JAIDS Journal of Acquired Immune Deficiency Syndromes:
Gender Differences in Discontinuation of Antiretroviral Treatment Regimens
Kempf, Mirjam-Colette MPH, PhD*; Pisu, Maria PhD†; Dumcheva, Anastasiya MD*; Westfall, Andrew O MS‡; Kilby, J Michael MD†; Saag, Michael S MD†
From the *Department of Epidemiology, School of Public Health, University of Alabama at Birmingham, Birmingham, AL; †Department of Medicine, School of Medicine, University of Alabama at Birmingham, Birmingham, AL; and ‡Department of Biostatistics, School of Public Health, University of Alabama at Birmingham, Birmingham, AL.
Received for publication November 10, 2008; accepted June 1, 2009.
Supported by University of Alabama at Birmingham Center for AIDS Research Grant and the MARY FISHER CARE Fund.
Presented at the 6th Annual North American Cohorts Meeting, November 9-10, 2005, Washington, DC; and at the 2nd North American Congress of Epidemiology, June 21-24, 2006, Seattle, WA.
Correspondence to: Mirjam-Colette Kempf, PhD, MPH, Department of Epidemiology, University of Alabama at Birmingham, RPHB 217G, 1530 3rd Avenue South, Birmingham, AL 35294-0022 (e-mail: firstname.lastname@example.org).
Background: Reported reasons for change or discontinuation of antiretroviral therapy (δART) include adverse events, intolerability, and nonadherence. Little is known how reasons for δART differ by gender.
Methods: In a retrospective cohort study, rates and reasons for δART alterations in a large University-based HIV clinic cohort were evaluated. Logistic regression analyses were used to evaluate the relationship between reasons for δART and gender. Cox proportional hazard models were used to investigate time to δART.
Results: In total, 631 HIV-positive individuals were analyzed. Women (n = 164) and men (n = 467) were equally likely (53.0% and 54.4%, respectively) to discontinue treatment within 12 month of initiating a new regimen. Reasons for δART, however, were different based on gender-women were more likely to δART due to poor adherence [adjusted odds ratio (OR), 1.44; 95% confidence interval (CI): 0.85 to 2.42], dermatologic symptoms (adjusted OR, 2.88; 95% CI: 1.01 to 8.18), neurological reasons (adjusted OR, 1.82; 95% CI: 0.98 to 3.39), constitutional symptoms (adjusted OR, 2.23; 95% CI: 1.10 to 4.51), and concurrent medical conditions (adjusted OR, 2.03; 95% CI: 1.00 to 4.12).
Conclusions: Although the rates of δART are similar among men and women in clinical practice, the reasons for treatment changes are different based on gender. The potential for unique patterns of adverse events and poor adherence among women requires further investigation.
Since the advent of antiretroviral (ARV) therapy, HIV-infected individuals live longer and better lives than ever before.1 Optimal treatment requires that patients adhere to their regimen, in addition to regular monitoring of therapy, to avoid side effects and virologic resistance. As a consequence HIV management consists of closely monitoring the reasons that lead to discontinuation or change of ARV regimens (δART).
Within HIV clinical care, the number and proportion of women infected with HIV is increasing in the United States and worldwide.2,3 Management of HIV-infected women is primarily based on efficacy and tolerability studies conducted in men as women were typically underrepresented in ARV clinical trials.4 Previous research on the effect of ARV therapy in women has predominantly relied on observational cohorts.5-11 However, these studies did not include comparisons of treatment outcomes between men and women. As a consequence, little is known with regard to gender differences in a number of important aspects of treatment, such as ARV tolerability, treatment duration, and reasons for δART.6,12
A better understanding of gender differences in treatment is warranted to provide most efficient care, which is less likely to result in therapy failure during follow-up. From the limited literature available on this topic, we know that women differ from men in a few important aspects of treatment. For example, women seem to delay antiretroviral therapy (ART) initiation more than men,13 they experience more side effects7,12 and seem to metabolize ARV drugs differently.14 Therefore, they may also δART more often than men, as well as being more likely to experience different reasons for changing regimens. The purpose of this study was to investigate gender differences in discontinuation rates and reasons for δART as reported by medical providers.
