目標：影響糖尿病的因素很多，多運動及增加蔬果攝取皆為降低風險的有效作為，但同時多運動也多吃蔬果的成效，則尚不詳。本研究目的在探討運動－蔬果攝取組合變項對降低台灣中老年人新發糖尿病的成效。方法：本研究分析「中老年身心社會生活狀況長期追踨調查」1999及2003年的資料庫。以1999年≥53歲4,440名個案為對象，排除已患糖尿病、心臟病、癌症、及BMI<18.5、四年間死亡、和資料不全者，實際有效個案為2,243名。以SPSS 12.0分析運動及蔬果攝取狀況單獨或組合，對隨後四年新發糖尿病的影響。結果：兩變項經組合，又以運動量及蔬果攝取頻率皆低為對照，運動量及蔬果攝取頻率皆高者，其隨後四年新發糖尿病風險降低61%（OR=0.39，95%CI=0.20－0.76，p=0.006），如果僅運動或蔬果攝取單獨一項高者，則不顯著。不組合時，則僅高運動顯著，高蔬果攝取不顯著。結論：本研究顯示在組合變項狀況下，同時高運動且也高蔬果攝取可有效大幅降低新發糖尿病的風險，如只高運動或高蔬果攝取，則不顯著。此結果凸顯平日多運動及多攝取蔬果對預防新發糖尿病的重要性，且可供國人糖尿病防治政策擬定之依據。

Objectives: Many factors impact the onset of diabetes. Higher levels of physical activity and fruit-vegetable (FV) consumption are considered to be effective in reducing the risk; however, the combined effect is largely unknown. This study attempted to determine the combined effect of a higher level of physical activity and higher FV consumption on the subsequent risk of new diabetes. Methods: Data for this analysis were taken from the 1999 and 2003 datasets of the "Taiwan Longitudinal Survey on Aging". A total of 4,400 participants (≥ 53 years old) completed the survey in 1999. After excluding those who had diabetes, heart disease, cancer, BMI<18.5, or incomplete data at baseline, those who died during 1999-2003, and those who failed to complete the 2003 survey, the net N was 2,243. Multivariate binary logistic regression analysis was performed to determine the independent or combined effect of physical activity and FV consumption on the risk of new diabetes four years later. Results: When the combination of high physical activity and high FV consumption was compared to no physical activity and low FV consumption, the risk of having new diabetes four years later was reduced to 39% (OR=0.39, 95% CI=0.20-0.76, p=0.006). All other combinations were not significant. When analyzed independently, only high physical activity was significant, high FV consumption was not. Conclusions: The combination of high physical activity and high FV consumption was highly effective in reducing the risk of new-onset diabetes in older Taiwanese. High physical activity alone was significant but high FV consumption alone was not. These results highlight the importance of having both daily physical activity and FV intake. Our results could provide the basis for a national policy on diabetes prevention.