Similar to other studies from Africa and other resource limited settings, HIV infected children in this study experienced a significant increase in height and weight after initiation of NNRTI based HAART [18–24]. The younger VF/IS group experienced a robust growth response despite failing to completely suppress virus on therapy. Our findings are similar to those of Ghaffari et al who found that the children on HAART with immunological treatment success, and with or without complete viral suppression had similar increases in weight and height z scores at 48 and 96 weeks post therapy . The authors postulate that although the virus may still replicate post therapy, its effect on growth may be less deleterious. Our study findings concur with the above observation and suggests that complete viral suppression may not be a requirement for the initial growth response in children on HAART  as long as there is a robust immune response. Children have high lymphoproliferative responses compared to adults which may help explain these findings. In contrast, Verweel et al found that the children on HAART who were viral responders had significant increases in height and weight while the non responders did not . Guillen et al also found that children with complete viral suppression had more significant increases in height and weight when compared to children with inadequate viral suppression . The differences noted in these two studies could be related to their small sample size, older age of the children, longer follow up period, or due to most of their study children being on second line therapy at the time of evaluation. In comparison, our cohort were all ART naïve, had a high rate of adherence to the antiretroviral fixed dose combination and the VF/IF group also had an initial modest response to therapy. However, other studies have also documented the significant increase in growth velocity while on HAART, without complete viral suppression [2, 10, 26]. Similar to our findings these studies also found no significant change in BMI on therapy and suggest that unlike adults children on therapy increase in both height and weight and therefore maintain the same BMI.
Older infected children in resource limited settings are usually long term survivors with severe immunosuppression and are less likely to have a rapid increase in height on HAART . In our study, the older VS/IF group had the poorest growth response to HAART suggesting that immune reconstitution may be more critical than complete viral suppression for initial growth response in children . When compared to the other treatment outcome groups in the study, these children had a higher baseline viral load, were more severely immunosuppressed and wasted and their growth response to therapy was modest. A longer follow-up may have provided more information on their potential for catch-up growth. However, considering findings from other studies with longer follow up, increases in height are less pronounced when HAART is initiated later [26, 28]. Therefore, to ensure that HIV infected children reach their full height potential, HAART needs to be initiated early before irreversible stunting occurs. This requires early identification of HIV infected children and could be achieved through scale up of national early infant HIV diagnosis (EID) programs using dried blood spots (DBS). From our data, children who were not showing a significant increase in height or weight were also failing immunologically. Another study found that height growth velocity was a predictor of survival independent of viral load, age and CD4 cell count in HIV infected children on HAART . In our cohort weight and height velocity over the 48 weeks did not predict virologic or immunologic treatment outcome. This may be related to the initial drop in viral load and modest increase in CD4 cell percent in most children leading to an initial growth response.
The children who initiated HAART with a low CD4 cell percent, regardless of age were less likely to have a robust increase in CD4 cells. It seems these children, in fact have lost their potential for 'good enough' immunological reconstitution and therefore they do not reach our threshold for 'immunological success' within the first 48 weeks on therapy. These findings are consistent with other studies where delay in initiation of HAART is more likely to lead to treatment failure in children and adults [19, 20, 28]. Our study further emphasize the importance of initiating HAART early in children so as to suppress virus, restore immune function and enable satisfactory somatic growth. It is critical to ensure that these initial growth responses in children from sub-Saharan Africa are sustained beyond the early years of therapy. Longer term follow up of children on HAART from Thailand, report younger age, higher baseline CD4 cell % and sustained viral suppression after 24 weeks as predictors of immune recovery at week 96 . In addition, a study from Cote d'Ivoire reported improved survival and sustained viral suppression with good immune recovery in children on HAART for a median follow up of 36 months .
Our documentation of early growth responses on HAART, despite inadequate viral suppression in children from a resource limited setting is encouraging. However, the continued replication of virus on HAART leads to selection of resistant mutations which may limit future treatment options [20, 31, 32]. Children in resource limited settings have even more limited treatment options and remaining on a first line regimen with adequate immune recovery and adequate growth response may be more important than switching to second line therapy to ensure complete viral suppression . Therefore it is important to develop strategies that ensure good adherence to the first line regimen and close clinical monitoring may be the better option for scaling up ART and improving survival of HIV infected children from resource limited settings. In the larger Ugandan population similar robust growth responses may not occur in all children initiating HAART because of the varying infectious diseases and socio-economic factors in the home.
There were certain limitations to our study including the small number of children in the failure groups and the lack of growth data for HIV negative control children which may have affected the results. Also, the follow-up duration was 48 weeks so we were unable to assess the long term impact of HAART on growth, despite inadequate viral suppression. However, the strength of this study is that it was a prospective cohort with longitudinal clinical and laboratory data collected consistently with no child lost to follow up, except for those who died. In addition, the children had high adherence levels to HAART and the majority experienced viral and immune treatment success with associated improvement in growth. Larger studies with longer follow up on therapy would be beneficial in documenting the sustained growth response of infected children and guiding future treatment strategies for children from resource limited settings.