The University of Alabama at Birmingham 1917 Clinic Cohort is an ongoing prospective observational cohort, which represents HIV-infected patients seeking care at the clinic since its inception in 1988. Physicians and nurse practitioners provide care for more than 1200 patients. Data from physician-patient interactions have been manually entered by trained staff or directly downloaded from the hospital laboratory system into the database. The data elements include demographics, medications (including name, dose, and start and stop date), clinical problems (including opportunistic infections), vital signs, key laboratory data [eg, CD4 cell counts (CD4), HIV-1 viral load (VL), and triglycerides/cholesterol], safety laboratory results, home health nursing, hospitalizations, and insurance status. Each month, data quality control protocols are run and reviewed for erroneous or missing data values. The University of Alabama at Birmingham Institutional Review Board has approved the database protocol.
For this study, the following inclusion criteria for patients were used: ≥16 years of age, highly active antiretroviral therapy (HAART)-naive at first clinic visit, entered into clinic for care between January 1995 and August 2004, and clinical follow-up evaluations that included more than 1 CD4 or VL measurement. A total of 631 HIV-positive individuals were included in the final study population. Four types of ART regimens were considered in this study for analysis: (1) Nucleoside reverse transcriptase inhibitors containing only HAART (NRTI-HAART), (2) Protease inhibitor containing HAART (PI-HAART), (3) nonnucleoside reverse transcriptase inhibitor containing HAART (NNRTI-HAART), and (4) other ARV drug regimens including monotherapy (other ART).
Chart reviews were performed to determine the reason(s) for discontinuation of each patient's initial ART regimen. A regimen was considered discontinued/changed if any ARV within the regimen was discontinued or if any additional ARV was added. Changes in therapy lasting less than 14 days were not considered in this analysis. Reasons for discontinuation of ART regimens were evaluated based on provider notes and all reasons given for a discontinuation event or change in ART were evaluated in this study. For the analysis clinically meaningful, composite variables for δART regimen were defined (Table 1).
Univariate analyses were conducted to investigate the association between δART due to specific reasons reported by providers (Table 1) and gender. Separate logistic regression models were fit to determine factors associated with the discontinuation of the regimen for reasons summarized by the composite variables reported in Table 1, with gender being considered as the primary predictor variable. Covariates considered in this analysis were: race, type of ART regimen (NRTI-HAART, PI-HAART, NNRTI-HAART, other ART), body mass index and CD4 count at ART initiation, and HIV transmission risk category (intravenous drug use and men having sex with men).
Cox proportional hazard models were used to assess factors associated with time to discontinuation of initial regimen. A patient's observation time lasted from the date of ART initiation at the clinic to the first discontinuation or change of any ARV drug or the last clinic visit date. The selected predictor variables for time to discontinuation/change included sex, race, type of ART regimen class (NRTI-HAART, PI-HAART, NNRTI-HAART, other ART), and CD4 count at ART initiation.
The T tests were used to evaluate differences in time off regimen per year of follow-up as number of regimens taken per year of follow-up stratified by gender. All analyses were performed using SAS/STAT software, Version 9.1.3 of the SAS System for Windows XP (SAS Institute Inc, Cary, NC).
Of the 631 individuals under study, 26% (n = 164 ) were women, of whom, 78% were black and 22% were white; 74% (467 individuals) were male patients, of whom, 40% were black and 60% were white. The average age at start of the first regimen was similar for women and men, whereas the mean CD4 count was higher for women than for men (P = 0.003), and the mean VL time was lower for women than for men (P = 0.015) (Table 2). Only 6% of all individuals were ARV therapy experienced at start of the first regimen at our clinic, and the majority initiated their first regimen between 1999 and 2004 (Table 3). There were no differences by gender on the timing and types of regimens received (Table 3).
In our cohort, 78% discontinued or altered their regimens during an average follow-up time of 4.2 years (SD ± 2.7) (Table 3). After 12 months of follow-up, about half of women (53.0%) and men (54.4%) discontinued/changed their initial regimen, with no significant differences seen when stratified by gender and race (African American females: 53.8%; African American males: 53.4%; white females: 50.0%; white males: 55.2%). Overall, women and men were equally likely to discontinue or change regimens. At the time of change/discontinuation, 1 of 3 women (34%) and men (31.1%) had achieved undetectable virus levels (<50 copies/mL). Although the number of regimens per year did not differ by gender, women experienced more days off therapy than men (Table 3). Furthermore, black women reported the highest number of days off regimen per year, 51.2 days per year followed by white women, 41.0 days per year; black men, 37.3 days per year; and white men, 31.4 days per year.
