Statistical examination to determine whether only 48-h value for serum concentration during high-dose methotrexate therapy is a predictor for clinical adverse events using ordered logistic regression analysis

Sustained elevation of serum methotrexate (MTX) concentrations (>1.0 μM) for 48 h (48-h value) has been found to have predictive significance for the development of toxicity. However, we sometimes experience severe adverse events during high-dose (HD) MTX therapy even if serum MTX concentrations comply with recommended values. We performed a retrospective study to identify predictors for occurrence of adverse events and examined whether only the 48-h value is a statistically significant predictor for clinical adverse events during HD MTX therapy. The subjects were 32 hematological patients (n = 58 episodes) treated with MTX at Kyoto Prefectural University of Medicine between February 2003 and July 2007. Ordered logistic regression analysis was used to identify predictors for occurrence of adverse events. The predictive factors identified were: 24-h continuous infusion therapy (24-h C-IV) (long infusion time) [odds ratio (OR) = 2.890, confidence interval (CI) =1.493–5.594; P = 0.0016] for fatigue, higher dose [OR = 2.282, CI = 1.287–4.046; P = 0.0048] and combination chemotherapy [OR = 2.177, CI = 1.059–4.477; P = 0.0344] for stomatitis, and 24-h C-IV [OR = 2.573, CI = 1.101–6.016; P = 0.0291] for neutropenia. We found that only the 48-h value was not a predictor for clinical adverse events for HD MTX therapy. A major limitation of the present study was the small number of participants. However, our findings suggest that there is evidence that a long infusion time is a significant predictor for general fatigue and neutropenia, while a higher dose and combination chemotherapy are predictors for stomatitis.