Long-term survival outcomes of HIV infected children receiving antiretroviral therapy: an observational study from Zambia (2003–2015)

The availability of Antiretroviral Therapy (ART) for children living with HIV and implementation of universal treatment of all pregnant and breastfeeding women living with HIV (Option B+) is a game changer in the global fight against HIV [1]. In 2017, about 1.8 million children were living with HIV globally and 90% of these children lived in sub-Saharan Africa [2, 3]. Although progress has been reported in the scale-up of access to treatment for children, only 52% of children living with HIV received lifesaving ART and only 51% of HIV-exposed infants were tested for HIV by the age of 2 months as recommended by the World Health Organization (WHO) guidelines [2]. These estimates fall short of the UNAIDS 90–90-90 treatment targets, a strategy to end the global HIV epidemic. This strategy aims to achieve the following by 2020: 1) 90% of people living with HIV will know their status, 2) 90% of all diagnosed people will be on ART, 3) 90% of people on ART will be virally suppressed [4]. Although some milestones have been achieved in the provision of ART, access to early HIV diagnosis and ART among infants and children remains a challenge in high HIV burden settings [2, 5].

Similarly, the pediatric HIV program in Zambia has made tremendous progress with over 64% of children living with HIV accessing ART by the end of 2017 [6]. With a population of 17 million people and an estimated HIV prevalence of 12.9%, 72,000 were children living with HIV and 8900 were newly infected in 2017 [6, 7]. This scenario creates a large pool of children living with HIV in need of treatment. An early study done in routine care settings in Zambia demonstrated that children were diagnosed at older ages with advanced WHO stage 3 or 4 disease. The same study reported that 57% of deaths occurred within the first 90 days of treatment initiation and loss to follow-up was high [8]. Early mortality was consistently associated with lower CD4 count, younger age, low weight for height and anemia at ART initiation [8–10]. Recent studies suggest that routine care settings in resource limited settings are still confronted with critical gaps in early infant diagnosis and pediatric HIV treatment, resulting in children initiating treatment with advanced disease and at high risk of death during the first 3–6 months of treatment [11, 12]. In addition, there are concerns around interpretation of treatment outcomes in routine care settings because of the high loss to follow-up and high numbers of transfers with no mortality ascertainment [11, 13].

Although, pediatric HIV treatment programs in resource limited settings like Zambia have been treating large numbers of HIV-exposed and infected children for over 15 years, long-term survival and associated factors including barriers to long-term adherence and retention in care are not adequately elucidated and therefore the wider impact of these interventions is not clearly understood [14–16]. The aim of this research is to describe treatment outcomes, measure mortality rates and assess predictors of mortality among children receiving ART in a routine care setting over a 12-year period (2003–2015) at Livingstone Central Hospital in Southern Province, Zambia.

Using a retrospective cohort study design, we evaluated treatment outcomes among children living with HIV who received ART between January 2003 and June 2015.

In this study carried out in a routine pediatric HIV treatment setting, we found that after 12 years of treatment (2003–2015) attrition due to death was low (7%; n = 71). Death was highest within the first 3 months of starting ART (mortality rate: 11.7/100 PYO; 95% CI:7.6–16.3). The hazard of death was highest among children aged less than 12 months (aHR: 3.1: 95% CI: 1.3–6.4), compared to the referent group 6–15 years. Loss to follow-up was high (16%; n = 164).

The mortality rate observed in this study was much lower than that observed in other parts of Zambia [8, 23]. In a study conducted in an urban area in Zambia, the mortality rate was 6.6/100 PYO over 3018 PYO [8] and another study in a rural area in Zambia found a high mortality of 14.4% after 6 months of treatment [23]. These 2 studies were characterized by high mortality within the first 3 months (17.4/100 PYO) and the associated risk factors were similar to what we observed [8]. Mortality rates in our study population were also similar to a study from a rural setting in Malawi where 12 months mortality rate was 6.6/100 PYO [95% CI: 5.5–7.9] and overall mortality was 3.4/100 PYO [95% CI:2.9–4] after 5 years of follow-up [24]. Our observed mortality rate is however higher than findings from developed countries where 10-year mortality was as low as 0.3/100 PYO in the Dutch cohort [25].

