The very high incidence of oesophageal squamous cell carcinoma (ESCC) in Golestan Province in northeastern Iran was suggested by studies in the 1970s as partly due to opium use, which is not uncommon in this area, but based on limited numbers. From December 2003 to June 2007, we administered a validated structured questionnaire to 300 ESCC cases and 571 controls, matched on neighbourhood of residence, age (+/-2 years), and sex. We used conditional logistic regression models to calculate odds ratios (ORs) and 95% confidence intervals (95% CIs) adjusted for potential confounders. Compared with those who used neither tobacco nor opium, risk of ESCC was increased in those who used tobacco only (OR, 95% CI: 1.70, 1.05-2.73), in those who used opium only (2.12, 1.21-3.74), and in those who used both tobacco and opium (2.35, 1.50-3.67). All forms of tobacco use (cigarettes, hookah, and nass) were associated with higher ESCC risk. Similarly, use of both crude opium and other forms of opium were associated with higher risk. Alcohol consumption was seen in only 2% of the cases and 2% of the controls, and was not associated with ESCC risk.