Benita Yip

Association of highly active antiretroviral therapy coverage, population viral load, and yearly new HIV diagnoses in British Columbia, Canada: a population-based study

BACKGROUND Results of cohort studies and mathematical models have suggested that increased coverage with highly active antiretroviral therapy (HAART) could reduce HIV transmission. We aimed to estimate the association between plasma HIV-1 viral load, HAART coverage, and number of new cases of HIV in the population of a Canadian province. METHODS We undertook a population-based study of HAART coverage and HIV transmission in British Columbia, Canada. Data for number of HIV tests done and new HIV diagnoses were obtained from the British Columbia Centre for Disease Control. Data for viral load, CD4 cell count, and HAART use were extracted from the British Columbia Centre for Excellence in HIV/AIDS population-based registries. We modelled trends of new HIV-positive tests and number of individuals on HAART using generalised additive models. Poisson log-linear regression models were used to estimate the association between new HIV diagnoses and viral load, year, and number of individuals on HAART. FINDINGS Between 1996 and 2009, the number of individuals actively receiving HAART increased from 837 to 5413 (547% increase; p=0.002), and the number of new HIV diagnoses fell from 702 to 338 per year (52% decrease; p=0.001). The overall correlation between number of individuals on HAART and number of individuals newly testing positive for HIV per year was -0.89 (p<0.0001). For every 100 additional individuals on HAART, the number of new HIV cases decreased by a factor of 0.97 (95% CI 0.96-0.98), and per 1 log(10) decrease in viral load, the number of new HIV cases decreased by a factor of 0.86 (0.75-0.98). INTERPRETATION We have shown a strong population-level association between increasing HAART coverage, decreased viral load, and decreased number of new HIV diagnoses per year. Our results support the proposed secondary benefit of HAART used within existing medical guidelines to reduce HIV transmission. FUNDING Ministry of Health Services and Ministry of Healthy Living and Sport, Province of British Columbia; US National Institute on Drug Abuse; US National Institutes of Health; Canadian Institutes of Health Research.

Intermittent use of triple-combination therapy is predictive of mortality at baseline and after 1 year of follow-up.

Objective To characterize the impact of intermittent use of triple drug antiretroviral therapy on survival. Design, setting and participants Population-based analysis of 1282 antiretroviral therapy naive HIV-positive individuals aged 18 years and older in British Columbia who started triple-combination therapy between August 1996 and December 1999. Therapy use was estimated by dividing the number of months of medications dispensed by the number of months of follow-up. Intermittent therapy was defined as the participant having obtained less than 75% of their medication in the first 12 months. Main outcome measure Cumulative all-cause mortality rates from the start of triple drug antiretroviral therapy to 30 September 2000. Results As of 30 September 2000, 106 subjects had died. Cumulative mortality was 3.9% (± 0.5%) at 12 months. In a multivariate model, after controlling for other variables that were significant in the univariate analyses each 100 cell decrement in baseline CD4 cell count and the intermittent use of antiretroviral drugs were associated with increased mortality with risk ratios of 1.31 [95% confidence interval (CI), 1.16–1.49;P < 0.001] and 2.90 (95% CI, 1.93–4.36;P < 0.001), respectively. In order to control for downward drift, intermittent use of therapy was measured over the first year whereas other factors were measured at the end of year 1. After adjusting for all other factors, those participants who used antiretroviral drugs intermittently were 2.97 times (95% CI, 1.33–6.62;P = 0.008) more likely to die. Conclusion Our study demonstrates that even after adjusting for other prognostic factors intermittent use of antiretroviral therapy was associated with increased mortality.

