Survival Status and Its Predictors Among Adult Human Immunodeciency Virus Infected Patients Attending Their Antiretroviral Treatment at Health Centers in Kirkos Sub City, Addis Ababa, Ethiopia. A Retrospective Cohort Study

Burden of HIV/AIDS is high in low- and middle-income countries including Ethiopia. In resource-poor countries like Ethiopia, the survival of patients with AIDS treated with ART depends on a variety of factors, which may also vary greatly with economic, demographic, behavioral risk, and health factors. Unlike other previous studies, this study was done at health centers since some of the clinical factors may differently affect the survival of the patients under ART.

illnesses since the start of the epidemic [1], [2]. People living with HIV/AIDS now live longer and healthier lives due to the greater availability of antiretroviral (ART) treatment. Moreover, where ART treatment is readily available, HIV/AIDS morbidity has shown a modest reduction. However, the mortality and morbidity are still higher in developing country including Ethiopia [3]. In Ethiopia, various activities have been done to improve the health quality of HIV-positive patients, quality of life, and extend the time interval from HIV-infection/AIDS-diagnosis to death [4]. However, around 11,000 people die from HIV related illness in Ethiopia [5].
The risk of mortality among patients with HIV/AIDS is approximately 5 times higher in patients with AIDS than in HIV-infected Patients without AIDS. Risk factors for excess mortality include a viral load greater than 400 Copies/mL, CD4 count less than 200 cells/mL, and cytomegalovirus retinitis [6]. Other contributing factors for mortality from HIV related illness were TB co-infection, WHO clinical stage, ART regimen, drug allergy, ART regimen change, and, hemoglobin level [7].

Methods And Materials
Study design, and setting A retrospective cohort study design was conducted to estimate mortality rate and its predictors among adult HIV positive patients who had been on ART follow-up. A study was done in health centers at Kirkos sub-city, Addis Ababa Ethiopia. Kirkos sub-city is one of the 10 sub-cities in Addis Ababa and located in the city center. According to the 2011 estimate, the total population in the sub-city was 235,441. There are a total of 7 health centers in Kirkos sub-city. ART clinic is one of the departments in the health centers which provide services for HIV/AIDS patients. The study was conducted on health centers in Kirkos subcity from January 2014 to October 2019 Study population and sampling All adult HIV positive patients on ART who have complete registration, intake, and follow-up forms in the selected health centers were included in the study. The sample size is calculated by using Epi-info, Version 7 considering tuberculosis (TB) co-infection which gives highest sample among several factors, a hazard-ratio (HR) of 2.9 and the percentage of outcome among unexposed was 3.1% [8], 95% level of con dence and 80% power, the computed sample size was 577. After adding 15% for the incomplete medical records, the nal sample size required for this study is 665. In this study, three health centers were randomly selected using the lottery method from seven health center found at Kirkos sub city. Then, study participants were selected proportionally from each of the three health centers based on the number of patients on the follow-up using a simple random sampling method. A schematic explanation is presented below in the Fig. 1 Data collection method and tool The selected health centers ART registration and follow up formats were the source of data and data was collected using a checklist which is prepared in English language after reviewing national registration book content and different kinds of literatures [7], [9]- [11], and it is designed to incorporate information extracted from electronic and paper-based ART registration and follow-up forms of the ART clinic.

Study variables
The dependent variable was time to death and divided as death (event) and censored outcomes which measured in months. Censored means Patients who were alive at the end of follow up (the 30th of October 2019) or lost-to-follow-up (transferred out).
The predictors include demographic characteristics: (age, sex, educational level, marital status, religion); Clinical characteristics: (Active TB during ART, Baseline weight, WHO clinical stage, CD4 count and Baseline hemoglobin and baseline BMI; and Treatment) and follow up related factors such as: ART regimen, drug allergy, Functional status, ART regimen change, drug adherence and co-trimoxazole therapy.
Data processing and Analysis: Data entered using Epi-Data version 3.1 and analysis was conducted by SPSS version 20 and STATA version 14.1. Descriptive statistics was used to summarize the characteristics of the cohort. Kaplan-Meier (KM) estimates and Log-rank tests were used to compare survival curves among the variables with categories. Both bivariate and multivariable Cox-proportional hazard model was also employed after checking the assumption of proportionality using Global test using STATA and both crude and adjusted Hazard Ratios were estimated. The result was presented by using texts, tables, charts, and graphs.

