社區糖尿病個案衰弱與營養因素探討

Abstract

背景：糖尿病為全球及台灣主要公共衛生議題之一，依國際糖尿病聯盟（IDF）及我國健康調查數據顯示，糖尿病盛行率逐年攀升，並高居我國十大死因之一。研究指出，糖尿病患者發生衰弱與肌少症之風險顯著高於非糖尿病者，且年齡愈高者其衰弱盛行率愈高。衰弱與營養不良之間存在交互影響關係，營養狀況不佳可能導致功能退化，而衰弱亦會影響飲食與營養吸收。此外，多篇研究指出，MNA評估結果與老年糖尿病患者之衰弱風險具顯著相關性，惟目前針對台灣社區高齡糖尿病個案之營養與衰弱相關性研究相對有限。 研究目的：本研究旨在探討某縣市社區65歲以上糖尿病個案之衰弱狀態與其營養狀況、臨床因素之關聯性，進一步建構預測模型，以辨識影響衰弱的主要因素，提供社區糖尿病整合照護之實務依據。 研究方法：本研究採相關性研究設計，資料來源為113年5月至10月彰化縣27家衛生所之次級資料，包含營養篩檢（MNA）、行動能力評估（SPPB）、憂鬱篩檢（PHQ-2）與基本資料（性別、年齡、生化數值、共病數等）。研究變項包含：依變項為衰弱狀態（以SPPB總分分組）；自變項為營養狀況（MNA總分與細項）；控制變項為年齡、性別、HbA1c、CKD分期與PHQ-2。統計分析使用SAS 9.4軟體，方法包含卡方檢定、單變項與多變項邏輯式迴歸分析，並以ROC曲線與AUC值檢視預測效能。 結果：共納入727位個案，其中女性佔56.1%，年齡以65~74歲為主、A1C以6.0–6.9%佔多數（59.7%），CKD以正常與初期分期（第1–3a期）為主。MNA結果顯示91.5%為正常營養狀態，8.5%具營養不良風險。依SPPB總分分為衰弱與非衰弱組，非衰弱組（10~12分）佔86.2%，衰弱組（0~9分）佔13.8%。邏輯式回歸分析顯示，女性、高齡及營養狀態（低MNA總分）與衰弱狀態具顯著相關性，模型AUC值達0.7以上，具中等預測效能。 討論：本研究發現營養狀況與糖尿病個案之衰弱狀態密切相關，部分MNA細項如體重減少、一天流質、自覺營養不良、同齡健康、低中臂圍、低小腿圍等，與衰弱風險具統計顯著性。儘管A1C與衰弱之間無顯著關聯，仍可觀察到在A1C過低與過高區間，其衰弱比率皆高於中間範圍，尤其是低A1C。研究亦指出，營養不良與衰弱具雙向關係，實務建議未來照護中如發現營養不良（含風險）個案，應同步評估衰弱狀態，反之亦然，以利及早介入。 結論：本研究顯示，社區糖尿病個案之營養狀況與衰弱風險具有顯著關聯性，營養指標可作為社區整合性健康評估與早期介入之參考依據。未來推動高齡糖尿病者照護時，應將營養評估與衰弱篩檢納入常規作業，並持續透過統計模型優化預測工具，提升照護效能與健康促進效果。

Background：Diabetes mellitus is a major public health issue both globally and in Taiwan. According to the International Diabetes Federation (IDF) and national health survey data, the prevalence of diabetes continues to rise annually, and it ranks among the top ten causes of death in Taiwan. Studies have shown that individuals with diabetes have significantly higher risks of frailty and sarcopenia compared to those without diabetes, and the prevalence of frailty increases with age. A bidirectional relationship exists between frailty and malnutrition—poor nutritional status may lead to functional decline, while frailty can affect dietary intake and nutrient absorption. Moreover, multiple studies have identified significant associations between MNA (Mini Nutritional Assessment) scores and frailty risk in older adults with diabetes. However, research focusing on the relationship between nutrition and frailty among community-dwelling older adults with diabetes in Taiwan remains limited. Objective：This study aimed to explore the relationship between frailty status and nutritional condition, along with clinical factors, among community-dwelling diabetes patients aged 65 and above in a county in Taiwan. The goal was to establish predictive models to identify key determinants of frailty and provide practical insights for integrated community diabetes care.

Methods：This correlation study utilized secondary data collected between May and October 2024 from 27 public health centers in Changhua County. The dataset included nutritional screening (MNA), physical performance assessment (SPPB), depression screening (PHQ-2), and basic demographic and clinical information (e.g., sex, age, biochemical markers, number of comorbidities). Study variables included: dependent variable—frailty status (based on SPPB total score); independent variables—nutritional status (MNA total score and items); control variables—age, sex, HbA1c, CKD stage, and PHQ-2 scores. Statistical analyses were performed using SAS 9.4, including chi-square tests, univariate and multivariate logistic regression, and ROC curve analyses with AUC to assess model performance. Results：A total of 727 participants were included, with 56.1% being female. Most were aged between 65 and 74 years. HbA1c values primarily ranged from 6.0% to 6.9% (59.7%), and CKD stages were mostly normal to early-stage (stage 1 to 3a). Based on MNA results, 91.5% of participants had normal nutritional status, while 8.5% were at risk of malnutrition. According to SPPB scores, 86.2% were classified as non-frail (10–12 points), and 13.8% as frail (0–9 points). Logistic regression analysis revealed that female sex, older age, and poor nutritional status (lower MNA scores) were significantly associated with frailty. The predictive model achieved AUC values above 0.7, indicating moderate predictive performance. Discussion：This study found a strong association between nutritional status and frailty in diabetes patients. Specific MNA items—such as weight loss, low fluid intake, self-perceived malnutrition, perceived health status compared to peers, low mid-arm circumference, and low calf circumference—showed significant associations with frailty risk. Although HbA1c was not significantly correlated with frailty, higher frailty rates were observed at both low and high extremes of HbA1c, especially in the low HbA1c group. The findings also emphasize the bidirectional relationship between malnutrition and frailty, suggesting that cases identified at risk for either condition should be evaluated for both to enable timely intervention. Conclusion：This study demonstrates a significant association between nutritional status and frailty risk among community-dwelling individuals with diabetes. Nutritional indicators may serve as practical references for integrated health assessments and early intervention strategies. It is recommended that routine care for older adults with diabetes include nutritional assessment and frailty screening, supported by predictive models to enhance care effectiveness and promote healthy aging.