India has witnessed a decline in mortality among people (including children) living with human immunodeficiency virus (HIV) who are on antiretroviral therapy (ART), during 2007-2011 [1]. This decline in mortality is argued to be attributed to the scale-up of ART in the country [2]. Globally, half the Children living with HIV (CLHIV) die of Acquired immune deficiency syndrome (AIDS) before their second birthday, and one-third during infancy in the absence of ART [3]. On the other hand, ART coverage for CLHIV had been scaled up significantly across the country during 2007 to 2013 [2]. Studies that examine the mortality/survival status of children who are on ART are scarce in India.

Data were analyzed using the Kaplan Meier survival and Cox proportional hazard model to measure survival and identify independent predictors of mortality of CLHIVs on ART. Cox proportional hazard ratios (HR) and adjusted hazard ratios (AHR) with 95% confidence intervals were used to assess the effect of baseline predictors on the survival of children on ART. Key variables used in the analysis included: gender, educational status (no schooling, attended school, information missing), CD4 count at ART initiation (£ 200, 201-250, 251-350, ³350), age at registration, age at start of ART (0-4 years, 5-9 years, 10-14 years), follow-up at most recent visit, CD4 count at ART initiation, WHO clinical stage (I, II, III, IV) and functional status (ambulatory, bed ridden and functional).

Of the 5874 children who were included in the analyses, 4461 (76%) children were alive, 336 (6%) died, and 760 (13%) were transferred out of the facility over the study period (Fig. 1). Table I depicts the characteristics of child-ren included in the study. The median (IQR) age at the start of ART was 8 (5,11) years. The median (IQR) CD4 count at the start of ART was 244 (153, 398)/mm3. The age-specific mortality rate among children on ART was highest among children younger than 5 years (Table II).

Fig. 1 Flow diagram of distribution of HIV +ve children included in the study.

TABLE I	Characteristics of Children  on Antiretroviral Therapy (N=5874) 

TABLE II  Age-specific Mortality Rates Among Children in Present Study

Children with CD4 count £200 were six times more likely to die (Adjusted Hazard Risk Ratio (AHR): 6.3; 95% CI 3.5-11.4) as compared to those with a higher CD4 count (³350) (Table III). Children who were in HIV clinical stage IV were three times more likely to die (AHR 3.2, 95% CI 1.9-5.3) as compared to those in the clinical stage I and II. Bedridden children were 4.5 times more likely to die (AHR 4.5; 95% CI 2.2-9.4) as compared to the children who were functional (Table III). There was no interaction effect between the child’s age and CD4 count. Baseline CD4 count had independent risk and the incidence was around 90, which is 3 times higher than the overall incidence density of 30. In the category with CD4 count 201-250, age (under 5 years) was a significant predictor, with mortality of 100 per 1000 person years, which was more than 3 times higher than the overall mortality of 30 per 1000 person years and also two and half times that (HR 2.621, 95% CI 1.05-6.52) observed in the 10-14 years age-group (P<0.05).

TABLE III  Factors Associated with Mortality Among Children on ART (N=5874)

Out of those children who died, 37% died within a month of starting ART, and 74% by six months. The overall estimated cumulative survival probability was 0.948 (95% CI 0.94-0.95) after 12 months, and 0.911 (95% CI 0.89-0.92) after 60 months. In the younger (<5 years) age group, cumulative survival probability at 12 months was 0.929 (95% CI 0.91-0.94) and 0.897 (95% CI 0.86-0.92) after 60 months. In the older age group (10 years and above), the cumulative survival probability after 12 months was 0.949 (95% CI 0.93-0.95) and after 60 months was 0.889 (95% CI 0.85-0.91) (Fig. 2).

Fig. 2 Survival curves of HIV-positive children by age.

Our findings indicate high levels of mortality among CLHIV. The incidence mortality rates are marginally higher amongst the youngest children (less than five years) than those among the older children. The corresponding probability of survival among children living with HIV is 95% after 12 months follow-up and 91% after 60-month from the start of their ART initiation. Results further suggest that CD4 count at ART initiation, WHO clinical staging and functional status of children at the time of baseline seems to independently determine the subsequent survival status of the children.

These results are comparable with earlier research from Gujarat (India) that suggested similar (but lower) survival probability of 86% after 36 months from the start of ART among children [10]. This suggests an improvement in child survival over time in India. The child mortality (30 per 1000 child years) in the current study found to be lower when compared to two different cohort studies conducted in Kenya which reported an overall mortality as 47 and 84 deaths per 1000 child years) [6,7]. The timing of deaths after start of ARI in this study is consistent with study results from northwest Ethiopia which showed that majority (90%) of the deaths occurred within the first year of treatment, and almost 50% within the first month [3,4,8].

The present study findings are important given the paucity of literature documenting the survival status of children who are on ART in India. However, the results must be considered in light of certain limitations. First, this is a secondary data analyses and is limited in terms of the number of variables or information available in the data base leading to several unexplained variations in the results. For example, the lack of key clinical data around weight, height, hemoglobin and body mass index (BMI) of children, and demographic information such as parental survival status, economic status limits our ability to interpret the data. Due to the retrospective nature of this study, several critical variable were missing for several children whose data could not be included. Since, the analyses was carried out taking a cohort of CLHIV who had been initiated on ART over a period six years, there may have been variations in survival status of children with the maturity of ART era. Future research could examine the survival status of children who are on ART at various time points during the ART era and also understand the reasons behind high mortality among specific sub-groups of children.

Despite these limitations, the present results document the mortality rates among a large number of children who have initiated ART during the early stage of ART introduction in India. The findings show that children less than five years, those with CD4 count less than or equal to 200/mm3 at the start of ART, and those in the clinical stage III or IV are more vulnerable to mortality. Thus, greater emphasis is needed to improve early HIV diagnosis and treatment in the young age group. Also, large-scale and long-term research is needed to confirm some of the current findings and also to ascertain the reasons for high mortality among children less than five years.