Risk factors for malnutrition in patients with diabetic foot ulcer and its association with prolonged length of hospitalization

Diabetic foot ulcer (DFU) is a common and serious complication of diabetes, affecting up to 25% of patients during their lifetime and constituting a frequent cause of hospitalization. Malnutrition—characterized by reduced body composition and cell mass due to insufficient nutrient intake—can be exacerbated in DFU by chronic hyperglycemia-driven muscle degeneration, increased resting energy expenditure, and heightened nutrient demands for wound healing. Malnutrition in DFU is associated with delayed healing, recurrence, increased readmissions, higher medical burden, and poorer quality of life. Prior studies report a wide range of malnutrition prevalence in DFU (15–62%), likely due to differing diagnostic methods and settings, and identify malnutrition as a predictor of adverse outcomes such as amputation. The present study aims to determine the incidence and risk factors for malnutrition in hospitalized DFU patients using the GLIM criteria and to evaluate the association between malnutrition and length of stay (LOS), a relationship previously underreported.

Previous research on nutrition in DFU shows variable malnutrition prevalence depending on assessment methods and populations: 15% (Gau), 29% (Rouland), 32% (Eneroth), 62% (Zhang), and 49% at risk of malnutrition (Xie). Nutritional deficiencies have been linked to poor outcomes, including amputation risk. Reported risk factors for malnutrition in DFU include low BMI, higher Wagner grade, and ulcer infection, though findings have been inconsistent. Since the 2019 GLIM guidelines, few studies have applied GLIM to DFU; one study (Lauwers et al.) found 24% malnutrition by GLIM but no short-term outcome differences. Notably, the association between nutritional status and hospital LOS in DFU inpatients had not been specifically examined, indicating a gap this study addresses.

Design and setting: Observational cohort study at two campuses of a hospital in Chongqing, China, from January 2021 to June 2023. Sample size planning (effect size 0.5, α=0.05, power 0.80, 1:1 ratio) indicated a minimum of 164; 219 were enrolled after exclusions. Ethics approval was obtained (Second Hospital of Chongqing Medical University, NO: 2022.30); informed consent was obtained. Participants: Inclusion criteria: adults ≥18 years with DFU per IWGDF guidelines, diabetes duration >1 year with ongoing glucose-lowering therapy. Exclusion: severe physical/mental conditions precluding assessment (e.g., massive edema precluding accurate weight, cognitive impairment), prior amputation or revascularization, tuberculosis, hyperthyroidism, malignancy, gastrointestinal disease or surgery, or hospital stay <24 h. Data collection: Trained staff collected data via standardized questionnaires and electronic medical records. Variables: demographics (age, sex), BMI (kg/m²), smoking/alcohol history, hypertension and dyslipidemia status, LOS, diabetes and DFU duration, treatment modality, ulcer infection (based on clinical signs and inflammatory biomarkers), diabetes complications, and SINBAD score. Vascular assessment by ankle-brachial index (ABI): normal ≥0.9; 0.4–<0.9 mild–moderate ischemia; <0.4 severe stenosis/occlusion. DPN diagnosed by Neuropathy Study Group criteria (symptoms/signs plus exam abnormalities); DN defined by persistent albuminuria or eGFR <60 mL/min/1.73 m² for >3 months. Labs: hemoglobin, HbA1c (HPLC), CRP, eGFR, albumin measured on morning after admission. Nutrition assessment: Within 24 h of admission, NRS-2002 was used for screening. GLIM criteria were applied for diagnosis, requiring at least 1 phenotypic (weight loss, low BMI, or reduced muscle mass) and 1 etiologic criterion (reduced intake or inflammation). Given DFU’s chronic inflammatory nature, all patients were considered to meet the inflammation criterion. Reduced muscle mass was assessed using calf circumference (CC) per Japanese sarcopenia standards (≤30 cm men, <29 cm women) as a proxy due to lack of body composition data. LOS definition: prolonged LOS defined as >17 days (mean hospitalization time). Statistical analysis: Continuous variables were summarized as mean±SD or median (IQR); categorical as n (%). Group comparisons used t-test, Mann–Whitney U, or chi-square. Variables with P<0.05 in univariate analyses entered multivariate binary logistic regression (Forward LR) to identify independent risk factors for malnutrition and for prolonged LOS. Odds ratios (ORs) with 95% CIs were reported. Model calibration assessed with Hosmer–Lemeshow test. Analyses performed with SPSS 26.

