Antiretroviral resistance and high-risk transmission behavior among HIV-positive patients in clinical care

Background: HIV-positive patients receiving antiretroviral therapy (ART) who engage in HIV transmission behaviors may harbor and transmit drug-resistant HIV. However, little is known about the risk behaviors of these patients, potential partners exposed and the relationship of these to ART resistance.

Objective: To determine the relationship of HIV drug resistance and continuing HIV transmission risk behavior among HIV-positive patients in care.

Methods: A retrospective, cross-sectional study of HIV transmission risk behavior and HIV drug resistance data from 333 HIV-positive patients.

Results: Among a diverse population of 333 HIV-positive patients, 75 (23%) had unprotected sex during the previous 3-months, resulting in 1126 unprotected sexual events with 191 partners of whom 155 were believed by patients to be HIV-negative or of unknown status. Eighteen of the 75 (24%) had resistant HIV and 207 unprotected sexual events, exposing 18% of the HIV- or status unknown partners. There was no difference in the proportion of patients engaging in unprotected sex who had undetectable viral load (VL) (22%): VL > 400 copies/ml without resistance (20%) and VL > 400copies/ml with resistance (26%). Resistance and risk behavior was predicted only by lower mental health scores (odds ratio, 10.3; 95% confidence interval, 1.7-18.6).

Conclusion: A substantial minority (23%) of patients in clinical care engaged in HIV sexual transmission risk behavior. A small subset of these also had ART-resistant HIV. However, this core group (approximately 5% of all patients) accounted for a large number of high-risk HIV transmission events with resistant virus, exposing a substantial number of partners.

Introduction

The prevalence of antiretroviral (ART)-resistant HIV in newly acquired infections in North America and Europe is estimated to range from 8 to 26% [^1-11]. Patients receiving ART in clinical care who carry resistant virus and engage in high-risk sexual HIV transmission risk behaviors (unprotected sex with HIV-negative persons), are likely to be a major source of new resistant infections. Little is known about these patients, the dynamics of their risk behaviors, partners exposed and the relationship of these to ART resistance. A better understanding of the dynamics of risk behavior and drug resistance among patients in clinical care is essential for the development of targeted prevention strategies to reduce transmission of both sensitive and resistant HIV.

To address this issue, we performed a study of ART resistance, risk behavior and the relationship of resistance and transmission risk in patients with HIV who are being followed in clinical care.

Methods

Participants

Patients who were already enrolled in the Options Project study, an ongoing longitudinal study of HIV transmission risk in HIV-positive patients in care, were recruited from two HIV clinics in Connecticut. The inclusion criteria for the ART resistance sub-study were: written informed consent, at least 18 years old, and healthy enough to complete the procedures. The study was approved by the Human Investigations Committees at the University of Connecticut, Hartford Hospital and Yale University.

Behavioral survey

At baseline before any behavioral intervention, participants completed the Assessment of HIV+ Risk Behavior (AHRB) survey. AHRB is a computer-administered self-interview with audio (ACASI) administered via laptop computers in private locations within the clinic setting. AHRB assessed for demographics, transmission risk categories, the constructs of the Information-Motivation-Behavioral Skills model [^12-14], mental and physical health functioning measured by the SF-12 [¹⁵], and sexual risk behavior over the last 3-months.

The following definitions were used.

No or low-risk sexual HIV transmission behavior: either no reported sexual events or 100% condom use.

Sexual HIV transmission risk behavior: unprotected sexual events (penile-vaginal and penile-anal for females and penile-vaginal, penile-anal and insertive penile-oral sex for men) with all partners (oral sex was restricted to partners considered to be HIV-negative or status unknown).

High-risk sexual HIV transmission behavior: transmission risk behavior with a partner believed by the subject to be HIV-negative or whose status was unknown.

Laboratory and resistance testing

HIV viral load (VL), treatment history, and CD4 cell counts were extracted from patients' medical records. HIV genotypic resistance tests were performed if the VL was > 400 HIV RNA copies/ml. The parent study did not require a plasma sample to be drawn at the time of the risk behavior interviews. Thus, we included only patients with a plasma sample available within a 3-month window of the risk survey.

