Diagnostic Agreement of GLIM Criteria with SGA for Malnutrition in Non-Dialysis Chronic Kidney Disease Authors: Shuai Chen, Jing Xue, Kaidi Zhang, Yicun Xu, Lina Yuan, Caixia Liu, Baoli Zhao, Lihua Zhao Published: 2026 ABSTRACT AbstractBackground and ObjectiveMalnutrition significantly contributes to the adverse prognosis of patients with chronic kidney disease (CKD). The Global Leadership Initiative on Malnutrition (GLIM) criteria were introduced to standardize malnutrition diagnosis globally. This study evaluates the diagnostic performance and agreement of the GLIM criteria compared with the Subjective Global Assessment (SGA) in non-dialysis CKD patients.MethodsIn this cross-sectional study, 248 non-dialysis CKD patients (stages G1-G5) were screened using the Nutritional Risk Screening 2002 (NRS-2002). Patients with a score <3 were classified as well-nourished. Those with a score ≥3 underwent GLIM assessment incorporating phenotypic and etiologic parameters. CKD was operationalized as fulfilling the GLIM disease-burden/inflammation etiologic criterion; therefore, GLIM-based prevalence was interpreted in light of this assumption. Concurrently, all patients were assessed using SGA (the reference standard). Diagnostic agreement was evaluated using Cohen’s Kappa, sensitivity, specificity, and the area under the receiver operating characteristic curve (AUC). Multivariable logistic regression identified factors independently associated with GLIM-diagnosed malnutrition.ResultsThe prevalence of malnutrition was 46.8% by SGA and 49.2% by GLIM. The GLIM framework demonstrated substantial agreement with SGA (Kappa = 0.812, P < 0.001). The GLIM criteria exhibited a sensitivity of 87.9%, a specificity of 84.8%, and an AUC of 0.864 (95% CI: 0.818–0.910). Malnutrition prevalence increased progressively with advanced CKD stages (P < 0.001). Multivariable analysis revealed that advanced CKD stage, older age, and lower serum albumin were factors independently associated with GLIM-diagnosed malnutrition.ConclusionThe GLIM criteria demonstrate good diagnostic agreement with the SGA in assessing malnutrition among non-dialysis CKD patients. While GLIM provides a standardized and structured framework, longitudinal studies are required to validate its predictive value for clinical outcomes.Keywords: global leadership initiative on malnutrition, subjective global assessment, nutritional risk screening 2002, chronic kidney disease, malnutrition, diagnostic performance Introduction Chronic kidney disease (CKD) is increasingly recognized as a major global public health concern. Recent epidemiological data suggest that the worldwide prevalence of CKD in the adult population is approximately 9% to 13%, translating to over 800 million affected individuals. The leading etiologies include diabetic nephropathy, hypertensive nephrosclerosis, and primary glomerulonephritis.1,2 As the global burden of non-communicable diseases escalates, CKD-associated morbidity and premature mortality are projected to become the fifth leading cause of death by 2040.3 Among the multifaceted complications of progressive renal impairment, impaired nutritional status stands out as a critical determinant of patient prognosis.4 Patients with CKD are exceptionally vulnerable to systemic malnutrition. While historically described under the umbrella of Protein-Energy Wasting (PEW) by the International Society of Renal Nutrition and Metabolism (ISRNM),5 it is crucial to recognize that standard malnutrition and PEW are related but conceptually distinct clinical entities. Malnutrition in CKD is multifactorial, driven by pathophysiological mechanisms such as uremic toxicity, which suppresses appetite, and metabolic acidosis alongside insulin resistance, which accelerate somatic protein degradation.6,7 Furthermore, gut dysbiosis and chronic low-grade inflammation synergistically disrupt nutrient assimilation and promote a hypercatabolic state.8 Unlike simple starvation, which can often be reversed by adequate caloric intake, the hypercatabolic state inherent to CKD-associated malnutrition fundamentally accelerates disease progression, exacerbates susceptibility to severe infections, heightens the risk of cardiovascular events, and significantly amplifies overall mortality rates, particularly as patients transition into advanced CKD stages.9 Therefore, early identification and timely clinical intervention for compromised nutritional status in the non-dialysis phase are imperative to mitigate these adverse outcomes and preserve residual renal function before renal replacement therapy becomes mandatory. Considerable heterogeneity exists in the reported prevalence of malnutrition among CKD cohorts, spanning from 11% to over 60%, a discrepancy largely attributable to the wide variation in the screening and assessment tools employed across different clinical settings.10 To accurately define PEW and establish an optimal nutritional care pathway, consecutive screening and rigorous diagnostic assessments are mandatory. For decades, the Subjective Global Assessment (SGA) has been widely supported in renal nutrition practice and