某醫學中心治療Candida tropicalis菌血症之抗黴菌藥品處方型態及療效分析

Abstract

背景： 念珠菌菌血症有高死亡率及罹病率，而Candida albicans是念珠菌菌血症中最常見的菌種，惟近年來非Candida albicans念珠菌菌血症發生率逐漸上升。臺灣地區的Candida tropicalis為非C. albicans中最常見的菌種，其治療首選藥物為fluconazole。雖然文獻顯示C. tropicalis對於fluconazole的抗藥性有升高的趨勢，但其抗藥性比例並不一致，因此fluconazole在臨床上是否仍適合做為首選藥物為待研究的議題。 研究目的： 分析C. tropicalis菌血症的抗黴菌藥物處方型態，包括比較fluconazole每日劑量除以最小濃度之比值與菌血症治療成功率、死亡率間的相關性。同時與最常見的C. albicans菌血症相比，評估感染C. tropicalis菌血症的危險因子與影響治療成功率、死亡率的因子。 研究設計： 本研究納入2011年4月1日至2012年3月31日期間於臺大醫院治療C. albicans或是C. tropicalis菌血症的住院病人，以病歷回顧的方式收集病人的基本資料、合併症、可能影響預後的潛在因子、菌血症發作前所使用的抗生素或抗黴菌藥物使用情形、相關的血液培養和對抗黴菌藥物的感受性結果、臨床表徵、所使用的抗黴菌藥物、相關的檢驗數值和治療結果。 利用SAS統計軟體（9.2版）整理資料及統計分析。使用卡方檢定、費雪精確檢定、t-test、Kruskal-Wallis test以及Mann-Whitney U test。存活曲線由Kaplan-Meier method繪製並以Log-rank test比較差異。感染C. tropicalis菌血症的獨立危險因子使用multinominal logistic regression、而死亡危險因子則使用單變項分析與stepwise logistic regression分析。 研究結果： 本研究共納入185位菌血症病人，其中110位為單純C. albicans、42位為單純C. tropicalis念珠菌菌血症以及33位多重菌株感染的病人，全體病人的年齡平均為63.0歲，男性占57.3%。病人以院內感染為主，三組病人的合併症皆以癌症為主。單純C. tropicalis菌血症病人中，33.3%為嗜中性白血球低下、使用過化學治療藥品的病人則佔61.9%，與其他兩組間有顯著的差異，p值分別為 < 0.01及0.01。單純C. tropicalis菌血症從抽血到報告為酵母菌的時間顯著短於單純C. albicans菌血症病人（p = 0.03）。大部分的病人在菌血症發作前皆曾接受抗生素治療，而三組病人在菌血症發作時的疾病嚴重程度並無統計上顯著的不同。multinominal logistic regression比較單純C. albicans菌血症與單純C. tropicalis菌血症發現，嗜中性白血球低下會顯著增加單純C. tropicalis菌血症的危險 （OR：5.166，p < 0.01）。 單純C. tropicalis菌血症的死亡率有較單純C. albicans菌血症的死亡率高的趨勢（第28天死亡率：59.5% vs. 43.6%）。菌血症發作後第28天兩者的治療成功率：單純C. albicans菌血症為57.3%，相較於單純C. tropicalis菌血症為38.1%有統計上顯著的不同（p = 0.03）。單純C. tropicalis菌血症病人中，未接受經驗性療法者於菌血症發作後第28天的死亡率顯著高於有接受經驗性治療者（69.7% vs. 22.2%，p = 0.02）。若分析所有未接受經驗性治療的病人，單純C. tropicalis菌血症於發作後第28天的死亡率顯著高於單純C. albicans菌血症（69.7% vs. 42.4%），p值為0.01。單純C. albicans菌血症有使用確切治療的病人治療成功率顯著高於單純C. tropicalis菌血症，在菌血症發作後的第28天兩組病人的死亡率有統計上顯著的不同（34.1% vs. 54.1%，p =0.04）。在抗黴菌藥品治療方面，使用echinocandin類藥品治療C. tropicalis菌血症的死亡率及治療成功率與fluconazole並無統計上顯著的不同；但對於持續性的單純念珠菌菌血症，echinocandin類藥品相對於fluconazole有治療成功率較高、死亡率較低的趨勢。 C. tropicalis之fluconazole的感受性顯著低於C. albicans：82.2%的C. tropicalis落於有感受性（MIC ≤ 2 μg/mL）的範圍、而C. albicans則是99.1%。在APACHE II score ≤ 17分者，fluconazole劑量除以最小抑菌濃度大於400時，C. tropicalis菌血症病人的預後較佳的趨勢。本研究中所有單純菌血症或是C. tropicalis菌血症於發作後第28天的獨立死亡因子皆為菌血症發作當天的APACHE II score > 17分，OR分別為3.904（p < 0.01）以及4.267（p = 0.03）。 結論： 與C. albicans菌血症相比：嗜中性白血球低下會顯著增加C. tropicalis的風險，且C. tropicalis之fluconazole的感受性顯著低於C. albicans。在菌血症發作後，C. tropicalis菌血症的死亡率有較C. albicans菌血症高的趨勢，而菌血症發作當天的APACHE II score > 17分為單純菌血症及C. tropicalis菌血症發作後第28天死亡的獨立危險因子。 關鍵詞： C. tropicalis、念珠菌菌血症、fluconazole、最小抑菌濃度、死亡因子

