High-Throughput Screening for Prescribing Cascades Among Real-World Angiotensin-Converting Enzyme Inhibitor Initiators.

Purpose: Angiotensin-converting enzyme inhibitors (ACEIs) are commonly prescribed, but their adverse effects may prompt new drug prescription(s), known as prescribing cascades (PCs). We aimed to identify potential ACEI-induced PCs using high-throughput sequence symmetry analysis.

Methods: Using claims data from a national sample of Medicare beneficiaries (2011-2020), we identified new ACEI users aged ≥ 66 years with continuous enrollment ≥ 360 days before and ≥ 180 days after ACEI initiation. We screened for initiation of 446 other (non-antihypertensive) "marker" drug classes within ±90 days of ACEI initiation, generating sequence ratios (SRs) reflecting proportions of ACEI users starting the marker class after versus before ACEI initiation. Adjusted SRs (aSRs) accounted for prescribing trends over time. For significant aSRs, we calculated the naturalistic number needed to harm (NNTH), and significant signals underwent clinical review for plausibility.

Results: We identified 308 579 ACEI initiators (mean age 76.1 ± 7.5 years; 59.6% female; 88.6% with hypertension). Of 446 marker classes evaluated, 81 signals were significant, and 42 (52%) classified as potential PCs after clinical review. The strongest signals ranked by lowest NNTH included corticosteroids (NNTH 313; 95% CI, 262-392) and serotonin type 3 (5-HT₃) antagonists (NNTH 496; 95% CI, 392-689); the strongest signals ranked by highest aSR included sympathomimetics (aSR, 1.97; 95% CI, 1.10-3.53) and other antianemic preparations (aSR, 1.87; 95% CI, 1.31-2.67).

Conclusion: Identified prescribing cascade signals were indicative of known and possibly underrecognized ACEI adverse events in this Medicare cohort. The findings are hypothesis-generating and require further investigation to determine the extent and impact of the identified PCs on health outcomes.