is included in the KDOQI Clinical Practice Guideline for Nutrition in CKD11 as a clinically useful tool for nutritional assessment. The SGA relies heavily on patient history and physical examination, capturing weight changes, dietary intake, gastrointestinal symptoms, and observable muscle or fat wasting. While robust, the SGA has been criticized for its inherent subjectivity and strong dependence on the clinician’s expertise, which may limit its inter-rater reliability and widespread integration into routine automated workflows.12 To overcome the lack of global consensus on the diagnostic criteria for malnutrition, four leading international clinical nutrition societies (ASPEN, ESPEN, FELANPE, and PENSA) collaboratively introduced the Global Leadership Initiative on Malnutrition (GLIM) framework in 2018.13 The GLIM criteria represent a standardized approach by establishing a universal, two-step diagnostic model: a mandatory initial nutritional risk screening utilizing validated tools such as the Nutritional Risk Screening 2002 (NRS-2002), followed by a comprehensive assessment identifying at least one phenotypic criterion (non-volitional weight loss, low body mass index [BMI], or reduced muscle mass) and at least one etiologic criterion (reduced food intake/assimilation or underlying disease burden/inflammation).14 Recently, studies have evaluated GLIM in dialysis-dependent and non-dialysis CKD cohorts; however, the reported performance has been heterogeneous, with agreement varying according to the reference tool, nutritional screening instrument, and the operational definition of reduced muscle mass. Therefore, evidence validating GLIM specifically against established clinical tools such as SGA in the broader, non-dialysis CKD population remains limited and occasionally conflicting with regard to diagnostic agreement and cut-off sensitivity.15,16 To bridge this substantial literature gap, the primary objective of this cross-sectional study was to robustly evaluate the diagnostic performance and agreement of the GLIM criteria, utilizing the NRS-2002 for the initial screening phase, with the SGA serving as the reference standard in an adult population with non-dialysis CKD. Furthermore, we aimed to investigate clinical factors independently associated with GLIM-diagnosed malnutrition. We hypothesized that the GLIM framework would show good diagnostic agreement with SGA and provide a structured diagnostic approach for identifying malnutrition in chronic kidney disease. The study was conducted and reported in strict accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.17 Materials and Methods Study Design and Setting This observational, single-center, cross-sectional study was prospectively designed and conducted at the nephrology inpatient unit of the First Hospital of Hebei Medical University, Shijiazhuang, China, between July 2023 and December 2024. The study protocol adhered to the ethical principles elucidated in the Declaration of Helsinki. Prior to study initiation, formal approval was granted by the Institutional Ethics Review Board of the First Hospital of Hebei Medical University (Approval No. S00438 [2023]). Written informed consent was systematically obtained from all participating patients or their legally authorized representatives prior to data collection. Sample Size Calculation To ensure sufficient statistical power for evaluating the diagnostic agreement (Cohen’s Kappa) between GLIM and SGA, an a priori sample size calculation was performed. Based on previous literature in non-dialysis CKD populations,16 estimating an anticipated malnutrition prevalence of 45% in non-dialysis CKD patients, a projected Kappa value of 0.75, a null hypothesis Kappa of 0.50, an alpha error of 0.05, and a power (1-β) of 0.90, the minimum required sample size was calculated as 215 patients. To account for a potential non-response or incomplete data rate of approximately 15%, the recruitment target was appropriately expanded to include at least 245 patients. Participants Patients were consecutively screened for eligibility upon admission to the nephrology ward. Inclusion Criteria: (1) Adult patients aged 18 to 85 years; (2) A confirmed diagnosis of CKD stage G1-G5 based on the KDIGO clinical practice guidelines (defined by structural or functional kidney abnormalities lasting >3 months, or an estimated glomerular filtration rate [eGFR] < 60 mL/min/1.73m2); (3) Not currently undergoing or having a history of maintenance renal replacement therapy (hemodialysis or peritoneal dialysis); (4) Fully conscious and cognitively capable of understanding. Exclusion Criteria: (1) Acute kidney injury or rapidly progressive glomerulonephritis requiring acute intervention; (2) Concomitant presence of active malignancies, hyperthyroidism, decompensated liver cirrhosis, or advanced heart failure (NYHA Class IV), which could independently severely alter nutritional status; (3) Patients who had already initiated enteral or parenteral nutrition support within four weeks prior to admission; (4) Severe limb edema, ascites, or amputations that technically precluded accurate