The "obesity paradox" in dosing: a comprehensive analytical review of cockcroft-gault versus CKD-EPI for predicting clearance of narrow-therapeutic-index drugs in patients with BMI > 40 kg/m².

BACKGROUND: Class III obesity (body mass index [BMI] ≥ 40 kg/m²) now exceeds 9% prevalence in the United States and is projected to reach nearly one in four adults in several states by 2030. This demographic shift collides with a half-century of drug-dosing infrastructure built largely on Cockcroft-Gault (CG) creatinine clearance estimates that use total body weight and were derived in a small, mostly male, non-obese cohort. OBJECTIVE: To evaluate whether CG (with total body weight [TBW], ideal body weight [IBW], adjusted body weight with a 40% factor [AdjBW40], or lean body weight [LBW]) or the de-indexed Chronic Kidney Disease Epidemiology Collaboration equation (CKD-EPI; 2009 and 2021 race-free) more accurately predicts the clearance of narrow-therapeutic-index (NTI) drugs - vancomycin, aminoglycosides, digoxin, lithium, low-molecular-weight heparins, direct oral anticoagulants and high-dose methotrexate - in patients with BMI > 40 kg/m². Scope is restricted to renal drug clearance for maintenance-dose selection. METHODS: A structured narrative review of original pharmacokinetic studies, prospective and retrospective cohorts compared against measured glomerular filtration rate (mGFR) by iohexol, iothalamate or 24-hour urinary creatinine clearance, and consensus guidelines published from 1976 to 2025. Study-level data were extracted into a structured spreadsheet and harmonised to a common percentage-bias scale; no pooling, meta-analysis or between-equation statistical testing was performed, and all figures are conceptual illustrations of published summary data ("Literature identification, data extraction and figure construction" section). RESULTS: CG using total body weight overestimated absolute renal clearance in class III obesity by approximately 19-34% across the included studies, while CG using ideal body weight underestimated clearance by 15-30%. CG with adjusted body weight (40% factor), CG with lean body weight, and de-indexed CKD-EPI all performed within ± 10-15% of mGFR and were statistically comparable. Adjusted body weight was originally derived by Bauer and colleagues in 1983 to correct the volume of distribution of aminoglycosides rather than clearance; its use as a clearance scalar is therefore empirical rather than mechanistic, and has been widely adopted because it approximates lean body weight and is trivial to compute from routinely available electronic health record variables. Cystatin C-based equations reduced bias further when available. Use of CG-AdjBW40, CG-LBW or de-indexed CKD-EPI rather than CG-TBW was associated with approximately half the proportion of vancomycin AUC₂₄ values exceeding 700 mg·h/L, with acute kidney injury rates paralleling this difference in the largest retrospective analyses. Similar magnitude effects were observed for aminoglycoside trough attainment (AUC/MIC is now the preferred pharmacodynamic target), anti-Xa attainment with low-molecular-weight heparins, and apixaban exposure. CONCLUSIONS: For narrow-therapeutic-index drugs in patients with BMI > 40 kg/m², neither CG using total body weight nor indexed CKD-EPI is appropriate as a stand-alone tool. De-indexed CKD-EPI 2021, CG with lean body weight, and CG with adjusted body weight (40% factor) are the appropriate first-line equations for maintenance-dose renal-function estimation, paired with therapeutic drug monitoring where such monitoring is standard practice. A pragmatic two-tier approach fits the reality of clinical care: a minimal-effort default that any prescriber can apply, and a specialist-level individualisation pathway for clinical-pharmacology and pharmacist-led services. Priorities for future research include prospective head-to-head comparisons of CG with lean body weight versus de-indexed CKD-EPI in specific NTI drug classes, validated cystatin C panels, and bedside-practical mGFR by dried-blood-spot iohexol assays.