Prescribing cascades in people with type 2 diabetes: analyses of two representative routine healthcare databases

Abstract

Abstract Background Prescribing cascades, defined as initiation of medications to treat adverse drug reactions caused by other medications, may contribute to polypharmacy in people with type 2 diabetes. We aimed to determine whether commonly used diabetes and cardiovascular medications were associated with subsequent initiation of treatments for their known adverse drug reactions, indicating potential prescribing cascades. Methods We conducted two parallel cohort studies using prescription sequence symmetry analysis to assess asymmetry in the initiation sequence of index medications and corresponding marker medications. Twelve index-marker medication dyads were selected a priori from published literature. Data were obtained from the Scottish Diabetes Research Network national diabetes registry and the Welsh Secure Anonymised Information Linkage Databank between January 2015 and December 2023. Adults aged 40 years or older with type 2 diabetes were eligible if they had incident use of both an index and marker medication within a 365-day period. Adjusted sequence ratios comparing initiation of marker medications after versus before index medications were calculated after adjustment for secular prescribing trends. Ratios greater than 1.0 indicated potential prescribing cascades. Sensitivity analyses examined age, sex, socioeconomic position, comorbidity, and shorter observation windows (90 and 180 days). Estimates from both datasets were pooled using random-effects meta-analysis. Results A total of 387,338 individuals from the Scottish Diabetes Research Network and 296,582 from the Welsh Secure Anonymised Information Linkage Databank were eligible. Seven of 12 medication dyads showed evidence of prescribing cascades: sodium-glucose cotransporter-2 inhibitors followed by urinary tract infection antibiotics (pooled adjusted sequence ratio 1.14, 95% confidence interval 1.10–1.17), antifungals (1.64, 1.58–1.69), and overactive bladder medications (1.16, 1.00-1.34); thiazolidinediones followed by loop diuretics (2.23, 1.28–3.89); calcium channel blockers followed by loop diuretics (1.80, 1.46–2.23); and statins followed by overactive bladder medications (1.32, 1.02–1.73) and hypnotics (1.45, 1.29–1.62). Findings were directionally consistent across demographic and clinical subgroups. Conclusions Prescribing cascades appear to be important contributors to polypharmacy among people with type 2 diabetes. Recognition of adverse drug reactions as a cause of new symptoms may identify opportunities to stop or modify the causative medication where appropriate, potentially reducing treatment burden and polypharmacy.

The authors' abstract, as published at the source. BMC Medicine, 2026 · DOI ↗