Published in:
Open Access 01-12-2022 | Sarcopenia | Research
Associations of geriatric nutrition risk index and other nutritional risk-related indexes with sarcopenia presence and their value in sarcopenia diagnosis
Published in: BMC Geriatrics | Issue 1/2022
Login to get accessAbstract
Objective
Standard modalities recommended for sarcopenia diagnosis may be unavailable in primary care settings. We aimed to comprehensively evaluate and compare associations of some better popularized nutritional risk-related indexes with sarcopenia presence and their value in sarcopenia diagnosis in community-dwelling middle-aged and elderly adults, including geriatric nutrition risk index (GNRI), albumin (ALB), calf circumference (CC), mid-arm circumference (MAC), triceps skinfold thickness (TST) and body mass index (BMI).
Methods
Based on the West China Health and Aging Trend study, the current study included participants aged 50 or older who were recruited in 2018. Sarcopenia-related assessment and diagnosis were in line with Asian Working Group for Sarcopenia 2019. For each single index, we assessed its association with sarcopenia presence by univariate and multivariate logistic regression analysis; we also computed diagnostic measures including the area under the receiver operating characteristic curve (AUC) and sensitivity, specificity, accuracy at the optimal cut-off value determined according to Youden’s index.
Results
A total of 3829 subjects were included, consisting of 516 and 3313 subjects in the sarcopenia and non-sarcopenia groups, respectively. Regarding the risk for sarcopenia presence, the fully adjusted odds ratios of GNRI, ALB, CC, MAC, TST and BMI per standard deviation decrease were 2.95 (95% CI 2.51–3.47, P < 0.001), 1.01 (95% CI 0.90–1.15, P = 0.816), 4.56 (95% CI 3.82–5.44, P < 0.001), 4.24 (95% CI 3.56–5.05, P < 0.001), 1.67 (95% CI 1.92–1.45, P < 0.001) and 4.09 (95% CI 3.41–4.91, P < 0.001), respectively. Regarding the value in sarcopenia diagnosis in the entire study population, their AUCs could be ordered as MAC (0.85, 95% CI 0.83–0.86) > GNRI (0.80, 95% CI 0.78–0.82), CC (0.83, 95% CI 0.81–0.85), BMI (0.81, 95% CI 0.79–0.83) > TST (0.72, 95% CI 0.70–0.74) > ALB (0.62, 95% CI 0.60–0.65). At the relevant optimal cut-off values, the sensitivity was the highest for CC (0.83, 95% CI 0.80–0.87) and MAC (0.80, 95% CI 0.77–0.84), while GNRI showed the highest specificity (0.79, 95% CI 0.78–0.81) and accuracy (0.78, 95% 0.76–0.79).
Conclusion
Overall diagnostic performance was the best for MAC, followed by GNRI, CC, BMI, and the worst for TST, ALB in distinguishing sarcopenia from non-sarcopenia in middle-aged and elderly adults in community-based settings. CC or MAC might do better in reducing missed diagnosis, while GNRI was superior in reducing misdiagnosis.