Cyclophosphamide, doxorubicin, vincristine, and prednisolone (CHOP) is used to treat patients with non-Hodgkin lymphoma. Interval decrease from 3 weeks of treatment (CHOP-21) to 2 weeks (CHOP-14), and addition of rituximab to CHOP-21 (R-CHOP-21) has been shown to improve outcome in elderly patients with diffuse large B-cell lymphoma (DLBCL). This randomised trial assessed whether six or eight cycles of R-CHOP-14 can improve outcome of these patients compared with six or eight cycles of CHOP-14. 1222 elderly patients (aged 61-80 years) were randomly assigned to six or eight cycles of CHOP-14 with or without rituximab. Radiotherapy was planned to sites of initial bulky disease with or without extranodal involvement. The primary endpoint was event-free survival; secondary endpoints were response, progression during treatment, progression-free survival, overall survival, and frequency of toxic effects. Analyses were done by intention to treat. The trial is registered on National Cancer Institute website, number NCT00052936 and as EU-20243. 3-year event-free survival was 47.2% after six cycles of CHOP-14 (95% CI 41.2-53.3), 53.0% (47.0-59.1) after eight cycles of CHOP-14, 66.5% (60.9-72.0) after six cycles of R-CHOP-14, and 63.1% (57.4-68.8) after eight cycles of R-CHOP-14. Compared with six cycles of CHOP-14, the improvement in 3-year event-free survival was 5.8% (-2.8-14.4) for eight cycles of CHOP-14, 19.3% (11.1-27.5) for six cycles of R-CHOP-14, and 15.9% (7.6-24.2) for eight cycles of R-CHOP-14. 3-year overall survival was 67.7% (62.0-73.5) for six cycles of CHOP-14, 66.0% (60.1-71.9) for eight cycles of CHOP-14, 78.1% (73.2-83.0) for six cycles of R-CHOP-14, and 72.5% (67.1-77.9) for eight cycles of R-CHOP-14. Compared with treatment with six cycles of CHOP-14, overall survival improved by -1.7% (-10.0-6.6) after eight cycles of CHOP-14, 10.4% (2.8-18.0) after six cycles of R-CHOP-14, and 4.8% (-3.1-12.7) after eight cycles of R-CHOP-14. In a multivariate analysis that used six cycles of CHOP-14 without rituximab as the reference, and adjusting for known prognostic factors, all three intensified regimens improved 3-year event-free survival (eight cycles of CHOP-14: RR [relative risk] 0.76 [0.60-0.95], p=0.0172; six cycles of R-CHOP-14: RR 0.51 [0.40-0.65], p<0.0001; eight cycles of R-CHOP-14: RR 0.54 [0.43-0.69], p<0.0001). Progression-free survival improved after six cycles of R-CHOP-14 (RR 0.50 [0.38-0.67], p<0.0001), and eight cycles of R-CHOP-14 (RR 0.59 [0.45-0.77], p=0.0001). Overall survival improved only after six cycles of R-CHOP-14 (RR 0.63 [0.46-0.85], p=0.0031). In patients with a partial response after four cycles of chemotherapy, eight cycles were not better than six cycles. Six cycles of R-CHOP-14 significantly improved event-free, progression-free, and overall survival over six cycles of CHOP-14 treatment. Response-adapted addition of chemotherapy beyond six cycles, though widely practiced, is not justified. Of the four regimens assessed in this study, six cycles of R-CHOP-14 is the preferred treatment for elderly patients, with which other approaches should be compared.