海洛因成癮患者長期健康影響：美沙冬治療、降血脂斯達汀類藥物與死亡之風險評估

Abstract

海洛因成癮是慢性疾病，影響個案的心理社會功能，且造成個案的長期健康影響包括各種疾病的共病性，如愛滋病感染。口服美沙冬為海洛因成癮的標準治療藥物，其最佳的治療劑量以及治療共病性對個案的長期健康相當重要。雖然愛滋病能夠以藥物有效治療，但個案可能處於長期的發炎狀態。斯達汀類降血脂藥物的使用，可能為愛滋病個案長期健康的重要可變因子。本研究論文以三個研究目標來進行分析：（1）估計國內美沙冬治療模式可能產生的健康效益；（2）以台灣疾病管制局的美沙冬治療資料庫串連死亡檔，分析美沙冬治療的藥物劑量、個案出席率、以及接受治療的總日數與死亡風險的相關；（3）以台灣健保資料庫，分析使用斯達汀類藥物與愛滋病個案死亡率的相關性。方法：第一部分，研究根據目前國內約十萬名海洛因成癮個案，接受美沙冬門診治療的費用、生活品質校正人年改善的程度，估算出美沙冬治療方式的遞增健康效益比（每增加一個健康校正人年所需的治療花費，ICER）。第二部分，研究使用台灣疾病管制局2006至2008年的美沙冬治療資料庫，總計33603名個案的全部出席與藥物劑量資料，並串連至衛生署的死亡資料庫勾稽個案至研究追蹤結束時的存活狀態。以Cox存活統計法分析美沙冬藥物劑量等治療因子與個案死亡的相關性。第三部分，研究使用台灣健保資料庫，並以1997-2008愛滋病個案領取重大傷病卡為對象，總計5183名，計算個案的斯達汀類藥物使用量，並以Cox存活統計法分析藥物使用與死亡風險的相關。結果：研究第一部分，個案的遞增健康效益比（ICER）約為4170美元/年至10420美元/年之間。研究第二部分，美沙冬門診個案的粗死亡率約為134.78/萬人年，而標準化死亡比約為4.68倍。以存活統計分析發現，年齡較大（勝算比，HR＝2.16）、愛滋病毒感染（HR＝2.14）明顯與高死亡率較有相關。較高的美沙冬藥物劑量組，死亡風險較低，且具有劑量效應（HR分別為0.96，0.75，及0.68）本研究曾分析一群2201名美沙冬門診個案，發現B型肝炎、C型肝炎、愛滋病毒、及梅毒感染率分別為26.4%, 78.2%, 10.8%, and 6.0%,。研究第三部分，分析總共5183名新感染愛滋病個案，總計追蹤28323人年，有576人死亡，使用time-dependent Cox proportional hazard分析，使用斯達汀類藥物組，校正後死亡風險（HR）約為0.42（0.31-0.56）。進一步針對使用斯達汀類藥物的總劑量（28-89、90-180、及180 DDD以上）分組，校正後死亡風險HR各為0.60、0.40及 0.16。結論：本論文探討了海洛因成癮個案的長期健康影響因子。研究發現在台灣使用美沙冬門診的治療方式，相當符合成本效益。個案接受美沙冬門診治療，在高劑量組的死亡風險較低。這些個案合併病毒感染的比率相當高。進一步以健保資料庫探討愛滋病毒感染個案使用斯達汀類藥物的狀況，分析發現服用這類藥物的個案，死亡風險比較低。在海洛因成癮的個案，尋找可改善個案長期健康狀況的治療因子相當重要，這對台灣也是相對比較新的經驗，值得朝此方向繼續努力。

