It has been reported that 7-ethyl-10-[4-(1-piperidino)-1-piperidino]carbonyloxy-camptothecin (CPT-11) and its active metabolite, 7-ethyl-10-hydroxy-camptothecin (SN-38), have absorption characteristics of weakly basic drugs, suggesting that alkalization of the intestinal lumen might reduce reabsorption and its attendant side effects. Furthermore, stasis of stools containing these compounds is thought to induce damage to the intestinal mucosa. The prevention of CPT-11-induced side effects by oral alkalization (OA) combined with control of defecation (CD) was estimated in a case-control study of lung cancer patients. Coinciding with day 1 of CPT-11 infusion and for 4 days thereafter, OA and CD were practiced utilizing orally administered sodium bicarbonate, magnesium oxide, basic water and ursodeoxycholic acid. OA involved the daily use of all four therapeutics, and CD required doses of up to 4.0 g/day of magnesium oxide and 2 L/day of excess basic water. From three ongoing prospective phase I/II studies, we selected 37 consecutive patients who were treated with CPT-11 in combination with cisplatin in the presence of OA and CD (group B). Thirty-two control subjects who were matched to the background characteristics of the case patients were treated with the same regimen in the absence of OA and CD (group A). Toxicities induced by the CPT-11/cisplatin combination were evaluated and analyzed in group A and group B in a case-control format. The use of OA and CD resulted in significantly higher stool pH (p < 0.0001), while reducing the incidence of delayed diarrhea (> or = grade 2: group A 32.3% versus group B 9.4%; p = 0.005), nausea (p = 0.0001), vomiting (p = 0.001) and myelotoxicity, especially granulocytopenia (p = 0.03) and lymphocytopenia (p = 0.034). In addition, dose intensification was well tolerated in patients receiving OA and CD, allowing dose escalation from 35.6 +/- 6.0 to 39.9 +/- 5.6 mg/m(2)/week (p < 0.001). Tumor response rates for non-small cell lung cancer were 59.3% (16/27 patients) in group B compared with 38.5% (10/26 patients) in group A. Multivariate analysis revealed that the risk of CPT-11-induced delayed diarrhea greater than grade 2 was associated with OA and CD (odds ratio for delayed diarrhea, 0.14 with use of OA and CD; 95% confidence interval, 0.05 to 0.4; p = 0.0002) and age (odds ratio, 1.08 per increase in age; 95% confidence interval, 1.02 to 1.15; p = 0.009). OA and CD appear to be useful in preventing the dose-limiting side effects of CPT-11 noted in clinical practice, mainly nausea, vomiting, granulocytopenia and especially delayed diarrhea. Risk factors statistically associated with delayed diarrhea include advanced age and the use of CPT-11 without OA and CD.