ABSTRACT Aims  The objective of this study was to identify factors associated with sustained injection cessation and to examine further the relationship between the occurrence of sustained injection cessation of injection drug users (IDUs) and prior injection frequency. Design and setting  IDUs in the Montreal St Luc Cohort who had at least three consecutive interviews between 1995 and 1999 were included. Sustained injection cessation was defined as a period of at least 7 consecutive months without injection. All IDUs completed interview‐administered questionnaires on socio‐demographic characteristics, drug and sexual behaviours and health‐related issues. Logistic regression was used for analyses. Findings  A total of 186/1004 (18.5%) IDUs reported a period of sustained injection cessation during the study period. In multivariate analysis, HIV‐positive status, ‘booting’ and cumulative time spent in prison were negatively associated with injection cessation, while injection initiation after 35 years of age and frequent crack use were positively associated with injection cessation. We found a negative association between the occurrence of injection cessation and the frequency of injection; the odds ratios (OR) for cessation were 0.49 95% confidence interval (CI): 0.03, 0.78 for IDUs who injected 30–100 times and 0.21 (95% CI: 0.10, 0.46) for IDUs who injected more than 100 times in the previous month. Attending needle exchange programmes (NEPs) or pharmacies appeared to be a modifier of the relation between cessation and prior injection frequency. The OR was 0.68 (95% CI: 0.42, 1.12) for IDUs who injected 30–100 times prior to injection and attended NEPs or pharmacies and was 0.07 (0.01, 0.30) for IDUs who did not use these services. Conclusions  Overall, a fifth of IDUs experienced at least one episode of injection cessation of 7 months or more during a period of 4.5 years. Our data suggest that NEPs and pharmacies may have played a role in inducing injection cessation episodes in a subgroup of IDUs. Research is needed to better identify the characteristics of IDUs who could benefit from an injection cessation intervention strategy. This information is important for social and health policy planning.