Background： Candidemia leads significant mortality and morbidity. Candida albicans is the most common species in patients with candidemia. It has been a documented increase in the proportion of candidemia caused by non-albicans species recently. Candida tropicalis is a predominating species in non-albicans in Taiwan, and fluconazole is the drug of choice in treating C. tropicalis infection. However, literatures reported rising resistance to fluconazole in C. tropicalis. Therefore, it is a concern that fluconazole is still the drug of choice in the treatment of C. tropicalis fungemia. Objective： The aims of this study are to evaluate patients’ characteristics, prognostic factors, antifungal therapy prescribing patterns, mortality rate and treatment outcome in patients with C. tropicalis fungemia, in comparison with patients with C. albicans fungemia. Study design： This retrospective study is part of a prospective study, and it entails reviewing medical record information on all patients who were diagnosed with fungemia due to C. albicans or C. tropicalis in the National Taiwan University hospital (NTUH) during April 1, 2011 to March 31, 2012. The exclusion criteria is that patients who were younger than 18 years old when candidemia onset. Clinical information was documented on case report forms, which included time and location of positive cultures, underlying diseases and severity of illness, laboratory data, and antifungal treatment of the given episode of fungemia, microbiological data, clinical and laboratory outcomes, demographic informations, gender, height, weight, and past medical history. The minimum inhibition concentration (MIC) was tested by Sensititre YeastOne Panels. Data were analyzed using SAS 9.2 software, and statistical methods included Chi-squre test, Fisher’s exact test, t test, Kruskal-Wallis test and Mann-Whitney U test. Survival curve shown by Kaplan-Meier method was analyzed with Log-rank test. Independent risk factors for C.tropicalis fungemia were examined using multinomial logistic regression; prognostic factors of monomicrobial candidemia and C. tropicalis fungemia were examined using univariate analysis and multiple logistic regression analysis. Results： A total of 185 patients were enrolled in this study, including 110 patients with C. albicans and 42 patients with C. tropicalis fungemia and 33 patients with polymicrobial infection. Their mean age was 63 years old, and 57.3% were male. Most of patients were diagnosed nosocominal fungemia. The most common underlying disease was malignancy in three subgroups. Statistically differences in patients with C. tropicalis fungemia by comparing with C. albicans fungemia and polymicrobial infection: 33.3% were neutropenic patients (p < 0.01), 61.9% had received chemotherapy within 30 days prior to candidemia onset (p = 0.01). The time from blood culture was drawen to report yeast is shorter in C. tropicalis fungemia than C. albicans fungemia (p = 0.03). The majority of the patients had received antibiotics treatment within 30 days prior to onset of candidemia, and there was no significant difference of disease severity among patients with different type of fungemia. In multinominal logistic regression, the only difference between C. tropicalis and C. albicans fungemia was a higher proportion of neutropenic patients in C. tropicalis fungemia (OR = 5.166, p < 0.01). C. tropicalis fungemia was more likely to have to higher mortality rate (59.5% vs. 43.6%, p = 0.08), and to have lower treatment success rate (38.1% vs. 57.3%, p = 0.03) on day 28 after candidemia onset. A higher proportion of receiving empirical antifungal therapy was in patients with C. albicans fungemia, and among C. tropicalis fungemia, patients who received empirical therapy had a lower mortality rate compared with those who didn’t. A significant higher mortality rate on day 28 in patients with fungemia due to C. tropicalis and didn’t receive empirical therapy comparing with patients with fungemia due to C. albicans (42.4% vs. 69.7%, p = 0.01). The successful response rate for candidemia due to C. albicans was higher than candidemia due to C. tropicalis in patients who received definitive antifungal therapy (67.0% vs. 43.2%, p = 0.01), and the mortality rate was statistically different on day 28 (34.1% vs. 54.1%, p = 0.04). The mortality rate and the response rate did not differ in patients with C. tropicalis fungemia who treated with different antifungal therapy The susceptibility of fluconazole was documented lower in C. tropicalis than C. albicans: 82.2% in C. tropicalis and 99.1% in C. albicans were susceptible (MIC ≤ 2 μg/mL). The successful response rate was higher in patients whose ratio of fluconazole daily dose/MIC is over 400 than whose ratio below or equal to 400 when their APACHE II score were ≤ 17 at candidemia onset. The multivariate analysis showed that APACHE II score > 17 was the independent risk factor for death in patients with monomicrobial fungemia (OR: 3.904, p < 0.01) or patients who with C. tropicalis fungemia (OR: 4.267, p = 0.03). Conclusion： The difference between candidemia due to C. tropicalis and C. albicans was that more neutropenic patients in C. tropicalis fungemia. The mortality rate of C. tropicalis fungemia was higher than C. albicans fungemia, and APACHE II score > 17 was the independent risk factor for death on day 28 in patients with candidemia or patients with C. tropicalis fungemia.