OP2-5 - Modifiable statin characteristics associated with potential statin-related prescribing cascades in a large commercial insurance claims data of adult statin initiators

Background

Clinicians sometimes prescribe new medications to treat drug-related adverse events, which constitutes a prescribing cascade. Prescribing cascades can generally be avoided by substituting an alternative medication class for the precipitating drug, yet in the case of statins, there are no therapeutically equivalent alternatives. Therefore, switching to other statins or reducing the dose may be considered when encountering statin-related adverse events. 

Objectives

To identify modifiable statin characteristics (e.g., intensity and individual statin medications) associated with the risk of statin-related prescribing cascades.

Methods

Using the 2015-2019 IBM MarketScan Commercial and Medicare Supplemental Insurance claims data, we conducted a case-only study with high-throughput sequence symmetry analysis to screen for potential statin-related prescribing cascades. . We included patients (aged ≥18) initiating statins and continuously enrolled between -720 days and +360 days of statin initiation. We assessed the initiation of individual statin-marker drug dyads including 524 other classes of medications (grouped by Anatomical Therapeutic Chemical [ATC], Level 4) within 90 days before or after statin initiation, using the adjusted sequence ratio to estimate the relative order of statin-marker drug initiation. Two investigators with clinical expertise independently evaluated each significant signal for biologic plausibility as a potential prescribing cascade. Among biologically plausible signals, we measured patients’ baseline characteristics ([-360, 0] days of statin initiation) and statin profile characteristics, including specific medication and intensity (i.e., low-, moderate-, and high-intensity) according to the ACC/AHA guideline. We used multivariable logistic regression in dyads with at least 500 patients to identify factors associated with each prescribing cascade, adjusting for baseline characteristics. Adjusted odds ratios (aOR) with 95% confidence intervals (CI) were reported to describe the association between the modifiable stain characteristics and the risk of stain-related prescribing cascades.

Results

Of the 2,290,011 patients in the study cohort, 19% were aged ≥65 years, and 51% were male. Simvastatin was the most commonly initiated statin (34%) and 71% of patients initiated a moderate-intensity statin. In total, 42 prescribing cascade signals were considered biologically plausible and had ≥500 patients. Compared to low-intensity statin initiators, patients initiating on high-intensity statins were associated with a higher risk to experience 22 (53%) of these prescribing cascades, including DPP-4 inhibitors (aOR= 1.27, 95% CI=1.15-1.40), GLP-1 analogues (1.33, 1.19-1.50), and opioids (1.18, 1.13-1.23). Individual statin medication also had a differential effect in 33 (79%) of the 42 evaluated prescribing cascades. Notably, compared with simvastatin initiators, lovastatin initiators were associated with a higher risk of initiating osmotically acting laxatives (1.09, 1.03-1.15) and opioids in combination with non-opioid analgesics (1.04, 1.01-1.07) after statin initiation, while atorvastatin initiators were associated a lower risk of initiating osmotically acting laxatives (0.92, 0.89-0.94) and opioids in combination with non-opioid analgesics (0.97, 0.95-0.98) after statin initiation.

Conclusions

In this high-throughput sequence symmetry analysis of commercially-insured adult statin initiators, high-intensity statin initiation is associated with an increased risk in statin-related prescribing cascades compared to low-intensity statin initiation. The risk of the statin-related prescribing cascade varied by the choice of individual statins.