思覺失調症病人首次住院時啟用抗精神病長效針劑藥物對於降低再住院風險的效果：2004至2017年的臺灣全國世代研究

Abstract

背景 使用抗精神病藥治療思覺失調症患者的一個主要挑戰是服藥不遵從。儘管長效針劑抗精神病藥物（以下簡稱長效針劑）已被證明具有更高的服藥遵從性，但在臨床實務中長效針劑並未被視為前線的用藥，且對其有效性的評估也易受到選擇性偏誤的影響。在最近一個在台灣首次入院思覺失調症患者世代中應用分層個體內 Cox擴展風險模型比較有效性的分析中，發現首次入院且出院後使用長效針劑對於預防再住院比使用口服抗精神病藥物更具有顯著效果。而尚不清楚的是，在第一次住院期間便使用長效針劑治療對於預防再住院方面是否也具有這種優勢。因此，本研究旨在比較台灣首次住院思覺失調症全國世代的患者在首次住院時便使用長效針劑治療對於降低再住院的有效性。而這些改善與臨床相關性也將被探索。 方法 本篇回溯性世代研究的患者來自國家健康保險研究資料庫中的精神科住院患者，納入標準為2004年至2017年期間發生首次精神科住院並在當次出院診斷中有思覺失調症或情感性思覺失調症診斷的患者。我們將這些患者依據首次住院期間以及出院後的治療分為三組：1）首次住院期間使用長效針劑且出院後未早期停藥；2）首次住院期間使用長效針劑但出院後早期停藥；3）僅使用口服抗精神病藥物。此外，我們回溯了入院前三年的抗精神病藥物處方記錄，以校正他們過去用藥經驗可能造成的干擾。我們使用Kaplan-Meier存活分析來估計精神科再住院風險。並且，由於變項可能不符合比例風險原則，我們同時使用了Cox比例風險模型和指數加速失效時間模型對潛在的干擾因素進行調整，分別估計出了再住院事件的調整風險比和調整時間比。 結果 在本研究納入的56211名首次住院思覺失調症患者中，3671名（6.5%）為使用長效針劑且未早期停藥組，5665名（10.1%）為使用長效針劑但早期停藥組，46875名（83.4%）則為僅接受口服抗精神病藥物組。以上三個暴露組的Kaplan-Meier存活曲線使用對數秩檢定後顯示有顯著差異（p < 0.0001），使用長效針劑且未早期停藥組的曲線高於僅接受口服抗精神病藥物組曲線，而使用長效針劑但早期停藥組的曲線低於僅接受口服抗精神病藥物組曲線。而與僅接受口服抗精神病藥物組相比，並調整性別、年齡、首次住院天數和過去抗精神病藥物用藥經驗後，使用長效針劑且未早期停藥組的調整風險比為0.88（95% CI：0.83-0.92）、使用長效針劑但早期停藥組的調整風險比為1.25（95% CI：1.21-1.30）；而相應的調整時間比分別為1.13（95% CI：1.07-1.18）和0.78（95% CI：0.75-0.81）。此結果在排除住院前使用過長效針劑治療的患者，以及改變中斷用藥的定義等敏感性分析中仍然相似。 結論 與首次住院期間僅使用口服抗精神病藥物治療相比，在此期間接受長效針劑治療且出院後未早期停藥的思覺失調症患者再住院風險降低12%，而住院期間接受長效針劑治療但早期停藥的患者再住院風險增加25%。我們的研究結果意味著促進長效針劑使用並預防早期中斷用藥對於降低思覺失調症患者的再住院風險具有重要意義。

Background A major challenge in the treatment of schizophrenia with antipsychotics is patients’ nonadherence. Although long-acting injectable antipsychotics (LAIs) have been demonstrated to have higher medication adherence, they have not been considered as front-line medications in clinical practice and the evaluation of their effectiveness has been confounded by this selection of recipients. In a recent comparative effectiveness analysis applying within-individual models among national cohorts of first-admission schizophrenia patients in Taiwan, the use of LAIs after the first admission was found to has notable advantages over oral antipsychotics in preventing readmission. What remains unknown is whether treatment with LAIs during the first admission would have had this advantage in preventing readmission. Hence, this study aims to compare the effectiveness of LAI treatment during the first admission in reducing readmission in national cohorts of first-admission schizophrenia patients in Taiwan. The clinical correlates of this improvement would be explored as well. Methods The retrospective cohorts examined in this study were derived from the National Health Insurance Research Database NHIRD for Psychiatric Inpatients with the inclusion criteria of patients having first psychiatric admission between 2004 to 2017 and being discharged with a diagnosis of schizophrenia or schizoaffective disorders. We divided first-admission schizophrenia patients into three groups by their treatment during the first admission and post discharge: 1) LAIs during the first admission without early discontinuation post discharge; 2) LAIs during the first admission with early discontinuation post discharge; and 3) solely on oral antipsychotics. In addition, we backtracked the three-year pre-admission for records of their antipsychotic prescription to adjust for potential confounding by their previous medication experiences. We conducted survival analysis to estimate the psychiatric readmission risk using Kaplan-Meier estimates. Because of the concern over non-proportionality, both Cox proportional hazards model and exponential accelerated failure time model were applied to estimate the adjusted hazard ratios (aHR) and time ratios (aTR), respectively, of readmission with adjustment for potential confounders. Results Among 56211 first-admission schizophrenia patients included in this study, 3671 (6.5%) received LAIs without early discontinuation, 5665 (10.1%) received LAIs with early discontinuation, and 46875 (83.4%) received solely oral antipsychotics. The Kaplan-Meier survival curves of the three exposure groups differed significantly using the log-rank test (p < 0.0001), with the curve of LAIs without early discontinuation higher than whereas that of LAIs with early discontinuation lower than that of solely on oral antipsychotics. Compared to the group of solely oral antipsychotics with adjustment for gender, age, length of first admission stay, and past antipsychotics experiences, the aHR was 0.88 (95% CI: 0.83-0.92) for the group of LAIs without early discontinuation and 1.25 (95% CI: 1.21-1.30) for the group of LAIs with early discontinuation, and the corresponding aTR were 1.13 (95% CI: 1.07-1.18) and 0.78 (95% CI: 0.75-0.81). The results of sensitivity analyses by excluding patients who had pre-admission LAIs treatment as well as changing the duration of discontinuation remain similar. Conclusions Compared to treatment solely on oral antipsychotics during the admission, first-admission schizophrenia patients who received treatment with LAIs during the period and without early discontinuation post discharge decreased the hazard of readmission by 12 %, whereas those received LAIs but had early discontinuation increased the hazard of 25%. Our findings have implications for both the promotion of LAIs and the prevention of their early discontinuation to reduce the readmission risk of schizophrenia patients.