anthropometric measurements; (5) Refusal to participate or incomplete clinical data. Clinical and Laboratory Data Collection A standardized electronic case report form was utilized to collect comprehensive baseline demographic and clinical data within 24 hours of hospital admission. Collected variables included age, sex, smoking history, alcohol consumption, primary etiology of CKD (diabetic nephropathy, hypertensive nephropathy, primary glomerulonephritis, or others), and duration of the disease. Fasting venous blood samples were drawn on the morning following admission. Standardized laboratory parameters, specifically renal profile indicators and nutritional biomarkers, were assessed using a fully automated biochemistry analyzer (Beckman Coulter AU5800, USA), including serum albumin (g/L), total protein (g/L), prealbumin (mg/L), and serum creatinine (μmol/L). The eGFR was meticulously calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) 2021 equation without the race variable,18 and patients were subsequently stratified into KDIGO stages G1 to G5. Anthropometric Measurements Anthropometric parameters were rigorously obtained by two specifically trained clinical dietitians to ensure inter-rater reliability. Body weight (kg) and height (cm) were measured with patients wearing light hospital clothing and barefoot, using a calibrated digital scale and stadiometer. Body mass index (BMI) was calculated as weight divided by the square of height (kg/m2). Calf circumference (CC), serving as a surrogate marker for muscle mass, was measured on the non-dominant leg with the patient in a relaxed sitting position, with the knee bent at a 90-degree angle. A non-stretchable measuring tape was placed horizontally around the maximum circumference of the calf without compressing the subcutaneous tissue.19 To minimize the confounding effect of fluid overload, patients with noticeable peripheral edema were excluded from the primary cohort as per the criteria. Nutritional Assessment Protocols Subjective Global Assessment (SGA) - The Reference Standard Within 48 hours of admission, a senior clinical nutritionist, blinded to the GLIM and NRS-2002 results, conducted the SGA. The SGA assesses five historical elements (weight change, dietary intake change, gastrointestinal symptoms, functional capacity, and metabolic demands of the underlying disease) and three physical examination components (loss of subcutaneous fat, muscle wasting, and presence of ankle/sacral edema). Based on these collective findings, patients were classified into three categories: SGA-A (well-nourished), SGA-B (mild to moderately malnourished), and SGA-C (severely malnourished). For the purpose of dichotomous statistical analysis, patients categorized as SGA-B or SGA-C were collectively merged and defined as having malnutrition.20 Step 1: Nutritional Risk Screening 2002 (NRS-2002) Conducted within the first 24 hours of admission, the NRS-2002 evaluates three components: impaired nutritional status (0–3 points), severity of disease (0–3 points, with CKD generally yielding 1–2 points depending on acute exacerbation), and an age adjustment (+1 point for patients ≥70 years). A cumulative score of ≥3 defines the patient as being “at nutritional risk,” mandating further assessment.21 Crucially, patients with an NRS-2002 score <3 bypassed the second step and were automatically classified as GLIM-negative (well-nourished). Step 2: GLIM Diagnostic Framework Only patients identified as having nutritional risk (NRS-2002 ≥ 3) proceeded to the GLIM diagnosis phase. According to the GLIM consensus,13 a positive diagnosis of malnutrition required the fulfillment of at least one phenotypic criterion and at least one etiologic criterion. Phenotypic criteria included: (a) Unintentional weight loss: >5% within the past 6 months, or >10% beyond 6 months; (b) Low BMI: <18.5 kg/m2 for patients <70 years old, or <20.0 kg/m2 for those ≥70 years old (applying Asian-specific BMI cut-offs as recommended); (c) Reduced muscle mass: Objectively defined by a calf circumference (CC) of <34 cm in males and <33 cm in females, in accordance with the established Asian Working Group for Sarcopenia (AWGS) 2019 criteria.22 Etiologic criteria included: (a) Reduced food intake or assimilation: Dietary intake ≤50% of requirements for >1 week, or any reduction for >2 weeks; (b) Disease burden or systemic inflammation: The presence of CKD was operationalized as fulfilling the GLIM disease-burden/inflammation etiologic criterion because CKD is a chronic disease frequently accompanied by low-grade inflammation. However, objective inflammatory biomarkers such as C-reactive protein or interleukin-6 were not routinely available in this cohort; therefore, the potential for diagnostic inflation related to this operational decision was specifically addressed in the Discussion. Statistical Analysis All statistical analyses were executed using SPSS Statistics for Windows, Version 26.0 (IBM Corp., Armonk, NY, USA) and R software (Version 4.2.1, R Foundation for Statistical Computing, Vienna, Austria). A complete case analysis approach was utilized as missing