Standard DNA sequencing of the HIV-1 pol gene was used to detect HIV genotypic resistance (ABI, Applied Biosystems, Foster City, CA, USA) [¹⁶]. A resistance mutation was defined utilizing the definitions of the International AIDS Society (2002) [¹⁷]. Resistance and baseline behavioral data were merged by coded identifier.

Data analysis

Data was analyzed using SPSS version 11.0.1 (SPSS, Inc., Chicago, Illinois, USA). ART-resistant individuals were characterized by demography, physical and mental health functioning, and the prevalence and amount of various types of risk over the preceding 3-month period. These characteristics were compared between those with and without resistant HIV, by univariate analyses using t-tests for continuous and chi-square tests for categorical variables. Differences were considered significant at the 0.05 or less level. Those variables that demonstrated significant univariate relations were then used in a multivariate logistic model [¹⁸] assessing the presence or absence of resistance, with associated odds ratios for each variable assessed. Sexual HIV transmission risk for the entire sample and for those with ART-resistant HIV was descriptively explored. The same strategy was used within the group of participants with resistant HIV, with and without reported risk behavior.

Results

Patient characteristics

There were 497 patients in the parent Options Project study, 404 of whom consented to the resistance sub-study. Seventy patients who agreed to the sub-study did not have a plasma sample available within the frame for resistance testing and thus were not included. These patients did not differ from the patients included in the analysis in respect to demography, clinical parameters and risk behavior. Of the 334 patients with a VL and a behavioral survey result 46% were female, 79% heterosexual, 40% African American, 33% Latino and 73% reported being on ART at the time of the survey. Heterosexual sex was reported by 47%, injection drug use by 41%, male-to-male sex by 9% and 3% reported blood transfusion as their mode of HIV acquisition.

ART resistance and sexual HIV transmission risk behavior

Overall, 53% (178 of 334) of the participants had an HIV VL > 400 copies/ml. Of the 178 patients with a detectable VL, 121 (68%) were on ART. Twenty-seven percent (89 of 334) of the entire sample had ART resistance. Most patients had resistance to a single class of antiretrovirals (54%), whereas 36% were resistant to two classes, and 10% to three classes.

Of the 334 participants, 333 provided complete responses to the baseline sexual risk behavior survey and 170 (51%) of these patients had engaged in any sexual activity in the previous 3-months. Of these, 49% (164 of 333) engaged in penetrative penile-vaginal or penile-anal sex. Twenty-three percent of patients (75 of 333 patients) engaged in sexual risk behavior, reporting one or more unprotected vaginal, anal or oral sex event over the preceding 3 months. There was no difference in proportion of patients engaging in unprotected sex between those with and without detectable VLs (23 versus 22%, P = 0.77). Further, a similar proportion of patients with a non-detectable VL (34 of 156 = 22%), a VL > 400 copies/ml without ART resistance (23 of 88 = 26%) and with ART resistance (18 of 89 = 20%) engaged in unprotected sex, P = 0.614.

The 75 patients engaging in sexual risk behavior reported a total of 1126 unprotected sexual events in the prior 3 months involving a minimum of 191 partners (Fig. 1). Eighteen of the 75 (24%) patients engaging in unprotected sex had ART-resistant virus. These 18 patients with ART resistance and sexual risk behavior had a mean CD4 T-cell count of 325 × 10⁶ cells/l (SD, ± 159) with a mean VL of 88 286 copies/ml (SD, ± 177 226). Eighteen percent of all reported unprotected sexual risk events (207 of 1126) were by patients who had resistant virus. Of the 207 reported unprotected events 166 (80%) involved unprotected vaginal or anal sex. Finally, these patients reported having unprotected sexual events with 16% of all partners engaging in unprotected sex (Fig. 1).

Fig. 1

With regard to partner HIV status, in a 3-month period 48 patients (14%) reported engaging in 703 high-risk sexual events with a total of 155 HIV-negative or status-unknown partners. Among these 48 patients, 15 (31%) had resistant HIV. These 15 patients had 149 unprotected high-risk sexual events and exposed 28 HIV-negative or status-unknown partners. Thus overall 4.5% of patients (15 of 333) had ART resistance and also engaged in high-risk sexual behavior.

