Cite this asEticha EM, Gemeda AB (2018) Predictors of mortality among adult patients enrolled on Antiretroviral Therapy in Hiwotfana specialized University Hospital, Eastern Ethiopia: Retrospective Cohort study. J HIV Clin Sci Res 5(1): 007-011. DOI: 10.17352/2455-3786.000027
Objective:Mortality of HumanImmunodeficiency Virus (HIV)- infected patients have been reduced substantially since the introduction of Antiretroviral Therapy (ART). However, many studies have shown high mortality. But the factors associated with this high mortality are poorly characterized. So this retrospective cohort study aims to determine mortality and identify predictors of it among patients on ART at Hiwotfana specialized university Hospital (HFSUH).
Result: A total of 61 (11.9%) deaths were observed giving an overall mortality rate of 2.8 per 100 person years. Advanced WHO stages (hazard ratio (HR) =3.44, 95% CI: 1.73 - 6.85), baseline CD4 count less than 200cells/mm3 (HR=3.18, 955 CI: 1.671- 6.06), regimen changed (HR=23.62, 95% CI: 9.95 - 56.05), and being anemic (HR=3.23, 95% CI: 1.71 - 6.11) were the independent predictors of mortality. Patients who started AZT based initial regimen had a lower risk of mortality (HR=0.29, 95% CI=0.13 - 0.64) when compared with the D4T based regimen.
Globally, the number of people living with human immunodeficiency virus (HIV) reached 38.8 million in 2015(1). Ethiopia is one of the sub-Saharan African countries hardest hit by the HIV epidemic. In Ethiopia, there were 786,040 HIV-infected people, 39,140 new HIV infections and 28,650 HIV/AIDS deaths in 2015 .
There has been a dramatic decline in morbidity and mortality in patients with HIV disease in all demographic sub-populations with the advent of highly active antiretroviral therapy (HAART) [2-4]. Global mortality peaked in 2005 at 1.8 million and subsequently fell by 5·5% per year to 1.2 million in 2015 . According to UNAIDS/Ethiopia report, approximately 20,000 HIV related deaths . Despite the increased availability of ART and promising efficacy, early mortality has been high after initiating ART [6,7]. Factors contributing to the high fatality rate and reduced quality of life in poor countries are poorly understood.
Multiples of factors have been associated with mortality among clients on ART [7-10]. However, those factors which contribute to the deaths of HIV-infected patients while on ART are not well explored particularly in eastern Ethiopia. Therefore, this study aimed to investigate the predictors of mortality among adult ART users in HFSUH, eastern Ethiopia.
Study period, setting and participants
The present study was conducted from January to June 2016 in HFSUH located eastern part of Ethiopia. The hospital starts ART service in 2005 and since then it has been providing, free of charge ART to eligible HIV-infected patients.
The treatment protocol for Ethiopia is implemented using World Health Organization (WHO) ART treatment guideline for HIV infection in adults and adolescents  and national guidelines for HIV prevention, care and treatment: Federal Democratic Republic of Ethiopia . According to the current treatment guidelines, HIV infected adults are eligible to start ART if their CD4cell count is < 500cells/mm3 irrespective of CD4 count or WHO clinical stage 3 or 4 irrespective CD4 cell count. Breast feeding women, pregnant women, and sero-discordant couples can start ART irrespective of CD4 cell count. For all HIV-infected children below 15 years of age, ART is recommended irrespective of WHO clinical stage and CD4 cell count.
A retrospective cohort study was conducted from January to June 2016 on treatment-naïve adult patients who enrolled on ART from 2005 to December 2015. Sample size calculation was done based on the assumption that the type I error of 5%, power by 80% and the exposure variable is CD4 count less than 200cells/mm3 (HR =5.4) which gives the maximum sample size among significant predictors collected from relevant literatures. Based on this, random sample of 513 patient’s medical records was drawn for data collection. Regarding the sampling technique, patients’ medical card numbers were generated from the computer database according to their entry time and adult patients were filtered using eligibility criteria, then we give a unique number for the remaining records and select each record for our study using systematic random sampling. Pregnant women started ART for prevention of mother to child transmission (PMCT), regimen change, and transfer in were excluded from the study.
