研究目的：主要是估計台灣地區中年人的身體質量指數軌跡圖，以及分析八年後與衰弱之間關係。 研究方法：在這個長期追蹤的前瞻性研究中，數據資料是從臺灣地區中老年身心社會生活狀況長期追蹤調查 ( Taiwan Longitudinal Study in Aging )所取出。這項研究是從1999年至2007年，追蹤8年期間。首先用群組化軌跡模式估計出最適當的身體質量指數曲線，最後軌跡分組共四組：持續正常偏低組 ; 持續正常偏高組 ; 持續過重組 ; 持續肥胖組。再以持續正常偏高組別當作參考組，使用邏輯式回歸來評估不同組別身體質量指數軌跡及是否增加未來的老年衰弱的風險。 研究結果：1999年參與者共1609人，體重位於持續正常偏低組共316人；持續正常偏高組共737人；持續過重組共449人；持續肥胖組共107人。在基本資料分析中，四組在性別比例、年齡分布、社會參與情形、抽菸狀況與慢性疾病罹病數量有顯著的差異。除上述變項調整外，另調整了教育程度、收入滿意度、自評健康狀況、喝酒等變項，進行多變項邏輯式回歸，結果顯示相較於持續正常偏高體重組，持續正常偏低組別[Odd ration(OR)= 1.794 (1.050-3.065) ; 95% confidence interval (CI) = 1.050-3.065]及持續肥胖組[Odds ratio (OR) = 2.013; 95% confidence interval (CI) = 1.013–4.001]有較高未來老年衰弱的風險，且統計學上有意義。 結論：台灣中年體重軌跡曲線趨向持續不變，多數人落在持續正常偏高組別，以此做參考組比較，持續正常偏低組及持續肥胖組老年衰弱風險增高，持續過重組則無統計學上意義。本研究並未發現肥胖矛盾情形(obesity paradox)，在台灣政策制定者對於中老年人體重控制仍以維持於正常範圍為佳，而過重之健康影響仍需更多研究證實。 Objectives: To evaluate midlife body mass index trajectory and the relationtionship with frailty eight years later in Taiwan. Method: In this long-term prospective cohort study, Data were retrieved from the Taiwan Longitudinal Study on Aging. This study was conducted from 1999 to 2007. Analysis was restricted to respondents who were not frail in 1999 and still survived in 2007 (n=1609). Frailty was defined using Fried criteria. Group-based model of trajectory was used to estimate body mass index trajectories among elderly. Logistic regression analysis was used to examine the association between body mass index change and frailty. Results: Four trajectory classes were identified and each remained no change across 8-year follow up. There were 316 participants (20.3%) in the low-normal weight group with baseline BMI=20.38kg/m2, 737 participants (44.7%) in the high-normal weight group with baseline BMI=23.22kg/m2, 449 participants (28.4%) were in the overweight group with baseline BMI=26.24kg/m2 and 107 participants (6.6%) in the obesity group with baseline BMI=30.65kg/m2. After adjustment of gender, age, level of education, income, social participation, self-rated health, health behaviors and major comorbidities, consistent low-normal weight group [Odd ration(OR)= 1.794 (1.050-3.065) ; 95% confidence interval (CI) = 1.050-3.065] and obesity group [Odds ratio (OR) = 2.013; 95% confidence interval (CI) = 1.013–4.001] were associated with increased frailty compared to high-normal weight group. Conclusion: Our results showed that body mass index trajectories of midlife elderly tended to be constant and either low-normal weight group or obesity group had more opportunity to develop frailty in elder stage. It may revealed that obesity paradox is not applicable when it comes to frailty. Therefore, for Asian elderly people, optimal weight targeting strategy should be considered for policy maker. Besides more studies will be needed for evaluating the health consequences of overweight group.
