Antiretroviral therapy in Sub-Saharan Africa: Evaluation of adverse effects

Objective of the study was to evaluate the prevalence and risk factors of ART-associated hepatotoxicity and anemia under d4T- and AZT-based regimens. The study was a multicenter retrospective observational study using pooled data from four AIDS-treatment sites in Mozambique and three in Malawi from 01.01.2006 to 31.08.2012. Computerized records of 10,567 HIV-infected, ART-naïve, non-pregnant adults, who started ART and had at least 12 months of follow-up after initiation of ART, were reviewed. Initial first-line ART regimens consisted of d4T- and AZT-based combinations; 95% of the patients received NVP as NNRTI. CD4 cell count, plasma HIV RNA (viral load, VL), body mass index (BMI), GOT and hemoglobin (HB) were measured before starting ART. GOT as indicator for hepatotoxicity and HB as indicator for anemia was measured after month 1, 2, 3, 6, 9 and 12. Prevalence of hepatotoxicity grade 1+2 and 3+4 and anemia grade 1+2 and 3+4 in d4T-based and AZT-based regimens during the 12-months follow-up were analyzed. The possible risk factors for prevalence increase “type of ART combination”, “CD4 count before ART”, “VL before ART”, “nutritional status before ART (BMI)”, sex and “malaria events after the start of ART” were identified by univariate and/or multivariate regression technique. Results The cohort of 10,567 patients was dichotomized in a “d4T-based arm”(n= 8231, 77.9%) and an “AZT-based arm” (n= 2306, 21.8%). Prevalence of hepatotoxicity 1+2 and 3+4 before starting ART was 18.1% and 0.5%. Prevalence of ART-associated hepatotoxicity 1+2 declined until month 3, and of hepatotoxicity 3+4 until month 6 of follow-up. In the “d4T-based arm”, prevalence of ART-associated hepatotoxicity 1+2 was during the follow-up higher than in the “AZT-based arm”. Prevalence of anemia 1+2 and 3+4 before starting ART was 28.5% and 6.7%. The prevalence of ART-associated anemia declined constantly during follow-up. The prevalence of ART-associated anemia 1+2 and 3+4 was during the complete follow-up (anemia 3+4 not in month 1) higher in the “AZT-based” than in the “d4T-based arm”. Evaluation of risk factors for increased hepatotoxicity prevalence In the univariate analysis, a d4T-based combination was during the complete follow-up a statistically significant risk factor for higher prevalence of hepatotoxicity grade 1+2 compared to AZT-based combinations, and never for higher prevalence of hepatotoxicity grade 3+4. In the multivariate analysis, a d4T-based combination was statistically significant associated at month 1, 2, 3, 6 and 9, and female sex at month 1, 2, 3, 6 and 12 with increased hepatotoxicity 1+2 prevalence. A pre-ART BMI>18.5 was a statistically significant protection factor against increased hepatotoxicity 1+2 prevalence in the first three months of ART. Pre-ART CD4 count and pre-ART VL were never statistically significant associated with increased hepatotoxicity prevalence. Evaluation of risk factors for increased anemia prevalence In the univariate analysis an AZT-based combination was a statistically significant risk factor for higher prevalence of anemia grade 1+2 during the complete follow-up, and for higher prevalence of anemia grade 3+4 at month 2, 3 and 6, compared to d4T-based combinations. In the multivariate analysis cumulative malaria was a statistically significant risk factor for an increased prevalence of anemia 1+2 in all months of the follow-up, and of anemia 3+4 in month 2 and 6. AZT-based combination was only in month 1statistically significant associated with increased anemia 1+2 prevalence, and with increased anemia 3+4 prevalence in month 2, 3 and 6. A pre-ART BMI>18.5 was a statistically significant protection factor against anemia 1+2 in month 1 and 2, and against anemia 3+4 in month 2. Pre-ART CD4 count was never and pre-ART VL only in month 1 statistically significant associated with increased anemia 1+2. An indirect comparison of overall, not only ART-associated hepatotoxicity and anemia prevalence in the d4T- and AZT- based arm during 12 months identified as most frequent statistically significant factors pre-ART VL>100,000 c/ml (risk factor) and pre-ART BMI>18.5 (protection factor). Pre-ART anemia was never and a high pre-ART CD4 count only once statistically significant associated with an increased hepatotoxicity and anemia prevalence. 19 of 22 statistically significant factors were identified in the first 3 months of ART. 17 of 22 statistically significant factors were identified in the “d4T-based arm”. In the “AZT-based arm” all 5 identified statistically significant factors were associated with anemia prevalence in the first 3 months of ART. Conclusion Hepatotoxicity prevalence: In sub-Saharan settings, a d4T-based combination and female sex are the most frequent statistically significant and independent risk factors for increased prevalence of ART-associated hepatotoxicity 1+2. A good nutritional status (pre-ART BMI>18.5) protects statistically significant against an increase of hepatotoxicity 1+2. In an indirect comparison of overall hepatotoxicity prevalence a pre-ART VL>100,000 c/ml was a statistically significant risk factor for hepatotoxicity 1+2 in the first 3 months of d4T-based ART. Anemia prevalence: In sub-Saharan settings, cumulative malaria increased significantly and independently the risk for anemia 1+2 during 12 months after the start of ART, and for anemia 3+4 at months 2 and 6. An AZT-based combination is only at month 1 a significant, independent risk factor for anemia 1+2, but until month 6 for anemia 3+4. A pre-ART BMI>18.5 was a significant and independent protective factor against anemia in the first two months of ART. In the indirect comparison of overall anemia prevalence a pre-ART VL>100,000 c/ml was a frequent statistically significant risk factor for anemia 1+2 and 3+4, mainly in the “d4T-based arm“. Less toxic first-line drugs than d4T, AZT and NVP must be introduced in sub-Sahara Africa. Also BMI and VL must be measured and considered when the start of ART and the initial first-line combination is decided.