Drug safety and effectiveness in relation to kidney function : a pharmacoepidemiological approach

Sammanfattning: Kidney function plays an important role in drug safety and effectiveness. As many medications are excreted by the kidneys, patients with reduced kidney function are at a higher risk of supra-therapeutic or toxic drug levels. At the same time, drug-induced nephrotoxicity is common due to the high filtration capacity and metabolic activity of the kidneys. Patients with chronic kidney disease (CKD) are at high risk for adverse drug event and drug overdosing. Therefore, randomized controlled trials (RCTs) have generally excluded patients with CKD or included only a small proportion that precludes strong conclusions about the safety and effectiveness in this segment of the population. Pharmacoepidemiological studies performed in real-world settings can help provide complementary evidence and expand findings of RCTs to the general population. However, existing observational studies are often limited in sample size, length of follow-up and inappropriate management of confounding and biases. The presented work aims to expand existing knowledge on drug safety and effectiveness of common cardiovascular and antidiabetic medications used in routine practice and to investigate differences in drug risk-benefit across levels of kidney function. Study I describes the frequency of hyperkalemia in a cohort of new users of mineralocorticoid receptor antagonist (MRA) identified from the Stockholm CREAtinine Measurement (SCREAM) project during 2007-2010. During the 1-year follow-up after treatment initiation, 18% of the patients experienced hyperkalemia in the overall cohort and 26% among patients with heart failure history. After hyperkalemia, 47% of patients discontinued the therapy and only 10% reduced the dose. CKD was common (28%) and it was a major risk factor for both hyperkalemia and MRA discontinuation. Study II examines safety and effectiveness associated with continuing vs stopping MRA treatment after an episode of hyperkalemia in routine care. A cohort of new users of MRA surviving an incident hyperkalemia during 2007-2018 was identified from the SCREAM project. Target trial emulation methods were applied to assess the association between treatment strategies (stopping vs continuing MRA within 6 months after hyperkalemia) and subsequent outcomes. Compared to the “continue MRA” strategy, patients who stopped MRA were at higher risk of cardiovascular events and mortality but lower risk of recurrent hyperkalemia. These associations were consistent across eGFR strata. Study III investigates the cardiovascular effectiveness associated to Glucagon-Like Peptide-1 Receptor Agonists (GLP-1 RAs) use, compared to a standard diabetic care, after an acute myocardial infarction (MI). A cohort of patients with diabetes surviving an acute MI during 2010-2017 were selected from the Swedish Web‐system for Enhancement and Development of Evidence‐Based Care in Heart Disease Evaluated According to Recommended Therapies (SWEDEHEART) registry. Results from the multivariable Cox regression showed a 28% relative risk reduction associated with GLP-1 RAs use compared with standard care. There was no suggestion of effect modification across stages of CKD. Study IV compares the risk of cardiorenal outcomes among patients with non-valvular atrial fibrillation (AF) initiating direct oral anticoagulants (DOAC) or vitamin K antagonists (VKA) treatment. Using data from the SCREAM project, we identified a cohort of patients who started oral anticoagulants (OAC) between 2011 and 2018. Propensity-score weighted Cox regression was used to estimate the treatment-outcomes associations adjusting for 50 measured confounders. Results showed a lower risk of CKD progression, acute kidney injury (AKI) and major bleeding associated with DOAC use compared to VKA treatment. No statistical difference was observed between treatment groups for the composite outcome of stroke/systemic embolism and mortality. The observed associations were mostly similar across levels of baseline kidney function. In conclusion, this work emphasizes the importance of pharmacoepidemiology in expanding trial evidence on the safety and effectiveness of medications in real-world settings. Moreover, this thesis also highlights the important role of kidney function in assessing the risk–benefit of medications.