臺灣大學: 流行病學與預防醫學研究所陳為堅郭千哲Kuo, Chian-JueChian-JueKuo2013-03-212018-06-292013-03-212018-06-292011http://ntur.lib.ntu.edu.tw//handle/246246/250338研究緣起: 自從1990年代初期, 甲基安非他命的濫用問題已成為全球之非法藥物成長最快速之一, 特別是在東南亞、澳洲, 及北美地區。甲基安非他命被報告會造成身體和精神之傷害; 前者包括心臟血管疾病, 後者則包括憂鬱、精神病症狀及自殺。甲基安非他命亦導致早期死亡的危險增加。然而, 關於死亡預後或死亡原因的實證資料在文獻中相當少。此外, 甲基安非他命已知對人體具有心臟毒性。心電圖為非侵犯性、低成本及廣泛使用之檢測工具, 本研究假設其可能成為評估甲基安非他命者各項傷害及預後的工具。 最近有回顧性文章顯示曾有8個使用世代族群探討安非他命死亡預後, 但只有其中的3個使用超過500個研究樣本。本研究計劃選擇一大型甲基安非者世代族群 (N=1254), 其均曾因甲基安非他命相關問題住院治療; 與死亡檔連結將可追蹤其死亡狀態。由巢式個案對照研究和標準化病歷回顧, 計劃分別探討自殺死亡和自殺死亡的危險因子。此外, 為探討此特殊族群的心電圖異常, 本研究將選擇一外部健康族群成為控制組, 以便與甲基安非他命組作比較。 目的: 目的一: 探討甲基安非他命者的長期死亡率及死因; 目的二: 探討甲基安非他命者自殺死亡的危險和保護因子; 目的三: 探討甲基安非他命者自然死亡的危險因子; 目的四: 探討甲基安非他命者的心電圖異常。 研究設計與方法: 第一部份: 本研究將收集於1990/1/1至2007/12/31, 在北部某一精神科專科醫院住院的所有甲基安非他命依賴(依DSM-III-R和DMS-IV診斷標準)的病患, 定義為世代族群和進行回溯性研究。由一初步簡要的病歷回顧收集每位個案的診斷和社會人口變項資料。經連結同時期的全國死亡檔, 以獲得死亡資料並計算標準死亡比(SMR); 估計總共約為150位死亡個案。 第二部份: 基於第一部份的資料連結, 有38位因自殺死亡, 定義為個案組; 利用密度抽樣(density sampling), 由原來世代族群中以1:2比例選取適當的控制組, 配對條件為性別、年齡不相差5歲及相同的指標住院年。之後使用詳細標準病歷回顧方式, 收集社會人口學、精神病理學及臨床相關變項。使用多變項邏輯式迴歸, 分析與自殺死亡相關的因子。 第三部份. 基於第一部份的資料連結, 有45位為因自然死亡, 定義為個案組; 利用密度抽樣(density sampling), 由原來世代族群中以1:4比例選取適當的控制組, 配對條件為性別、年齡不相差5歲及相同的指標住院年。由詳細標準病歷回顧方式, 收集相關臨床變項及實驗室指標。使用多變項邏輯式迴歸, 分析與自然死亡相關的因子。 第四部份. 根據甲基安非他命者的性別和年齡分層, 選擇一個曾經參與身體健檢之健康族群 (N=300)成為對照組, 以探討心電圖的異常。為對心電圖標準判讀, 本研究將二組的心電圖送往Epidemiological Cardiology Research Center (EPICARE), Department of Public Health Sciences, Wake Forest University School of Medicine, Winston-Salem, North Carolina, USA; 且採用Minnesota分類作標準判讀。判讀時該機構不清楚研究樣本之分組(甲基安非他命組或正常組)。 第四部份. 本研究收集二個樣本, 分別選擇由上述世代族所選擇的一個甲基安非他命(N=387), 以及一個健康族群(n=300); 之後依Minnesota分類標準以進行詳細心電圖的比較。利用一標準病歷回顧方式, 收集每位個案的社會人口學和臨床資訊。標準判讀心電圖的實驗室(Epicare)對所有的臨床和分組資料均不知。心電圖根據下列五個主要的分類區分為正常和不正常(輕度和重度): 心肌梗塞或缺氧, 傳導缺損, 心律不整, 心房或心室肥大, 以及心室再極化延長。甲基安非他命與健康組間的心電圖標準化差異, 為根據此二組之特質差異進行調整。根據巢式個案對照研究設計, 本研究亦分別探討心電圖異常與所有死因、自然死因、非自然死因和心臟血管疾病死亡的相關性。 結果: Part I. 在研究過程中, 有130位個案死亡; 其中有63.1%為非自然死亡, 而36.9%為自然死亡。相對於一般族群, 本研究世代族群有過多死亡(SMR=6.02), 而女性比男性在非自然死亡部份, 有較高的SMR (p=0.001)。 Part II. 有關在研究期間的最後一次住院(latest admission), 財務獨立(financial independence)降低自殺危險性(adjusted risk ratio [aRR] = 0.33, p<.05), 而視幻覺(visual hallucinations)增高自殺危險性(aRR=2.57, p<.05)。對於出院後的期間(post-discharge period), 財務獨立仍降低自殺危險性(aRR] = 0.11, p<.05), 而自殺企圖(aRR = 8.78, p < .05)及憂鬱症候群(aRR = 3.28, p = .059)則增加自殺危險性。 Part III. 對於指標住院(index admission)的臨床特質, 以自然死亡為預後的多變項邏輯式迴歸顯示: 相對於AST正常組(<40 U/L), AST明顯上升組(>80U/L)的調整後危險比為其10.69 (p < .001), 而輕度上升組(40-80 U/L)之危險比則為4.71 (p < .01)。若以ALT而言, 明顯上升組與輕度上升組的5.22 (p < .01) and 5.13 (p < .01)倍。若以研究期間的最後一次住院, 使用抗精神病藥物則與低危險性有相關性。 Part IV. 在調整甲基安非他命組與正常組之間的差異, 連續心電圖指標中, 只有三個指標的調整z分數大於0.5 (medium effect size), 包括PR duration (-0.9422), QTc duration (0.9062) and QTI (0.9021)。這些指標的差異性均約在一個標準差。 甲基安非他命組比控制組有較高的心律不整比例(17.6% vs. 8.7%, p=0.001), 而其他四個指標則在二組無明顯差異。甲基安非他命組比控制組較易有心電圖重度異常(4.4% vs. 1.3%, p=0.015), 也容易有輕度異常(47.8% vs. 39.7%, p=0.003)。 此外, 在甲基安非他組中, 若依甲基安非他命使用的狀態, 與最近使用者(recent users)比較, 現在使用者(current users)除有較快的心跳外, 在其他指標並無明顯異常。心電圖的異常與甲基安非他命使用長短(duration of use)無明顯差異。 結論: Part I. 在死亡者中, 有不少為自然死亡, 但大多為死於非自然死亡。此發現提供明顯的證據對此特殊族群有價值的洞察。 Part II. 在此族群中所發現的自殺死亡保護和危險因子, 將可作為臨床介入和預防的重要參考。 Part III. 此研究強調甲基安非他命造成多器官系統損壞, 可能與自然死亡相關; 對特定具有升高AST/ALT的個案需要密切追蹤。 Part IV. 此研究確認在甲基安非依賴者的心電圖異常, 尤其是PR interval縮短和QT interval延長。此外, 建議能篩檢具有心電圖異常者, 此可能為潛在可治療心臟毒性的個案。