念珠菌菌血症之危險因子以及臨床治療結果分析：著重於抗黴菌藥物之治療對於預後之成效影響

Abstract

目的： 分析近年臺大醫院念珠菌菌血症的（candidemic）流行病學，並評估念珠菌菌血症病人治療之處方模式、感染念珠菌菌血症之死亡率及影響死亡的危險因子分析。另外也針對使用不同的抗黴菌藥物對念珠菌菌血症病人預後之影響。 研究設計、地點及研究對象： 此研究於國立臺灣大學醫學院附設醫院，位於台灣北部的醫學中心，以蒐集病歷資料進行單醫學中心的回溯性世代分析研究。自2009年1月1日至2009年6月30日間確定有感染念珠菌菌血症的病人，只納入第一次感染事件，並排除未滿18歲或病歷資料不全的病人。 研究方法： 以研究開始前設計之病歷個案報告表查閱病人紙本和電子病歷記錄研究對象之相關資料，包含病人基本資料、合併症及造成感染的潛在因子、黴菌血症發作前30天是否有細菌或黴菌感染之情形以及治療方式和抗生素或抗黴菌藥物使用情形、黴菌血症發作時之血液培養結果和菌株對抗黴菌藥物敏感性測試結果、臨床表徵、相關檢驗數據、治療結果。本研究的主要觀察點是病人於30天內的死亡率。 統計方法包含卡方檢定（χ2 test）、無母數檢定（Mann-Whitney U test）、T檢定（t-test）、費雪精確檢定（Fisher’s exact test）。存活分析使用Kaplan-Meier method繪製存活曲線並以Log-rank test比較差異。死亡危險因子分析利用單變項分析及多變項羅吉斯逐步回歸（Logistic stepwise regression）分析。 結果： 本研究共納入126位感染念珠菌菌血症的病人，其中單純黴菌感染的病人有89位。病人平均年齡為64歲，範圍為18歲到98歲，男與女比例為1.42：1.0（74 vs. 52），有95.2%的病人為院內感染，Charlson’s comorbidity score中位數為6分，病人合併症最多的為癌症（72.2%）、其次為心血管疾病（44.4%）、腸胃道疾病（39.7%），平均住院天數為44天；在菌種分布方面，有六成以上的感染是由C. albicans（61.1%）引起，然後發生率由高到低依序為C. tropicalis（19.0%）、C. glabrata（18.3%）、C. parapsilosis（8.7%）、C. krusei（0.8%）。黴菌血症發作時臨床表徵以敗血性休克最多（52.4%），原發性黴菌血症佔31.0%，黴菌血症發作時的Pitt bacteremia score中位數為3分、APACHE II score中位數為22分。 在治療處方模式方面，針對感染C. albicans、C. tropicalis和C. parapsilosis黴菌血症的病人，治療藥物以fluconazole為主，疾病程度較為嚴重、治療一段時間後臨床症狀未改善或嗜中性白血球低下者可能會選用polyenes類或echinocandins類藥物作為治療選擇；針對感染C. glabrata的病人，會根據抗黴菌藥物對菌株的最小抑菌濃度（MIC）作為選擇治療藥物的參考；感染C. krusei黴菌血症的病人，由於本研究中僅有一位病人的黴菌血症是由C. krusei所引起（使用的抗黴菌治療藥物為amphotericin B），故無法看出其大致上的處方模式為何；且因為受限於病人數不足，也無法看出針對特定菌種感染，使用不同抗黴菌藥物對預後的影響。 本研究中全體病人在念珠菌菌血症發作後產生的併發症主要有以下幾種：眼內炎（8.7%）、腦膜炎（4.0%）、心內膜炎（4.0%）、急性腎臟衰竭（27.0%）、呼吸衰竭（17.5%）等，不過產生這些併發症與否對於30天死亡率並沒有顯著的影響（p＝0.36）。 本研究針對全體病人在感染後30天內死亡率為57.1 %（單純黴菌血症感染者為50.6%）；分析影響30天內死亡率的獨立危險因子，較差的腎功能（OR：1.01；95% C.I.：1.00-1.02；p＝0.0148）、Charlson's comorbidity score分數愈高（OR：1.17；95% C.I.：1.02-1.35；p＝0.03）、在此次念珠菌菌血症感染前曾經有過念珠菌的感染或移生（OR：3.33；95% C.I.：1.39-7.97；p＝0.0071）、念珠菌菌血症發作時伴隨有敗血性休克（OR：4.94；95% C.I.：1.97-12.43；p＝0.0007）、念珠菌菌血症發作後未移除或更換血液導管（OR：5.80；95% C.I.：2.61-12.86；p＜0.0001），為造成念珠菌菌血症30天內死亡率愈高的獨立危險因子。除此之外，使用適當且劑量足夠之抗黴菌藥物來治療念珠菌菌血症會使30天內死亡率較低（OR：0.02；95% C.I.：0.003-0.10；p＜0.0001）。 結論： 感染念珠菌菌血症的病人，30天內的死亡率和病人本身情況的嚴重程度（Charlson’s comorbidity score、腎功能）以及黴菌血症發作時之嚴重程度（敗血性休克）有關；另外，在此次念珠菌菌血症感染前曾經有過念珠菌的感染或移生、發作後未移除或更換血液導管也都是使30天死亡率升高的原因；反之，使用適當且劑量足夠之抗黴菌藥物來治療念珠菌菌血症會使30天內死亡率較低。

