Therapeutic drug monitoring of raltegravir C_trough levels was carried out in the setting of the Raltegravir Switch for Toxicity or Adverse events (RASTA) trial, a randomized pilot study exploring a 48-week safety and efficacy of treatment switch to raltegravir associated with tenofovir/emtricitabine or abacavir/lamivudine in patients with regimens with optimal virologic control. Blood sampling for measurement of raltegravir plasma levels was carried out at weeks 4, 12, 24, 36 and 48. Plasma samples were analysed by a recently developed and validated UPLC-MS method. A total of 164 samples from 39 patients were assayed. Analysis for intra- and inter-subject variability was restricted to those patients with 4 or more determinations, including 30 patients and 142 determinations. The intra- and inter-subject variability measures were 85.9 and 124.6%, respectively, with an intra-/inter-subject variability ratio of 69%. We also analysed data from a subset of patients with well-documented adherence to protocol, defined as protocol compliant population, including 21 patients and 93 determinations. In this subpopulation, we estimated intra- and inter-subject variability of 79.87% and 110%, respectively, with an intra-/inter-subject variability ratio of 72.6%. This study confirms the notion that raltegravir is a highly variable drug according to the European Medicines Agency criteria. While this condition does not favour the adoption of therapeutic drug monitoring in the clinical practice, the latter is deemed useful in patients with drug plasma concentrations below or near the threshold level of efficacy (since intracellular raltegravir levels might be as low as 5% of the corresponding plasma levels), or to identify drug-drug interactions of potential clinical relevance.