One thousand two hundred twenty-two patients treated in the Rituximab with CHOP over age 60 years (RICOVER-60) trial were examined for central nervous system (CNS) disease developing during first-line therapy or after a complete or partial remission had been achieved. Patients received 6 or 8 courses of CHOP (cyclophosphamide, adriamycin, vincristine, prednisone) administered every 2 weeks (CHOP-14) with or without rituximab. CNS prophylaxis for patients with involvement of bone marrow, testes, upper neck, or head consisted of intrathecal (i.th.) methotrexate (days 1 and 5 of first 2 courses). Fifty-eight cases of lymphoma in the CNS were observed (36/609 patients in the CHOP-14 and 22/608 patients in the arituximab-CHOP-14 [R-CHOP-14] arm). The estimated 2-year incidence of CNS disease was 6.9% (confidence interval [CI] 4.5; 9.3) after CHOP-14 and 4.1% (CI 2.3; 5.9) after R-CHOP-14. R-CHOP reduced the relative risk for CNS disease to 0.58 (95% CI 0.3; 1.0, P = .046). Cox regression analysis identified "involvement of more than 1 extranodal site" and "B-symptoms" as significant risk factors for CNS disease. Patients treated with R-CHOP-14 did not show any benefit from i.th. methotrexate. We conclude that elderly patients with aggressive CD20-positive lymphoma show a significantly lower incidence of CNS disease if treated with R-CHOP-14 instead of CHOP-14. Intrathecal methotrexate has no role in preventing CNS disease for patients treated with combined immunochemotherapy (R-CHOP-14)--with the possible exception of patients with testicular involvement.