SETTING:Reducing pain from intramuscular injection of kanamycin (KM) could improve the tolerability of multidrug-resistant tuberculosis (MDR-TB) treatment. Lidocaine has been shown to be an effective anaesthetic diluent for some intramuscular injections, but has not been investigated with KM in the treatment of adult patients with MDR-TB.OBJECTIVE AND DESIGN:We performed a randomised single-blinded crossover study to determine if lidocaine reduces KM injection-site pain. We recruited patients aged ⩾18 years on MDR-TB treatment at two TB hospitals in Cape Town, South Africa. KM pharmacokinetic parameters and a validated numeric pain scale were used at intervals over 10 h following the injection of KM with and without lidocaine on two separate occasions.RESULTS:Twenty participants completed the study: 11 were males, the median age was 36 years, 11 were HIV-infected, and the median body mass index was 17.5 kg/m2. The highest pain scores occurred early, and the median pain score was 0 by 30 min. The use of lidocaine with KM significantly reduced pain at the time of injection and 15 min post-dose. On multiple regression analysis, lidocaine halved pain scores (adjusted OR 0.5, 95%CI 0.3–0.9). The area under the curve at 0–10 h of KM with and without lidocaine was respectively 147.7 and 143.6 μg·h/ml.CONCLUSION:Lidocaine significantly reduces early injection-site pain and has no effect on KM pharmacokinetics.