The primary goals of antiretroviral therapy are preventing HIV-related morbidity and mortality, and improving quality of life by restoring immunologic function through suppression of viral load [10]. Many studies have reported high early mortality among patients starting antiretroviral treatment in sub-Saharan Africa [4]. One of the reasons cited for this is that despite better availability of ART, people are still diagnosed late and thus start ART at later stages of the HIV disease [11].

Long-term retention of patients in antiretroviral treatment is a prerequisite for achieving any adherence at all. Various studies have shown that mortality during the first 6months after initiating ART is much higher than in developed countries and retention of patients in programs is poor [4,13]. However, most longitudinal studies conducted in Africa have been either short-term or have involved small numbers of participants.

In our study, the overall mortality rate of 2.03 deaths per 100 person years is lower or comparable to that reported elsewhere; however the mortality rate at the first year (55.67; 95% CI 43.5–71.3; or 4.3%) is high. In the ART-LINC Collaboration, which analyzed data from 18 cohorts across the developing world, mortality averaged 4.2% across all 18 cohorts in the first year after initiation [11]. In Yaounde, Cameroon, from 312 patients, the incidence rate of mortality rate was 21.2 per 100 person years (95% CI 15–31) [14]. In southern Ethiopia, the mortality rate was 15.4 per 100 person-years of observation [12]. The majority or 74.1% of the deaths occurred in the first year after treatment. The study highlights the high early mortality in ART cohorts in resource-limited settings that has been observed by other groups in similar settings [4,15,16].

The primary causes of death in AIDS patients could be the causes such immune reconstitution syndrome and opportunistic infections as a result of very weak immunity levels. In our study one of the factors associated with early mortality is late presentation for ART. This may also account for the high rate of death in the first year. According to reports, patients that start ART at WHO Stage III and IV are at an increased risk of dying [17,18]. Furthermore, early mortality risk is higher among those with low CD4 cell count [4,19]. The CD4 count is a proxy indicator of severity of disease which corresponds to the functional status and reflects the immune state of patients [20]. In our study, the functional status of patients at the entry level had a positive correlation with their disease stage and negative correlation with CD4 count. In a report from Hong Kong, there was a 79% reduction in mortality among ART taking patients with CD4 counts of less than 200/ml [21]. The majority of morbidity and mortality seen among individuals starting ART with low baseline CD4 cell count occurs during the first 3–6months on treatment [2,3,15,16,22]. In one study patients initiating ART with a base-line CD4 cell count of less than 50 cells/mm3 had a 3.2-fold higher mortality rate (p 0.004) compared with patients with a CD4 cell count between 51 and 200 cells/mm3 at the time of ART initiation [23]. Patients starting treatment with CD4 cell count below 100 cells/mm3 were at significantly greater risk of death during the follow-up period (OR 2.69; 95% CI 1.12–6.44) [12].

Nutritional and physical status can predict early mortality. In our study, 10% weight loss determines mortality during the course of treatment. In a district hospital in Ethiopia, weight loss was seen in about a third of patients who survived up to the fourth week, and it was associated with increased death [18]. Higher body weight at baseline was found to be associated with lower risk of mortality with a HR of 0.58 for weight groups of 40–50 kilograms compared with the reference of less than 40 kilograms, and HR of 0.25 for groups ≥60 kilograms compared with the reference groups (p<0.013). The two year patient survival was significantly related with co-trimoxazole initiation at or before treatment, clinical stage IV disease, working functional status and weight greater than 60 kilograms [24]. In another study, baseline body mass index (BMI) of less than 18.5 was independently associated with early mortality risk [19].

About 61% of patients in this study had education below secondary school. According to reports, low educational level among patients is a contributing factor to late presentation for ART [25]. This is understandable since the more educated a patient is the better their understanding of the disease state and comprehension of instructions given on drug usage. These could enhance treatment outcomes [26]. A report from British Columbia reported a protective effect of educational level on mortality from all causes among intravenous drug users receiving anti-retroviral treatment [27,28]. Most reports suggest that low educational level has consistently been associated with higher mortality, both overall and cause-specific [29,30]. In our study lower educational level was associated with a higher risk of mortality. However, patients with primary education seem to be at higher risk of death compared to those with no education. This is an interesting finding for which we could not find explanation due to the nature of our data, however it needs further investigation in future research.

This study has limitations. CD4 cell counts and HIV RNA measurements were not available for all patients because of cost issues. Diagnostic tests that would confirm the presence of certain opportunistic infections were limited; as a result we were not able to include these in our analysis. The retrospective cohort study design limited our ability to gather data about factors that may influence the risk of mortality, for instance factors such as lack of social supports networks, disclosure of infection status and depression. Data was collected from those who were attending ART centers mostly through self-reporting and hence may have reporting and recall bias. In ART programs of developing countries like Ethiopia, poor ascertainment of deaths and recording of information on losses to follow up may lead to underestimation of mortality rates.