Estimated GFR. Comparison of five algorithms:implications for drug dosing

Area of Science:

• Nephrology
• Pharmacology
• Biostatistics

Background:

• Glomerular filtration rate (GFR) is crucial for drug dosing, especially for medications with narrow therapeutic windows.
• Creatinine-based estimations of GFR (eGFR) are common but introduce uncertainty compared to direct measurements.
• Accurate eGFR requires normalization and individualization using anthropometric data.

Purpose of the Study:

• To compare common algorithms for estimating normalized and individual GFR.
• To assess the feasibility of these algorithms for optimizing drug dosing.
• To evaluate the impact of different eGFR estimation methods on drug dosage calculations.

Main Methods:

• Simulation of four normalized eGFR algorithms and one individual eGFR algorithm.
• Comparison of algorithm outcomes using a simulated Caucasian male profile.
• Interconversion of normalized and individual eGFR using published anthropometric data.

Main Results:

• Differences between eGFR algorithms were within the diagnostic uncertainty range.
• Three of four normalized eGFR algorithms showed discontinuity at extreme creatinine levels.
• Conversion of normalized to individual eGFR underestimated values by 5-40% compared to the Cockcroft-Gault method.

Conclusions:

• Algorithm-based eGFR estimations increase result uncertainty.
• Differences among algorithms are minimal, but interindividual variation poses challenges for global staging.
• Dosage adjustments based on normalized eGFR conversions may lead to underdosing compared to Cockcroft-Gault estimations.