Result
Social-demographic characteristics of the participants A total of 665 HIV infected adult patients were included in this study with and 52 were excluded due to incomplete medical record. Of these, 329 (53.7%) were females. The mean age was 38.65 (SD ± 10.32) years and more than half 320 (52.2%) of participants age was 35-50 years and the majority (96.7%) were urban residents. Regarding educational level, 195 (32.3%) attended primary (1-8) education, More than half (52.6%) of the patients were daily laborer and majority (50.8%) were single and majority (62.3%) were orthodox religious followers (Table 1). Concerning Co-trimoxazole therapy, more than half (57.1%) have not received Co-trimoxazole therapy ( Table 2).

Discussion
In this study, the mortality of patients enrolled on ART, and factors that predict mortality of patients under ART follow up was assessed. Among 613 adult HIV infected patients under follow up, 55(8.5%) were died giving an incidence death rate of 3.25 per 100 person-years, Similar nding was reported from a study done at Aksum hospital, northern Ethiopia in which overall mortality rate of 3.2 per 100 person-years [9].
And lower than studies done in Addis Ababa, Ethiopia, the mortality incidence was 3.8/100 person-years, 3.9 per 100mperson year in Harar, Ethiopia and another higher mortality rate of 10.3%; 5.4 deaths/100 person-years was reported from a study done at seven universities based on national ART follow-up data and 3.4 deaths per 100 person-years in Dilla [10], [12]- [14]. And present study nding is higher than study ndings done in southern Ethiopia 2.03 per 100 person-years [7], eastern Ethiopia 1.89 deaths per 100 person-years [15], and study done in Uganda [16]. The reason for such difference could be due to differences in study setting or difference due to the difference in the quality of care and service provision.
And in this study, the majority of (46%) of death occurred within 6 months of their follow up and 11(20%) died within the rst 3 months. A consistent nding was demonstrated in a hospital-based retrospective study in western Ethiopia. According to the study, most of the deaths (60%) were within the rst 6 months [15]. And another multi-site prospective cohort study done in Ethiopia also revealed that seventy percent of the deaths occurred within six months of starting ART [12]. But in some other studies, the majority of deaths occurring within three months of starting ART [12], [14], [17]. The reason could be the fact that usually HIV patients come to institution after advanced clinical stage and at early time, lack of drug adherence could contribute to high mortality.
Regarding predictors of mortality, age, active TB, WHO clinical staging, baseline functional status, cotrimoxazole therapy, baseline BMI, and, baseline CD4 + count were independent predictors of mortality in the adjusted Cox regression model.
In the present study, patients with age category above 50 years have a 4.89 times higher hazard of death as compared to 17-34 years. A consistent nding was reported from previous studies done in Uganda [16] and western Ethiopia [15]. The reason could be the fact that old age is associated with immunologic suppression, exposure to infectious diseases, psychosocial comorbidities, and the other factors of disease progression. But, it was not reported as a signi cant predictor in studies done in other parts of Ethiopia [18], [19]. This could be because of the predominantly young age of patients and study setting and sample difference.
A study done in Uganda revealed that being male increases the risk of mortality, as well as single marital status was related to a higher risk of HIV mortality [16]. And sex, educational status, and marital status of the Patients were reported as a signi cant predictor in studies done in Ethiopia [17], [19]. However, there was no signi cant difference in sex, educational status, and marital status of the participant in our study; which is in agreement with the study done at Fiche Hospital in North Shao, Oromia. education, sex, and marital status were not reported as signi cant predictors [18]. And a similar nding was reported from other studies done in Ethiopia, including a study done in Zewditu hospital in Addis Ababa [7], [10]- [12], [14], [20].
In the present study, it is demonstrated that Patients with TB comorbidity was highly associated with an increased risk of mortality and the hazard of death among patients with TB comorbidity was 3.46 times higher than without active tuberculosis. A consistent nding was reported from another similar investigation from Ethiopia, And the risk of death for patients who lived with tuberculosis was about 2.872-fold times higher than those patients who were negative also reported from a study done on two hospitals and six health centers at Illubabor and Buno Bedele zone [11]. and report from Uganda showed that mortality from TB co-infection was 1.81 times higher [16]. And based on a research report from eastern Ethiopia, TB co-infection at baseline or later was also associated with increased risk of mortality [17].A similar nding was reported from other studies [8], [18]. The reason is the fact that TB co-infection lowers the immunity of the patient.
