Abstract Prophylactic high-dose methotrexate (HD-MTX) is often used for diffuse large B-cell lymphoma (DLBCL) patients at high risk of central nervous system (CNS) relapse, despite limited evidence demonstrating efficacy or the optimal delivery method. We conducted a retrospective, international analysis of 1384 patients receiving HD-MTX CNS prophylaxis either intercalated (i-HD-MTX) (n = 749) or at the end (n = 635) of R-CHOP/R-CHOP-like therapy (EOT). There were 78 CNS relapses (3-year rate 5.7%), with no difference between i-HD-MTX and EOT: 5.7% vs 5.8%, P = .98; 3-year difference: 0.04% (−2.0% to 3.1%). Conclusions were unchanged on adjusting for baseline prognostic factors or on 6-month landmark analysis (n = 1253). In patients with a high CNS international prognostic index (n = 600), the 3-year CNS relapse rate was 9.1%, with no difference between i-HD-MTX and EOT. On multivariable analysis, increasing age and renal/adrenal involvement were the only independent risk factors for CNS relapse. Concurrent intrathecal prophylaxis was not associated with a reduction in CNS relapse. R-CHOP delays of ≥7 days were significantly increased with i-HD-MTX vs EOT, with 308 of 1573 (19.6%) i-HD-MTX treatments resulting in a delay to subsequent R-CHOP (median 8 days). Increased risk of delay occurred in older patients when delivery was later than day 10 in the R-CHOP cycle. In summary, we found no evidence that EOT delivery increases CNS relapse risk vs i-HD-MTX. Findings in high-risk subgroups were unchanged. Rates of CNS relapse in this HD-MTX-treated cohort were similar to comparable cohorts receiving infrequent CNS prophylaxis. If HD-MTX is still considered for certain high-risk patients, delivery could be deferred until R-CHOP completion.