OBJECTIVE: To identify patients at high risk for major toxicity after theophylline intoxication who might benefit from early charcoal hemoperfusion. DESIGN: A 67-month prospective study. SETTING: Massachusetts Poison Control System. PATIENTS: 249 consecutive patients referred after theophylline intoxication (defined by a peak serum theophylline concentration > or = 167 mumol/L [30 mg/L]). INTERVENTIONS: Uniform, protocol-directed management recommendations. MAIN OUTCOME MEASURES: Identification of risk factors for major toxicity. RESULTS: 119 patients (48%) not receiving theophylline therapy had acute intoxication; among those receiving such therapy, 92 (37%) had theophylline intoxication because of chronic overmedication and 38 (15%) had acute intoxication. Major toxicity developed in 62 patients (25%); 13 patients (5%) died. Major toxicity was more common in patients with intoxication due to chronic overmedication than in those with acute intoxication who were not receiving theophylline therapy (49% compared with 10%, risk ratio, 4.85; 95% CI, 2.96 to 7.94), even though the former group had lower peak serum theophylline concentrations (283 mumol/L compared with 777 mumol/L, P = 0.001). Logistic regression analysis identified two major factors associated with the development of major toxicity: 1) peak serum theophylline concentrations in cases of acute intoxication and 2) patient age in cases of chronic overmedication. Receiver-operating characteristic curve analysis indicated that major toxicity occurred in patients with a peak serum theophylline concentration of greater than 555 mumol/L (100 mg/L) after acute intoxication and in patients older than 60 years (regardless of peak serum theophylline concentration) after chronic overmedication. CONCLUSIONS: Predictors for major toxicity after theophylline intoxication differ by type of overdose.