Survival analysis of people living with HIV on antiretroviral therapy in Nepal.

Research Report.

SUMMARY: Background: The national Centre for AIDS and STD Control (NCASC) has estimated that about 40,723 people were living with HIV (PLHIV) in Nepal 2013 with adult HIV prevalence of 0.23% HIV Prevalence in some key affected population (KAP) groups has remained consistently above 5% in some parts of the country. HIV epidemic in Nepal is changing over time and Nepal is categorized as a country facing concerntrated HIV epidemic. People who inject drugs (PWID), female sex workers (FSWs) and their clients, migrant workers and their spouses and men who have sex with men(MSW) are KAP at higher risk of spreading the epidemic in the country. In Nepal, at the time of study Antiretroviral Therapy (ART) services were provided from 53 sites. Previously, a similar study was carried out in Far-West region of Nepal. However, so far at the national level, no studies have been conducted to analyze mortality among PLHIV on ART and factors contributing to the mortality. This study was based on a large sample drawn from five large ART sites selected from five development regions of Nepal. These sites were Bheri Zonal Hospital (BZH) BP koirala Institute of Health Sciences (BPKIHS), Seti Zonal Hospital(SZH), Sukra Raj Tropical and Infectious Disease Diseases Hospital (STIDH) and Western Regional Hospital(WRH). The objectives of the study were: (i) to measure mortality among adult PLHIV on ART, and (ii) to identify the determinants of mortality among adult PLHIV on ART. Methodology: A restrospective cohort study design was used Data on 6,977 Art patients, who started ART between Jan 2004 and Dec. 2013, from any of the five ART sites selected for the study were collected using Census and Survey Processing System(CSPro) Version 5.0 software in laptops. In total 3,799 case met the eligibility criteria of the study and complete data were collected for the cases only. Data analysis was performed using Statistical package for the Social Sciences (SPSS) version 17.0 and Stat SE version 12.0 software. Firstly, frequency distribution of all selected variables was analyzed and proportion of deaths by the categories of these variables was calculated. Secondly, mortality rates per 100 person-years at risk were calculated at the national as well as ART site level. Such mortality rates were calculated by follow-up time intervals(i.e duration on ART). Moreover, mortality rates per 100 person-years at risk were calculated separately for male and female cohorts. Kaplan-Meier (KM) survival curves were plotted for 13 selected variables. Thirdly, Cox proportional hazard regression analysis was performed at the national as well as ART site level. In these analysis (KM curves and regression) only 13 selected variables were used in a sample of 2,906 ART patients. Cases from the total 3,799 had to be excluded because of missing data in some variables. Results: In this study, data from a total, 3,799 eligible PLHIV on ART was analyzed. The highest number of cases were in STIDH, Kathmandu (41.3%) compared to other four ART sites included in the study. The mean age of eligible ART patients in the sample was 35.75+_ 8.62 years. In this study around 60% of the PLHIV were male; for more than half (56.3%) of PLHIV on ART, usual place of residence was different from the district where they received ART, and 42.5% of PLHIV on ART 'upper cast' groups. Similarly, 43% ART patients illiterate and only about 22% were employed. Overall proportion of death in the sample was 19.6% (24.1% in males and 12.9% in female). The proportion of death was consistently high among males compared to females in all categories of the selected socio-demographic, personal history, family history and clinical characteristics variables. For, example proportion of death in all four age groups(15-24;25-34;35-44; and 45 years) was consistently higher among males. Study has estimated mortality rate per 100 person-years at risk to be 6.98(95% CI: 6.46-7.54) at the national level. Mortality rate per 100 person-years is the number of deaths that will take place if 100 people at risk are observed for one year. Compared to high income countries this estimate is a bit higher but is comparable to lower middle income countries. Mortality rates were almost double among males (9.14;95% CI:8.36-9.99) compared to female (4.11; 95% CI;3.53-4.79) patients and differed by ART sites - highest in BZH, Banke( 11.46;95% CI;9.11-14.42) and lowest in BPKIHS, Sunsari(3.89; 95% CI: 2.75-5.50). Moreover, mortality increase with decrease in the duration of follow-up time intervals, Analysis showed mortality rate as high as 44.64(95% CI: 40.15-49.63) among ART patients whose follow0up time interval was 0-3 months compared to 6.98(95% CI: 6.46-7.54) among ART patients whose follow-up time interval was 4-8 years. Kaplan-Meier survival curve for 2,9.6 adult HIV -positive patients on ART showed a increasing trend of survival probability among adult HIV- positive patients over follow-up time (i.e. duration on ART). The survival probability of patients at 3 month, 6 month , 1 year , 2 year, 5 year and at 8 years was 91.51%, 89.57%, 87.79%, 83.59%, 79.75% and 75.98% respectively. The analysis also indicated that survival curves by the categories of age (p> 0.001), sex (p<0.001), place of residence (p< 0.001) WHO clinical staging (p<0.001), baseline performance scale (p<0.001), body weight (p<0.001), CD4 count (p<0.001), TB treatment during ART (p<0.001), treatment switched to 2nd line drugs (p+0.008) and ART sites (p=0.003) differed significantly. Cox-Regression analysis revealed several factors being associated with death among ART patients. Out of the 13 predictor variables selected for regression analysis ten were significantly associated with mortality in bivariate models. Out of these 10 factors, only eight factors (ART site sex, ART patient's usual place of residence, WHO clinical stage, baseline performance scale, bodyweight of patient at the start of ART , CD4 count at the start of ART and whether treatment was switched to 2nd line) showed significant associations with mortality in multivariate model. Recommendations and program and policy implications:This study has recommended that ART programs should bring PLHIV patients to ART services before their performance scale becomes too poor. As bodyweight has significant association with risk of death among PLHIV on ART, it should be closely monitored by ART programs and should start ART before their body weight is too low. Similarly, those PLHIV whose WHO staging at the start of ART was III or IV had higher risk of death, so ART programs should strengthen their capacity to monitor WHO staging of the PLHIV who are in their contact so that they can start ART before their WHO clinical stage deteriorates and minimize all possible delays for enrollment. Monitoring of CD4 count of PLHIV should be one of the active components of ongoing ART programs. This would allow ART programs to start treatment before CD4 count is too low. It was found that many cases were transferred out from the large ART sites selected for the study. As a consequence, capacity of the new ART sites that have been started in periphery of old ART centers need be strengthened in terms of handling the transferred out cases. Finally, the issue of poor recording of data at the ART sites should be addressed by allocating sufficient human resource and by providing refresher training to the staff on data recording.