To assess the effect and toxicity of hypotonic cisplatin treatment (HPT) consisting of the intrapleural administration of cisplatin in distilled water for malignant pleural effusion in patients with non-small-cell lung cancer (NSCLC). Non-small-cell lung cancer patients with cytologically proven and previously untreated malignant pleural effusion were enrolled into this study. Firstly, the lung was fully re-expanded by a tube thoracostomy, and then 25 mg cisplatin in 500 ml of distilled water was instilled through a chest tube and then the tube was clamped. After 1 h, the tube was declamped and allowed to drain. The chest tube was removed when the pleural effusion volume decreased to 200 ml or less per day. A complete response (CR) was considered to occur when the pleural effusion disappeared. A partial response (PR) was determined to occur when the volume of pleural effusion remained under (1/4) of hemithorax. The response at 4 weeks was evaluated by an extramural review. Out of 84 patients enrolled from February 1998 to August 2002, 80 patients were eligible and analysed in the present study. The toxicity of HPT was acceptable. Neither a haematological toxicity of any grade nor grade 4 nonhaematological toxicity was observed. Grade 3 nonhaematological toxicities were observed, including nausea (4%), vomiting (3%), pyothorax (1%) and dyspnoea (1%). The median time of drainage from HTP was 4 days. Twenty-seven (34%) and 39 (49%) patients achieved CR and PR, respectively, for an overall response rate of 83% (95% confidence interval, 74-91%). The median duration of the response was 206 days. The median survival time of all patients was 239 days. Hypotonic cisplatin treatment for malignant pleural effusion of NSCLC is therefore considered to be feasible and effective. A phase III study of HPT is thus warranted.