clinical data accounted for <2% of variables. Normality of continuous variables was assessed employing the Kolmogorov–Smirnov test. Normally distributed continuous variables were expressed as mean ± standard deviation (SD) and compared using independent sample Student’s t-tests or Analysis of Variance (ANOVA) followed by Bonferroni post-hoc analysis. Non-normally distributed variables were presented as median (interquartile range, IQR) and analyzed using the Mann–Whitney U-test. Categorical variables were presented as frequencies and percentages, with group differences evaluated using the Pearson Chi-square (χ2) test or Fisher’s exact test where appropriate. To evaluate diagnostic agreement, Cohen’s Kappa (κ) coefficient was calculated between the GLIM criteria and the SGA reference standard. Kappa values were interpreted as follows: <0.20 slight, 0.21–0.40 fair, 0.41–0.60 moderate, 0.61–0.80 substantial, and 0.81–1.00 almost perfect agreement.23 Diagnostic performance metrics, including Sensitivity, Specificity, Positive Predictive Value (PPV), and Negative Predictive Value (NPV) with their respective 95% Confidence Intervals (CI), were calculated. A Receiver Operating Characteristic (ROC) curve was plotted, and the Area Under the Curve (AUC) with 95% CI was determined to quantify diagnostic accuracy. Furthermore, univariable logistic regression was initially performed to identify potential clinical factors associated with GLIM-diagnosed malnutrition. Variables exhibiting a P-value <0.10 in the univariable analysis were subsequently incorporated into a multivariable logistic regression model using a forward stepwise selection method to identify factors independently associated with GLIM-diagnosed malnutrition. Variance Inflation Factor (VIF) was evaluated to rule out multicollinearity prior to multivariable regression. Results were expressed as Odds Ratios (OR) with 95% CI. All statistical tests were two-tailed, and a P-value of <0.05 was considered statistically significant. Results Baseline Demographic and Clinical Characteristics During the study period, a total of 280 patients were initially screened. Following the application of inclusion and exclusion criteria, 248 non-dialysis patients with CKD were formally enrolled in this study (Figure 1). The final cohort comprised 152 males (61.3%) and 96 females (38.7%), with an average age of 52.4 ± 12.1 years (range: 24 to 82 years). The mean body mass index (BMI) was 22.8 ± 3.5 kg/m2. The primary etiologies of CKD included diabetic nephropathy in 95 patients (38.3%), primary glomerulonephritis in 82 patients (33.1%), and hypertensive nephropathy in 54 patients (21.8%). According to the KDIGO staging system, the distribution of patients was as follows: Stage G1, 28 (11.3%); Stage G2, 54 (21.8%); Stage G3, 67 (27.0%); Stage G4, 54 (21.8%); and Stage G5, 45 (18.1%). The median duration of disease was 14.5 months (IQR: 6.0–36.0). Detailed baseline demographic, clinical, and laboratory characteristics of the study population are summarized in Table 1. Study flow diagram illustrating the patient screening, inclusion, and exclusion process in accordance with the STROBE reporting guidelines. The dotted arrow indicates that all enrolled patients, including those not at nutritional risk by NRS-2002, underwent SGA reference standard assessment. Table 1Baseline Demographic and Clinical Characteristics of Non-Dialysis Chronic Kidney Disease Patients (N=248)CharacteristicsTotal (N = 248)Well-Nourished by GLIM (n = 126)Malnourished by GLIM (n = 122)P valueAge (years), mean ± SD52.4 ± 12.148.3 ± 12.156.8 ± 10.5<0.001Sex, n (%)0.245 Male152 (61.3)72 (57.1)80 (65.6) Female96 (38.7)54 (42.9)42 (34.4)BMI (kg/m2), mean ± SD22.8 ± 3.524.5 ± 2.821.0 ± 3.2<0.001Smoking status, n (%)38 (15.3)16 (12.7)22 (18.0)0.241Alcohol consumption, n (%)24 (9.7)10 (7.9)14 (11.5)0.347CKD Etiology, n (%)0.562 Diabetic nephropathy95 (38.3)45 (35.7)50 (41.0) Primary glomerulonephritis82 (33.1)45 (35.7)37 (30.3) Hypertensive nephropathy54 (21.8)27 (21.4)27 (22.1) Others17 (6.9)9 (7.1)8 (6.6)CKD Stage (KDIGO), n (%)<0.001 Stage G128 (11.3)24 (19.0)4 (3.3) Stage G254 (21.8)41 (32.5)13 (10.7) Stage G367 (27.0)39 (31.0)28 (23.0) Stage G454 (21.8)10 (7.9)44 (36.1) Stage G545 (18.1)12 (9.5)33 (27.0)Disease duration (months), median (IQR)14.5 (6.0–36.0)10.0 (5.0–24.0)24.0 (12.0–48.0)<0.001Serum albumin (g/L), mean ± SD38.0 ± 5.140.5 ± 4.235.4 ± 4.6<0.001Prealbumin (mg/L), mean ± SD248.5 ± 58.4265.3 ± 51.2231.1 ± 60.5<0.001Notes: Continuous variables with normal distribution are presented as mean ± standard deviation (SD) and were analyzed using the independent Student’s t-test. Non-normally distributed variables are presented as median (interquartile range, IQR) and were analyzed using the Mann–Whitney U-test. Categorical variables are presented as frequencies (n) and percentages (%) and were analyzed using the Pearson Chi-square (χ2) test.Abbreviations: BMI, Body Mass Index; CKD, Chronic Kidney Disease; GLIM, Global Leadership Initiative on Malnutrition; KDIGO, Kidney Disease: Improving Global Outcomes. Baseline Demographic and Clinical Characteristics of