Laboratory signals as predictors of antiretroviral treatment change

Abstract

Background: Adverse drug reactions (ADRs) secondary to the use of antiretroviral therapy (ART) are the most common reasons for medication non-adherence. This study aimed to characterise laboratory signals associated with ADRs and investigate whether they could be an effective pharmacovigilance (PV) tool for predicting ADRs in patients receiving ART.
Methods: A quantitative, analytical, retrospective study was conducted using data from patients who commenced ART from 1 January 2013 to 31 December 2017 at a private practice in Pretoria, South Africa. ADRs were graded according to the Division of AIDS (DAIDS) table. Data were analysed using frequencies, percentages, graphs, tables, and univariate and multivariable logistic regression using Stata 17.
Results: A total of 267 patient files were reviewed. Almost three-quarters (73%) had at least one ART change. There were 36 patients (13.6%) who experienced renal toxicity: two-thirds (24/36, 66.7%) were mild/moderate, while one-third (12/36, 33.3%) were either severe or potentially life-threatening. Liver and lipid toxicities were rare and mild when they did occur. A treatment change was significantly associated with the choice of non-nucleoside reverse transcriptase inhibitor (NNRTI), efavirenz (EFV) (47/176, 26.7%) versus any other NNRTI (3/39, 7.7%) (p = 0.011), and protease inhibitor (PI), lopinavir (LPV) (15/18, 83.3%) versus any other PI (12/26, 46.2%) (p = 0.013), as well as the presence of creatinine toxicity (p = 0.039).
Conclusion: This study demonstrated that the creatinine level was the only strong laboratory predictor of treatment change. This suggests that a targeted laboratory-based PV system may be sufficient to detect ADRs in this population of patients on ART.