Data collection and quality control
A Standard checklist containing study variables were developed from the patient registry card which was developed by the Ethiopian Federal Ministry of Health (FMOH). During data collection, the most recent laboratory results before starting ART were generally used as baseline values. If there is no pre-treatment laboratory test, however, the results obtained within one month of ART initiation were used. The data collection/extraction process was supervised by the principal investigator. Two trained data collectors were involved in the study; both were nurses working in the ART clinic. All completed data were examined for clarity and consistency on a daily basis.
Study variables and measurements
Patients’ survival time (time to death) was the outcome variables of this study. In some patients who lacked the time of death, the last follow up time was used. Mortality/death was defined as any recorded AIDS-related deaths, including deaths due to opportunistic infections secondary to delayed ART initiation, treatment failure and poor adherence to both ART regimen and prophylactic antibiotics.
The independent variables include: age, sex, marital status, past opportunistic infection (OI), baseline WHO clinical stages, baseline CD4 counts, nutritional status (measured by body mass index (BMI)), baseline opportunistic infection, IPT status, initial HAART, Hemoglobin level (Patients were called anemic if the hemoglobin level is (<12mg/dl)), HIV/TB co-infection, Isoniazed prophylaxis therapy (IPT) status, and regimen change. Regimen change excludes transfer-out and defaulters.
The collected data were cleaned, categorized, coded, entered and analyzed by using STATA 12. Kaplan-Meier survival analysis was used to estimate survival probability. We applied Cox-proportional hazards regression to identify risk factors associated with outcome variable (time of regimen change). Variables found to be associated with the outcome in the univariate analysis assuming a significance threshold of 20% were included in the multivariate analysis. The multivariate analysis results showed a significant effect on the outcome considering the significance thresholds of 5% were described.
Ethical clearance was obtained from Ethical clearance board of Haramaya University, college of Health and Medical science (reference number sop 792/1/2016) and data access permission was obtained from the medical director of HFSUH. We simply extracted anonymized data from the patient’s medical registry and no participant was involved in the study.
Characteristics of study participants
A total of 513 patients medical records were used to extract the information. The socio-demographic characteristics of the cohort are summarized in table 1. The mean age of the patients was 34.75 +7.19 years. Up on enrolment, 48.34% of the patients were in WHO clinical stage III or IV and the mean CD4 count were 221 cells ⁄mm3 (+111.32). The initial HAART regimens were changed for 251(48.93%) of the patients during the follow up period.
iIPT, Isoniazid prophylaxis therapy, WHO clinical stage I indicates asymptomatic and persistent generalized lymphadenopathy; WHO clinical Stage 3 was defined if one of the following is present: weight loss of >10% body weight, chronic diarrhea for >1 month, fever for >1 month, oral candidiasis, or pulmonary Tb within the previous year, or severe bacterial infections; WHO clinical Stage 4 was defined if one of the following is present in an HIV diagnosed patient: HIV wasting syndrome, PCP, toxoplasmosis of the brain, cryptosporidiosis with diarrhea for >1 month, cytomegalovirus disease, herpes simplex virus infection, progressive multifocal leukoencephalopathy, candidiasis, extra-pulmonary Tb, lymphoma, kaposi’s sarcoma.
Prevalence and trends of mortality
Patients were followed for a total of 2123.9 person years (PY) with a median follow up of 48.48 months. The overall mortality rate was 2.8 per 100 PY of follow-up (95% CI=2.22 - 3.69). In this cohort, 11.9% (n=61) of the patients were died, most of the deaths, 27 (44.26%), occurring within the first six months. The estimated survival probability at 6, 12, 18 and 24 months were 94.6%, 93.5%, 92.3% and 89.7%, respectively.
Predictors of mortality
The results from the multivariable Cox proportional hazards regression analysis found patients who were on WHO clinical stage III or IV (HR=3.44, 95% CI: 1.73 - 6.85), regimen changed (HR=23.62, 95% CI: 9.95 - 56.05), baseline CD4 count less than 200 cells/mm3 (HR=3.18, 955 CI: 1.671- 6.06) and anemic patients (HR=3.23, 95% CI=1.71 - 6.11) had higher odds of death compared with their respective counterparts shown in table 2. Patients who started AZT based initial regimen had 71% less risk of mortality (HR=0.29, 95% CI=0.13 - 0.64) when compared with the D4T based regimen.
