Background The combination of combined active antiretroviral therapy (cART) with chemotherapy in the treatment of lymphoma in human immunodeficiency virus (HIV)-positive patients has improved the overall survival of these patients. However, drug–drug interactions between antineoplastic agents and the antiretroviral agents non-nucleoside reverse transcriptase inhibitors (NNRTIs) and protease inhibitors (PIs) can occur by influencing the activity of the CYP3A4 enzyme. So far, little is known about the clinical relevance of this interaction: the effect on the efficacy and toxicity of the chemotherapy. Also, there is no general consensus which cART is preferable in combination with antineoplastic drugs. Objective To compare PI-based with NNRTI-based cART on the efficacy and toxicity of chemotherapy in lymphoma patients. Setting The Onze Lieve Vrouwe Gasthuis, located in Amsterdam, The Netherlands. Method A retrospective observational cohort study including all patients with HIV and lymphoma over a 10-year period. Clinical outcome (response to chemotherapy and survival) and toxicity of chemotherapy (renal, hepatic and bone marrow toxicity as well as dose reduction, treatment delay and discontinuation) was compared in patients with PI based and NNRTI-based cART. Main outcome measure: Response to chemotherapy and survival. Results Patients using PI-based cART (n = 22) had a significantly lower 1 year survival compared to NNRTI-based cART (n = 21). No significant differences were observed in reaching complete remission after chemotherapy. No overall significant differences in toxicity and discontinuation of the chemotherapy were observed. However, there was a trend towards more severe bone-marrow toxicity in patients with PI-based cART. In addition, patients with PI-based cART received earlier dose-reduction and treatment delay, indicating increased toxicity in PI-treated patients. Conclusion This retrospective study shows that PI-based cART is inferior in combination with chemotherapy to NNRTI-based cART: a lower 1 year survival is observed and dose-reduction and treatment delay occur earlier, possibly based on an earlier onset of toxicity.