Non-Dialysis Chronic Kidney Disease Patients (N=248) Notes: Continuous variables with normal distribution are presented as mean ± standard deviation (SD) and were analyzed using the independent Student’s t-test. Non-normally distributed variables are presented as median (interquartile range, IQR) and were analyzed using the Mann–Whitney U-test. Categorical variables are presented as frequencies (n) and percentages (%) and were analyzed using the Pearson Chi-square (χ2) test. Abbreviations: BMI, Body Mass Index; CKD, Chronic Kidney Disease; GLIM, Global Leadership Initiative on Malnutrition; KDIGO, Kidney Disease: Improving Global Outcomes. Prevalence of Nutritional Risk and Malnutrition Following the initial screening utilizing the NRS-2002 tool, 142 patients (57.3%) scored ≥3, indicating they were at nutritional risk and thereby required further formal diagnostic assessment. Using the SGA as the established reference standard, 116 patients (46.8%) were diagnosed with malnutrition (combined SGA-B and SGA-C). Concurrently, when applying the two-step GLIM framework, 122 patients (49.2%) were diagnosed with malnutrition. Regarding severity categories, among the 116 SGA-defined malnourished patients, 85 were classified as SGA-B and 31 as SGA-C. According to GLIM severity grading, 89 patients had moderate malnutrition and 33 had severe malnutrition. These severity distributions are summarized in Supplementary Table 1. Notably, the prevalence of malnutrition progressively and significantly increased with advancing CKD stages regardless of the diagnostic tool applied. For instance, based on the GLIM criteria, malnutrition prevalence escalated from 14.3% in CKD Stage G1 to a striking 73.3% in CKD Stage G5 (Cochran-Armitage trend test, P < 0.001), as illustrated in Figure 2. Furthermore, malnourished patients were significantly older (56.8 ± 10.5 vs 48.3 ± 12.1 years, P < 0.001) and presented with significantly lower serum albumin (35.4 ± 4.6 vs 40.5 ± 4.2 g/L, P < 0.001), prealbumin (231.1 ± 60.5 vs 265.3 ± 51.2 mg/L, P < 0.001), and BMI (21.0 ± 3.2 vs 24.5 ± 2.8 kg/m2, P < 0.001) compared to their well-nourished counterparts. Prevalence of malnutrition evaluated by the Global Leadership Initiative on Malnutrition (GLIM) criteria and Subjective Global Assessment (SGA) across different Chronic Kidney Disease (CKD) stages (G1–G5). Notes: The prevalence of malnutrition showed a statistically significant increasing trend as CKD stage advanced (Cochran-Armitage trend test, P < 0.001). Diagnostic Performance and Agreement Between GLIM and SGA A cross-tabulation of malnutrition diagnoses formulated by the GLIM criteria versus the SGA revealed a substantial overlap. Of the 116 patients diagnosed with malnutrition by SGA, 102 were correctly identified by the GLIM criteria. Conversely, of the 132 patients deemed well-nourished by SGA, 112 were correctly classified as such by the GLIM framework. Cohen’s Kappa statistic demonstrated an almost perfect agreement between the GLIM criteria and the SGA reference standard (Kappa = 0.812, 95% CI: 0.745–0.879, P < 0.001). The cross-tabulation and overall diagnostic performance metrics are detailed in Table 2. Table 2Diagnostic Performance and Agreement Between GLIM Criteria and Subjective Global Assessment (SGA)GLIM CriteriaSGA (Reference Standard)TotalDiagnostic MetricsValue (95% CI)MalnourishedWell-NourishedMalnourished10220122Sensitivity87.9% (80.6–93.2%)Well-nourished14112126Specificity84.8% (77.6–90.5%)Total116132248PPV83.6% (75.8–89.7%)NPV88.9% (82.1–93.8%)Cohen’s Kappa0.812 (0.745–0.879)P value (Kappa)<0.001Notes: Malnourished status by SGA combines SGA-B (mild-to-moderate) and SGA-C (severe). The GLIM diagnosis was established following positive NRS-2002 screening (≥3 points) as the first step. Exact P-values < 0.05 denote statistical significance.Abbreviations: CI, Confidence Interval; GLIM, Global Leadership Initiative on Malnutrition; NPV, Negative Predictive Value; PPV, Positive Predictive Value; SGA, Subjective Global Assessment. Diagnostic Performance and Agreement Between GLIM Criteria and Subjective Global Assessment (SGA) Notes: Malnourished status by SGA combines SGA-B (mild-to-moderate) and SGA-C (severe). The GLIM diagnosis was established following positive NRS-2002 screening (≥3 points) as the first step. Exact P-values < 0.05 denote statistical significance. Abbreviations: CI, Confidence Interval; GLIM, Global Leadership Initiative on Malnutrition; NPV, Negative Predictive Value; PPV, Positive Predictive Value; SGA, Subjective Global Assessment. When evaluating diagnostic performance using SGA as the reference standard, the GLIM framework achieved a sensitivity of 87.9% (95% CI: 80.6–93.2%) and a specificity of 84.8% (95% CI: 77.6–90.5%). The Positive Predictive Value (PPV) was 83.6% (95% CI: 75.8–89.7%), and the Negative Predictive Value (NPV) was 88.9% (95% CI: 82.1–93.8%). Binary ROC analysis further indicated good diagnostic accuracy of the GLIM criteria, yielding an Area Under the Curve (AUC) of 0.864 (95% CI: 0.818–0.910, P < 0.001) (Figure 3). Subgroup analyses demonstrated that GLIM maintained robust