Estimating prevalence and characteristics of statin intolerance among high and very high cardiovascular risk patients in Germany between 2017–2020. Issue 4 (22nd December 2022)

• Record Type: Journal Article
• Authors: Parhofer, K G; Anastassopoulou, A; Calver, H; Becker, C; Singh Rathore, A
• Is Part Of: European heart journal. Volume 3:Issue 4(2022)
• Publication Date: 2022-12-22
• Subjects: Medical informatics -- Periodicals; Medical technology -- Periodicals; Cardiovascular system -- Periodicals; 616.100284
• DOI: 10.1093/ehjdh/ztac076.2789
• Languages: English
• ISSNs: 2634-3916

Abstract

Abstract: Background: Statins remain the backbone of lipid management. Nevertheless, the degree to which statins can be used and dosed in clinical practice remains a great challenge, also due to statin intolerance (SI). The lack of widely accepted SI definition leads to poor understanding of the condition and SI patient profile Purpose: To estimate the current SI prevalence and understand better the patient characteristics, using machine learning techniques Methods: Retrospective cohort study, based on representative sample of electronic patient records from outpatient setting in Germany. Patients were included if they had high CV risk, atherosclerotic cardiovascular disease (ASCVD) or hypercholesterolemia (HC), between 2017 and 2020. Patients were categorized as having "absolute" (history of SI events and permanent statin discontinuation) or "partial" (history of SI events while treated with statins) SI. Machine learning techniques were utilized to calibrate the prevalence estimates and to identify patient clusters. Estimates of SI prevalence were derived based on different rules and confidence levels (high, moderate and low). The low confidence estimates contain the most uncertainty in identifying SI Results: The study population consisted of 292, 603 patients (57.3% aged >70 years; 55.6% male). Of these, ∼24% had high CV risk, ∼56% had ASCVD, and ∼20% had HC. After deploying machine learning, the SI identification improved by ∼27% in absolute SI and by ∼57% in partial SI patients, resulting in a maximum estimate of 12.5% SI with high/moderate confidence and further 11.8% with low confidence (absolute SI 15.8%, partial SI 8.5%). The low confidence group may contain patients with insufficient statin treatment due to reasons other than SI (e.g. clinical inertia). Statistically significant risk factors for SI were hypothyroidism, vitamin D deficiency, liver and chronic kidney disease. Cramps, muscle spasms, myalgia and myopathy were the most common statin associated muscle symptoms (SAMS) observed in the SI population. Atorvastatin 40mg was the most frequently down-titrated statin, while simvastatin to atorvastatin was the most predominant class switch in SI patients. Machine learning techniques applied on high confidence SI patients characteristics and the most commonly observed cluster for patients over 60 years showed predominant musculoskeletal disorders, concomitant high SAMS incidence and high use of multiple statins. In males under 60 years, depression and somatoform disorders along with musculoskeletal disorders, pain, and gastric events were common, while females under 60 years had predominant depressive episodes, along with musculoskeletal, mental, and metabolic disorders Conclusion: Addressing the complexity in defining SI using advanced analytics, this study provides prevalence estimates and describes distinct patients clusters that may inform diagnosis and optimal treatment pathways for SI patients in Germany Funding Acknowledgement: Type of funding sources: Private company. Main funding source(s): Daiichi Sankyo Europe GmbH