The most common reason for δART regimens was virologic failure (composite variable); about 50% of both men and women had virologic failure documented as one of the reasons for changing regimens (Fig. 1). Individuals who started NNRTI-HAART instead of NRTI-HAART [adjusted odds ratio (OR), 0.46; 95% confidence interval (CI): 0.24 to 0.88] had lower rates of δART due to virologic failure. Individuals on other ARVs had a higher probability of discontinuing a regimen for virologic failure than those on NRTI-HAART (adjusted OR, 13.10, 95% CI: 5.55 to 30.96). As one of the reasons for virologic failure, although numbers involved were small, more men than women discontinued/changed their regimen due to documented resistance (6.9% vs. 2.4%, P = 0.04).
Women were more likely (44%) than men (36%) to discontinue/change their regimen because of reported nonadherence to treatment (composite variable) (adjusted OR, 1.44; 95% CI: 0.85 to 2.42). Individuals who started on NNRTI-HAART (adjusted OR, 0.44; 95% CI: 0.22 to 0.88) compared with those on NRTI-HAART were significantly less likely to discontinue their regimen because of documented poor adherence, whereas individuals on other ARVs (adjusted OR, 2.32; 95% CI: 1.53 to 3.95) were more likely to discontinue their regimens due to nonadherence to treatment. Patients who reported intravenous drug use as the mode of HIV transmission were significantly more likely to discontinue treatment for adherence reasons (adjusted OR, 2.13; 95% CI: 1.15 to 3.95) than those who did not, whereas older patients were less likely to discontinue their treatment due to nonadherence (adjusted OR, 0.96; 95% CI: 0.94 to 0.98).
Multivariate logistic regression analyses demonstrated that women were more likely to δART for neurologic (adjusted OR, 1.82; 95% CI: 0.98 to 3.39) and dermatologic symptoms (adjusted OR, 2.88; 95% CI: 1.01 to 8.18), for constitutional symptoms (adjusted OR, 2.23; 95% CI: 1.10 to 4.51), and for other concurrent medical conditions (adjusted OR, 2.03; 95% CI: 1.00 to 4.12) (composite variables) (Fig. 1). Within the composite category of neurologic reasons (Table 1 for definition of composite variables), 15% of women vs. 7% of men discontinued their regimen because of peripheral neuropathy (P = 0.002), whereas about 2% of women vs. 0.4 % of men changed/discontinued their treatment due to experiencing vertigo or dizziness (P = 0.04). Within the composite category of constitutional symptoms, 9% of women and 4% of men discontinued because of weight loss (P = 0.02) and/or fatigue (P = 0.05). Within the composite category of central nervous system related, psychiatric reasons, women were more likely to discontinue because of depression than men (7% vs. 4%, P = 0.07). In addition, although the difference was not statistically significant, women tended to discontinue/change regimens more often for rash (7.9% vs. 4.5%, P = 0.09). As expected, women were more likely to change/discontinue ART due to pregnancy (4.3%) or delivery (2.4%) of a child (categorized as a concurrent medical condition).
Time to discontinuation/change of the first regimen was not significantly different for men and women even after adjusting for race, type of regimen, and CD4 count at start of regimen. The only independent factor that had an effect on time to discontinuation/change of the first regimen was the type of regimen that was prescribed (Table 4).
In the presented cohort of 631 HAART-naive patients receiving outpatient care in a single urban clinic in the Southeast of the United States, women and men were equally likely to discontinue or change their initial HAART regimen, regardless of race, types of regimens, and CD4 count at therapy initiation. However, women spent more time off therapy than men, and the data suggest that women were more likely to discontinue or change their regimens due to neurologic, dermatologic, or constitutional toxicities or more specifically due to symptoms, such as rash, peripheral neuropathy, fatigue, weight loss, and feelings of vertigo/dizziness.
Previous studies have reported conflicting results with regard to the association of gender and discontinuation of therapy. A study by Mocroft et al8 reported that women were less likely to discontinue HAART than men, whereas d'Arminio Monforte et al7 reported that women were more likely to discontinue therapy due to toxicities. Differences between these studies in the characteristics of their populations, the types of regimens, and the definition of adverse event outcomes might explain some of the conflicting results. For example, African American women have been shown to be more likely to discontinue ART than white women.5 However, despite the high proportion of African American women (78% of women) in our cohort, we did not see a difference in discontinuation by gender or race.
Although general differences in ARV therapy response have been described for both men and women,7,10,15 specific reasons for discontinuation or change of therapy as documented by the provider and how they differ by gender have not been reported in detail. When focusing on specific ARVs or side effects, previous studies have shown gender-specific differences in ARV side effects, such as rash16 and depression.17 However, these studies were focusing on specific adverse events as outcomes of interest, rather than exploring the relative contributions of different events to the overall discontinuation rates. No studies to our knowledge have investigated the association of gender differences in therapy discontinuation due to peripheral neuropathy, fatigue, or weight loss, although the importance of these side effects as predictors of therapy failure has been established.10 In our study, these were documented more frequently for women than men as reasons for therapy discontinuation or change.
