In a study conducted at St Bartholomew's Hospital between 1972 and 1982, using moderately intensive therapy (OPAL/HEAV'D), a low blast count at presentation (less than 10 x 10(9) 1(-1)) and common ALL (C-ALL) phenotype correlated favourably with duration of remission. Fifty-four patients (age range 15-57, median 32) subsequently received a modification of the previous treatment programme which included high-dose ara-C 2 g m-2 b.d. for 6 days as cycle 3 (OPAL + HD ARA-C). CR was achieved in 36/54 (67%) patients, response correlating favourably with younger age (15-30 years vs 31-57 years, P = 0.02). Three patients died in CR. Overall, there was no difference in survival or remission duration between patients who received high dose ara-C and those in the control group. However, in contrast to the early results, there was a reversal in the relevance of the prognostic factors with a trend in favour of high blast count (greater than 10 x 10(9) 1(-1)) and T-cell phenotype in terms of remission duration. Moreover, comparison of duration of remission for the previously defined prognostic groups according to therapy suggests that the prognosis of patients with 'high risk' disease (T, B, null ALL or high blast count) is improved with more intensive therapy. In contrast, those with 'low risk' disease (C-ALL and low blast count) have a better prognosis with less intensive therapy. These observations confirm those of others and allow for individualization of therapy on the basis of pre-treatment variables.