Background:  The protease inhibitor (PI) darunavir, boosted by ritonavir (DRV/r), is virologically effective and well tolerated in adults. Although co-formulated lopinavir (LPV/r) is currently the first-line PI for HIV-1 infected children in the UK, DRV/r has potential utility as once daily first- or second-line PI after previous PI failure. Identifying the prevalence of DRV/r resistance associated mutations (RAM) in children is important for determining clinical utility.

Methods:  Data during 2000 to 2007 combine that from the Collaborative HIV Pediatric Study (CHIPS, a cohort of ~95% reported HIV-1 infected children in UK/Ireland since 1996), and the UK HIV Drug Resistance Database. DRV/r RAM were identified from the 2008 IAS mutations list (V11I, V32I, L33F, I47V, I50V, I54L/M, T74P, L76V, I84V, L89V) with additional mutations (I47A, G73S/T/C, I84A/C, V82F) from the Stanford database. The prevalence of RAM was estimated in (i) PI-naive children, using the first PI-naive test, and (ii) PI-experienced children, using cumulative resistance at the last test on a PI. Associations between type and duration of PI exposure and area under the viremia curve since the start of PI with number of RAM was analyzed by multivariate Poisson regression. Susceptibility to DRV/r was defined using the Stanford algorithm.

Results:  In this study, 344 children were given a PI-naïve resistance test and 14/344 (3%) had a single RAM (2 V11I, 2 V32I, 1 I47A, 7 I50V, 1 G73S, 1 L89V); none had multiple RAM. In addition, 12 (86%) were non-B subtype virus. One-hundred fifty-six children had a resistance test on PI (55 (35%) prior LPV/r only, median (IQR) 2.6 (1.2 to 5.0) years on PI); 21(13%) had one RAM, 5 (3%) had 2, and 3 (2%) had 3. In a multivariate model, a higher number of DRV/r RAM was independently associated with increased time on PI (P =0.04), larger area under the viremia curve since the start of PI (P =0.01), and any exposure to a PI other than LPV/r (P =0.02 vs LPV/r only). However, only 3 (2%) PI-experienced children had intermediate level resistance to DRV/r using Stanford.

Conclusions:  PI-naïve children in the UK, and those whose only prior PI is LPV/r, have little DRV/r resistance, although the number of DRV/r RAM increases slightly with longer time and increased viremia on PI. Few PI-experienced children have more than one DRV/r RAM and susceptibility to DRV/r is high. This suggests that once daily DRV/r has utility both as a second PI, as well as an alternative first PI, particularly if it can be co-formulated with a PI booster.