SummaryBackground Maintenance therapy with 5-aminosalicylic acid (5-ASA) is a key strategy for preventing relapse in many patients with inflammatory bowel disease (IBD). Factors which disrupt 5–ASA delivery, such as non-adherence and 5-ASA switches, may destabilise symptom control.Aim To investigate the impact of non-adherence and medication switches on stable symptom control in UK patients with IBD.Methods A retrospective cohort study was conducted using a UK dispensing database. Adherence was analysed in randomised matched samples for each of the six leading oral mesalazine formulations, measured by medication possession ratio (MPR); MPR ≥80% was classified as adherent. Relationships among adherence, switch and relapse were analysed over 18 months in patients receiving continuous mesalazine therapy throughout a 6–month baseline period (primary subgroup analysis). Relapses of active ulcerative colitis were identified using a doubling of MPR as a proxy.Results Only 39% of patients in the matched samples (n = 1200) were classed as adherent. No significant differences in adherence were observed among mesalazine formulations. In the primary subgroup analysis (n = 568), non-adherent patients had a significantly greater risk of relapse than adherent patients (RR = 1.44, 95% CI = 1.08–1.94; P = 0.014). Among adherent patients (n = 276), those who switched had a 3.5-fold greater risk of relapse than those who did not switch (95% CI = 1.16–10.62; P = 0.008).Conclusions Both non-adherence and mesalazine switches in adherent patients were associated with significant increases in the risk of relapse, suggesting that disruption of mesalazine maintenance therapy may destabilise symptom control. These findings provide evidence to advocate caution when considering mesalazine switches for stable patients.