以群組化軌跡模式分析老年族群之長期精神用藥趨勢及相關不良結果

Abstract

研究背景：老年人長期精神藥物之多重用藥的問題與其臨床不良結果的相關性近年來十分受到關注。然而綜觀過去文獻，多數研究以單一類別精神藥物為主並採用橫斷性設計，未能考量用藥隨時間變動的特性，以及同時使用其他精神藥物的影響及整體精神藥物的劑量累積效應，然而這些因素都可能會影響精神藥物與臨床不良結果相關性的判讀。 研究目的：使用群組化軌跡模型分析新使用精神藥物之老年族群，在用藥隨時間變化影響下，有哪些不同的長期精神藥物使用模式(5年)，比較不同使用模式之間的族群特性與精神藥物使用狀況的差異，並探討不同使用模式與臨床不良結果之相關性。 研究方法：本研究為一回溯性世代研究設計，利用台灣健保資料庫的老年特殊需求檔，篩選2004年內年滿65歲(含)新使用口服精神藥物的門診病人作為研究對象。本研究定義研究對象自新使用精神藥物起的五年為軌跡觀察期，先分析研究對象於軌跡觀察期前一年住院次數、共病症以了解每位研究對象之背景資料，之後以定義每日劑量為單位，紀錄研究對象於軌跡觀察期間每月精神藥物的使用量，並以軌跡化群組模式進行分析，找出研究族群有幾種不同的長期精神藥物使用模式後，依據分析結果進行分組。為了能完整記錄軌跡觀察期的精神藥物使用狀況，若研究對象於軌跡觀察期間死亡者則予以排除。分組完成後，以軌跡觀察期結束後三年作為不良結果觀察期，本研究之臨床不良結果為全死因死亡與非預期性住院事件，使用Kaplan-Meier survival curves及Cox proportional hazard model進行分析，探討不同長期精神藥物使用模式與臨床不良結果的相關性。 研究結果：本研究共納入39803位於門診新使用口服精神藥物的老年人，經群組化軌跡模式分析後，研究族群長期精神藥物使用模式可分為五組，分別為infrequent users (51.61 %)、increasing users(15.67 %)、decreasing users(13.45 %)、continued users(12.14 %)、sustained intense users(7.13 %)。 分析第一筆精神藥物處方，以鎮靜安眠藥物的開方比例最高，占57.92 %。進一步分析開方科別與相關診斷發現，任一組別皆有超過90 %的處方由非精神科專科醫師開出，並且該筆處方上未有任何精神疾病或失眠相關診斷者超過80 %。 精神藥物使用之長期趨勢方面，sustained intense users在軌跡觀察期間使用精神藥物的時間最長，其中又以鎮靜安眠藥物者的使用時間最長，達24.39個月。分析整體族群精神藥物多重用藥趨勢，有40.44 %的人曾併用兩種精神藥物，8.64 %的人曾併用三種精神藥物，若比較五個組別間差異，皆以sustained intense users比例最高，併用兩種有85.81 %，併用三種有44.77 %，並且大部分都是同時使用benzodiazepine及z-drug兩種鎮靜安眠藥物為主 不論在一年或三年觀察期中，皆發現長期精神藥物使用模式與全死因死亡及非預期性住院事件有正相關性，以infrequent users作為對照組時，sustained intense users發生兩種事件的風險皆為最高(1-year unplanned hospitalization: HR=1.38 (95 % CI: 1.23-1.55); 1-year death HR=1.26 (95 % CI: 1.08-1.48); 3-year unplanned hospitalization: HR=1.35 (95 % CI: 1.25-1.45); 3-year death HR=1.28 (95 % CI: 1.16-1.39)) 研究結論：台灣地區老年族群其長期精神藥物使用模式可分為五種，各模式間在族群特性、長期精神藥物使用趨勢皆有差異，因此在老年族群不能僅以短期精神藥物使用狀況來推測長期趨勢。長期精神藥物使用模式顯示與非預期性住院事件及全死因死亡有正相關性，其中又以sustained intense users風險最高，因此針對這群老人需要特別密切注意不良事件的發生，並且積極介入進行用藥評估。

Background: Serious concerns regarding long-term use of multiple psychotropic agents (or psychotropic polypharmacy) in the elderly have been raised considering their vulnerability to psychotropic agents-associated adverse effects, such as hospitalization or mortality. However, existing studies on the use of psychotropic agents in the elderly are limited to cross-sectional design, one-time measurement of drug-exposure, or focusing on single category of psychotropic agent. In addition, no study has investigated the cumulative dose effect. Objectives: This study aims to identify distinct trajectories of longitudinal psychotropic agents use and to examine the association between risk of mortality and unplanned hospitalization, according to distinct trajectories. Method: Adult who aged 65 years or older and newly initiated their oral psychotropic agents in 2004 were identified from Taiwan’s National Health Insurance Research Database. The date of first prescription of psychotropic agents was defined as the index date for each identified subject. Monthly consumption of psychotropic agents by defined daily dose (DDD) during the 5 years of follow-up since index date were retrieved. Group-based trajectory modeling with a zero-inflated Poisson distribution and third order polynomials was used to identify distinct group of longitudinal psychotropic agents use over 5 years. Clinical adverse outcomes of interest were unplanned hospitalization and all-cause mortality. Kaplan-Meier survival curves and Cox proportional hazard model were used to examine the association between distinct trajectories of longitudinal psychotropic agents use and clinical adverse outcomes. Result: Of the 54192 eligible older people, we identified 5 trajectories of longitudinal psychotropic agents use over a 5-year follow-up: sustained intense users (7.13 %), continued users (12.14 %), decreasing users (13.45 %), increasing users (15.67 %), and infrequent users (51.61 %). Most of the elderly used sedative drug as their first psychotropic agents (57.92 %). More than 90 % elderly get their first psychotropic agents from non-psychiatrist and more than 80 % of them were without any diagnosis of mental illness or insomnia. The pattern of long-term psychotropic agents use were different between trajectories groups. Sustained intense users had the longest time of using psychotropic agents, especially when they were using sedative drugs during the 5 years of follow-up since index date. About psychotropic polypharmacy, 85.81 % sustained intense users had been used 2 psychotropic agents, at the same time and 44.77 % of them had been used 3 psychotropic agents at the same time. The most common combination were a benzodiazepine and a z-drug. The positive associations between unplanned hospitalization and all-cause mortality and different trajectories were statistically significant, after adjusting age, sex and baseline comorbidities (1-year unplanned hospitalization: HR=1.38 (95 % CI: 1.23-1.55); 1-year death HR=1.26 (95 % CI: 1.08-1.48); 3-year unplanned hospitalization: HR=1.35 (95 % CI: 1.25-1.45); 3-year death HR=1.28 (95 % CI: 1.16-1.39)). Conclusion: we identified 5 distinct trajectories of longitudinal psychotropic agents use in the elderly in Taiwan and the psychotropic polypharmacy is very common in Taiwanese elderly, especially for those sustained intense users. Sustained intense users was associated with highest risks of death and unplanned hospitalization.