Aims To describe injection drug use among inmates, and to identify correlates of drug injection while incarcerated. Design Cross-sectional survey. Setting Six provincial correctional centres in Ontario, Canada. Participants Face-to-face interviews were conducted with a random sample of 439 adult males and 158 females. Measurements Inmates were asked about drug use in their lifetime, outside the year prior to their current incarceration, and while incarcerated in the past year. Among the 32% (189 / 597) with a prior history of drug injection, independent correlates of injection while incarcerated in the past year were identified using multiple logistic regression. Findings Among all inmates while incarcerated in the past year, 45% (269 /597) used drugs and 19% (113 / 596) used non-cannabis drugs. Among those with a prior history of injecting, 11% (20 / 189) injected while incarcerated in the past year. Rates of injection with used needles were the same pre-incarceration as they were while incarcerated (32%). Independent correlates of drug injection while incarcerated were injection of heroin (OR = 6.4) or other opiates (OR = 7.9) and not injected with used needles (OR = 0.20) outside in the year prior to incarceration, and ever being incarcerated in a federal prison (OR = 5.3). Conclusions The possibility of transmission of human immunodeficiency virus (HIV), hepatitis C (HCV) or other blood-borne diseases exists in Ontario correctional centres. In this setting, drug injection while incarcerated is primarily related to opiate use prior to incarceration. The correlation between injecting and extensive incarceration history suggests missed opportunities to improve inmates' health. (abstract from http://www.blackwell-synergy.com/links/doi/10.1046%2Fj.1360-0443.2003.00...)