Tolerability of integrase inhibitors in a real-life setting

Abstract

Background: Integrase inhibitors have shown better tolerability than other drugs in clinical trials, but some postmarketing data have suggested potential differences among them. Aims: We compared rates and reasons for discontinuation of raltegravir-, elvitegravir- and dolutegravir-based regimens in a large cohort of HIV-infected patients. Methods: Retrospective analysis of a prospectively followed cohort including all antiretroviral-naive and all virologically suppressed antiretroviral-experienced patients prescribed a first regimen containing raltegravir, elvitegravir or dolutegravir with at least one follow-up visit. Major outcomes were early discontinuation (≤1 year) due to any reason and more specifically due to toxicity. Incidence was calculated as number of episodes per 1000 person-years. Risk factors for discontinuation were assessed by multivariate Cox models. Results: Early discontinuations due to any reason were 271 (raltegravir), 168 (elvitegravir) and 264 (dolutegravir) per 1000 patient-years (P=0.0821). Early discontinuations due to toxicity were 76 (raltegravir), 103 (elvitegravir) and 81 (dolutegravir) per 1000 patient-years (P=0.6792). Overall, the most common toxicities leading to discontinuation were neuropsychiatric, osteomuscular or digestive. Most frequent neuropsychiatric manifestations reported at discontinuation were insomnia, dizziness, headache and anxiety irrespective of the integrase inhibitor. Among discontinuations due to toxicity, neuropsychiatric effects were more common with dolutegravir than with raltegravir or elvitegravir (P=0.0046). Age (HR 1.04, 95% CI 1.02-1.07, P=0.0007) was the only independent risk factor for early discontinuation due to toxicity. Conclusions: Discontinuations due to any reason tended to be less common with elvitegravir, but discontinuations due to toxicity did not differ among integrase inhibitors. Neuropsychiatric toxicity leading to drug discontinuation was more frequent with dolutegravir.