Decline in deaths from AIDS due to new antiretrovirals

We determined whether availability of new antiretroviral treatments has had any impact on the rate of death for people with HIV-1 infection. Distribution of antiretroviral drugs in British Columbia, Canada, is free of charge through the Centre for Excellence in HIV/AIDS Treatment Programme. For physicians to prescribe antiretrovirals, they must complete a participant enrolment form that serves as the drug prescription. Individuals infected with HIV-1 are eligible to receive antiretroviral therapy from this programme if they have at least one CD4 cell count less than 0·5 10/L. Until December, 1995, monotherapy was made available to participants with CD4 counts less than 0·5 10/L, while double combination therapy was made available to those with CD4 counts of less than 0·35 10/L. After December, 1995, double combination treatment was made available to everyone with CD4 counts less than 0·5 10/L. Viral-loaddriven antiretroviral treatment and triple combination therapy became available after June, 1996. Of the five new medications introduced in 1996, lamivudine became available in January, saquinavir in June, stavudine in July, and indinavir and ritonavir in September. Patterns of mortality were assessed by comparing changes in death rates for those individuals on antiretroviral therapy by quarter and CD4-count groupings (0·1 10/L, 0·1–0·34 10/L, and 0·35–0·49 10/L). Mortality data were obtained through regular surveillance and computerised record linkages with Division of Vital Statistics of the British Columbia Ministry of Health. Population figures were based on the number of programme participants actively on antiretroviral therapy. Rates were expressed as deaths per 1000 active participants and were calculated over a 3-year period from January, 1994 to December, 1996. There were 604 deaths during this period among individuals ever on antiretroviral therapy; of these, 475 deaths (79%) were attributed to participants with CD4 counts less than 0·1 10/L. There was a significant decline in programme mortality rates since the first quarter of 1994 (trend test p<0·001). On average, the rate of death for those on antiretroviral treatment declined at a rate of 1·7 deaths per 1000 participants per quarter or from 18·9 deaths per 1000 participants in the first quarter of 1994 to 5·7 deaths per 1000 participants in the last quarter of 1996. As shown in the figure, the greatest decline in mortality was experienced in those participants with CD4 counts less than 0·1 10/L (trend test p<0·001). On average, the death rate for participants with CD4 counts of less than 0·1 10/L declined at a rate of 3·5 deaths per 1000 participants per quarter—ie, from 62·0 to 19·8 deaths per 1000 participants from the first quarter of 1994 to the last quarter of 1996. Although there was a decline in the death rates for other two CD4-count groups, the rates were not statistically significant. Delayed reporting was not likely to affect our analysis. The vast majority of deaths were reported through active follow-up. In this analysis, data on 536 (89%) deaths were obtained through physician and hospital reports and data on 68 (11%) were obtained through linkages. Furthermore, in a subanalysis of 179 deaths obtained through physician and hospital reports over a 1-year period ending on June 30, 1996, we found that the median follow-up time between the actual date of death and the date of reporting was 7 days (interquartile range 5–11 days). Our data show a substantial decrease in AIDS-related mortality in the province of British Columbia. The decline in mortality coincides with the availability of lamivudine in the province through open access and with the expanded use of double combination antiretroviral therapy. We believe this mortality trend will likely continue as protease inhibitors and non-nucleoside reverse transcriptase inhibitors are used to greater extent within the treatment programme.

Predictors of HIV Drug-Resistance Mutations in a Large Antiretroviral-Naive Cohort Initiating Triple Antiretroviral Therapy

OBJECTIVE The objective of this study was to systematically characterize the incidence and determinants of antiretroviral resistance in the HOMER (Highly Active Antiretroviral Therapy [HAART] Observational Medical Evaluation and Research) cohort of 1191 human immunodeficiency virus-infected, antiretroviral-naive adults initiating HAART in British Columbia, Canada. METHODS All plasma samples with plasma virus loads (pVLs) >1000 copies/mL collected during the first 30 months of follow-up were genotyped for drug resistance. The primary outcome measure was time to the first detection of major drug-resistance mutation(s). Cox proportional hazard regression was used to identify factors significantly associated with the detection of drug-resistance mutations. RESULTS Drug-resistance mutations were detected in 298 subjects (25%). Factors significantly associated with detection of drug-resistance mutations included high baseline pVL (multivariate hazard ratio [HR], 1.59; P<.001) and adherence (estimated using prescription-refill data and/or untimed plasma drug-concentration measurements). When compared with subjects with low (0%-<20%) prescription-refill percentages, subjects at an elevated risk of harboring drug-resistance mutations were those with relatively high but imperfect prescription-refill percentages (80%-<90%; multivariate HR, 4.15; P<.001) and those with essentially perfect (>/=95%) refill percentages but with 2 plasma drug concentrations below the steady-state trough concentration minus 1 standard deviation (multivariate HR, 4.57; P<.001). Initial use of nonnucleoside reverse-transcriptase inhibitor-based HAART was significantly associated with multiclass drug resistance (multivariate HR, 1.84; P=.001). CONCLUSION High baseline pVLs and substantial but imperfect levels of adherence were major predictors of antiretroviral resistance.

