Background:  The dual HIV and TB pandemics in Southern Africa have lead to a dramatic increase in co-infected children. Treatment of HIV-TB co-infection is complicated by metabolic drug interactions and overlapping toxicities, particularly in children, for whom treatment options are limited. Current South African guidelines advocate lopinavir/ritonavir-based HAART for children <3 years old, and Efavirenz-based HAART for older children. We hypothesized that interactions between lopinavir/ritonavir and anti-tuberculosis therapy may decrease viral suppression rates among co-infected children.

Methods:  Virologic suppression rates among 892 children who received HAART at 2 medical centers in KwaZulu Natal, South Africa were analyzed retrospectively and the impact of co-treatment for tuberculosis on viral suppression was assessed.

Results:  In this study, 324 children (36.3%) received simultaneous HIV and TB treatment. The overall rate of viral suppression (<400 copies/mL) was 86% at 6 months and 88% at 12 months. Among subjects who received concurrent treatment for HIV and TB, viral suppression rates were lower (81% at 6 months and 82% at 12 months) than among those without TB (88% and 90%; P =0.009 for both). When stratified by initial ARV regimen, subjects who received NNRTI-based HAART had similar rates of viral suppression whether they received concurrent TB therapy (87% and 85% at 6 and 12 months) or not (91% and 90%; P =0.199 and P =0.123, respectively). In contrast, children who received PI-based first line therapy had significantly lower viral suppression rates with TB co-treatment (70% and 76% at 6 and 12 months) than without (82% and 90%; P =0.026 and 0.019). Among children who received TB therapy while on PI-based HAART, rates of viral suppression at 6 months did not statistically differ whether the regimen contained ritonavir (n = 23; 57%), standard lopinavir/ritonavir (n = 51; 71%) or double-dosed lopinavir/ritonavir (n = 36; 72%).

Conclusions:  Concurrent treatment for TB is associated with lower rates of viral suppression among children on HAART, but this impact is limited to those receiving PI-based antiviral regimens (in South Africa, generally infants and children <3 years). This may be due to pharmacologic interactions (CYP450 induction), decreased adherence due to overlapping toxicities or high pill burden, or biologic interactions. Guidelines for the care of young HIV-TB coinfected infants should be continually evaluated, as PI-based ARV may not provide optimal viral suppression in this population.