diagnostic consistency across different clinical profiles. Specifically, sensitivity and specificity remained high irrespective of age categories (<60 vs ≥60 years), sex, and early (G1-G3) versus advanced (G4-G5) CKD stages, with detailed subgroup performance metrics comprehensively presented in Table 3. Receiver Operating Characteristic (ROC) curve evaluating the diagnostic accuracy of the GLIM criteria using the Subjective Global Assessment (SGA) as the reference standard. The red dot marks the observed diagnostic operating point for GLIM relative to SGA (sensitivity 87.9%, specificity 84.8%); the dashed diagonal line indicates chance-level discrimination. Table 3Subgroup Diagnostic Performance of the GLIM Criteria Compared with SGASubgroupsNSensitivity, % (95% CI)Specificity, % (95% CI)AUC (95% CI)KappaAge Categories <60 years17085.4% (75.8–92.2)86.7% (78.6–92.5)0.860 (0.803–0.917)0.724 ≥60 years7891.3% (79.2–97.5)81.2% (63.5–92.7)0.863 (0.768–0.958)0.741Sex Male15288.6% (79.5–94.7)85.5% (75.6–92.5)0.871 (0.812–0.929)0.742 Female9686.7% (73.2–94.9)83.6% (71.2–92.2)0.852 (0.771–0.932)0.704CKD Stage Early (G1–G3)14981.8% (64.5–93.0)86.8% (79.2–92.4)0.843 (0.762–0.925)0.627 Advanced (G4–G5)9990.4% (81.9–95.7)75.0% (50.9–91.3)0.827 (0.725–0.928)0.669Notes: Subgroup Area Under the Curve (AUC) and Kappa statistics demonstrate consistent diagnostic performance across varying demographic and clinical profiles. SGA was utilized as the reference standard.Abbreviations: AUC, Area Under the Curve; CI, Confidence Interval; CKD, Chronic Kidney Disease; GLIM, Global Leadership Initiative on Malnutrition; SGA, Subjective Global Assessment. Subgroup Diagnostic Performance of the GLIM Criteria Compared with SGA Notes: Subgroup Area Under the Curve (AUC) and Kappa statistics demonstrate consistent diagnostic performance across varying demographic and clinical profiles. SGA was utilized as the reference standard. Abbreviations: AUC, Area Under the Curve; CI, Confidence Interval; CKD, Chronic Kidney Disease; GLIM, Global Leadership Initiative on Malnutrition; SGA, Subjective Global Assessment. Clinical Factors Independently Associated with GLIM-Diagnosed Malnutrition To systematically investigate factors independently associated with malnutrition (as defined by the GLIM criteria) in non-dialysis CKD patients, univariable and multivariable logistic regression analyses were conducted. The results are presented in Table 4. In the univariable analysis, variables such as age, disease duration, advanced CKD stage (G4-G5 vs G1-G3), serum albumin, and prealbumin were significantly associated with the presence of malnutrition (all P < 0.05). After adjusting for potential confounders in the multivariable stepwise forward logistic regression model, advanced CKD stage (G4-G5) showed the strongest independent association (OR = 2.14, 95% CI: 1.45–3.15, P < 0.001). Additionally, older age (adjusted OR = 1.05 per year, 95% CI: 1.01–1.09, P = 0.021) and lower serum albumin (adjusted OR = 1.14 per 1 g/L decrease, 95% CI: 1.05–1.22, P = 0.002) were independently associated with GLIM-diagnosed malnutrition. These associated factors and their corresponding effect sizes are visually depicted in the forest plot (Figure 4). Table 4Univariable and Multivariable Logistic Regression Analysis of Factors Associated with GLIM-Diagnosed MalnutritionVariablesUnivariable AnalysisMultivariable AnalysisOR (95% CI)P valueAdjusted OR (95% CI)P valueAge (per 1-year increase)1.07 (1.04–1.10)<0.0011.05 (1.01–1.09)0.021Sex (Male vs Female)1.43 (0.84–2.45)0.187––BMI (per 1 kg/m2 decrease)1.41 (1.26–1.58)<0.001––Disease duration (≥12 vs <12 months)2.05 (1.23–3.42)0.0061.32 (0.78–2.24)0.301CKD Stage (G4–G5 vs G1–G3)4.58 (2.61–8.04)<0.0012.14 (1.45–3.15)<0.001Serum albumin (per 1 g/L decrease)1.25 (1.16–1.35)<0.0011.14 (1.05–1.22)*0.002Prealbumin (per 10 mg/L decrease)1.12 (1.06–1.18)<0.0011.03 (0.95–1.12)0.458Notes: Variables with a P value < 0.10 in the univariable analysis were included in the multivariable stepwise forward logistic regression model. BMI was excluded from the multivariable model as it is already an intrinsic phenotypic component of the GLIM criteria (to avoid mathematical coupling/multicollinearity). *Serum albumin was modeled per 1 g/L decrease; therefore, an adjusted OR >1 indicates increased odds of GLIM-diagnosed malnutrition with lower albumin. Variance Inflation Factor (VIF) values for all variables were < 2.5, indicating no multicollinearity.Abbreviations: BMI, Body Mass Index; CI, Confidence Interval; CKD, Chronic Kidney Disease; OR, Odds Ratio. Univariable and Multivariable Logistic Regression Analysis of Factors Associated with GLIM-Diagnosed Malnutrition Notes: Variables with a P value < 0.10 in the univariable analysis were included in the multivariable stepwise forward logistic regression model. BMI was excluded from the multivariable model as it is already an intrinsic phenotypic component of the GLIM criteria (to avoid mathematical coupling/multicollinearity). *Serum albumin was modeled per 1 g/L decrease; therefore, an adjusted OR >1 indicates increased odds of GLIM-diagnosed malnutrition