Effect of Medication Adherence on Survival of HIV-Infected Adults Who Start Highly Active Antiretroviral Therapy When the CD4+ Cell Count Is 0.200 to 0.350 × 109 cells/L

Context Highly active antiretroviral therapy (HAART) improves outcomes in HIV-infected patients. When is the best time to start this therapy? Contribution 1422 HIV-infected adults were followed for 2 to 6 years after starting HAART. Adherent patients who started treatment with lower (0.200 to 0.349 109 cells/L) and higher ( 0.350 109 cells/L) CD4+ cell counts had statistically similar mortality rates. Nonadherent patients had higher mortality rates than adherent patients, regardless of baseline CD4+ cell count of 0.200 to 0.349 109 cells/L or 0.350 109 cells/L or greater. Implications In HIV-infected patients, adherence, rather than when therapy is initiated before a CD4+ cell count of 0.200 109 cells/L, may be the most important determinant of survival. The Editors The benefits of highly active antiretroviral therapy (HAART) in the management of HIV disease are well established. By suppressing plasma HIV-1 RNA, HAART decreases morbidity and mortality in HIV-infected patients (1, 2). However, the optimal time to initiate HAART is uncertain. As a result, expert recommendations on the optimal time to initiate antiretroviral therapy widely differ (3-6). Several studies have suggested that only patients who initiated therapy when the CD4+ cell count had declined below 0.200 109 cells/L were at increased risk for disease progression, regardless of the baseline HIV RNA level (7-9). In 1 of these studies, further examination with additional duration of follow-up suggested that mortality may be elevated in patients who initiated therapy after the CD4+ cell count declined below 0.350 109 cells/L (10). In fact, a growing number of studies have suggested that delaying HAART after the CD4+ cell count declines below 0.350 109 cells/L may be unsafe (11-14). However, previous studies (7, 9, 11-14) did not adjust for patient adherence. This limitation is critically important because incomplete adherence is associated with increased mortality (15-17). Therefore, we sought to evaluate the effect of baseline CD4+ cell count and adherence on survival rates after the initiation of HAART. Methods The HAART Observational Medical Evaluation and Research (HOMER) study, conducted through the British Columbia Centre for Excellence in HIV/AIDS Drug Treatment Program, has been described in detail elsewhere (7, 17). Briefly, the Centre is the only free source of antiretroviral medications in British Columbia, Canada; pharmaceutical sales suggest that less than 1% of HIV-infected patients in this province obtain antiretroviral agents outside the program (18). For all participants, the program maintains a complete prospective profile of antiretroviral therapy. In the present study, we restricted analyses to HIV-infected men and women who had been antiretroviral naive until triple-drug antiretroviral therapy was prescribed between 1 August 1996 and 31 July 2000 and were followed through 31 March 2002. Study participants were initially prescribed HAART regimens that included 2 nucleoside reverse transcriptase inhibitors and either a protease inhibitor or a nonnucleoside reverse transcriptase inhibitor (the enrolling physician, in consultation with the patient, chose the regimen). End Points The primary end point in this analysis was time to death. Deaths occurring during the follow-up period were continuously identified from physician reports and through record linkages with the British Columbia Division of Vital Statistics. In the primary analysis, we evaluated all-cause mortality; in subanalyses, we censored deaths from accidental causes at the time of death and classified them as nonevents (7). Statistical Analysis KaplanMeier Analyses For the KaplanMeier analyses, we stratified patients into 5 baseline CD4+ cell count strata on the basis of previous studies and the recommendations of therapeutic guidelines: less than 0.200, 0.200 to 0.249, 0.250 to 0.299, 0.300 to 0.349, and greater than or equal to 0.350 109 cells/L (3-5, 7-9). We further stratified patients into adherent and nonadherent categories according to adherence to prescription refills (19). The definition of adherence was based on the ratio of time that medication dispensed would last as a proportion of follow-up time. This calculation was restricted to each patients first year of therapy in order to limit the potential for reverse causation in patients who ceased antiretroviral therapy after they became too sick to take medication (15, 20). We have previously demonstrated that adherence defined in this way strongly predicts virologic response and mortality and that it can be used to adjust for the potentially confounding effect of treatment interruption (15, 17, 21, 22). In the primary analysis, we wanted to evaluate how moderate adherence to HAART affected survival. Therefore, to estimate the effect of moderate adherence, a priori we categorized patients as nonadherent only if they received