急診鮑氏不動桿菌菌血症感染病人之處方型態分析:預後因子與治療結果

Abstract

研究背景： 鮑氏不動桿菌常為伺機性感染，於院內感染與使用呼吸器病人居多。而臺灣近十年的研究指出，多重抗藥性鮑氏不動桿菌有上升的趨勢。雖然鮑氏不動桿菌為伺機性感染，但回顧論文指出，鮑氏不動桿菌也具有社區感染的潛力，只是目前相對於院內感染的研究，社區型研究的比例還是偏低。於臺大醫院急診處經細菌學檢驗為鮑氏不動桿菌菌血症中有25.2%為抗藥性菌株。因此本研究選擇進行急診鮑氏不動桿菌菌血症的研究，同時針對社區型鮑氏不動桿菌菌血症及多重抗藥性病株進行分析。 研究目的： 分析急診就診病人感染之社區型、醫療照護型及院內感染的鮑氏不動桿菌菌血症之差異 感染多重抗藥性鮑氏不動桿菌菌血症之危險因子分析 鮑氏不動桿菌菌血症之死亡因子分析 經驗性抗生素與治療結果的關聯性 研究設計、地點及對象： 本研究收納自2008年1月1日到2012年12月31日期間到國立臺灣大學醫學院附設醫院急診室就診的病人，研究對象為細菌學檢驗發現血液培養有鮑氏不動桿菌之成年病人。 研究方法： 收入於急診血液培養為鮑氏不動桿菌菌血症的病人，以病歷回顧的方式進行分析。所收集的資料分成三個區塊，分別是鮑氏不動桿菌菌血症發作前、發作時及發作後。發作前以病人的基本資料、潛在性疾病為主，發作時以臨床表徵、實驗數值紀錄及藥物治療為主，發作後記錄病人的預後。以發病後第三十天的死亡率做為治療結果的評估，也有紀錄第七天、第十四天以及出院時的死亡率做為分析。此外，也針對非多重抗藥性與多重抗藥性菌株之病人資料進行比較。統計方法包含單變項及多變項的羅吉斯回歸分析，對於存活分析以Log-rank進行比較，統計分析的工具以Microsoft Excel 2013及SAS（9.3版）進行資料處理。 研究結果： 本研究中，122位病人共發生123個菌血症事件，有78個事件為單一菌種菌血症；整體病人的平均年齡為66歲，男性比例為61%；病人的Charlson’s comorbidity index為3.4±2.5、Pitt bacteremia score為 2.1±2.4，發生敗血性休克的比例為23.6%，過去一個月內有使用過抗生素的比例為34.2%；經驗性治療使用合併抗生素治療者只有10例，而確切性治療使用合併抗生素治療者也只有12例。多重抗藥性菌血症占了25.2%，全部123個菌血症及78個單一菌種菌血症個案於菌血症發作後的第三十天死亡率分別為19.7%及16.9%。 對單一菌種而言，經驗性使用之單一或合併抗生素治療屬適當者分別為45.8%及50.0%，並沒有達到統計上顯著的差異（P=1.000）。而經驗性單一或合併抗生素對於第七天、第十四天及第三十天之死亡率都沒有達到統計上顯著差異。 由多變項羅吉斯回歸分析發現，整體病人感染多重抗藥性鮑氏不動桿菌之危險因子，包含：護理之家住民（OR 23.49，P=0.0004）、腎臟病末期病人（OR 12.78，P=0.0107）、住加護病房者（OR 3.57，P=0.0473）、癌症病人（OR 0.206，P=0.0089）；而其中單一菌種菌血症，則包含：護理之家住民（OR 33.74，P=0.0002）、住加護病房者（OR 6.188，P=0.0305）、過去一個月內有動過手術者（OR 21.60，P=0.0192）。 對於第三十天之死亡率，經由多變項羅吉斯回歸分析，就全體病人而言，影響第三十天之死亡危險因子，包含了敗血性休克（OR 10.828，P<0.0001）及多重抗藥性菌株（OR 3.419，P=0.0259）；就單一菌種菌血症病人而言，影響第三十天死亡之危險因子也包含了敗血性休克（OR 9.618，P=0.0019）及多重抗藥性菌株（OR 6.06，P=0.0161）；而對於社區型發作型之病人而言，其預後因素只有敗血性休克（OR 6.216，P=0.0018）。經驗性抗生素合併與否及經驗性抗生素治療適當與否，對於預後沒有達到統計上顯著差異。 結論 於本研究中，第三十天死亡率與多重抗藥性與敗血性休克有關；此外，住在護理之家、住在加護病房、過去一個月內有動過手術為多重抗藥性鮑氏不動桿菌菌血症之危險因子。經驗性抗生素適當與否或經驗性抗生素合併與否，於菌血症發作後第三十天死亡率並無顯著影響。