Background: Heroin addiction is a chronic disease with impact on the patient's long-term health outcome, including HIV infection and other comorbidities, in addition to psychosocial decompensation. Methadone treatment is the standard maintenance treatment. Optimal treatment of heroin addiction with methadone service and treatment of comorbidity is important in overall health outcome of these patients. Although the survival of Human Immunodeficiency Virus (HIV) improved significantly in the era of highly active anti-retroviral treatment, chronic inflammation may be persistent. Statins may have protective effects against chronic virus infection, and hence may be an important modifiable treatment-related factor in the long-term course of HIV infection patients as well as those comorbid with heroin addiction. By using methadone treatment registry by the Centers for Disease Control and reimbursement database of National Health Insurance (NHI) in Taiwan, we conducted explorations with three specific aims: (1) to estimate the utility of nationwide methadone treatment service; (2) to investigate whether dosage, attendance rate, and overall treatment duration are associated with the reduction of mortality risk in heroin addiction patients; (3) to investigate if there is an association of statin use with mortality risk reduction of HIV patients. Methods: In the first part, we conducted an estimation of utility gained from the nationwide methadone treatment service, based on the parameters obtained from current methadone service in Taiwan. In the second part, we conducted a population-based cohort study from Taiwan's Methadone Registry maintained by the Centers for Disease Control. A total of 33,603 patients were included as the study cohort, with complete records of methadone treatment from 2006 to 2008. All-cause mortality was tracked by the linage to national mortality registry from the Department of Health. Cox proportional hazard regressions were employed to calculate the hazard ratios (HRs) for the association between the methadone dosing and overall mortality risk. In the third part, we used the National Health Insurance Research Database to analyze the health outcome of HIV infection patients, because of the difficulty in identifying heroin addiction in the NHI Research Database. A total of 5183 HIV-infected patients were included as the study cohort. Each subject was individually tracked from 1997 to 2008 to identify mortality since 1999. Subsequent use of statins was identified. Cox proportional hazard regressions were employed to calculate the hazard ratios (HRs) for the association between use of statins and occurrence of mortality in the HIV-infected cohort. Results: In the first part, range of Incremental Cost-Effectiveness Ratio (ICER) was estimated, with sensitivity analysis of difference in gain of Quality Adjusted Life-Year. As the annual cost for administration of methadone program was about 40,000 NTD, which leads to an incremental cost of 133,300 to 333,300 NTD, or 4,170 to 10,420 USD (1 USD=32 NTD) per QALY, the MMT strategy may be cost-effective. In the second part, the overall crude mortality rate is 134.78/10,000 person-years, and the standardized mortality ratio in reference to age-specific general population is approximately 4.68. In the adjusted model of Cox proportional hazard analysis, older age (adjusted hazard ratio, HR = 2.16, P<0.001) and positive HIV infection status (adjusted HR = 2.14, P<0.001) were associated with higher mortality risk. For treatment-related factors, the higher-dosage group has lower mortality risk than the lower-dosage group (adjusted HR＝0.96, 0.75, and 0.68 to reference group, P = 0.03). In one methadone clinical sample study with total 2201 patients, positive rate of HBV, HCV, HIV, and syphilis were 26.4%, 78.2%, 10.8%, and 6.0%, respectively. In the third part, a total of 5183 newly infected patients with HIV were included, with 28323 person-years of follow-up; 576 deaths and 184 cancer diagnoses were recorded during the follow-up period. Among the 480 statin users, the adjusted hazard ratios of all-cause mortality were 0.42 (0.31-0.56) relative to statin non-users in time-dependent Cox regression analysis. The adjusted hazard ratios in 28-89, 90-180, and more than 180 total DDDs are 0.60 (0.24-1.49), 0.40 (0.12-1.28), and 0.16 (0.07-0.38), in a dose-response pattern following adjustments for confounders. Conclusion: This thesis explores the long-term health outcome of heroin addicted patients. The result from the utility estimation of heroin addicted patients reveals that methadone significantly improved the overall utility of the patients and is quite cost-effective in current methadone clinic service in Taiwan. The result from the exploration of methadone treatment database reveals that higher methadone dosage was associated with lower all-cause mortality, indicating the health benefit of methadone service in terms of mortality as an outcome. The result from the exploration of a local methadone clinic confirms a significant proportion of HCV and HIV infection comorbidity in heroin addicted patients. By using the HIV patients as a surrogate in NHI database, the exploration of statin use in HIV patients reveals an association of lower mortality risk in statin users. Since heroin addiction is a major health problem in addiction psychiatry and public health in addition to social welfare, modifiable factors related to the treatment for these patients are important in designing and providing adequate treatment setting, which is new to Taiwan.