Objectives: The aim of this study is to evaluate the epidemiology of candidemia in National Taiwan University Hospital and the impact of adequate antifungal therapy on clinical outcomes. To analyze mortality and prognostic factors of candidemia is another goal of this study. Study design and study populations: A retrospective cohort study was performed by charts reviewing for all adult patients admitted to the National Taiwan University Hospital（NTUH）, a medical center in northern Taiwan, with Candida bloodstream infection since January 1st, 2009 to June 30th, 2009. Only the first episode of each patient was included. Patients younger than 18 or whose medical records were incomplete were excluded. Methods: Data were collected from paper medical reports and hospital computerized databases, including patients’ profiles, underlying diseases, comorbidities, predisposing factors for infection, previous infection history, antibiotics or antifungal agents exposure before fungemia onset, clinical presentation when fungemia onset, relevant laboratory data, antifungal therapy during treatment period, treatment outcome. The primary endpoint was 30 day all-cause mortality. Statistical methods included Chi-square test（χ2 test）, Mann-Whitney U test, T test, Fisher’s exact test. Survival curves shown by Kaplan-Meier method were analyzed with Log-rank test. Prognostic factors were examed using univariate analysis and multiple logistic regression analysis. Results: One hundred and twenty-six patients with Candida bloodstream infection were enrolled in this study, and 89 patients were pure fungemia among all the patients. The average age was 64 years old (ranged from 18 to 98). The proportion of male to female patients was approximately 1.42:1.0. One hundred and twenty (95.2%) patients were classified as nosocomial infection. Median of Charlson’s comorbidity score was 6. The most common underlying diseases were malignancy (91 episodes, 72.2%), followed by cardiovascular diseases (56 episodes, 44.4%) and gastrointestinal disorders (50 episodes, 39.7%). The average length of stay in hospital ranged from 2 to 1291 days (median, 44 days). The clinical manifestation, septic shock, at fungemia onset was the most common (66 episodes, 52.4%). Median of Pitt bactermia score was 3 and APACHE II score was 22. The distribution of Candida species in National Taiwan University Hospital was showed as follows: C. albicans (61.1%), C. tropicalis (19.0%), C. glabrata (18.3%), C. parapsilosis (8.7%), C. krusei (0.8%). As for treatment of candidemia, patients with C. albicans, C. tropicalis, or C. parapsilosis infection would be mostly treated with fluconazole. But for critically-ill, neutropenia, or clinically deteriorate after a period time of treatment patients, polyenes or echinocandins would be drugs of choice. And for patients with C. glabrata infection, minimum inhibitory concentration (MIC) would guide the antifungal treatment. Since there was only one patient infected with C. krusei in this study, the treatment rule of C. krusei could not be told. Besides, we could not evaluate the impact of different antifungal agents on prognosis due to inadequate patient number of this study as well. Acute renal failure (27.0%), respiratory failure (17.5%), endophthalmitis (8.7%), meningitis (4.0%), and endocarditis (4.0%) were complications of candidemia. However, these complications were all statistically insignificantly related to 30 day mortality. The 30 day all-cause mortality rate was 57.1% (72/126). Multivariate analysis on the 30 day all-cause mortality showed that poor renal function (OR: 1.01; 95% C.I.: 1.00-1.02; p=0.0148), higher Charlson’s comorbidity score (OR: 1.17; 95% C.I.: 1.02-1.35; p=0.03), previous infection or colonization of Candida species (OR: 3.33; 95% C.I.: 1.39-7.97; p=0.0071), septic shock (OR: 4.94; 95% C.I.: 1.97-12.43; p=0.0007) and unchange of blood catheters (OR: 5.80; 95% C.I.: 2.61-12.86; p＜0.0001) were factors that made 30 day mortality rate higher. On the other hand, adequate antifungal therapy (OR: 0.02; 95% C.I.: 0.003-0.10; p＜0.0001) would make 30 day mortality rate lower. Conclusions: Poor renal function, Charlson’s comorbidity score, septic shock, previous infection or colonization of Candida species, unremove or unchange of blood catheters, and inadequate antifungal therapy were found to be associated with poor prognosis by multivariate analysis in patients with candidemia.