Background: Since the 1990s, the abuse of methamphetamine has become one of the fastest growing illicit drug problems worldwide, particular in South-east Asia, Australia, and North America. Methamphetamine (MAMP) was reported to induce the physical and mental harms; the earlier ones include cardiovascular disorders and the later include depression, psychosis, and suicide. Methamphetamine use leads to increased likelihood of premature death. However, the empirical information regarding the mortality outcome and causes of death is limited in the literature. Besides, methamphetamine has a well-known cardiotoxicity in the humankind. Given the non-invasiveness, low cost, and wide availability, the electrocardiogram (ECG) may be a useful tool for assisting in the clinical assessment of various harms and prediction of future outcome in patients with methamphetamine dependence. A recent systematic review highlighted only eight cohort studies that investigated the mortality of amphetamine users in the literature, and three of these studies had a sample size of 500 or larger. The project was proposed to follow up a large cohort of inpatients (N=1254) with methamphetamine dependence who had ever been hospitalized due to methamphetamine-related problems by means of a record-linkage study for obtaining their mortality information. Based on the nested case-control study design and standardized chart reviewing process, the risk factors for suicide mortality and natural death separately are explored. Additionally, a healthy study group will be enrolled as the control group for the comparison of electrocardiographic abnormalities with the methamphetamine-dependent group. Design and Methods: Part I. In a consecutive series of 1,254 inpatients with methamphetamine dependence (diagnosed according to DSM-III-R and DSM-IV criteria) admitted to a psychiatric center in northern Taiwan from January 1, 1990, through December 31, 2007, was defined as the cohort and retrospectively studied. Diagnostic and sociodemographic data for each subject were extracted from the medical records based on a brief chart review process. Mortality data were obtained by linking to the National Death Certification System (1990-2007) and standardised mortality ratios (SMRs) were estimated. Totally, around 150 deceased subjects are estimated to be found. Part II. Based on the link process in part I, 38 deaths due to suicide were identified as cases, and 76 controls were randomly selected using risk-set density sampling in a 2:1 ratio, matched for age, sex, and the year of index admission. A standardized comprehensive chart review process was adopted to collate sociodemographic, psychopathological and clinical information for each study subject. Multivariate conditional logistic regression analysis was used to identify correlates of suicide among these patients. Part III. Based on the link process in part I, forty-five cases died of natural causes, each of which was then matched with four living controls on age, sex and the year of index admission using risk-set sampling. Clinical data and laboratory results during hospitalization were obtained through a standardized review of medical records. Part IV. This study enrolled two groups, a large-size roster of subjects (N=387) derived from a cohort of inpatients with methamphetamine dependence and a healthy control group (n=300), for the comparison of comprehensive ECG indices based on the Minnesota code classification. A standardized chart review process was adopted to collate sociodemographic and clinical information for each study subject. The laboratory (Epicare) for standardized ECG readings was blinded to all of clinical data and grouping information. ECG readings were classified into normal and abnormal (minor, major) by 5 categorical indices: myocardial infarction or ischemia, conduction defect, arrhythmias, atrial enlargement or ventricular hypertrophy, and prolonged ventricular repolarization. Differences of indices of electrocardiograms between the two groups were standardized against those of a healthy control group, with adjustment for covariates derived from the differences between methamphetamine and control groups. Based on the nested case-control study, this study also explored the ECG abnormality in association with all-cause, natural, unnatural