antiretroviral medications less than 75% of the time during the first year of therapy (15, 17). In a subanalysis, we estimated the risk for death from HIV disease among patients who were highly adherent to HAART. We defined adherence more rigorously for this analysis; patients were considered nonadherent only if they received antiretroviral medications less than 95% of the time during the first year of therapy (21, 23). We also censored deaths from accidental causes at the time of death and classified them as nonevents (7). For both analyses, we evaluated the cumulative mortality rates among the predefined baseline CD4+ cell count strata using KaplanMeier methods. Survival curves were compared by using the log-rank test. Cox Regression Analyses We were aware that additional confounding may persist if the distribution of baseline AIDS diagnoses or baseline HIV RNA levels differed between patients with various baseline CD4+ cell counts (9). Therefore, we performed Cox proportional-hazards regression analyses to calculate the adjusted relative hazards of mortality (24). In these analyses, we stratified patients into combined low (<0.200 109 cells/L), medium (0.200 to 0.349 109 cells/L), and high ( 0.350 109 cells/L) CD4+ cell count strata (3-5, 7-9). To derive adjusted relative hazards of mortality among adherent and nonadherent patients in the various CD4+ cell count strata, we built fixed models with indicator variables for each adherence and CD4+ cell count strata while adjusting for baseline HIV RNA level and other relevant covariates. Variables examined in these analyses included protease inhibitor use in the initial regimen (yes or no), a previous clinical diagnosis of AIDS (yes or no), age, sex, physician experience ( 6 patients previously enrolled in the program) (17), date of therapy initiation (before or after July 1997) (25), and baseline HIV RNA levels (log10-transformed). The assumption of proportional hazards was validated by inspection of log (log [survival function]) estimates against log time plots. We fit all multivariate models by adjusting for all variables that were statistically significant (P < 0.05) in univariate analyses. Role of the Funding Sources The funding sources had no role in the design, conduct, or reporting of the study or the decision to submit the manuscript for publication. Results Between 1 August 1996 and 31 July 2000, 1583 antiretroviral-naive participants 18 years of age and older began triple-drug therapy. Of these, 161 (10%) were excluded from this analysis because baseline CD4+ cell count and plasma HIV-1 RNA levels were not available within 6 months before the start of antiretroviral therapy. Therefore, the study sample was based on 1422 (90%) participants (1198 men [84%] and 224 women [15%]). Sex, presence of AIDS at baseline, and subsequent mortality did not differ between the study sample and the persons excluded. However, excluded persons were more likely to be younger (P = 0.04) and to be taking protease inhibitors (P = 0.02). The overall median follow-up time was 40.1 months (interquartile range, 27.7 to 52.9 months). At baseline, the median patient age was 37 years (interquartile range, 32 to 44 years), the median CD4+ cell count was 0.270 109 cells/L (interquartile range, 0.130 to 0.420 109 cells/L), and the median plasma HIV RNA level was 120 000 copies/mL (interquartile range, 38 000 to 300 000 copies/mL). Initial therapy was a protease inhibitor in 983 (69.1%) patients and a nonnucleoside reverse transcriptase inhibitor in 439 (31%) patients. During the study period, 193 patients died: 25 (13%) as a result of an accident or suicide and 14 (7%) as a result of an illicit drug overdose. Compared with patients who were less than 75% adherent, patients who were at least 75% adherent were more likely to be male (78.0% vs. 59.4%) and older (37.9 years vs. 35.4 years) and were less likely to have a history of injection drug use (78.3% vs. 65.5%) (P < 0.05 for all comparisons). In the primary analysis that considered all-cause mortality and 75% adherence, 45 of the 632 patients with at least 75% adherence who initiated HAART before the CD4+ cell count reached 0.200 109 cells/L had died by 31 March 2002; the crude mortality rate was 7.1% (Figure 1, left). Conversely, among the 264 patients who started HAART when the CD4+ cell count was 0.200 109 cells/L or greater and who were less than 75% adherent, 40 had died (crude mortality rate, 15.2%) (Figure 1, right). The absolute difference in crude mortality between adherent and nonadherent patients with a baseline CD4+ cell count of 0.200 109 cells/L or greater was 8.1 percentage points (Table 1). In KaplanMeier analyses, when patients were considered according to the CD4+ cell count strata (<0.200, 0.200 to 0.249, 0.250 to 0.299, 0.300 to 0.349, and 0.350 109 cells/L), starting HAART at a CD4+ count of 0.200 109 cells/L or greater had no survival benefit in the 632 patients who were at least 75% adherent (Figure 1, left). Table 2 shows the log-rank P values and estimated CIs comparing patients with

