BACKGROUND Carcinoma of the esophagus is a relatively uncommon but lethal cancer that continues to kill over 90% of its victims within 5 years. Surgery is the treatment of choice for most localized esophageal cancer patients. However, despite curative resection, the 5-year survival rate ranges from 15% to 39%. The failure of surgery to cure clinically localized esophageal cancer is because of the advanced state of the disease before symptoms occur, high frequency of lymph node involvement, and the common occurrence of submucosal spread and extension to surrounding structures. Preoperative chemotherapy has been used in an attempt to decrease tumour activity, increase resectability, and improve disease-free and overall survival. A number of studies have investigated whether preoperative chemotherapy followed by surgery leads to an improvement in cure rates, but the individual reports have not been encouraging. The role of preoperative chemotherapy in the treatment of resectable thoracic esophageal cancer remains undefined. OBJECTIVE The objective of this review is to determine the role of preoperative chemotherapy on overall survival and/or quality-of-life for patients with resectable thoracic esophageal carcinoma. METHODS Trials were identified by searching the Cochrane Controlled Trials Register (Issue 2 - 2000), MEDLINE (1966 - 2000), EMBASE (1988 - 2000) and CancerLit (1993 - 2000). The references of all identified studies, review articles, and standard textbooks were examined. Members of the Cochrane UGPD Group and experts in the oncology field were contacted and asked to supply details of any outstanding clinical trials and relevant unpublished materials. There were no language restrictions. The searches were updated in June 2000. The clinical trial registers of the National Cancer Institute and the Radiation Therapy Oncology Group were consulted for ongoing trials. METHODS Types of studies Studies (published or unpublished) that randomised patients with potentially resectable carcinoma of the esophagus (of any histologic type) to chemotherapy or no chemotherapy before surgery were included in this review. Studies were excluded if they were not truly randomised (phase I or II trials), earlier versions of updated trials, if other treatment modalities (e.g. radiotherapy, hyperthermia) were used, or if there was not a surgery alone control arm. Types of participants The participants consisted of patients with potentially resectable thoracic esophageal carcinoma (of any histologic type). Trials involving patients with carcinoma of the cervical esophagus were excluded. Types of interventions Trials that compared chemotherapy before surgery (esophagectomy) with surgical resection alone (esophagectomy). Types of outcome measures The primary outcome was death at yearly intervals. Morbidity (complications), and quality-of-life were secondary outcomes. METHODS Overall mortality at yearly intervals was determined by extracting the total number of patients randomised to the treatment and control groups and the number of deaths in each group. All analyses were carried out on intention-to-treat that is patients were analyzed according to their allocated treatment, irrespective of whether they received that treatment. Mortality at 1, 2, 3, 4 and 5 years were used as endpoints of clinical relevance. If survival numbers at the specified time intervals were not given, they were estimated from the published survival curves. The number of deaths in the treatment groups (preoperative chemotherapy plus surgery) was compared to the number of deaths in the control groups (surgery alone). Treatment modalities as well as patient demographics and characteristics and side-effects were also recorded. Trials meeting the inclusion criteria were evaluated by two independent reviewers using the Jadad method RESULTS A total of 14 randomised controlled trials and 1 meta-analysis of preoperative chemotherapy versus surgery alone for esophageal carcinoma were identified to be potentially eligible for review. This review is based on 7 randomised trials and 1653 patients. At 1 year the Peto odds ratio based on the fixed-effects models showed no difference in mortality between preoperative chemotherapy and surgery alone (OR = 1.03). At 2 years there was a 20% significant decrease in mortality for preoperative chemotherapy (OR = 0.80; 95% C.I. 0.65 to 0.99) but the results were not robust. The results at 3, 4, and 5 years found odds ratios tending to favour preoperative chemotherapy, but wide confidence intervals that included 1. None of the published trials reported on quality-of-life outcomes. There appeared to be an increased risk of morbidity with chemotherapy. CONCLUSIONS The results of this review suggest that there is no strong evidence to recommend preoperative chemotherapy in the treatment of surgically resectable carcinomas of the thoracic esophagus. (ABSTRACT TRUNCATED)