老年人長期抗膽鹼用藥負荷與臨床不良結果：不同測量工具之比較

Abstract

研究背景：Anticholinergic Risk Scale (ARS)及Anticholinergic Cognitive Burden Scale (ACB)抗膽鹼用藥評分系統為臨床上最常用於估算老年人抗膽鹼用藥負荷之評估工具，然而目前兩個評分系統與不良結果之關聯性由於現有研究的限制，如樣本數少、橫斷性研究設計或是觀察時間短，無法提供醫療人員足夠的資訊以判定何者於臨床上最為適用。 研究目的：比較ARS及ACB兩評分系統與不良結果之相關性，並進一步分析在相同抗膽鹼用藥負荷之下，不同抗膽鹼藥品組合對於不良事件之影響是否相同。 研究方法：本研究為一回溯性世代研究，利用台灣健保資料庫2000年及2005年之百萬人承保歸人檔，篩選於2001年1月1日年滿65歲的門診病人作為研究族群。於2001年期間先分析研究對象之基本特質，之後以一個月為單位評估研究對象2002年至2011年期間每個月抗膽鹼用藥負荷的變化以及不良結果的發生(總計120個月)。抗膽鹼用藥負荷以ARS及ACB評分系統計算，並分析抗膽鹼藥品之組合；而不良結果則包括急診就診事件、全部住院事件、骨折住院事件與失智症。本研究以廣義估計方程式(generalized estimating equations, GEE)分析抗膽鹼用藥負荷與不良結果之相關性，並分成不同年齡層(65-74歲、75-84歲、85歲以上)探討兩者之相關性，同時校正性別、抗膽鹼藥品平均使用劑量以及共病症之影響。第一部分會比較ARS及ACB與不良結果之相關性，第二部分進行抗膽鹼用藥負荷總分組合分析，探討在相同抗膽鹼用藥負荷總分之下，不同的藥品組合對於不良結果的影響。 研究結果：本研究共納入116,043位老年人，其中年齡65-74歲占整體研究族群之68.07 %，75-84歲者占27.82 %，而85歲以上僅占4.11 %。於追蹤起始時，整體研究族群平均一個月抗膽鹼用藥負荷分別為0.26 (ARS)與0.60 (ACB)。於十年追蹤期間，ARS及ACB總分皆呈現上升趨勢，並以ACB增加的幅度較多。 ARS及ACB總分皆與急診(e.g. 65-74 y/o, ARS 1-4+, aOR ranged from 1.55-2.04; ACB 1-4+, aOR range from 1.41-2.25)、住院(e.g. 65-74 y/o, ARS 1-4+, aOR ranged from 1.44-1.95; ACB 1-4+, aOR range from 1.32-1.92)、骨折住院(e.g. 65-74 y/o, ARS 1-4+, aOR ranged from 1.61-1.96; ACB 1-4+, aOR range from 1.10-1.71)以及失智症(e.g. 65-74 y/o, ARS 1-4+, aOR ranged from 3.87-10.41; ACB 1-4+, aOR range from 3.13- 10.01)之間呈現顯著的相關性。其中ACB總分與急診、住院、骨折住院以及失智症等不良結果之間皆呈現數值-風險正向增加的關係(e.g.急診：65-74 y/o, ACB 1, aOR 1.41 [95% CI 1.37-1.45]; ACB 2, aOR 1.71 [1.66-1.75]; ACB 3, aOR 1.82 [1.78-1.87]; ACB 4+, aOR 2.25 [2.19-2.31])；而ARS總分僅與骨折住院風險呈現數值-風險正向增加的趨勢(e.g.骨折住院：65-74 y/o, ARS 1, aOR 1.61 [1.47 -1.76]; ARS 2, aOR 1.64 [1.48-1.81]; ARS 3, aOR 1.64 [1.49-1.81]; ARS 4+, aOR 1.96 [1.72-2.23])，在其他不良結果的部分，則是呈現U型關係(e.g.急診：65-74 y/o, ARS 1, aOR 1.90 [1.86-1.95]; ARS 2, aOR 1.55 [1.51-1.59]; ARS 3, aOR 1.65 [1.61- 1.70]; ARS 4+, aOR 2.04 [1.97-2.11])。 進一步分析不同抗膽鹼藥品組合之影響，多種低抗膽鹼活性藥品的累積效應比單一種高抗膽鹼活性藥品之影響更大，發生不良結果之風險更高(即1+1 > 2，1+1+1 > 1+2 > 3) (e.g.急診：65-74 y/o, ACB 1+1, aOR 1.73 [1.68-1.77]; ACB 2, aOR 1.62 [1.53-1.71]; ACB 1+1+1, aOR 2.05 [1.99-2.12]; ACB 1+2, aOR 2.04 [1.91-2.17]; ACB 3, aOR 1.62 [1.57-1.66])；不過在骨折住院以及失智症部分，則是以ACB評分越高分的藥品對老年人的影響更大(即2 > 1+1，3 > 1+2 > 1+1+1) (e.g.失智症：65-74 y/o, ACB 1+1, aOR 3.02 [2.71-3.36]; ACB 2, aOR 6.23 [5.18- 7.48]; ACB 1+1+1, aOR 3.30 [2.84-3.84]; ACB 1+2, aOR 5.84 [4.59-7.41]; ACB 3, aOR 9.15 [8.38-9.99])。 研究結論：兩評分系統相比之下，ACB評分系統較能區別對老年人有不同不良影響程度的藥品，在所有老年族群中能夠辨別較高風險的老年人，提供臨床介入之依據；而ARS評分系統雖然也與不良結果之間呈現顯著的相關性，但較ACB評分系統不具選擇性。即使在相同抗膽鹼用藥負荷總分之下，不同的抗膽鹼藥品組合對於臨床不良結果之影響程度並不相同。

