Purpose: To compare patients treated in or outside clinical protocols, using de novo acute myeloid leukemia (AML) as a model disorder.
Patients and methods: We retrospectively compared the characteristics of all patients with de novo AML diagnosed in the referral area of our university hospital between 1985 and 1994.
Results: Of a total of 170 AML patients, 45% were included in a phase III trial for the treatment of AML and 55% were treated outside a protocol. Another 45 patients were registered only at diagnosis but were treated elsewhere. Nonstudy patients differed significantly from patients included in clinical trials with respect to age and performance status at clinical presentation, comorbidity, and type of AML. The great majority of patients excluded from trial participation showed distinct exclusion criteria, such as advanced age and severe comorbidity. Study patients were treated significantly more often with curative intent and achieved better response and survival. Patients treated in an equivalent manner but outside a protocol showed no significant difference in survival compared with patients enrolled onto a trial.
Conclusion: Study patients were not representative for the entire population of patients with AML; many patients were excluded from phase III trial participation for failure to meet stringent entry criteria. Therefore, results of phase III studies may not be extrapolated to all AML patients but should only be applied to patients who do not differ in substantial characteristics from the study population.