with lower albumin. Variance Inflation Factor (VIF) values for all variables were < 2.5, indicating no multicollinearity. Abbreviations: BMI, Body Mass Index; CI, Confidence Interval; CKD, Chronic Kidney Disease; OR, Odds Ratio. Forest plot displaying the results of the multivariable logistic regression analysis for factors independently associated with GLIM-diagnosed malnutrition in non-dialysis patients with chronic kidney disease. Notes: Squares indicate adjusted odds ratio point estimates, and horizontal lines indicate 95% Confidence Intervals (CI). Odds ratios for age and serum albumin are modeled per 1-year increase and per 1 g/L decrease, respectively. Discussion Malnutrition represents a pervasive and insidious complication in the natural trajectory of chronic kidney disease (CKD), significantly impairing physical function and magnifying cardiovascular and all-cause mortality. The cornerstone of reversing this pathological decline lies in early, accurate, and standardized nutritional assessment. The present cross-sectional study is among the few rigorously designed investigations to evaluate the novel two-step GLIM framework—incorporating the NRS-2002 as the mandated screening phase—against the nephrology reference standard Subjective Global Assessment (SGA) in a substantial cohort of non-dialysis CKD patients. Our findings demonstrated that the GLIM criteria showed good diagnostic accuracy (AUC = 0.864) and substantial-to-almost-perfect agreement (Kappa = 0.812) with SGA. These results suggest that GLIM may provide a standardized framework for identifying malnutrition in hospitalized non-dialysis CKD patients, although external validation is required before broader generalization. In our cohort of 248 non-dialysis CKD patients, the prevalence of malnutrition was markedly high, reaching 46.8% via SGA and 49.2% via GLIM. This aligns closely with contemporary epidemiological data indicating that malnutrition is common in CKD cohorts, although prevalence estimates vary across disease stages, clinical settings, and diagnostic criteria.24,25 We observed a striking and continuous escalation in malnutrition prevalence commensurate with the deterioration of renal function, culminating in over 70% of patients in CKD Stage G5 being classified as malnourished. Beyond the physiological impact of uremia, the high prevalence of malnutrition observed in our cohort may also be influenced by dietary, socio-economic, and lifestyle-related factors. Restrictive dietary advice frequently given to CKD patients to control potassium, phosphorus, sodium, and protein intake may unintentionally reduce overall caloric and protein adequacy if not individualized. In addition, food insecurity and limited access to renal-appropriate foods or timely nutrition counselling may delay early intervention, allowing subclinical wasting to progress unchecked.26 It is important to conceptually distinguish the GLIM framework from the traditional ISRNM criteria for PEW. While PEW uses kidney-specific diagnostic domains, including biochemical indicators, reduced body mass, reduced muscle mass, and reduced dietary intake, to capture protein and energy depletion in acute or chronic kidney disease, GLIM provides a broader, universal diagnosis of malnutrition applicable across various chronic diseases.5 The SGA has long been supported in renal nutrition practice and is included in the KDOQI Clinical Practice Guideline for Nutrition in CKD11 as a clinically useful tool for assessing nutritional status. However, the SGA inherently relies on subjective clinical judgment regarding historical weight changes and physical signs of muscle/fat wasting, demanding specialized training and rendering it susceptible to inter-observer variability.12 In light of the heterogeneous findings reported in dialysis cohorts, our results should be interpreted as specific to non-dialysis CKD patients. Nevertheless, broadly in line with the acceptable diagnostic performance reported for adapted GLIM against SGA in hemodialysis patients by Aziz et al27 our data revealed a sensitivity of 87.9% and specificity of 84.8% for GLIM relative to SGA, confirming that GLIM accurately captured the vast majority of true malnutrition cases without producing excessive false positives. Furthermore, our multivariable logistic regression analysis reinforced the profound interplay between clinical parameters and nutritional decline. Advanced CKD stage, older age, and lower serum albumin were identified as factors independently associated with malnutrition. Older age may increase vulnerability to CKD-associated malnutrition through age-related loss of muscle mass, reduced appetite, multimorbidity, and inflammatory burden.28 Notably, lower serum albumin was also independently associated with GLIM-diagnosed malnutrition. Although serum albumin is influenced by inflammation, hydration status, and comorbidity rather than nutritional intake alone, its independent association with GLIM-diagnosed malnutrition may reflect the combined effects of inflammatory burden and phenotypic wasting in CKD.29 Additionally, as eGFR declines and serum