First, as our cohort was selected retrospectively,

it was not completely homogeneous in terms of antiretroviral experience and duration of ATV-based therapy. Secondly, as in clinical practice TDM is requested on the basis of the judgement of individual clinicians, criteria for its application may be heterogeneous and this could have FXR agonist introduced potential biases. Thirdly, most patients showed an undetectable baseline viral load, so the threshold we identified may primarily be applicable to patients on stable antiretroviral therapy to reduce the risk of virological rebound or to patients with undetectable viral load switching to ATV-based regimens during treatment simplification (e.g. for reasons of toxicity, reduction of pill burden, or simplification to once-daily regimens). ATV plasma C12 h appeared to be weakly correlated with unconjugated bilirubin level. This finding highlights the point that factors other than drug concentration, such as genetic predisposition, contribute to the extent of bilirubin elevation [13]. Genetic variability could be one of the explanations of our inability to identify a toxicity cut-off in the studied population. We found high inter-individual variability in ATV concentration in clinical practice and investigated several factors

that could explain this, focusing particularly on drug interactions. As expected, ATV plasma concentration was higher in patients receiving boosted ATV regimens and lower in those concomitantly taking acid-reducing agents. ATV is usually recommended with ritonavir boosting [14,15]. However, when boosted check details with ritonavir, ATV shows

a higher risk of hyperbilirubinaemia, gastrointestinal intolerance and dyslipidaemia [16]. In such cases, TDM could be used to determine whether switching to an unboosted ATV regimen could be an option to manage toxicity without exposing the patient to suboptimal drug levels. As ATV requires an acid gastric pH for dissolution and absorption, coadministration of acid-reducing agents (antacids, proton pump inhibitors and H2-receptor antagonists) should be limited to selected agents and staggered, as some subjects could develop Carbohydrate subtherapeutic drug levels: in these cases TDM could be used to determine whether the potential drug–drug interaction was clinically relevant in the individual patient. Overall, we did not observe different ATV plasma levels in subjects for whom tenofovir was part of the combination regimen. However, patients receiving tenofovir were more frequently administered boosted ATV (as currently recommended) and this counterbalanced the potential interaction. Indeed, this was confirmed by the finding that, in the subgroup of patients receiving unboosted ATV, concomitant tenofovir use was associated with lower ATV plasma levels.