Adaptation of drug dosage to kidney function is a common problem in general practice. The aim was to describe adaptation of cardiovascular drugs and metformin according to renal function and its association with mortality with regard to metformin in a cohort of elderly patients. This was an ancillary study to the S.AGES cohort made up of patients over 65 years of age managed by their general practitioner under real-life conditions and followed up prospectively for 3 years. The medications studied were digoxin, spironolactone and metformin. Adaptation of their daily dose according to renal function (eGFR according to CKD/EPI) was compared to that recommended in the summaries of product characteristics (SPCs) or international scientific societies (ISS). A total of 900 patients were included, including 588 on metformin. At baseline, dose adjustment according to renal function was 100% and 87.6% (95% CI: 82.6–92.6) for patients on digoxin and spironolactone respectively. For metformin, only 71.3% (95% CI: 67.6–74.9) or 78.1% (95% CI: 74.7–81.4) of patients had their dosage adapted at inclusion according to their renal function depending on whether the SPCs or ISS recommendations were considered. During the 3-year follow-up period, 42/588 patients died (none from lactic acidosis). At inclusion, a metformin dosage not adapted for renal function according to ISS was not associated with an increase in all-cause mortality (OR 1.7; 95% CI 0.6–5.0, p = 0.32). In conclusion, approximately one-quarter of elderly patients treated with metformin do not have their dosage adapted for renal function according to ISS although there is no increase in mortality after follow-up for 3 years.