Background：Anticholinergic Risk Scale (ARS) and Anticholinergic Cognitive Burden Scale (ACB) are two anticholinergic drug scoring systems that are most commonly used to measure anticholinergic burden in the elderly. However data are limited regarding which scoring system is more clinically relevant. Studies comparing ARS and ACB were limited by small sample sizes, cross-sectional design or short follow-up periods that could not provide enough information about which scoring system should be used. Objectives：To compare the association between different scoring systems (ARS/ACB) and adverse outcomes, and to investigate whether different compositions of score would have different impact on adverse events under the same anticholinergic burden. Methods：A retrospective cohort study was conducted by using two sets of Longitudinal Health Insurance Database (LHID 2000 and 2005) from Taiwan’s National Health Insurance Research Database. The study cohort comprised outpatients who aged 65 years old and older on January 1, 2001. Baseline characteristics were collected in 2001, and the changes of anticholinergic burden and adverse events were measured monthly from 2002 to 2011 (total 120 months). Anticholinergic burden was estimated by ARS and ACB, and the compositions of score were also examined. Adverse outcomes included emergency room visit, all-cause hospitalization, fracture-specific hospitalization and incident dementia. Generalized estimating equations (GEE) were used to examine the association between anticholinergic burden and adverse outcomes, and stratified by age (65-74, 75-84, 85+ y/o). The models were further adjusted for sex, time-varying comorbidities and average daily dosage of anticholinergic agents. In the first part of the study, we compared the association between ARS/ACB and adverse outcomes, and in the second part, we investigated the impact of different compositions of score on adverse events under the same anticholinergic burden. Results：A total of 116,043 elderly were included in the study (aged 65-74: 68.07 %; aged 75-84: 27.82 %; aged 85+: 4.11 %). At baseline, mean anticholinergic burden based on ARS and ACB was 0.26 (/p/m) and 0.60 (/p/m) respectively. During ten-year follow-up, both ARS and ACB scores increased gradually with more increase in ACB. Both ARS and ACB scores were all significantly associated with increased odds of emergency room visit (e.g. 65-74 y/o, ARS 1-4+, aOR ranged from 1.55-2.04; ACB 1-4+, aOR range from 1.41-2.25), all-cause hospitalization (e.g. 65-74 y/o, ARS 1-4+, aOR ranged from 1.44- 1.95; ACB 1-4+, aOR range from 1.32-1.92), fracture-specific hospitalization (e.g. 65- 74 y/o, ARS 1-4+, aOR ranged from 1.61-1.96; ACB 1-4+, aOR range from 1.10-1.71) and incident dementia (e.g. 65-74 y/o, ARS 1-4+, aOR ranged from 3.87-10.41; ACB 1-4+, aOR range from 3.13- 10.01). ACB scores showed dose-response relations with all adverse outcomes (e.g. emergency room visit：65-74 y/o, ACB 1, aOR 1.41 [95% CI 1.37-1.45]; ACB 2, aOR 1.71 [1.66-1.75]; ACB 3, aOR 1.82 [1.78-1.87]; ACB 4+, aOR 2.25 [2.19-2.31]), but ARS scores showed dose-response relation only with fracture- specific hospitalization(e.g. fracture-specific hospitalization：65-74 y/o, ARS 1, aOR 1.61 [1.47-1.76]; ARS 2, aOR 1.64 [1.48-1.81]; ARS 3, aOR 1.64 [1.49-1.81]; ARS 4+, aOR 1.96 [1.72-2.23]). There were U-shaped relations between ARS scores and other outcomes (e.g. emergency room visit：65-74 y/o, ARS 1, aOR 1.90 [1.86 -1.95]; ARS 2, aOR 1.55 [1.51-1.59]; ARS 3, aOR 1.65 [1.61-1.70]; ARS 4+, aOR 2.04 [1.97-2.11]). After analyzing the compositions of score, we found that the cumulative effects of multiple medications with low anticholinergic activity were greater than one medication with high anticholinergic activity (1+1 > 2, 1+1+1 > 1+2 > 3) (e.g. emergency room visit：65-74 y/o, ACB 1+1, aOR 1.73 [1.68-1.77]; ACB 2, aOR 1.62 [1.53-1.71]; ACB 1+1+1, aOR 2.05 [1.99-2.12]; ACB 1+2, aOR 2.04 [1.91-2.17]; ACB 3, aOR 1.62 [1.57-1.66]). In contrast, medications with higher points in ACB had greater impact on fracture-specific hospitalization and incident dementia (2 > 1 + 1, 3 > 1 + 2 > 1 + 1 + 1) (e.g. incident dementia：65-74 y/o, ACB 1+1, aOR 3.02 [2.71-3.36]; ACB 2, aOR 6.23 [5.18-7.48]; ACB 1+1+1, aOR 3.30 [2.84-3.84]; ACB 1+2, aOR 5.84 [4.59-7.41]; ACB 3, aOR 9.15 [8.38-9.99]). Conclusion：ACB was more selective in capturing medications that likely to cause adverse outcomes and identifying high risk populations for intervention. Although ARS was also associated with adverse outcomes, it was less selective. Even under the same anticholinergic burden score, different compositions of score would have different impact on adverse events.