The observation that hazards of death were highest within the first few months after ART initiation and then declines after the first 6 months of treatment is consistent with studies done in other parts of Zambia and in similar settings [8, 26–28]. Other studies have attributed the high early mortality to late presentation to care [29]. Studies done in routine clinic settings in resource limited settings found that the cumulative incidence of mortality during the first year of treatment among older children between 5 and 10 years of age was less than 2% to more than 45% among infants aged less than 12 months with severe disease [10]. Once children get past the initial 6 months of ART, their risk of mortality declines to very low rates especially when adherence to treatment and follow-up is optimized [11, 27, 30].

Our finding that the hazards of mortality were highest among infants and children aged less than 12 months of age is consistent with predictive models that have been done in both developed and developing countries and in fact motivated the universal ART policy by WHO to test and treat all HIV-exposed infants and children [31]. Results of a modelling study done in six countries in Sub-Sahara Africa suggested that mortality is higher among perinatally HIV-infected children than those infected through breastfeeding [32]. Perinatally HIV-infected infants have a higher risk of mortality and disease progression [5]. Early diagnosis of HIV infection and early initiation of ART improves the outcomes of infants and children [33]. Although early infant diagnosis of HIV is critical, there are still challenges in resource limited settings to diagnose children. Children who are missed by the PMTCT programs are at highest risk of mortality because they usually present to the hospital after an illness when they are already immunocompromised with high risk of mortality [34].

We assessed the causes of death in our study but only found death certificates for 27 children and no postmortem or thorough investigations had been done to ascertain cause of death. However, causes of death recorded on the few death certificates were similar to the causes of childhood deaths in Zambia, which are diarrhea, severe pneumonia and protein energy malnutrition [35]. This setting has a very high infant mortality rate (65/1000 in 2015) [36]. In the pre-ART era, AIDS related deaths led to an increase in infant mortality in high HIV burden countries. Antiretroviral therapy improved child survival and evidence is now showing that early initiation of ART leads to better clinical outcomes [33]. In a study that pooled results of clinical trial data from Zimbabwe and Uganda, the mortality risk was attributed to pre-ART risks that persist until the antiretroviral drugs reach their maximal effectiveness [27]. These early deaths have been attributed to suboptimal management of malnutrition, Tuberculosis and other related medical conditions during the early treatment stage including Immune Reconstitution Syndrome (IRIS) [11]. However, the role of IRIS in mortality during the early months of treating HIV seropositive children with ART is poorly understood [37]. Our study had insufficient evidence to support the role of IRIS as a risk factor of early mortality.

The major strength of our study is that it was done in a routine clinic setting among children commencing ART in a high HIV burden setting. This provides a real-world effectiveness of pediatric HIV treatment outcomes and provides information on important predictors of mortality. Our sample size was large and we had enough outcomes to make reasonable conclusions.

A major limitation is that there was a high proportion of loss to follow-up (16%) and many transfers (20%). This made ascertainment of mortality very challenging because some of the children who were lost to follow-up may be deceased and therefore misclassified and the mortality rates that we observed may be an underestimate. The estimated mortality rates under the “worst case scenario” assumption that all children who were lost to follow-up had died was 7 times higher after 1 year of follow-up and 5 times higher after 10 years of follow-up. A study from Malawi found that only 11% of the children who were lost to follow-up had died [13]. This study was done in an urban setting unlike our study setting which is both urban and rural. Our sensitivity analysis approach allowed us to estimate maximum mortality rates which was more informative in our setting.

Transfers are high in this clinic, which is motivated by Ministry of Health policy that encourages people to seek care at the health facility nearest to them. Studies have showed that this is an effective strategy as patients can additionally benefit from support groups and other programs that may be available at their nearest health facility [38, 39]. However, this makes it difficult to study long-term treatment outcomes because patient tracking is impossible once they transfer to another facility.

We observed low attrition due to mortality among children on ART in Zambia. Loss to follow-up was high (16%) and could underestimate the mortality. Mortality was highest during the first 3 months of treatment. Children aged less than one year and children with advanced WHO disease stage had higher mortality. We recommend to program managers and clinicians to develop effective interventions to improve retention in care and strengthen early infant diagnosis of HIV. Pretreatment screening and treatment of opportunistic infections among children commencing ART needs to be strengthened to reduce early mortality.