Interpretation of Drug Interactions between Dolutegravir and Artemether-Lumefantrine or Artesunate-Amodiaquine

LETTER

Walimbwa and colleagues investigated drug-drug interactions between dolutegravir, an antiretroviral integrase strand transfer inhibitor, with artesunate-amodiaquine (ASAQ) and artemether-lumefantrine (AL), two commonly used artemisinin-based combination therapies for malaria (1). Their study is of public health importance given the rapid uptake of dolutegravir-based antiretroviral therapy as first-line treatment in settings where both HIV and malaria are endemic. They reported that coadministration of AL and ASAQ with dolutegravir significantly lowered the steady-state trough concentrations of dolutegravir, and that ASAQ but not AL also reduced the overall dolutegravir exposure.

We agree with their interpretation of a drug-drug interaction between dolutegravir and ASAQ and with their conclusion that the impact of ASAQ on dolutegravir trough concentrations would likely not result in a clinically significant effect, as antimalarial treatment is coadministered with dolutegravir over a short duration (3 days) and the dolutegravir trough concentrations were above 300 ng/ml, which is the accepted minimum concentration derived from a phase 2b study (2).

However, we feel that a drug interaction is an unlikely explanation for their observed 37% decrease in dolutegravir trough concentrations at 24 h (C₂₄) with coadministration of AL. The plot of dolutegravir mean concentrations over time, when administered alone and with AL (see Fig. 1 in their article), shows that dolutegravir concentrations at all time points were almost identical in the dolutegravir-alone and the dolutegravir-AL periods, except for higher mean C₂₄ in the dolutegravir-alone period. This higher mean C₂₄ in the dolutegravir-alone period was well above that of the mean predose dolutegravir concentration, unlike the dolutegravir-AL period in which predose concentrations and C₂₄ were very similar. In the Discussion in their article, the authors state that the observed difference in C₂₄ “may have been driven in part by an unexplained rise in dolutegravir C₂₄ in some participants.” The most likely explanation for these influential observations appears to be that some participants in the dolutegravir-alone period took the 24-h dose of dolutegravir before the C₂₄ sample without the knowledge of study personnel. Individual concentration-time plots would have been informative. The authors state that additional dose taken by participants was unlikely as participants were given clear instructions and exact numbers of pills, “which precluded such additional intake.” However, perfect adherence by participants to study protocol instructions does not always occur.

We encountered a similar problem of presumed intake of the 24-h dose before the C₂₄ sample by some participants in a study of linezolid pharmacokinetics with drug-resistant tuberculosis (3). We developed a rule to exclude C₂₄ data from participants if this exceeded the predose concentration and was >50% of the concentration of linezolid at the prior sampling point (6 or 8 h) based on the half-life of linezolid. We imputed the C₂₄ from the predose concentration in these participants with evidence of intake of linezolid before the C₂₄ sample. We suggest that the authors consider this approach as a way of handling biologically implausible C₂₄ values.