creatinine rises (key renal profile indicators), the retention of uremic solutes directly suppresses appetite and alters gut microbiome integrity, accelerating this nutritional deterioration.30 Despite these findings, several limitations within our study warrant careful consideration. First, as a single-center, cross-sectional study, we cannot infer strict causality between clinical variables and the onset of malnutrition, nor can we assess the impact of GLIM-diagnosed malnutrition on longitudinal clinical outcomes such as mortality or time to dialysis initiation. Second, to comply with the GLIM etiologic framework, the presence of CKD was automatically considered a chronic disease burden. While methodologically justifiable, this automatic fulfillment may introduce a degree of diagnostic inflation, potentially overestimating malnutrition prevalence in early-stage (G1-G2) patients compared to advanced stages. Third, the lack of objective inflammatory biomarkers (such as CRP or IL-6) in our routine data collection slightly weakens the precision of the etiologic assessment. Fourth, the GLIM criteria mandate the assessment of reduced muscle mass. While dual-energy X-ray absorptiometry (DEXA) or bioelectrical impedance analysis (BIA) are considered optimal, these are not always feasible in resource-constrained settings. We utilized calf circumference (CC) as a validated surrogate. While pragmatic for daily clinical practice, CC is a limited surrogate compared to advanced imaging, and subclinical edema prevalent in CKD could theoretically overestimate CC, masking true muscle depletion.22 Furthermore, inter-rater reliability for anthropometric measurements and SGA was not formally evaluated. Future multi-center, prospective cohort studies incorporating advanced imaging modalities and specific longitudinal outcome validation are imperative to fine-tune the optimal phenotypic cut-offs for the CKD demographic. Conclusion In conclusion, our study demonstrates that the Global Leadership Initiative on Malnutrition (GLIM) framework exhibits substantial diagnostic agreement with the reference standard Subjective Global Assessment (SGA) for identifying malnutrition in non-dialysis chronic kidney disease patients. Because GLIM uses predefined phenotypic and etiologic parameters, it may reduce—but cannot eliminate—the subjectivity inherent in traditional clinical assessment and may be considered for structured nutritional evaluation in nephrology settings. However, the present findings should be interpreted in light of the automatic fulfillment of the etiologic criterion by CKD and the single-center hospitalized cohort design. While it offers a standardized and structured approach, future multicenter and longitudinal validation studies are required to assess its prognostic value for hard clinical outcomes and to clarify its relationship with specific ISRNM PEW criteria. Funding Statement This study was financially supported by Medical Science Research Project of Hebei (Grant No. 20241957). The funding body had no role in the design of the study, data collection, analysis, interpretation of data, or in writing the manuscript. Abbreviations AUC, Area Under the Curve; AWGS, Asian Working Group for Sarcopenia; BMI, Body Mass Index; CC, Calf Circumference; CI, Confidence Interval; CKD, Chronic Kidney Disease; eGFR, estimated Glomerular Filtration Rate; GLIM, Global Leadership Initiative on Malnutrition; IQR, Interquartile Range; KDIGO, Kidney Disease, Improving Global Outcomes; NPV, Negative Predictive Value; NRS-2002, Nutritional Risk Screening 2002; OR, Odds Ratio; PPV, Positive Predictive Value; ROC, Receiver Operating Characteristic; SD, Standard Deviation; SGA, Subjective Global Assessment; STROBE, Strengthening the Reporting of Observational Studies in Epidemiology. Data Sharing Statement The datasets generated and/or analyzed during the current study are not publicly available due to institutional data protection policies and patient privacy regulations but are available from the lead corresponding author, Lihua Zhao, on reasonable request. Ethics Approval and Consent to Participate This study was conducted in accordance with the principles outlined in the Declaration of Helsinki and was strictly approved by the Ethics Committee of First Hospital of Hebei Medical University (Approval No.: S00438 [2023]). Written informed consent was systematically obtained from all participants or their legal proxies prior to their inclusion in the study. Consent for Publication Not applicable. The manuscript does not contain identifiable individual person’s data in any form. Author Contributions All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work. Lead contact author: Lihua Zhao. Disclosure The authors declare that they have no competing financial or non-financial interests. References Associated Data Data Availability Statement The datasets generated and/or analyzed during the current study are not publicly available due to institutional data protection policies and patient privacy regulations but are available from the lead corresponding author, Lihua Zhao, on reasonable request.