Discontinuation of antihypertensive and lipid-lowering medication in primary care: a systematic review of observational data and sociodemographic differences

Technical Summary: Discontinuation of Antihypertensive and Lipid-Lowering Medication in Primary Care

Problem Statement
Cardiovascular disease (CVD) remains the leading cause of global mortality and morbidity. While clinical trials demonstrate that antihypertensives (AHTs) and lipid-lowering therapies (LLTs) can reduce CVD risk by up to 30% in primary and secondary prevention, these benefits are often unrealized in practice due to frequent medication discontinuation. Existing literature on discontinuation is fragmented: previous reviews are often outdated (≥10 years old), rely on randomized controlled trial (RCT) data which lacks generalizability, focus on adherence rather than discontinuation, or lack granular data on socio-demographic variations (specifically ethnicity and socio-economic position). Furthermore, no prior study has simultaneously synthesized aggregated data on AHT and LLT discontinuation patterns to identify shared or divergent drivers.

Methodology
This study is a systematic review of observational data from electronic health records (EHRs) conducted according to PRISMA guidelines.

• Data Sources: PubMed, EMBASE, Web of Science, and CINAHL were searched for studies published between 2010 and 2025.
• Inclusion Criteria: Studies were required to use EHRs to assess discontinuation rates in adults (≥18 years) initiating first- to third-line AHTs or LLTs in primary care settings. RCTs were excluded.
• Selection Process: From 5,756 unique records, 44 reports were included (29 on LLTs, 15 on AHTs), covering a total of 11,645,049 patient registrations.
• Analysis: Data extraction focused on discontinuation and restarting prevalences, as well as associations with age, sex, socio-economic position (SEP), and ethnicity. Risk ratios were estimated using logistic models where raw counts were available. Study quality was assessed using the ROBINS-E tool. Due to significant methodological heterogeneity (e.g., varying definitions of discontinuation gaps and switching), a meta-analysis of effect estimates was precluded; instead, data were visualized using forest plots.

Key Contributions

• First Simultaneous Synthesis: This is the first review to simultaneously aggregate real-world evidence on both AHT and LLT discontinuation, allowing for comparative insights into drivers of non-persistence across these two cornerstone therapies.
• Real-World Focus: By exclusively utilizing EHR data from primary care, the study captures discontinuation patterns reflective of actual clinical practice, avoiding the artificial constraints of trial protocols.
• Socio-Demographic Granularity: The review specifically quantifies disparities in discontinuation based on age, sex, income-based SEP, and ethnicity, addressing a gap in previous literature.
• Restarting Data: The study provides rare aggregated data on medication restarting rates following discontinuation.

Results

• Discontinuation Prevalence:
  • LLTs: The median discontinuation prevalence was 43% (IQR 33%–59%). Restarting prevalence was 58% (IQR 40%–61%).
  • AHTs: The median discontinuation prevalence was 41% (IQR 30%–49%). Restarting prevalence was 28%.
  • Variability: Rates varied significantly based on the definition of discontinuation (e.g., whether medication switching was counted) and the indication (primary vs. secondary prevention).
• Age: Patients aged approximately 65 years were less likely to discontinue than those at the extremes of age (<60 or >75 years). A U-shaped curve was observed in several studies, where the youngest and oldest groups showed higher discontinuation rates.
• Sex:
  • LLTs: Women generally discontinued more often than men.
  • AHTs: Men were more likely to discontinue AHTs for primary prevention, though this pattern reversed or diminished in mixed prevention populations.
• Socio-Economic Position (SEP): Income-based measures were associated with discontinuation (lower income linked to higher discontinuation), but composite SEP measures (e.g., area-level deprivation indices) showed inconsistent or non-significant associations.
• Ethnicity: Minority ethnic groups were consistently more likely to discontinue both LLTs and AHTs compared to White/European groups. In UK data, South Asian and African/African-Caribbean patients had significantly higher odds of discontinuation.

Significance and Claims
The authors claim this review identifies high levels of discontinuation and distinct inequalities across age, sex, and ethnicity in real-world primary care settings. They posit that these patterns suggest specific drivers:

• Sex Differences: The divergence between LLT (higher in women) and AHT (higher in men for primary prevention) discontinuation suggests that sex-specific adverse drug events (e.g., muscle symptoms for statins in women; sexual dysfunction for AHTs in men) may be key causative factors.
• Ethnic Disparities: The consistent pattern of higher discontinuation across minority ethnic groups for both drug classes implies compounded, modifiable disadvantages potentially driven by structural issues, interpersonal racism, or mistrust in healthcare systems, rather than biological factors alone.

The paper concludes that awareness of these specific discontinuation patterns is crucial for targeting healthcare interventions. It emphasizes that addressing discontinuation in the highest-risk, most vulnerable groups could improve health equity. The authors note limitations, including the inability to perform meta-analysis due to heterogeneity and the lack of data from low- and middle-income countries, calling for future qualitative research to elucidate the specific patient-level drivers of these disparities.