Although we identified significant gender differences in reasons for discontinuation or change of therapy, gender differences in the toxicities described may be also explained by other factors that were significantly different between men and women, that is, race and progression of HIV disease as shown by VL and CD4 count. Both race and CD4 counts have been described as predictors of adverse events in response to therapy, such as efavirenz or nevirapine.18,19
Women were more likely to discontinue or change their medications than men due to poor adherence although this difference was not significant in multivariate analysis. In particular, providers were more likely to report that women self discontinued their medicine-this might explain why women were found to be more days off therapy than men. The number of days off therapy was higher in black women than white women and higher than in white and black men. This emphasizes the importance of strengthening ART adherence strategies and interventions especially in African American women. Previous studies have also found that nonadherence to ART among women was associated with African American race. Other factors included suffering from depression, reluctance to take medications openly at home, and socioeconomic status.5,20 Although we did not evaluate socioeconomic status or barriers to ART adherence, women in our study were more likely to be African American and to have depression as a reported cause of regimen discontinuation.
The only reason for discontinuation/change in therapy that seemed to be more frequent in men than women was viral resistance. Because women were more likely to be nonadherent and to have more days off therapy than men, this difference may be more likely related to primary drug resistance of the transmitted virus rather than a response to the therapy administered. Although most cohort studies have not identified gender as a predictor for primary drug resistance in recently infected patients,21-24 the Canadian HIV strain and drug resistance surveillance program reported a higher frequency of primary drug resistance in white, male, homosexual populations than any other population group. In our study, white homosexuals accounted for the majority of the male patient population, therefore, it is plausible that the resistance observed in men was primary rather than secondary to suboptimal ARV adherence. However, we did not have data that allowed us to distinguish between primary and secondary drug resistance, nor was resistance testing administered systematically.
Although we found no differences in the discontinuation rate by gender, we found that type of regimen was associated with time to initial therapy change or discontinuation. In particular, in comparison to PI-based regimens, NNRTI-based regimens fared the best in delaying the time to regimen change. Therefore, depending on the type of regimen administered, rates of discontinuation may be found to be higher in women than in men. In our study, ART types did not differ significantly by gender, with about 33% of individuals being on PI-based regimens, 33% being on NNRTI-based regimens, and 9% being on NRTI-based regimens.
The design of this study had several limitations, which need to be considered when interpreting the results. Adverse events leading to discontinuation/change of therapy were those reported by providers and were retrospectively summarized for this study. Therefore, some information on reasons for discontinuation/change may be missing or misclassified in the medical records reviewed. However, all reasons for change/discontinuation were documented by rigorously reviewing the entire record of each patient to minimize the extent of missing information. Furthermore, questionable records were discussed with the providers for clarification. Only adverse events/reasons that led to discontinuation/change of therapy were documented for this study; therefore, the frequency of adverse events as a response to therapy may be higher assuming that not every adverse event will lead to therapy change/discontinuation. Therefore, this study did not investigate the overall frequency of specific drug toxicities by gender.
In conclusion, given the increasing predominance of women and in particular African American women in the HIV infected population, documented therapy gaps in women, especially African Americans, is a matter of concern and warrants further detailed investigation. Interventions targeting the prevention of the described adverse events and toxicities need to be developed to minimize therapy discontinuation in women. Such interventions will also have an indirect positive effect on adherence patterns given that toxicities might lead to suboptimal adherence to therapy. As suggested by Ofotokun and Pomeroy,25 individualized drug dosages based on drug plasma levels may reduce the disparities in adverse drug events and thus in discontinuation of therapy. Therefore 2 types of interventions may be needed-one at the provider level-adjusting drug dosage levels to avoid adverse events in response to therapy-and another on the patient level-providing education and resources to stress and support adherence to ART, especially among women who lack the support for taking their medications in their daily life.
