Almost all patients with malignant mesothelioma eventually have disease progression after first-line therapy. Previous studies have investigated maintenance therapy, but none has shown a great effect. We aimed to assess the efficacy and safety of switch-maintenance gemcitabine in patients with malignant mesothelioma without disease progression after first-line chemotherapy. We did a randomised, open-label, phase 2 trial in 18 hospitals in the Netherlands (NVALT19). We recruited patients aged older than 18 years with unresectable malignant mesothelioma with no evidence of disease progression after at least four cycles of first-line chemotherapy (with platinum and pemetrexed), who had a WHO performance status of 0-2, adequate organ function, and measurable or evaluable disease. Exclusion criteria were active uncontrolled infection or severe cardiac dysfunction, serious disabling conditions, symptomatic CNS metastases, radiotherapy within 2 weeks before enrolment, and concomitant use of any other drugs under investigation. Patients were randomly assigned (1:1), using the minimisation method, to maintenance intravenous gemcitabine (1250 mg/m2 on days 1 and 8, in cycles of 21 days) plus supportive care, or to best supportive care alone, until disease progression, unacceptable toxicity, serious intercurrent illness, patient request for discontinuation, or need for any other anticancer agent, except for palliative radiotherapy. A CT scan of the thorax or abdomen (or both) and pulmonary function tests were done at baseline and repeated every 6 weeks. The primary outcome was progression-free survival in the intention-to-treat population. Safety was analysed in all participants who received one or more doses of the study drug or had at least one visit for supportive care. Recruitment is now closed; treatment and follow-up are ongoing. This study is registered with the Netherlands Trial Registry, NTR4132/NL3847. Between March 20, 2014, and Feb 27, 2019, 130 patients were enrolled and randomly assigned to gemcitabine plus supportive care (65 patients [50%]) or supportive care alone (65 patients [50%]). No patients were lost to follow-up; median follow-up was 36·5 months (95% CI 34·2 to not reached), and one patient in the supportive care group withdrew consent. Progression-free survival was significantly longer in the gemcitabine group (median 6·2 months [95% CI 4·6-8·7]) than in the supportive care group (3·2 months [2·8-4·1]; hazard ratio [HR] 0·48 [95% CI 0·33-0·71]; p=0·0002). The benefit was confirmed by masked independent central review (HR 0·49 [0·33-0·72]; p=0·0002). Grade 3-4 adverse events occurred in 33 (52%) of 64 patients in the gemcitabine group and in ten (16%) of 62 patients in the supportive care group. The most frequent adverse events were anaemia, neutropenia, fatigue or asthenia, pain, and infection in the gemcitabine group, and pain, infection, and cough or dyspnoea in the supportive care group. One patient (2%) in the gemcitabine group died, due to a treatment-related infection. Switch-maintenance gemcitabine, after first-line chemotherapy, significantly prolonged progression-free survival compared with best supportive care alone, among patients with malignant mesothelioma. This study confirms the activity of gemcitabine in treating malignant mesothelioma. Dutch Cancer Society (Koningin Wilhelmina Fonds voor de Nederlandse Kankerbestrijding) and Stichting NVALT studies.