Kidney function and clinical recommendations of drug dose adjustment in geriatric patients

BACKGROUND: In elderly patients chronic kidney disease often limits drug prescription. As several equations for quick assessment of kidney function by estimating glomerular filtration rate (eGFR) and several different clinical recommendations for drug dose adjustment in renal failure are published, choosing the correct approach for drug dosage is difficult for the practitioner. The aims of our study were to quantify the agreement between eGFR-equations grouped by creatinine-based or cystatin C-based and within the groups of creatinine and cystatin C-based equations and to investigate whether use of various literature and online references results in different recommendations for drug dose adjustment in renal disease in very elderly primary care patients. METHODS: We included 108 primary care patients aged 80 years and older from 11 family practices into a cross-sectional study. GFR was estimated using two serum creatinine-based equations (Cockroft-Gault, MDRD) and three serum cystatin C-based equations (Grubb, Hoek, Perkins). Concordance between different equations was quantified using intraclass correlation coefficients (ICCs). Essential changes in drug doses or discontinuation of medication were documented and compared in terms of estimated renal function as a consequence of the different eGFR-equations using five references commonly used in the US, Great Britain and Germany. RESULTS: In general, creatinine-based equations resulted in lower eGFR-estimation and in higher necessity of drug dose adjustment than cystatin C-based equations. Concordance was high between creatinine-based equations alone (ICCs 0.87) and between cystatin C-based equations alone (ICCs 0.90 to 0.96), and moderate between creatinine-based equations and cystatin C-based equations (ICCs 0.54 to 0.76). When comparing the five different references consulted to identify necessary drug dose adjustments we found that the numbers of drugs that necessitate dose adjustment in the case of renal impairment differed considerably. The mean number of recommended changes in drug dosage ranged between 1.9 and 2.5 per patient depending on the chosen literature reference. CONCLUSIONS: Our data suggest that the choice of the literature source might have even greater impact on drug management than the choice of the equation used to estimate GFR alone. Efforts should be deployed to standardize methods for estimating kidney function in geriatric patients and literature recommendations on drug dose adjustment in renal failure.