Analysis of Time-to-Death Survival Function in a Nationally Representative Random Sample of HIV-seropositive Treatment-experienced Adult Patients from Malawi – A Historical Cohort Sample of 2004-2015 HIV Data

Abstract

ABSTRACTHow rapid HIV infection progresses to AIDS and to death is affected by different factors. This study explores survival times and associated survival factors from treatment initiation to death or censoring in antiretroviral therapy-experienced HIV-seropositive adults in Malawi from 2004-2015.A multicentre non-concurrent, retrospective cohort study was undertaken from eight ART Centres where patients’ medical records (PMRs) of HIV-positive adult patients aged 15+ years were reviewed. A life table, the Kaplan-Meier log-rank, and Cox Proportion Hazard regression were used to calculate survival time-to-death and its correlates, respectively. Hazard ratio with 95%CI and p<0.05 were used to declare statistical significance.Data for (n=9,953) patients were abstracted from PMRs. Patients median age was 40 (IQR: 33-48 years). 60.8% were females, 45.2% were aged 20-39 years, and 78.8% were married. At treatment initiation, 48.1% had advanced HIV disease clinical stage III, 24.5% had WHO stage IV, whereas 27.5% were asymptomatic – of which, 24.9% and 2.6% initiated ART due to low CD4+ count and PMTCT’s Option-B+ eligibility criteria, respectively. Survival function findings revealed that each patient had a single entry into the study. Exit time ranged from 1 to 9,224 days with the mean value of 2,421.9 days, occurring at the rate of 0.00004883 event-failure per-person-day. Time-to-death was observed at the rate of 1.78/100 person-years-at-risk (PYAR). 213 deaths (18.1%) occurred early in year-one post-ART-initiation. Deaths occurred more among persons of 20-39 years (N=470, 39.97%), and of 40-54 years (N=483, 41.07%), and was mostly due to mycobacterial pathogenic conditions (N=106, 37.3%) in particular TB infection (N=103); most of which were PTB cases (N=69, 66.9%). Mortality was high in Southern region (63.1%, N=743) but was least in Northern region (N=313) [p<0.0001]. In a multivariate Cox regression predictive model, males gender (aHR=1.42), patients age-groups of 20-39 years (aHR=1.63), 40-54 years (aHR=1.71), and 55+ years (aHR=2.66), Mzuzu Central hospital ART centre (aHR=2.66), Thyolo District hospital ART centre (aHR=3.02), semi-rural areas (aHR=1.30), urban areas (aHR=0.80), being single (aHR=0.86), chronic cough and/or breathlessness (aHR=1.19), chronic diarrhoea or weight loss (aHR=1.43), chronic fever and/or severe headache (aHR=1.30), skin or oral lesion (aHR=1.33), WHO clinical stage III (aHR=17.90), WHO clinical stage IV (aHR=20.09), low baseline CD4 count <250 cells/µL, (aHR=1.17), high baseline VL>1,000 copies/mL (aHR=2.46), Nevirapine-based therapies (aHR=1.14), and HIV duration of 3-5 years (aHR=1.17), 6-10 years (aHR=1.19) and >10 years (aHR=1.16) were all statistically significantly associated with time-to-death.This study has demonstrated survival factors associated with time-to-death among HIV-positive adults in Malawi. In order to effectively reduce AIDS mortality and win the war against AIDS-related death, the need to critically address and carefully prioritise the identified factors in HIV/AIDS management is great and cannot be overemphasised.