A large UK study of metformin adherence investigates which patients may be likely to discontinue metformin therapy.
In 1999, the British Journal of Clinical Pharmacology published “The UK Prospective Diabetes Study” (UKPDS), which showed that metformin use results in a 42% reduction in diabetes death and a 36% decrease in all-cause mortality, and a 32% reduction in any diabetes-related endpoint. The study results were so significant that metformin quickly became the United Kingdom‘s first line anti-diabetic agent and is now used in 80-90% of patients with diabetes. As long term, consistent therapy is essential with metformin for effectiveness in lowering A1c and hospitalization rates, a UK based study set out to investigate patterns that may lead to discontinuation of metformin therapy.
Researchers utilized the Clinical Practice Research Datalink (CPRD) to gather patient data for a retrospective cohort study of type 2 diabetes patients. The CPRD is a database managed by the UK Department of Health that consists of over 13 million patient records entered by primary care physicians. This study’s inclusion criteria required that patients are 18 years of age or older with a diagnosis of type 2 diabetes, and have filled at least one metformin prescription in 2013. Patients were excluded if they received a secondary diabetes diagnosis such as gestational or type 1 diabetes, had polycystic ovarian syndrome or did not have sufficient medical records. A total of 111,185 patients were chosen for the study, 10.1% of whom were new metformin users, and 89.9% were ongoing metformin users. Patient factors such as age, baseline medication use, A1c, BMI, and comorbidities were then analyzed in the new and ongoing metformin users to find trends in discontinuation, which was defined as a treatment gap of ≥ 90 days.
The study found that after 12 months, treatment was discontinued in 35.9% of new metformin users, compared to 23.1% of ongoing users. Ongoing metformin users had even lower rates of discontinuation if they were of advanced age, had comorbidities such as macrovascular disease, or were on lipid-lowering therapy. Both groups had better adherence in smokers (OR 1.188; 95% CI 1.140-1.237) and those with comorbidities such as chronic kidney disease, depression, anxiety, dementia, or chronic obstructive pulmonary disease. Adherence rates were lower in patients of both groups that had obesity or had an A1c of 8-8.9% (OR 0.862; 95% CI 0.813-0.913). Patients in both groups who were using 5-10 medications before the addition of metformin were also more adherent to therapy (OR 1.073; 95% CI (1.035-1.112). Patients using sulfonylureas were explicitly shown to have better metformin adherence (OR 1.091; 95% CI 1.037-1.105). Age or sex had no statistically significant impact on metformin therapy compliance.
Study strengths included being a retrospective cohort study with a large sample size that drew patient records from reliable, nationalized databases. However, despite the large sample size, the study did not adjust for race, educational level, socioeconomic status, or lifestyle-related factors. Missing demographic information makes it difficult to generalize results to other populations. It also did not distinguish between those who may have negligently discontinued metformin use compared to those who switched to a different class of anti-diabetic medication due to ineffectiveness or intolerance. This may have skewed results to make it appear that new metformin users are less compliant overall. Finally, the study would have benefitted from distinguishing results between different formulations of metformin, such as immediate versus extended-release.
Despite the limitations, the study was able to show that adherence rates are different in new compared to ongoing metformin users, and these results are further impacted by factors such as comorbidities, smoking status, A1c, and BMI. Healthcare professionals can utilize the results of this study by emphasizing the importance of consistent metformin therapy to patients at higher risk of discontinuation, such as those with higher BMIs and A1c 8-8.9%. Future studies occurring over a more extended time and analyzing demographic factors of race and education level could further elucidate trends in metformin therapy adherence, allowing for better generalizability.
- New metformin users and all patients who have obesity or have higher A1c’s are at higher risk of non-compliance with therapy and may require additional counseling on the importance of medication adherence.
- Smokers and those with comorbidities tend to be more adherent to new medications.
- Those who are already on long-term medications are more likely to be adherent to a new therapy, especially if those medications are anti-lipidemic drugs or sulfonylureas.
Tang, Yuexin, et al. “Metformin Adherence and Discontinuation among Patients with Type 2 Diabetes: A Retrospective Cohort Study.“ Journal of Clinical & Translational Endocrinology, vol. 20, April 13. 2020, p. 100225., doi:10.1016/j.jcte.2020.100225.
Sameen Khan, Pharm.D. Candidate, USF College of Pharmacy