Continued improvement in survival among HIV-infected individuals with newer forms of highly active antiretroviral therapy.

Objective: To characterize the temporal changes in mortality and life expectancy among HIV-positive individuals initiating antiretroviral therapy in British Columbia, Canada, from 1993 to 2004. Methods: This analysis was restricted to 2238 antiretroviral-naive HIV-positive individuals who started antiretroviral therapy between January 1993 and September 2004. The primary analysis endpoint was all-cause mortality stratified by four time periods: 1993–1995, 1996–1998, 1999–2001, and 2002–2004. Cox proportional hazard models, with associated 95% confidence intervals (CI), were used to estimate the hazard of death. Abridged life tables were constructed to compare life expectancies at the age of 20 years. Results: Product limit estimates of the cumulative mortality rate at 12 months after therapy initiation decreased from 15.8% (± 1.6%) in 1993–1995 to 6.1% (± 1.1%) in 2002–2004. Life expectancy at the age of 20 years has increased from 9.1 years (± 2.3 years) in 1993–1995 to 23.6 years (± 4.4 years) in 2002–2004. Subjects in 1993–1995 were more likely to die than those who started therapy in 2002–2004 (hazard ratio 2.78; 95% CI 1.92–3.85). Patients who initiated dual therapy or therapies containing three or more antiretroviral drugs were, respectively, 1.49 (95% CI 1.23–1.82) and 2.56 (95% CI 2.13–3.13) times less likely to die than those who started on monotherapy. Conclusion: A significant and progressive decrease in mortality and increase in life expectancy were observed over the 12-year study period. The increase in life expectancy and decrease in mortality were directly associated with the use of modern forms of HAART.

The combined effect of modern highly active antiretroviral therapy regimens and adherence on mortality over time.

Objective:To characterize the impact of longitudinal adherence on survival in drug-naive individuals starting currently recommended highly active antiretroviral therapy (HAART) regimens. Methods:Eligible study participants initiated HAART between January 2000 and November 2004 and were followed until November 2005 (N = 903). HAART regimens contained efavirenz, nevirapine, or ritonavir-boosted atazanavir or lopinavir. Marginal structural modeling was used to address our objective. Results:The all-cause mortality was 11%. Individual adherence decreased significantly over time, with the mean adherence shifting from 79% within the first 6 months of starting HAART to 72% within the 24- to 30-month period (P value <0.01). Nonadherence over time (<95%) was strongly associated with higher risk of mortality (hazard ratio: 3.13; 95% confidence interval (CI): 1.95 to 5.05). Nonadherent (<95%) patients on nonnucleoside reverse transcriptase inhibitor (NNRTI)-based and boosted protease inhibitor-based regimens were, respectively, 3.61 times (95% CI: 2.15 to 6.06) and 3.25 times (95% CI: 1.63 to 6.49) more likely to die than adherent patients. Within the NNRTI-based regimens, nonadherent individuals on efavirenz were at a higher risk of mortality. Conclusions:Incomplete adherence to modern HAART over time was strongly associated with increased mortality, and patients on efavirenz-based NNRTI therapies were particularly at a higher risk if nonadherent. These results highlight the need to develop further strategies to help sustain high levels of adherence on a long-term basis.

A Simple, Dynamic Measure of Antiretroviral Therapy Adherence Predicts Failure to Maintain HIV-1 Suppression

BACKGROUND High levels of antiretroviral therapy adherence are important for human immunodeficiency virus type 1 (HIV-1) suppression, yet the magnitude of adherence required to maintain it is less well characterized. Furthermore, methods to accommodate changes in adherence over time are lacking. In the present study, our objective was to determine the magnitude of antiretroviral therapy adherence needed to maintain HIV-1 suppression by use of a time-updated adherence measure that has the potential to be of use in a clinical setting. METHODS We examined a population-based cohort of HIV-1-infected subjects > or =18 years of age, residing in British Columbia, Canada, who started receiving antiretroviral therapy between 1 August 1996 and 30 September 2003, who had at least 2 consecutive viral loads <500 copies/mL and who had prescriptions filled at least 3 times during a follow-up period ending 30 September 2004. Virological failure was defined as the second of 2 consecutive viral loads >1000 copies/mL. Cox proportional hazards model was used to determine the relationship between virological failure and refill-based, time-updated surrogate measure of adherence. RESULTS Among the 1634 participants > or =18 years of age who initiated triple combination therapy during the study, 606 virological failure events were identified. In multivariate analyses, subjects with < or =95% adherence were 1.66 (95% confidence interval, 1.38-2.01) times more likely to experience virological failure than those with >95% adherence. CONCLUSIONS The highest levels of antiretroviral therapy adherence are associated with higher rates of maintained virological suppression. This simple, dynamic surrogate measure of adherence overcomes the limitation of single-point-in-time calculations of adherence and may be useful in real time to determine whether an individual is exhibiting incomplete adherence.

