To identify treatment factors that may affect the survival of children with inv(16)(p13.1q22), we compared the outcomes of 19 patients with this genetic feature treated at our institution during two treatment eras. Nine patients were treated during era 1 (1980 to 1987), and 10 were treated during era 2 (1988 to 1996). All entered complete remission (CR) with induction therapy. Eight of the nine children treated in era 1 died, seven of relapsed leukemia. In contrast, three of 10 patients treated during era 2 have died, all of non-disease-related causes. Event-free survival (EFS) estimates were significantly higher for patients treated during era 2 than for those treated during era 1 (P = 0.03); the 6-year estimates were 70 +/- 15% (s.e.) and 11 +/- 7%, respectively. Era 2 treatment protocols differed from those of era 1 in their use of higher doses of cytarabine and etoposide during induction and consolidation chemotherapy and in their use of 2-chlorodeoxyadenosine (2-CDA). These results suggest that dose intensification of cytarabine benefits children with AML and inv(16), as is the case in adults. They also suggest that dose intensification of etoposide and addition of 2-CDA may also offer an advantage. This study underscores the dependence of the prognostic impact of cytogenetic features on the efficacy of treatment.