This 10-year retrospective cohort study of HIV/AIDS patients on HAART gives an insight into survival and its predictors in hospital settings in Ethiopia. In this cohort, 1/10th (11.9%) of the patients died and the majority of the deaths (44.26%) occurred within the first six months. Studies from other parts of Ethiopia, Shashamane referral hospital, Jinka hospital and Ethiopian Somali region reported comparable results, 10.3%, 10%, 11.1% respectively [13-15]. In contrary, studies from other regions of Ethiopia, Nekemte Referral Hospital (7.2%) and Aksum Hospital (8.85%) reported lower mortality [9,16].
The estimated survival probability of our cohort at 6, 12, 18 and 24 months were 94.6%, 93.5%, 92.3% and 89.7%, respectively. This shows a better survival compared with Ethiopian Somali  and Malawi studies . This better survival in our cohort group may be due to majority of this cohort started HAART at WHO stage I or II (51.66%) and at a CD4 count > 200 cells/mm3 (57.5%).
Previous studies done in Ethiopia show that advanced WHO clinical stage was one of the major predictors of mortality for patients enrolled on ART [17-19]. This study was consistent with those findings. A Cohort study from Cameroon , indicated patients with stage III (2 times) and stage IV (3.79 times) more likely to die than patients with stage I or II. Previous studies in Africa, Tanzania, South Africa and Senegal [7,21,22] also showed that advanced stage of the disease was associated with more than doubling in the hazard of death.
In this cohort study, clients with baseline CD4 count < 200 cells/mm3 had higher hazard of mortality than patients with CD4 count >200 cells/mm3. Cohort study in university of Gondar indicated patients presented with CD4 count < 200cells/mm3 had 5 times higher risk of mortality than those with >200cells/mm3 . In line with the present study, different studies found twice or more risk of mortality among patients with CD4 count of < 200 cells/mm3 (7,8,24). Advanced immunodeficiency was associated with opportunistic infection, thereby increasing the likelihood of death.
Patients who were anaemic at baseline were 3 fold higher risk of mortality than those who doesn’t have. Previous studies done across many parts of Africa and in industrialized countries reported that anaemia at baseline was independently associated with higher mortality among patients enrolled on ART [7,20,25]. A Similar study from Ethiopia by Tadesse and colleagues , mortality was higher among patients with anaemia than their comparator. Although there is no concrete evidence on causal association between anaemia and mortality, the incidence of anemia, increased with the progression of HIV disease [7,26,27].
The initial regimen change was a significant predictor of mortality in this cohort study. Different studies confirm that ART regimen change causes diminishing the clinical and immunological benefit of treatment [28,29] failing virological suppression, increase drug resistance  and thereby increase morbidity and mortality due to HIV/AIDS [30,31]. This cohort also showed that patients on Zidovudine based regimens were shown to have less risk of mortality (HR: 0.34; 95% CI: 0.16 - 0.76) than D4T based regimen.
This study has revealed an overall low mortality rate, but the high mortality rate in the first six months of HAART initiation. Advanced WHO stages, lower CD4 count, ART regimen changed, D4T based regimen and anemia were the independent predictors of mortality. For this reason, early enrolment of patients on HIV care service and treatment is very crucial to improve patients’ survival.
We acknowledge the limitations of the current study, including: 1) This study includes only patients’ medical records with complete baseline information, so this might have made a selection bias; 2) all deaths documented were considered as HIV/AIDS related deaths due to lack of available records on the cause of death; 3) it also shares limitations of retrospective studies.
The authors acknowledge HFSUH, College of Health and Medical Science for providing access to the data. We were also grateful for the data collectors and supervisors for the carefully undertaking of their tasks.
No specific funding was received. This study was conducted for the partial fulfilment of a Bachelor of Degree in Pharmacy at School of Pharmacy, College of Health and Medical Science, Haramaya University.
EM: conceived, designed and participated in data collection, conducted the data analysis and interpretation, developed the first draft and revised subsequent drafts. MH: participated on data collection, data analysis and interpretation. TF: reviewed and commented on successive drafts. AB: commented on successive drafts.
Ethical clearance board of Haramay University, College of Health and Medical science, ethically approved all the study methods and protocols and responded with a letter reference number sop792/1/2015. Informed consent was not taken from the patients, as the information was extracted from anonymized data.
Availability of data and material
All relevant data are within the paper. The STATA data of individual patients are not permitted to be provided to other bodies, as indicated on ethical clearance. However, researchers who need further clarification can obtain anonymised data from the corresponding author on reasonable request.
Subscribe to our articles alerts and stay tuned.