| 摘要: | 本研究的目的在於探討促使糖尿病患者死亡的主要因子，將藉由糖尿病個案管理收案之病患，由已死亡案例和現存活案例中，找出兩者間的差異，以評估該因子是否為死亡率的指標。
方法:
利用院內糖尿病照護登錄系統2002~2006年間糖尿病個案管理病例，已滿一年完成糖尿病衛教課程收案中之已死亡案例和現存活案例作為研究對象。所收案之基本資料包括年齡、性別、罹病時間長短、血糖用藥控制的方式和是否有其他共病疾患包括：高血壓和高血脂症等。生化臨床檢驗項目有糖化血色素(HbA1c)、血脂肪：總膽固醇(TC)、三酸甘油脂(TG)、低密度膽固醇(LDL-C)和微量白蛋白尿排出率(ACR)。臨床檢查包括腳踝與上臂血壓的比值(ABI)、脈波傳播速率(PWV)、非散瞳視網膜檢查(NMRP)、神經病變檢查(Michigan neuropathy score)等所有項目。
收集之資料以Microsoft Excel進行資料建檔處理，再以SPSS. 12.0 中文版軟體做統計分析。使用的統計方法：比較兩組連續變項，使用獨立樣本t–test檢定；類別變項利用描述性統計卡方檢定；再以對數迴歸多變項分析找出導致死亡原因的危險因子，以0.05作為界定是否達到統計顯著性的標準。將微量白蛋白尿排出率(ACR)用接受器操作特定性曲線(ROC curve)找出預測數值中的截斷點（cut-off point），來作為預測的指標。
結果:
由院內糖尿病照護登錄系統中，四年間所收案的糖尿病個案7207位，其中死亡案例數，由497位個案中篩選出189位，平均年齡為73.58±9.58歲，男性為98位(51.9%)；女性為91位(48.1%)。所有的取樣是經年齡配對後現存活案例和死亡案例4：1的比例篩檢，以抽樣的方式由現存活案例數中篩選出754位案例數做分析，平均年齡為71.98±9.11歲，男性為341位(45.2%)；女性413位(54.8%)。
分析結果顯示:性別、罹患糖尿病時間的長短、血脂肪、糖化血色素（HbA1c）和脈波傳導速率（PWV）等未達到統計上顯著的判定標準。而年齡（p =0.033）、微量蛋白尿排出率（ACR）（p <0.001）和腳踝與上臂血壓的比值（ABI）（p <0.001）、糖尿病病患的用藥控制方式（p =0.033）和有否神經病變（p <0.001）等皆達統計學上顯著差異。最後，微量白蛋白尿排出率（ACR）以接受器操作特定性曲線(ROC curve)分析絕對預測值的截斷點（cut-off point）為64.9，其約佔63.1﹪的預測機率（p <0.001）。
結論:
藉由糖尿病病患死亡危險因子分析出高危險因素，進而找出可降低糖尿病患者死亡率的臨床指標，針對其高危險因素做積極的管理控制，期望降低死亡的風險。
由本研究結果可以得出結論即微量蛋白尿排出率（ACR）和有神經病變這兩個變項，更能有效預測出影響糖尿病病患死亡的高危險因子，與病患的年齡、用藥控制方式和疾病的嚴重程度是成正相關的。而微量白蛋白尿排出率（ACR）和神經病變則是所有糖尿病患在大、小血管病變中最早發生的，尤其是微量白蛋白尿排出率（ACR）的數值在研究結果中發現，當糖尿病患者檢驗數值為64.9 µg/mg以上時，其糖尿病併小血管病變早已發生，將會有63.1%的機率轉變為慢性腎病變。所以，早期預測糖尿病病患合併症產生的同時，加強並提升糖尿病個案管理之效能與品質之依據，進而延長病患的存活機率。
Objectives :
This study investigated the clinical factors that were associated with death from diabetes. Among the patients enrolled in diabetes case management, those who had died were compared with the surviving cases to evaluate potential factors that predict the mortality. Findings can be adapted to improve the care of diabetic patients.
Methods :
The study subjects consisted of dead patients and the surviving cases who had completed at least a year of diabetes education at our hospital’s diabetes case management program in 2002 - 2006. The baseline personal data included age, gender, disease duration, medications, and co-morbidity (hypertension and hyperlipidemia). The laboratory measures were extracted from medical records including hemoglobin A1c (HbA1c), lipid profile (total cholesterol, triglyceride, and low density lipoproteins cholesterol) and microalbumin rate (ACR), and clinic examines included ankle brachial index (ABI), pulse wave velocity (PWV), non-mydriatic retino-photo (NMRP) and Michigan neuropathy score. Differences between dead subjects and surviving subjects were compared using t-test and Chi-square test. Linear Logistic regression was used to identify risk factors related to the death. P < 0.05 was defined to statistic significance. The receiver operating characteristics (ROC) curve was used to establish ACR positive predictive value for determining the cut-off point value.
Results :
Among 7207 patients participating in the diabetes case management program, 189 cases were selected from 497 dead cases and 754 surviving subjects were selected matched with age. The mean ages were 73.6 years (SD, 9.58) in dead subjects (51.9% males, 48.1% females) and 72.0 years (SD, 9.11) in surviving subjects(45.2% males, 54.8% females). The results showed that the distributions in gender, DM duration, lipid profile, HbA1c and PWV were not significantly different between the 2 groups. Age, ABI , medication use and neuropathy were significantly higher in the dead group than in the survivals. Mean albumin creatinine ratio and neuropathy score were significantly better in the survivals. In death subjects, the elevated albumin creatinine ratio indicated a more complicated status. Data analysis showed an elevated mortality in patients with micro-vascularpathy. The ROC curve established the ACR positive predictive value with a cut-off point at 64.9µg/mg and 63.1% of predictive rate.
Conclusion :
Diabetic patients with chronic nephropathy are at an elevated risk of mortality, particularly when ACR increases to 64.9 µg/mg. Diabetic care with adequate risk factor control and case management may reduce the risk of death. |
