In a consecutive study of 14 boys and 17 girls with non-B-cell ALL who were > or = 1 year of age at diagnosis, the degree of myelosuppression during the last year of MTX/6MP maintenance therapy was analyzed in relation to the erythrocyte concentration of MTX polyglutamates and 6-thioguanine nucleotides (E-MTX and E-6TGN, the respective major cytotoxic metabolites of MTX and 6MP). For each patient, E-MTX and E-6TGN levels were measured 2-15 (median, 6) and 2-17 (median, 7) times, respectively. From these measurements, arithmetic means of E-MTX and E-6TGN were calculated (mE-MTX and mE-6TGN, respectively). Since MTX and 6MP probably work synergistically, the product of mE-MTX and mE-6TGN was calculated for each patient (mE-MTX x 6TGN). The degree of myelosuppression was registered as the mean WBC determined following cessation of the therapy minus the mean WBC measured during the therapy (mWBCshift). The mean WBCs measured on therapy (mWBC(on)) and off therapy were highly correlated (r = 0.48, P = 0.009). The median mWBCshift was 2.7 x 10(9)/l (range, 1.4-4.8 x 10(9)/l). In a multivariate regression analysis, the best-fit model to predict the mWBCshift included mE-MTX x 6TGN, age at drug withdrawal, and mWBC in the order given [mWBCshift = 4.3 + 0.00089 x (mE-MTX x 6TGN) - 0.097 x age - 0.41 x mWBC(on); global rs = 0.66, P = 0.0002]. Thus, the patients with higher mE-MTX x 6TGN values, the younger patients, and the patients with the lowest WBC during therapy had the most pronounced degree of myelosuppression as measured by mWBCshift. These results indicate that E-MTX and E-6TGN may give a better reflection of the treatment intensity than do the WBCs alone.