Is zidovudine first-line therapy virologically comparable to tenofovir in resource-limited settings?

OBJECTIVE: To compare virologic success between adult patients on tenofovir (TDF) and zidovudine (AZT)-containing first-line antiretroviral (ART) regimens in 10 rural clinics in Lesotho, Southern Africa. METHODS: Multicentre cross-sectional study, patients ≥16 years, on first-line ART ≥6 months, receiving AZT/lamivudine (3TC) or TDF/3TC combined with efavirenz (EFV) or nevirapine (NVP). Patient characteristics and clinical/therapeutic history were collected on the day of blood draw for viral load (VL). Analysis was stratified for non-nucleoside reverse transcriptase inhibitor (EFV or NVP). A logistic regression model weighted for patients' baseline characteristics was used to assess the likelihood of virologic success (<80 copies/ml) in patients with TDF- as compared to AZT-backbones. RESULTS: In total 1539 patients were included in the analysis. Most were clinically and immunologically stable (clinical failure: 2.7% (AZT) and 2.8% (TDF); immunological failure: 4.6% (AZT) and 4.8% (TDF)). In EFV-based regimens (n = 1162), TDF was significantly associated with higher rates of virologic suppression than AZT (93.8% vs. 88.1%; weighted odds ratio: 2.15 (95% CI: 1.29-3.58; P = 0.003)). In NVP-based regimens, a similar trend was observed, but not significant (89.4% vs. 86.7%; 1.99 (0.83-4.75, P = 0.121)). CONCLUSION: These findings support the WHO recommendation to use TDF/3TC/EFV as first-line regimen. They do, however, not support the recommendation that patients who are clinically stable on AZT should continue on this first-line regimen.

Higher rates of metabolic syndrome among women taking zidovudine as compared to tenofovir in rural Africa: preliminary data from the CART-1 study.

INTRODUCTION: Due to its side effects stavudine (D4T) has been replaced by zidovudine (AZT) and tenofovir (TDF) in most low- and middle-income countries (LMICs). In 2014 about 38% of adult first-line regimens contain AZT and 62% TDF [1]. Whereas the unfavourable metabolic outcomes of D4T in comparison to TDF have been described extensively, studies from LMICs comparing metabolic profiles between patients on AZT and TDF are scarce. Given the high number of patients in LMICs still taking AZT, data on their metabolic profile are needed. We present rates of metabolic syndrome (MS) in adult patients taking either AZT- or TDF-containing first-line, non-nucleoside reverse transcriptase (NNRTI)-based regimens. MATERIALS AND METHODS: Data derived from a cross-sectional multi-disease screening conducted in ten facilities in two rural districts of Lesotho, Southern Africa [2]. Patients were eligible if aged ≥25 years and on NNRTI-containing first-line ART ≥6 months. The MS definition for Africa of the International Diabetes Federation was applied [3]. Assessed potential predictors for MS were age, time on ART, virologic suppression, body-mass index (BMI), alcohol consumption, wealth quintile, NNRTI (nevirapine (NVP) or Efavirenz (EFV)), history of previous D4T exposure and ART-backbone (AZT or TDF). Statistical analyses - stratified for sex - comprised univariate logistic regression for each predictor variable with subsequent construction of a multivariate model including all predictors with an association to MS at a significance level<0.1 in univariate analysis. RESULTS: Out of 1026 patients, 660 (64.3%) were female. MS prevalence was 9.8% (95% CI 6.9-13.4) in men and 22.9% (19.7-26.3) in women. In women, aged ≥35 years, AZT-backbone, NVP-base, BMI ≥25kg/m2 and taking ART for ≥4.5 years were associated with MS in univariate analysis. In the multivariate model only AZT (adjusted odds-ratio: 2.2, 95% CI 1.4-3.6; p=0.001) and BMI ≥25kg/m2 (9.8; 2.8-34.1, p<0.001) were associated with MS. For men, age, higher wealth quintile, history of D4T exposure and BMI were associated with MS in univariate analysis. In the multivariate model only a BMI ≥25kg/m2 was associated with MS (8.9; 3.8-20.9, p<0.001). CONCLUSIONS: In rural Lesotho, Southern Africa, the use of AZT instead of TDF among women who are on ART for ≥6 months predisposes to the development of metabolic syndrome. Given that, still 38% of first-line regimens in LMIC contain AZT, this finding needs to be verified in other settings in Sub-Saharan Africa.

Clinical and socio-demographic predictors for virologic failure in rural Southern Africa: preliminary findings from CART-1.

INTRODUCTION: In 2013, the World Health Organization (WHO) recommended scaling up of routine viral load (VL) monitoring for patients on antiretroviral therapy (ART) in resource-limited settings [1]. During the transition phase from no VL-testing at all to routine VL-monitoring, targeted VL for groups at particular risk of virologic failure (VF) may be an option [2]. We present socio-demographic and clinical risk factors for VF in a cohort in rural Lesotho with no access to VL prior to the study. MATERIALS AND METHODS: Data derive from a cross-sectional study providing multi-disease screening as well as VL testing to adult patients (≥16 years old) on first-line ART ≥6 months [3]. VF was defined as VL≥1000 copies/mL. Assessed potential predictors of VF were: (1) socio-demographic (sex, age, wealth-quintile, education, employment status, disclosure of HIV status to environment, travel-time to facility); (2) treatment history (history of treatment interruption >2 days, previous drug substitution within first-line ART, time on ART, ART-base and -backbone); (3) adherence (pill count) and (4) clinical (clinical or immunological failure as defined by WHO guidelines [1], presence of papular pruritic eruption (PPE)). All variables with association to VF in univariate analysis were included in a multivariate logistic regression reporting adjusted Odds ratios (aOR). RESULTS: Data from 1,488 patients were analyzed. Overall VF-prevalence was 6.9% (95% CI 5.7-8.3). In univariate analysis, the following were associated with VF: age <30, lower wealth-quintile, no primary education, history of treatment interruption, nevirapine-base, zidovudine-backbone, history of drug substitution, travel-time to clinic ≥2 hours, disclosure of HIV status to <5 persons, clinical failure, presence of PPE and immunological failure. In multivariate analysis, 6 out of the above 12 variables were independent predictors: age <30 years (aOR: 2.4; 95% CI 1.1-5.3, p=0.029), history of treatment interruption (2.5; 1.3-4.7, p=0.005), PPE (6.9; 2.5-18.9, p<0.001), immunological failure (11.5; 5.7-23.2, p<0.001), history of drug substitution (1.9; 1.0-3.7, p=0.043), disclosure of HIV status to <5 persons (1.8; 1.1-3.1, p=0.03). CONCLUSION: In this cohort in rural Lesotho, several socio-demographic and clinical predictors were associated with VF. Particularly age <30 years, history of treatment interruption, PPE and immunological failure were strongly associated with VF. These patients may be prioritized for targeted VL-testing.