquisition.^27,28 Previous research has also reviewed the prevalence of syphilis infection among HIV-infected individuals¹ and identified the increased risk for HIV and syphilis diagnoses among MSM.^{2,16,17,29,30} However, it is vital to concurrently examine the risks and timing of syphilis infection among those that are HIV positive so that STI prevention may be prioritized among HIV-positive persons. Enhancement of targeted STI prevention is promising to reduce acquisition because interventions among HIV positive have been demonstrated to significantly reduce risk behaviors.³¹ Improved targeting of STI prevention among HIV-positive persons also has direct benefits in further maximizing limited public health resources, as outlined in CDC’s High-Impact HIV Prevention Strategy.³² The results of this study indicate that targeting HIV-positive African Americans, youth, and MSM with STI prevention messaging may be beneficial to reducing comorbidity. Although similar studies have often examined coinfection morbidity 1 to 5 years after HIV diagnosis,^2,18,29 this study adds to the growing body of literature that syphilis coinfection occurs over many years after HIV infection. Over the course of this study’s 10-year follow-up period, almost 41% of subsequent syphilis diagnoses occurred 5 years past the date of HIV diagnosis. This finding confirms that STIs are of concern long after HIV diagnosis, a similar conclusion drawn in a recent review of coinfection literature with studies of shorter follow-up periods.¹

In light of the resurgence of syphilis among MSM,³³ the CDC recommends that all MSM be screened for syphilis and other STI every 3 to 6 months.³⁴ Public health officials have also recommended an increased frequency of syphilis testing in HIV-infected patients to every 3 or 4 months, concurrent with every CD4 T-cell count or plasma HIV viral load test.⁹ Although STI screening among those living with HIV is vital to interrupting disease transmission, population-based research on risks and timing of syphilis infection assists in prioritization of STI prevention efforts. In a resource-constrained public health environment, activities to prevent STI acquisition among the HIV-infected may be first indicated for the populations most at risk identified in this study.