Predictors of mortality among HIV infected children on anti-retroviral therapy in Mekelle Hospital, Northern Ethiopia: a retrospective cohort study

The introduction of antiretroviral viral therapy (ART) in 1996 has improved survival and quality of HIV-infected patients in the developed countries. This survival benefit of highly active antiretroviral therapy (HAART) in reducing HIV related deaths has been well studied in the developed world. In resource-poor settings however, where such treatment was started recently, knowledge on treatment results is limited. Moreover, mortality has been high in the first month of ART initiation. But factors contributing to high mortality are poorly investigated [1–3]. About 1,400 children under the age of 15 are infected with HIV every day and approximately 90% of these infections occur in sub-Saharan Africa. Without appropriate care and treatment, more than 50% of newly infected children die before celebrating their second birthday [4–6].

The use of ART has dramatically reduced HIV-associated mortality and morbidity among children in resource-rich settings. Pediatric HIV has been transformed from rapidly fatal infection of early childhood to a manageable chronic disease. For this reason, many HIV infected children are aging into adolescence and early adulthood [3–5]. But according to American Academy of Pediatrics, there might be differences from country to country in many ways such as nutritional status, racial or ethnic and gender differences in total lymphocyte count and CD4+ cell count [6].

Findings from studies in Africa and other low income countries also showed that there was good result of the ART programs i.e. the mortality of children on ART has decreased [7–10]. But in sub-Saharan Africa including Ethiopia, there is substandard information about treatment results and the impact of ART on mortality of children. The independent predictors of survival among children living with HIV and on ART remain poorly characterized in Ethiopia. Therefore, the available evidence in Ethiopia regarding the determinants of survival among children less than 15 years old is inadequate. For this reason, this study dares to investigate the predictors of survival among under 15 year old children on ART.

The findings of this study indicated that from the registered cohort, there were 20 deaths in 1186.25 person-years of retrospective follow up, providing an incidence density of 16.85 deaths per 1000 child-years of death. About 23(5.53%) patients were lost to follow up. Mortality was higher in the first 12 months after ART initiation (2.39 death per 1000 child months) But the over all mortality was 1.40 deaths per 1000 child-months. The over all mortality rate was actually lower when compared with two different studies done in Kenya which reported an overall mortality of the cohort of children on ART as 47 and 84 deaths per 1000 child years [12, 13].

Early mortality (death <18 months after ART initiation) was higher than late mortality (death >= 18 months after ART initiation) in this study. This finding is consistent with a study done in USA and ten European countries and South Africa which showed that most of the deaths occurred in the first 6 months following ART initiation [14–16]. Children in WHO clinical stage IV and with severe immunodeficiency had a significant risk of early death at 3 or 6 months [17–19]. This indicated that early HIV diagnosis and early antiretroviral therapy might not be well adopted as recommend which could have reduced early infant mortality by 76% in this study [20, 21].

Mortality rate among <= 18 months old children on ART was 56.80 deaths per 1000 child years which was higher as compared with the age 120 months or above i.e. 16.06 deaths per 1000 child-years. This study was in line with a study done in Zambia that showed, as age increases mortality rate will decreased [22]. Moreover, a study done in Malawi ,South Africa and Democratic Republic Congo also revealed that younger age (<18 months) and late WHO stage were more likely to die [23–25]. However, this finding contrasts with studies conducted in Uganda and Zambia that showed children initiated ART at an older age were associated with higher mortality [26, 27].

Children with hemoglobin level < 8 g/l were four times more likely to die than those with ≥ 8 g/dl. When results of this study was compared against a 4 years clinical cohort study in Kenya and Zambia among <15 years aged children, children with baseline hemoglobin level of ≤ 8 g/l was two times more likely to die than children with hemoglobin level >8 gm/dl [13, 22].

Children who initiated ART at WHO clinical stage (III/IV) were associated with higher mortality. Studies done in South Africa pointed out that, advanced disease (WHO stage III&IV) had an increased risk of mortality [23, 28]. Moreover, studies from Cote d’Ivoire, Malawi, and Ethiopia stated that children with advanced HIV disease and very low CD4% have been found to be at highest risk of early mortality following HAART initiation [18, 29, 30].

Hemoglobin level ≤ 8 gm/dl, CD4 count percent ≤50/ul was found to be significant predictors. This is consistent with studies conducted in Addis Ababa, Ethiopia at Zewditu memorial hospital and All African Leprosy and Rehabilitation Centre (ALERT hospital), showed that severe wasting, severe immunodeficiency (absolute CD4) and low hemoglobin value and, lower age and immunosuppression (low CD4 count) were associated with mortality [31, 32].

The sensitivity analysis was performed including the composite endpoint of lost to follow-up (LTFU) and death. The proportion of the lost to follow-up included as being dead were 10.34% (3.02 per 1000 child-months) almost doubled the mortality rate. Age less than 18 months, hemoglobin level, CD4 < 10%, were significantly associated with death. Similar study conducted in South Africa displayed that age less than 18 months, CD4 < 10% ,weight for age Z-score, and female gender were significantly associated with mortality [16]. A study from ten treatment programs in 4 countries (South Africa, Malawi, Mozambique, and Zimbabwe) showed that ignoring LTFU led to substantial underestimation of mortality [33]. Finally data on conrimoxazol prophylaxis was not collected in this study because it was started 3 years after the start date of this study. But a study done among Zambian children on ART reported that Daily cotrimoxazole prophylaxis has been shown to substantially reduce non–Pneumocystis jiroveci–related deaths and hospital admissions in children after infancy, and it is a recommended standard of care for all HIV-infected children. For this reason, this study recommends a further study on added impact of cotrimoxazole prophylaxis among children who are receiving ART [34].

The main weakness of the study is the composition of the study participants with almost half of the children being 5 years or more. As the mortality of HIV infected children is higher in lower age groups this may underestimate the estimates because many younger patients with poor prognosis were probably not included as they died early or were not linked to the program. Finally retrospective cohort study design also limits our ability to gather data about factors that may influence the risk of mortality, for example factors such as lack of family supports, cares given at family level etc.

In conclusion, findings of this study indicated that mortality in this cohort was found to be 4.8% deaths or 1.40 per 1000 child-months on ART. Finally, age less than 18 months, WHO clinical stage III & IV, chronic diarrhea and hemoglobin < 8 g/dl at baseline remained independent predictors of mortality. The high early mortality in the present study would support the value of an earlier treatment initiation before development of signs of immunodeficiency. Generally, it is sensible to give due emphasis about comprehensive early HIV treatment, care and support for all children on ART to improve survival of children on ART. Finally further study is needed about the effect of cotrimoxazol prophylaxis on mortality of children on ART.