Background： Acinetobacter baumannii infections play an important role in opportunistic infection. They have become common in hospitalized patients or patients on mechanical ventilator. According to a longitudinal surveillance on A. baumanniiin Taiwan, the number of multidrug resistant（MDR）strain has climbed gradually over the last decade. Although A. baumannii infections were considered opportunistic, they could cause community-acquired infection with rare occasions. A. baumannii bacteremia cultured in the emergency department（ED）at National Taiwan University Hospital exhibited 25.2 % with MDR strains. Therefore, it is essential to conduct a study in ED patients to analyze the clinical characteristics of the community-acquired A.baumannii infections and multidrug resistance. Study objective： 1. To discover the differences between community-acquired infections, healthcare-associated infections and nosocomial infections. 2. To identify risk factors of MDR resistant A. baumannii bacteremia in the emergent department. 3. To analyze prognostic factors of 30 day（D30）mortality of A. baumannii bacteremia 4. To evaluate the correlation between empirical antibiotics and treatment outcome Study location、design and study population A retrospective study was conducted in National Taiwan University Hospital. Patients with A. baumannii bacteremia visiting ED between January 2008 and December 2012 were identified from the medical records of the clinical microbiology laboratory. Methods Data were collected from medical charts and the hospital computerized database. Data retrieved for each patient included patients’ demographic profiles, underlying diseases, previous A. baumannii infection or colonization history, clinical presentation, antibiotics regimens at onset of bacteremia, management and clinical response after bacteremia. The primary endpoint was D30 all-cause mortality. Risk factors of MDR A. baumannii bacteremia were also analyzed. Risk factors and clinical outcomes were examined using univariate analysis and multivariate logistic regression analysis. Survival curves shown by Kaplan-Meier method were compared with Log-rank test. Microsoft 2013 and SAS（ver. 9.3）were used for data management. Result One hundred and twenty three episodes in 122 patients were included in this study. Among them, 78 were monomicrobial infections and 45 were polymicrobial infections. Patients’ average age was 66 years old, and 61% of them were male. The average Charlson’s comorbidity index was 3.4±2.5. Previous exposure to antibiotics within 1 month was 34.2%. Patients’ average Pitt bacteremia scores were 2.1±2.4. There were 85.4% episodes presenting sepsis with 28.6% of septic shock. Empirical combination therapy was given only in 10 episodes （8.4%）and definitive combination therapy was used in 12 episodes （11.5%）.There were 31 （25.2%）A. baumannii episodes. The overall D30 mortality was 19.7%. It was 16.9% for monomicrobial infections. Empirical monotherapy and combination therapy provided comparably adequate therapy（45.8% vs. 50.0 %, P=1.0000）with similar D30 mortality in each regimen （0% vs. 18.1%, P=1.000）. Multivariate logistic regression model identified the following variables as significant independent risk factors for monomicrobial MDR A. baumannii bacteremia: nursing home residents（OR 33.74, P=0.0002）, intensive care unit stay （OR 6.188, P=0.0305）and surgery within 1 month（OR 21.60, P=0.0192）. Following factors were independently associated with D30 mortality in monomicrobial group: septic shock（OR 9.618, P=0.0019）and multi-drug resistant pathogen（OR 6.06, P=0.0161）.Empirical therapy was not significantly associated with mortality in all patients and monomicrobial group. Conclusion: Septic shock and MDR bacteremia were significantly associated with D30 mortality in patients with A.baumannii bacteremia in the ED. Living in the nursing home, ICU stay, surgery within 1 month played important roles in acquisition of MDR A. baumannii in these patients. Empirical treatment did not affect mortality of A.baumannii bacteremia.