Characteristics of patients with ART resistance with and without sexual HIV transmission risk behavior

The patients with resistant HIV with and without risk behavior were compared by demographics, physical and mental health functioning, and clinical parameters. There were no differences by gender, ethnicity, stability of housing, years of being HIV-positive, mean CD4 cell counts, or VL. However, those with resistance reporting sexual risk were younger (40.11 versus 44.21 years, respectively; P = 0.036) and reported higher average levels of education (1.22 versus 0.73, respectively; P = 0.07; 1 = a high school education). Bisexual orientation was more common in the group with resistance and high-risk behavior (22 versus 1.4%, P = 0.001), although the number of patients with this characteristic was small. Finally, participants with resistance and sexual risk scored significantly lower on mental health functioning on the SF-12 than non-risk resistant participants (P = 0.008). Relevant variables were entered into a multivariate logistic regression model (age, mental health scores, sexual orientation, educational attainment, and yearly income). Only mental health functioning appeared as an independent predictor of sexual risk behavior (odds ratio, -10.34; 95% confidence interval, -18.20 to -1.72; P = 0.02). Thus, for each unit decrease in mental health functioning, the odds of being classified in the resistant virus with HIV transmission group increased 10-fold.

Discussion

This study provides one of the first descriptions of sexual HIV transmission risk behaviors among HIV-positive patients in care, with both antiretroviral sensitive and resistant virus. The results indicate that there is substantial opportunity for transmission of both sensitive and resistant HIV to at-risk partners from patients in care. Among these HIV-positive patients 23% engaged in unprotected sexual risk behavior during the previous 3 months, with the total number of unprotected sexual events exceeding 1000. Further, a substantial proportion of these events were with partners who were thought by the subjects to be HIV-negative or of unknown HIV status. When the resistance and behavioral risk data was linked, 24% of patients who engaged in unprotected sex did so with resistant virus. Thus, patients who had both resistance and engaged in high-risk sexual behaviors represented only a small proportion of the entire study population (approximately 5%). Although this proportion is small, the total number of potential transmission events and partners at risk of acquisition of resistant HIV is quite substantial. This study also offers insight into the characteristics of patients with antiretroviral resistance reporting sexual risk behaviors. When assessed in a multivariate model the only independent predictor of risk behavior was lower mental health functioning. Thus, of those variables that might differentiate these patients from others and thus direct prevention efforts, lower mental health functioning, probably in the form of depression, appears to be of special importance. Recent studies have also demonstrated that ongoing risk behavior may be linked with other behaviors or conditions. Researchers have found an association between risk behavior and a history of trading sex for money and drugs [¹⁹], being female [²⁰] and, among HIV-positive men who have sex with men, depression and sildenafil use [^21,22].

Interestingly, in this study patients with non-detectable VLs and those with and without resistance had the same levels of sexual risk behavior, suggesting that, in this clinic population, the presence of VL or resistance status did not appear to affect the likelihood of engaging in risk behavior. This is consistent with a recent meta-analysis of HIV sexual risk behavior studies by Crepaz and colleagues, who reported that patients receiving ART did not exhibit increased sexual risk behavior, even when achieving an undetectable VL; however, patients beliefs about ART and VL was associated with risk behavior [²³]. Our study is limited in that it was a retrospective, cross-sectional analysis of behavior and resistance in patients from a single geographic region. Nevertheless, all risk groups were represented and the demography and risk profile is typical for urban areas of the United States where the HIV epidemic is mature [⁵]. Another limitation is that we did not identify and test exposed partners to determine true rates of resistance transmission, which would be critical in determining the transmissibility of resistant virus.

HIV-positive patients in care and engaging in risk behaviors should be a major focus of new, targeted prevention strategies that integrate prevention and clinical care [^24,25]. Results from this study suggest that a likely source of resistant infections is a small core group of patients within the clinic setting that have both resistance and high risk HIV transmission behaviors (approximately 5% of the population). Of those variables that might differentiate antiretroviral-resistant patients engaging in risk behavior from resistant patients who are not, mental health functioning appears critical, and addressing mental health issues may be the key to reducing transmission risk behaviors.

Acknowledgements

We would like to thank all the patients who participated in the study and the HIV clinicians who referred their patients.

Sponsorship: This work received financial support from the University of Connecticut NIH (grant: NIMH 1R01 MH59473-02) (J.F.) and a Veterans Administration Career Development Award (M.J.K.).