- Sample: 219 DFU inpatients (mean age 67±12 years), after excluding 25. Malnutrition prevalence by GLIM: 38.36% (84/219). Prolonged LOS: 42% (92/219); median LOS 16 days (IQR 15–20). - Univariate associations with malnutrition included lower BMI, dyslipidemia, ulcer infection, DPN, DN, higher SINBAD, lower ABI, elevated CRP, lower Hb and albumin, higher HbA1c, and longer LOS (all P<0.05). - Independent risk factors for malnutrition (multivariable): lower BMI (P<0.001), lower albumin (P=0.002), higher HbA1c (P<0.001), presence of ulcer infection (P<0.001), longer LOS (P=0.010), and lower ABI (P=0.024). In discussion, ulcer infection was associated with markedly higher odds of malnutrition (OR≈8.847; P<0.001). Model fit: Hosmer–Lemeshow P=0.927; R²=0.716. - Factors associated with prolonged LOS (multivariable): DPN (P=0.033), DN (P<0.001), and malnutrition (P=0.001). Malnutrition predicted prolonged LOS with OR 2.857 (95% CI 1.497–5.450; P=0.001). Model fit: Hosmer–Lemeshow P=0.310; R²=0.295. - Group differences for prolonged LOS included lower BMI, higher rates of ulcer infection, DPN, DN, lower ABI, lower Hb and albumin, higher HbA1c, and higher malnutrition prevalence (all P<0.05).

Applying GLIM criteria, malnutrition was common in hospitalized DFU patients and independently associated with markers of poorer metabolic control (higher HbA1c), systemic protein status (lower albumin), vascular compromise (lower ABI), active ulcer infection, and longer LOS. These findings address the study question by identifying modifiable clinical factors linked to malnutrition and demonstrating that GLIM-defined malnutrition is a robust predictor of prolonged hospitalization. The observed threefold increase in odds of prolonged LOS among malnourished patients underscores the clinical and resource implications of nutritional status in DFU care. Mechanistically, the bidirectional relationship between infection and malnutrition may drive a vicious cycle of inflammation, nutrient loss, impaired healing, and further infection risk. Poor glycemic control and hypoalbuminemia likely exacerbate tissue hypoxia and impaired nutrient absorption, respectively. Compared with prior GLIM-based work reporting no short-term outcome differences, this study reveals a significant association with LOS, highlighting the utility of early GLIM-based nutritional assessment for risk stratification and management planning in DFU.

Malnutrition occurred in 38.36% of hospitalized DFU patients using GLIM criteria. Lower BMI, lower albumin, higher HbA1c, ulcer infection, longer LOS, and lower ABI were independent risk factors for malnutrition. GLIM-defined malnutrition was associated with approximately threefold higher odds of prolonged hospital stay. These results support routine nutritional screening and assessment with GLIM and prompt nutritional interventions in DFU management. Future research should include larger, multicenter cohorts, longitudinal monitoring of nutritional status, and grading of malnutrition severity to refine prognostication and intervention strategies.

- Nutritional assessment was performed only at admission; no follow-up assessments to capture changes during or after hospitalization. - Some historical data (e.g., weight loss) relied on patient recall, introducing potential recall bias. - Not all comorbidities or complications during hospitalization were captured, which may affect accuracy and generalizability. - The degree (severity) of malnutrition per GLIM (moderate vs severe) was not evaluated. - Single-center, observational design limits causal inference; larger, multicenter studies are warranted.