Advanced WHO clinical stages (stage IV) have been consistently reported as risk factors for mortality in several studies from Ethiopia [12], [13], [17], [19] this study demonstrated similar nding thus, WHO stage IV were at increased risk of death as compared to their counterparts (stage I) and the hazard of death was 4.1 times higher. ART should be started (initiated) as early as possible, before advanced clinical stages. Mortality among patients with AIDS was nearly halved in the HAART era but remains approximately 5 times higher in patients with AIDS than in HIV-infected Patients without AIDS [6]. And this study demonstrated that drug adherence as a signi cant predictor of mortality. A study done in western Ethiopia revealed similar ndings in which non adherent participants had a mortality of 42.5 deaths per 100 person-years and were two times as likely to die as adherent participants [15]. Another similar study also reported consistent ndings (19). Some of the similar previous cohort studies demonstrated that higher CD4 cell count would reduce morbidity and mortality [17]- [19]. This study found that patients whose CD4 cell count < 200cells/ mm3 had a higher risk of death compared to patients with a CD4 cell count ≥ 500cells/mm3. The study showed that the CD4 cell count was an independent predictor of AIDS progression, and was also consistent with the other research results done outside of Ethiopia [21], [22] which indicated that AIDS progression to death was clustered among patients starting therapy with a lower CD4 cell count. Thus, emphasizing the need to early diagnose, link, and engage patient into the comprehensive ART care program.
Another important independent predictor of mortality identi ed in the current study was baseline BMI.
Patients with BMI < 18.5 kg/m 2 had greater hazard of death as compared to ≥ 18.5 kg/m 2 and a consistent nding was reported from a study in Somalia. Thus, the risk of death in patients with a BMI < 18.5 Kg/m2 was more than two times higher compared ≥ 18.5 kg/m 2 [17]. Also, another study conducted in Cameroon showed that BMI < 15 kg/m 2 had a three times higher risk of death than BMI > 18.5 kg/m. [23]. This could be due to malnutrition suppress the immunity and result of the aggregate effects of malnutrition-induced immune system dysfunction, a higher burden of opportunistic infections, metabolic derangement and anthropometric variations [24] In the present study, Adult HIV-infected patients who were bedridden at ART initiation had a higher risk of mortality compared to the patients with working functional status at treatment initiation. the nding is consistent with a study done in Eastern Ethiopia [17], southern Ethiopia [10], and those described elsewhere [19], [25], [26]. However, it was not reported as a signi cant predictor in a study done in western Ethiopia [15].
Co-trimoxazole is a feasible, well-tolerated, and inexpensive intervention for people living with HIV to reduce HIV-related morbidity and mortality [27]. However, the current study demonstrated not taking cotrimoxazole prophylaxis was signi cantly associated with mortality. Thus, patients who had not taken have 2.56 times higher risk of death compared to its counterpart. A consistent result was demonstrated from a study done in Harar, Ethiopia. Thus, not taking cotrimoxazole Prophylaxis Treatment (CPT) at the baseline has a higher risk of death [13]. The nding was also in line with the study done in Uganda [16] and other parts of Ethiopia [7].

Limitation of the study
Even though incomplete and missing data was minimized by cross checking electronic data base and patients medical record, due to retrospective nature of the data, selection bias is possibly introduced because patients with incomplete records of variables were excluded. And for those patients lost to follow-up, mortality results might be an underestimation because the status of the patient's status cannot be known.

Conclusion
This study showed that the incidence of death was 3.25/100 person-years with the majority of deaths occurring within six months of ART initiation. Ethics approval and consent to participate Ethical clearance was obtained from the Ethical Review Committee of Debre Berhan University.
Permission letter was also obtained from the selected Health Centers. All collected data were kept