The Global Leadership Initiative on Malnutrition (GLIM) consensus standardized malnutrition diagnosis, but incorporating low Body Mass Index (BMI) as a phenotypic criterion introduces a circularity bias in weight-stratified analyses and conceals "hidden hunger" in high obesity populations. Using nationally representative data from the U.S. NHANES 1999 2020 (N = 23,363 adults), we propose a "GLIM-adapted" framework that excludes BMI relying solely on weight loss and DXA-derived appendicular lean mass index (ALMI). Standard GLIM identified an 18.9% malnutrition prevalence, but low BMI was the sole indispensable phenotypic driver for only 233 individuals (5.2%). The GLIM-adapted framework maintained a highly comparable prevalence of 17.9% (n = 4,038). Thus, low BMI exhibits low diagnostic sensitivity in modern U.S. cohorts. Excluding BMI successfully bypasses circularity, unmasks qualitative muscle wasting under excess body weight, and establishes a robust, weight-independent standard for clinical and epidemiological research.