Severely delayed elimination of methotrexate (MTX) is difficult to predict in patients treated with high-dose MTX (HD-MTX), but it may cause life-threatening toxicity. It has not been defined how an increase in plasma creatinine can be best used as a predictor for severely delayed MTX elimination, thus providing a guide for therapeutic interventions to minimize renal toxicity. Pharmacokinetic data were retrospectively collected on 218 Danish children with acute lymphoblastic leukemia treated with HD-MTX 5 or 8 g/m2 on the NOPHO2000 protocol. Moderately delayed MTX elimination was defined as 42-hour plasma MTX ≥ 4.0-9.9 μM, and severely delayed elimination was defined as 42-hour plasma MTX ≥ 10 μM. Median 42-hour plasma MTX was 0.61 μM (interquartile range, 0.4-1.06 μM). Of 1295 MTX infusions with 5 g/m2 (n = 140 patients) or 8 g/m2 (n = 78 patients), 5.1% were severely (1.5%) or moderately (3.6%) delayed. The risk of having delayed elimination was highest in the first of eight infusions with MTX 5 g/m² (7.4% vs 0.0 to 4.1% for subsequent MTX infusions) (P < 0.02). A 25 μM increase or a 1.5-fold increase in plasma creatinine within 36 hours from start of the MTX infusion had a sensitivity of 92% (95% CI, 82%-97%) and a specificity of 85% (95% CI, 83%-87%) for predicting 42-hour MTX ≥4.0 μM. A 25 μM increase or a 1.5-fold in plasma creatinine within 36 hours after start of an HD-MTX infusion can predict delayed MTX elimination, thus allowing intensification of hydration and alkalization to avoid further renal toxicity and promote the elimination of MTX.