綠膿桿菌菌血症之預後因子及感染多重抗藥性綠膿桿菌菌血症之危險因子分析：著重於抗生素治療之影響

Abstract

研究目的： 評估合併或單一抗生素治療以及經驗性治療適當與否對於綠膿桿菌菌血症患者之臨床治療效果，並辨別導致病人死亡之危險因子；另外也針對多重抗藥性綠膿桿菌菌血症患者進行危險因子分析。 研究設計、地點及對象： 此研究於國立臺灣大學醫學院附設醫院-臺灣一家醫學中心進行，為單中心、回溯性世代分析研究。研究對象為西元2007年1月1日至2007年11月31日期間感染綠膿桿菌菌血症之成年病患。 研究方法： 自行設計之個案報告表為本研究收集資料之基本工具，記錄紙本病歷以及電子病歷所記載之相關變項，包含病人基本資料、合併症、造成感染之潛在因子等，並於綠膿桿菌菌血症發作前三十天內，記錄是否有細菌感染之情形以及所使用的抗生素；在綠膿桿菌菌血症發作時，記錄可能感染源及臨床表徵、相關檢驗數據、細菌檢驗結果及感染嚴重程度之評估；於綠膿桿菌菌血症發作後，記錄使用之抗生素、後續細菌培養結果以及相關之預後因子，並以菌血症發作後之三十天（D30）死亡率為觀察終點。另外也針對感染多重抗藥性綠膿桿菌菌血症之病人進行危險因子之分析。統計方法包含單變項分析以及多變項羅吉斯迴歸分析，存活曲線以Kaplan-Meier method繪製，並以Log-rank test進行比較。 研究結果： 本研究共收集了186個事件，有120事件（65％）為單一菌種感染；病人之平均年齡為61.2±18.4歲，男性佔56％；病人之平均Charlson’s comorbidity index為4.09±2.37分，過去一個月內有使用過抗生素及抗綠膿桿菌抗生素之比例分別為76.2％及43.8％。菌血症發作當下病人之平均Pitt bacteremia score為3.8±3.5分，39.8％之患者有敗血性休克；經驗性治療使用合併治療者僅有14事件（7.7％），確切治療使用合併治療者有38事件（24.4％）；由菌血症發作至使用適當抗生素之平均時間為34±46小時。有27事件（14.2％）為多重抗藥性綠膿桿菌菌血症感染。全體及單一菌種病人之D30總死亡率分別為41.4％及46.7％。 以單一菌種感染來說，經驗性合併或單一治療對於經驗性治療適當與否並無顯著之關係（77.8％ vs. 61.5％，P=0.48），而經驗性合併或單一治療對於死亡率也無顯著影響（44.4％ vs. 46.8％，P=1.00）；另外，合併或單一治療對於是否會減少重疊感染、抗藥性菌種之產生於統計上並未有顯著之差異。 由多變項羅吉斯迴歸分析發現，感染MDRPA之危險因子有過去六個月內曾有MDRPA感染或移生（OR 28.27，P<0.0001）、過去三十天內有使用過carbapenem（OR 12.2，P=0.0001）、過去三十天內使用過fluoroquinolone（FQ）（OR 9.30，P=0.005）、腦血管疾病（OR 5.45，P=0.02）以及此次菌血症感染前之住院天數（OR 1.01，P=0.02）。 由多變項羅吉斯迴歸分析發現影響全部病人第七天（D7）死亡率的獨立危險因子有肝膽疾病（OR 5.15，P=0.0124）、Pitt bacteremia score （OR 1.93，P<0.0001）以及年齡（OR 1.04，P=0.02）；影響單一菌種感染病人之D7死亡率的危險因子有：急性腎衰竭（OR 6.12，P=0.02）以及Pitt bacteremia score（OR 1.97，P<0.0001）。而影響全部病人D30死亡率之獨立危險因子有敗血性休克（OR 4.15，P=0.008）、肝膽疾病（OR 3.82，P=0.02）、之前使用過抗綠膿桿菌抗生素（OR 3.11，P=0.02）、免疫抑制劑之使用（OR 2.98，P=0.02）、Pitt bacteremia score （OR 1.52，P<0.0001）以及Charlson’s comorbidity index（OR 1.22，P=0.04），而多重菌種感染為保護因子（OR 0.31，P=0.02）；影響單一菌種感染病人D30死亡率的危險因子有：院內感染（OR 7.64，P=0.01）、敗血性休克（OR 4.97，P=0.011）、免疫抑制劑之使用（OR 4.37，P=0.006）以及Pitt bacteremia score（OR 1.46，P=0.007）。經驗性合併治療與否以及經驗性治療適當與否於分析中並未顯著影響預後。 結論： 經驗性治療並不會影響到綠膿桿菌菌血症患者之預後；D30死亡率僅與病人本身之情況（Charlson’s comorbidity index、肝膽疾病、是否有使用免疫抑制劑）以及菌血症發作時之嚴重程度（Pitt bacteremia score、敗血性休克）有關；另外，過去有使用過FQ或是carbapenem者、住院較久、腦血管疾病者、先前已有多重抗藥性綠膿桿菌感染或移生者為易感染多重抗藥性綠膿桿菌菌株之危險族群。