Highly Active Antiretroviral Therapy and Survival in HIV-Infected Injection Drug Users

CONTEXT Highly active antiretroviral therapy (HAART) is often withheld from injection drug users (IDUs) infected with the human immunodeficiency virus (HIV) based on the belief that their unstable lifestyles may predetermine a markedly inferior outcome with HAART. However, long-term evaluations of HIV treatment outcomes among IDUs in comparison with other risk groups are not available. OBJECTIVE To compare survival rates among HIV-infected patients initiating HAART with and without a history of injection drug use. DESIGN, SETTING, AND PATIENTS Population-based, prospective cohort study (HAART Observational Medical Evaluation and Research [HOMER]) of 3116 antiretroviral-naive HIV-infected patients in a province-wide HIV/AIDS treatment program in British Columbia, Canada. Of the 3116 patients, 915 were IDUs (29.4%), 579 were female (18.6%), and the median age was 39.4 years (interquartile range, 33.3-46.4 years). Treatment with HAART was initiated between August 1, 1996, and June 30, 2006. The median duration of follow-up was 5.3 years (interquartile range, 2.8-8.3 years) for IDUs and 4.3 years (interquartile range, 2.0-7.6 years) for non-IDUs. Patients were followed up until June 30, 2007. Data were analyzed between November 1, 2007, and May 26, 2008. MAIN OUTCOME MEASURE All-cause mortality. RESULTS Overall, 622 individuals died (20.0%) during the study period (232 IDUs and 390 non-IDUs), for a crude mortality rate of 20.0% (95% confidence interval [CI], 18.4%-21.5%). At 84 months after the initiation of HAART, the product limit estimate of the cumulative all-cause mortality rate was similar between the 915 IDUs (26.5%; 95% CI, 23.2%-29.8%) and 2201 non-IDUs (21.6%; 95% CI, 16.9%-26.2%) (Wilcoxon P = .47). In multivariate time-updated Cox regression, the hazard ratio of mortality was similar between IDUs and non-IDUs (1.09; 95% CI, 0.92-1.29). CONCLUSION In this study population, injection drug use was not associated with decreased survival among HIV-infected patients initiating HAART.

Emergence of Drug Resistance Is Associated with an Increased Risk of Death among Patients First Starting HAART

Background The impact of the emergence of drug-resistance mutations on mortality is not well characterized in antiretroviral-naïve patients first starting highly active antiretroviral therapy (HAART). Patients may be able to sustain immunologic function with resistant virus, and there is limited evidence that reduced sensitivity to antiretrovirals leads to rapid disease progression or death. We undertook the present analysis to characterize the determinants of mortality in a prospective cohort study with a median of nearly 5 y of follow-up. The objective of this study was to determine the impact of the emergence of drug-resistance mutations on survival among persons initiating HAART. Methods and Findings Participants were antiretroviral therapy naïve at entry and initiated triple combination antiretroviral therapy between August 1, 1996, and September 30, 1999. Marginal structural modeling was used to address potential confounding between time-dependent variables in the Cox proportional hazard regression models. In this analysis resistance to any class of drug was considered as a binary time-dependent exposure to the risk of death, controlling for the effect of other time-dependent confounders. We also considered each separate class of mutation as a binary time-dependent exposure, while controlling for the presence/absence of other mutations. A total of 207 deaths were identified among 1,138 participants over the follow-up period, with an all cause mortality rate of 18.2%. Among the 679 patients with HIV-drug-resistance genotyping done before initiating HAART, HIV-drug resistance to any class was observed in 53 (7.8%) of the patients. During follow-up, HIV-drug resistance to any class was observed in 302 (26.5%) participants. Emergence of any resistance was associated with mortality (hazard ratio: 1.75 [95% confidence interval: 1.27, 2.43]). When we considered each class of resistance separately, persons who exhibited resistance to non-nucleoside reverse transcriptase inhibitors had the highest risk: mortality rates were 3.02 times higher (95% confidence interval: 1.99, 4.57) for these patients than for those who did not exhibit this type of resistance. Conclusions We demonstrated that emergence of resistance to non-nucleoside reverse transcriptase inhibitors was associated with a greater risk of subsequent death than was emergence of protease inhibitor resistance. Future research is needed to identify the particular subpopulations of men and women at greatest risk and to elucidate the impact of resistance over a longer follow-up period.