and cardiovascular deaths separately. Results: Part I. During the study period, 130 patients died. Of them, 63.1% died from unnatural deaths, while the remaining 36.9% died from natural deaths. Relative to the general population, the cohort had excessive mortality (SMR of 6.02) and women had a higher SMR for unnatural deaths than men (p = 0.001). Part II. For the sociodemographic and symptom profiles at the latest admission, financial independence lowered the risk (adjusted risk ratio aRR = 0.33, p < .05) whereas visual hallucinations elevated the risk (aRR = 2.57, p < .05) for suicide. For the profiles during the post-discharge period, financial independence (aRR = 0.11, p < .05) remained associated with reduced risk to suicide, whereas suicide attempt (aRR = 8.78, p < .05) and depressive syndrome (aRR = 3.28, p = .059) were associated with increased risk of suicide. Part III. For clinical characteristics at the index admission, multivariate conditional logistic regression revealed that the adjusted risk ratio for natural death was 10.69 (p < .001) in the group with markedly elevated aspartate aminotransferase (> 80 U/L) and 4.71 (p < .01) in the mildly elevated group (40–80 U/L), with reference to the normal group (< 40 U/L). As for alanine aminotransferase, the adjusted risk ratios were 5.22 (p < .01) and 5.13 (p < .01). For the profiles at the latest admission during the study period, the use of antipsychotic drugs was identified to be associated with lowered risk. Part IV. After adjusting the covariates regarding the differences between two groups, only three indices of continuous indices had the adjusted z score greater than 0.5 (medium effect size), including PR duration (-0.9422), QTc duration (0.9062) and QTI (0.9021). The differences for these indices were around 1.0 standard deviation. Methamphetamine-dependent group had a higher proportion of arrhythmia than the healthy control group (17.6% vs. 8.7%, p=0.001), whereas other 4 indices were not significant different between two groups. Methamphetamine-dependent group were likely to have major abnormality of electrocardiogram compared with control group (4.4% vs. 1.3%, p=0.015), along with minor abnormality (47.8% vs. 39.7%, p=0.003). Additionally, stratified by the status of methamphetamine use, current users were not significantly different from recent users except higher heart beat rates. The electrocardiographic abnormalities were not associated with duration of methamphetamine use in methamphetamine group. Conclusions. Part I. A substantial proportion of the deceased died of natural deaths, but most of them died of unnatural deaths. The findings show significant evidence to provide a valuable insight for the premature deaths in the specific population. Part II. Both protective and risk factors for suicide mortality were found among inpatients with methamphetamine dependence, and the findings have implications for clinical intervention and prevention. Part III. This study highlights the damage of methamphetamine use to multiple organ systems, which is associated with natural death, along with the necessity of intensive follow-up among individuals with elevated aspartate or alanine aminotransferases. Part IV. This study confirmed the electrocardiographic abnormality in methamphetamine-dependent groups, predominantly shortened PR and prolonged QT intervals. Furthermore, it is highly suggestive for screening methamphetamine users with electrocardiographic abnormality, as a potentially treatable or manageable form of cardiotoxicity.1121425 bytesapplication/pdfen-US甲基安非他命預後死亡危險因子自殺自然死亡心電圖Methamphetamineoutcomemortalityrisk factorsuicidenatural deathelectrocardiogramabnormality[SDGs]SDG3thesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/250338/1/ntu-100-D94842001-1.pdf