Objectives: This study is aim to evaluate the impact of antibiotic combination therapy and adequate empirical treatment on clinical outcomes and to identify the prognosis factors for Pseudomonas aeruginosa bacteremia. The other goal of this study is to investigate the risk factors for multidrug-resistant Pseudomonas aeruginosa (MDRPA) bacteremia. Study location、design and study populations: A retrospective cohort analysis was performed for all adult patients admitted to the National Taiwan University Hospital (NTUH) with positive P. aeruginosa blood culture between 1 January 2007 and 31 November 2007 . Methods: Data were collected from medical records and hospital computerized databases. The data retrieved for each patient included patients’ profile, underlying diseases, previous hospitalization history, previous P. aeruginosa infection or colonization history, antibiotics exposure before bacteremia onset, clinical presentation, antibiotics regimens at bacteremia onset, management and clinical response after bacteremia. The primary endpoint was 30 day all-caused mortality. Risk factors of MDRPA bacteremia were also analysed. Risk factors and clinical outcomes were examed using univariate analysis and multivariate logistic regression analysis. Survival curves shown by Kaplan-Meier method were compared with Log-rank test. Result: One hundred and eighty six episodes were included in this study, 120 (65%) were monomicrobial infections and 66 polymicrobial infections. Patients’ average age was 61.2±18.4 and 56% of them were male. The average Charlson comorbidity index of included patients was 4.09±2.37. Previous exposure to antibiotics and anti- Pseudomonas antibiotics were 76.2% and 43.8% respectively. Patients’ average Pitt bacteremia scores were 3.8±3.5. All episodes were presented with sepsis with 39.8% of septic shock. After the onset of symptoms, it took an average of 34±46 hours to initiate an adequate therapy. Empirical combination therapy was given only in 14 episodes (7.7%) and definitive combination therapy was used in 38 episodes (24.4%).There were 27 (14.2%) MDRPA episodes. The overall D30 mortality was 41.4%, and 46.7% for monomicrobial infections. Empirical monotherapy and combination therapy provided comparably adequate therapy (77.8% vs. 61.5 %, P=0.48) with similar D30 mortality in combination therapy and monotherapy (44.4% vs. 46.8%, P=1.00). Also, the superinfection and emergence of resistance after the bacteremia onset did not differ significantly between two regimens. Multivariate logistic regression identified the following variables as significant independent risk factors for MDRPA bacteremia: P. aeruginosa infection or colonization within previous 6 months (OR 28.27, P<0.0001), fluoroquinolone usage (OR 9.30, P=0.005), carbapenem usage (OR 12.2, P=0.0001), cerebrovascular accidents (OR 5.45, P=0.02) and lengh of hospital stay before bacteremia (OR 1.01, P=0.02). Each of the following factors were independently associated with 30-day mortality in all patients: septic shock (OR 4.15, P=0.008), hepatobiliary diseases (OR 3.82, P=0.02), exposure to anti- Pseudomonas agents (OR 3.11, P=0.02), immunosuppressant (OR 2.98, P=0.02), Pitt bacteremia score (OR 1.52, P<0.0001) and Charlson’s comorbidity index (OR 1.22, P=0.04). Polymicrobial infection was a protective factor (OR 0.31, P=0.02). Empirical therapy was not significantly associated with mortality in all and monomicrobial patients. Conclusion: This study showed empirical therapy did not affect mortality in P. aeruginosa bacteremia patients. Patients’ comorbidity and severity of illness at bacteremia onset played important roles on mortality. Exposure to carbapenem and fluoroquinolone were associated with MDRPA bacteremia acquisition.