Information about the etiology of childhood myeloid leukemia is limited. A population-based nested case-control study of prenatal and neonatal risk factors for childhood myeloid leukemia was performed with the use of the Swedish National Cancer Register and the Swedish Birth Register. A total of 98 cases of myeloid leukemia were identified in successive birth cohorts from 1973 through 1989. From the Birth Register, five controls were matched to each case. Fourteen of the 98 cases with myeloid leukemia and none of the controls had Down syndrome [odds ratio (OR) = infinity; 95% confidence interval (CI) = 21.0-infinity]. The risk for myeloid leukemia also increased among children who had physiological jaundice (OR = 2.5; 95% CI = 1.2-5.0; children who had been treated with phototherapy (OR = 7.5; 95% CI = 1.8-31.9); or who had been treated in an incubator (OR = 3.5; 95% CI = 1.2-10.2). Excluding cases with Down syndrome, however, decreased these risks, so that their 95% lower confidence interval included the no-effect value. Maternal age < 20 years old (OR = 2.5; 95% CI = 1.1-6.0), hypertension (OR = 2.4; 95% CI = 1.2-5.0), Cesarean section (OR = 2.5; 95% CI = 1.3-4.9), maternal smoking (OR = 2.4; 95% CI = 0.9-6.5), and being one of a multiple birth (OR = 3.6; 95% CI = 1.1-11.3) increased the risk for myeloid leukemia among those without Down syndrome. When the analyses were repeated, by restricting the cases to those with acute myeloid leukemia, the risk associated with young maternal age declined and became nonsignificant.(ABSTRACT TRUNCATED AT 250 WORDS)