1. Bhaskaran K, Hamouda O, Sannes M, et al. Changes in the risk of death after HIV seroconversion compared with mortality in the general population. JAMA
3. UNAIDS. 2008 Report on the Global AIDS Epidemic. Geneva, Switzerland: UNAIDS; August 2008.
4. Floridia M, Giuliano M, Palmisano L, et al. Gender differences in the treatment of HIV infection. Pharmacol Res
5. Ahdieh-Grant L, Tarwater PM, Schneider MF, et al. Factors and temporal trends associated with highly active antiretroviral therapy discontinuation in the Women's Interagency HIV Study. J Acquir Immune Defic Syndr
6. Boulassel MR, Morales R, Murphy T, et al. Gender and long-term metabolic toxicities from antiretroviral therapy in HIV-1 infected persons. J Med Virol
7. d'Arminio Monforte A, Lepri AC, Rezza G, et al. Insights into the reasons for discontinuation of the first highly active antiretroviral therapy (HAART) regimen in a cohort of antiretroviral naive patients. I.CO.N.A. Study Group. Italian Cohort of Antiretroviral-Naive Patients. AIDS
8. Mocroft A, Youle M, Moore A, et al. Reasons for modification and discontinuation of antiretrovirals: results from a single treatment centre. AIDS
9. Moore AL, Sabin CA, Johnson MA, et al. Gender and clinical outcomes after starting highly active antiretroviral treatment: a cohort study. J Acquir Immune Defic Syndr
10. O'Brien ME, Clark RA, Besch CL, et al. Patterns and correlates of discontinuation of the initial HAART regimen in an urban outpatient cohort. J Acquir Immune Defic Syndr
11. Yuan Y, L'Italien G, Mukherjee J, et al. Determinants of discontinuation of initial highly active antiretroviral therapy regimens in a US HIV-infected patient cohort. HIV Med
12. Currier JS, Spino C, Grimes J, et al. Differences between women and men in adverse events and CD4+ responses to nucleoside analogue therapy for HIV infection. The Aids Clinical Trials Group 175 Team. J Acquir Immune Defic Syndr
13. Mocroft A, Gill MJ, Davidson W, et al. Are there gender differences in starting protease inhibitors, HAART, and disease progression despite equal access to care? J Acquir Immune Defic Syndr
14. Dickinson L, Back D, Chandler B, et al. The impact of gender on saquinavir hard-gel/ritonavir (1000/100mg BID) pharmacokinetics and PBMC transporter expression in HIV-1 infected individuals. Paper presented at: 6th International Workshop on Clinical Pharmacology of HIV Therapy; April 28-30, 2005; Quebec City, Quebec, Canada.
15. Murri R, Lepri AC, Phillips AN, et al. Access to antiretroviral treatment, incidence of sustained therapy interruptions, and risk of clinical events according to sex: evidence from the I.Co.N.A. Study. J Acquir Immune Defic Syndr
16. Bersoff-Matcha SJ, Miller WC, Aberg JA, et al. Sex differences in nevirapine rash. Clin Infect Dis
17. Rabkin JG, Johnson J, Lin SH, et al. Psychopathology in male and female HIV-positive and negative injecting drug users: longitudinal course over 3 years. AIDS
18. Schackman BR, Ribaudo HJ, Krambrink A, et al. Racial differences in virologic failure associated with adherence and quality of life on efavirenz-containing regimens for initial HIV therapy: results of ACTG A5095. J Acquir Immune Defic Syndr
19. van Leth F, Andrews S, Grinsztejn B, et al. The effect of baseline CD4 cell count and HIV-1 viral load on the efficacy and safety of nevirapine or efavirenz-based first-line HAART. AIDS
20. Sayles JN, Wong MD, Cunningham WE. The inability to take medications openly at home: does it help explain gender disparities in HAART use? J Womens Health (Larchmt)
21. Booth CL, Garcia-Diaz AM, Youle MS, et al. Prevalence and predictors of antiretroviral drug resistance in newly diagnosed HIV-1 infection. J Antimicrob Chemother
22. Grubb JR, Singhatiraj E, Mondy K, et al. Patterns of primary antiretroviral drug resistance in antiretroviral-naive HIV-1-infected individuals in a midwest university clinic. AIDS
23. Oette M, Kaiser R, Daumer M, et al. Epidemiology of primary drug resistance in chronically HIV-infected patients in Nordrhein-Westfalen, Germany, 2001-2005 [in German]. Dtsch Med Wochenschr
24. Palma AC, Araujo F, Duque V, et al. Molecular epidemiology and prevalence of drug resistance-associated mutations in newly diagnosed HIV-1 patients in Portugal. Infect Genet Evol
25. Ofotokun I, Pomeroy C. Sex differences in adverse reactions to antiretroviral drugs. Top HIV Med
antiretroviral therapy discontinuation; antiretroviral therapy change; gender; HIV; HAART; women
© 2009 Lippincott Williams & Wilkins, Inc.
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Highlight selected keywords in the article text.
Data is temporarily unavailable. Please try again soon.