The aim of this study was to evaluate the role and potential benefit of G-CSF administered following standard regimen chemotherapy (CHT) in non-Hodgkin lymphomas. Twenty patients with NHL were given CHOP or CHOP/CVP CHT every 21 days. None was given G-CSF after the first cycle. After each cycle, G-CSF was administered only for: 1) ANC < 1 x 10(9)/L between cycles; or 2) delay in cycle schedule due to ANC < 1 x 10(9)/L on the planned day of treatment; or 3) development of a febrile syndrome or a documented infection, regardless the ANC. Once administered, G-CSF was maintained in the following cycles. Nineteen patients required administration of G-CSF (5 micrograms/Kg B.W.), but for different reasons only 16 were treated (a mean of 10 +/- 3 doses/cycle). Comparing 48 cycles where G-CSF was not administered, with 50 where it was, in this last group we observed: 1) a ANC significantly higher at day 7 (p < 0.0001), day 14 (p < 0.0001) and day 21 (p = 0.0030); 2) a significantly lower (p = 0.0001) incidence of neutropenias (6 vs 29); 3) a trend (p = 0.1040) in favour of lower incidence of febrile neutropenias of infections (1 vs 6); 4) a significantly lower (p < 0.0001) incidence of cycle delays (1 vs 22) with a median of 8 days (1 to 20); and 5) a significantly higher (p < 0.0001) dose intensity (99.5% vs 87.8%).(ABSTRACT TRUNCATED AT 250 WORDS)