影響發燒反應之相關因素研究：以院內血流感染患者為例

A study on factors that affect the febrile response in patients with nosocomial bloodstream infection

呂淑華(Shu-Hua Lu)

指導教授 ： 戴玉慈

國立臺灣大學/醫學院/護理學研究所/博士(2009年)

https://doi.org/10.6342/NTU.2009.00353

摘要

背景：
發燒是感染的重要徵象，然而臨床上某些個案身受感染，卻缺乏明顯發燒反應，稱之模糊鈍化發燒反應，可能因此而延誤診斷及治療。而關於影響發燒反應的研究，多以老化或年齡為主軸進行探究。關於非年齡因素僅止於論述性敘說缺乏實證性研究。故本研究以發燒反應病理生理機制及體溫調節機轉為理論依據，探討身體遭受急性感染時，影響發燒反應之相關因素。
目的：
確立影響發燒反應即感染開始期體溫與發燒強度之相關因素。
設計：
採回溯性、重複測量、比較和相關預測的設計。
方法：
研究計劃通過倫理委員會審議後，以北部某醫學中心的感染控制中心為資料庫，回溯2007.01.01 ~ 2008. 6.30之住院病人經血液培養結果陽性，確認院內血流感染者為潛在研究對象，以臨床醫療病歷作為資料來源，自擬病歷資料萃取單為研究工具，以白血球數量、年齡、共病症嚴重度（查爾森合併症指數）、身體活動功能（巴氏指數）、營養狀態（體重、身體質量指數）及菌種類型（血液細菌培養結果）為自變項；探討其對發燒反應的影響，所得結果運用SPSS 15.0版統計軟體分析，以描述性統計呈現個案基本屬性及發燒參數資料分佈，再以t-檢定、單因子變異數分析、雙變項相關分析、回歸分析及廣義線性模式確立影響發燒反應相關因素。
結果：
經隨機選樣獲得個案600位，最終有效樣本230位，個案屬性以男性居多（58.8%），年齡自18至94歲，主要診斷以腫瘤最多（46.1%），排名前四類的感染菌種是克雷白氏肺炎菌、金黃色葡萄球菌、大腸桿菌及念珠菌屬。體溫基礎值平均為36.33°C（S.D. 0.32°C），具性別差異性，但不具年齡差異性。感染開始期體溫平均值為38.33°C，感染開始期體溫< 38°C者有48位，佔總個案數20.86% (48/230)；多元回歸分析結果白血球數量、共病症嚴重度、體重、菌種類型此4個變項僅能聯合預測感染開始期體溫7.4%的變異量。本研究顯示87.3％（n=199）個案，在感染期間僅有1次發燒事件，考慮時間因素使用廣義估計方程式分析結果顯示白血球數量、年齡、查爾森合併症指數、與菌種類型為影響發燒強度顯著性因子。
結論與建議：
本研究以臨床資料為基礎，非僅單一考量年齡因素，納入學理上可能影響因素一同分析，結果發現白血球數量、年齡、合併症嚴重度、身體活動、體重、菌種類型對發燒反應在統計上皆有顯著的影響，由於發燒反應是一種複雜的生理適應反應，因此有許多疑點需要未來研究繼續澄清，尤其是白血球數量與發燒強度之關係方向性；本研究結果可提供臨床觀察急性感染患者發燒反應強弱重要評估參考，並作為未來發燒反應前瞻性研究之基礎，延伸護理研究於深入生理變項及生物護理研究範疇。

關鍵字：發燒反應、院內血流感染、感染開始期體溫、發燒強度、病歷回顧

關鍵字

發燒反應 ； 院內血流感染 ； 感染開始期體溫 ； 發燒強度 ； 病歷回顧法

並列摘要

Background:
Fever is considered a hallmark sign of infection; however, some patients suffer from infection without manifesting a febrile response. Blunted febrile response can delay diagnosis and treatment and thus result in higher mortality. Most studies have focused on the influence of aging on the febrile response. Few empirical studies of febrile response have addressed non-age-related factors. Therefore this study was based on pathophysiological mechanism of febrile response and body temperature thermoregulation, and explored the factors that affect the febrile response to hospital acquired bloodstream infection.
Purpose:
The aims of this study were to examine factors that affect febrile response including temperature at the onset, and magnitude of fever.
Design:
The study used a retrospective, repeated measures, comparison, correlational and predictive design.
Method:
After obtaining ethics committee approval, the investigator used
an infection control database of a university hospital in Taipei to select random sample of 600 medical records of bloodstream infection cases reported on patients hospitalized between January 1, 2007 to June 30, 2008. Research instrument was chart abstraction form designed by the author. The independent variables were white blood cell count (WBC), age, severity of co-morbidity (Charlson Comorbidity Index), physical function (Barthel Index), nutritional status (body weight, body mass index, and albumin), and microbial classification (blood culture results). To determine the factors that affect febrile response SPSS 15.0 statistical software package was used to analyze the data. The investigator used descriptive statistics to characterize the demographics and fever parameters of the sample cases. T-tests, ANOVA, Pearson correlation, multiple regression and General Estimating Equation (GEE) were used to analyze the predictive factors.
Results:
The cases ranged from 18 to 94 years old, 58.8% were male and 46.1% had a primary diagnosis of cancer. The top four categories of identified pathogens were staphylococcus aureus, Escherichia coli, klebsiella pneumoniae and candida spp. The mean baseline temperature was 36.33°C (SD 0.32°C). Although women had slightly higher baseline temperatures compared to men, there was no difference related to age. The mean temperature at the onset of fever was 38.33°C (SD 0.77°C); 48 cases (20.86%) had temperature at the onset lower than < 38°C. Multiple regression analysis results indicated white blood cell count, severity of co-morbidity, body weight, and microbial classification were significant predictors of temperature at the onset. However, these factors predicted only 7.4% of the variance. GEE analysis demonstrated white blood cell count, age, severity of co-morbidity, physical function and microbial classification were significant predictors of the magnitude of fever. Most cases (87.3%) had only one fever episode related to their hospital acquired bloodstream infection episode.
Conclusions and suggestion:
The findings revealed that white blood cell count, age, severity of co-morbidity, physical function, body weight, and microbial classification have different effects on the febrile response. Febrile response is a complex biophysical adaption. For the reason many issues need to be continued investigating in the future. Especially in the relationship between white blood cell count and magnitude of fever is need to research. The findings of the study can provide clinical practitioners information to use the parameters of fever response to assess febrile patients, and can establish the base for prospective studies of the biophysical factors related to fever responses.

Key words: febrile response, nosocomial bloodstream infection, temperature at the onset, magnitude of fever, chart review

並列關鍵字

febrile response ； nosocomial bloodstream infection ； temperature at the onset ； magnitude of fever ； chart review

參考文獻

張玉坤（1996)．GEE之敏感度分析－偵測高影響之觀察值．中華公共衛生雜誌， 15（5），403-410。

施智源、陳瀅淳、劉美芳（2007）．美國疾病管制中心2004年院內感染定義中譯．感染控制雜誌，17（1），11-44。

Holtzclaw, B. J. (2002). Use of thermoregulatory principles in patient care: fever management. Online Journal of Clinical Innovations, 5(5), 1-64.

Gould, D. (1993). Homeostasis: the key to normal function. In S. M. Hinchliff, S. E. Norman & J. E. Schober (Eds.), Nursing practice and health care (2nd ed.). London E. Arnold.

Smith, M. A., Nitz, N. M., & Stuart, S. K. (2006). severity and comorbidity. In R. L. Kane (Ed.), Understanding health care outcomes research (2nd ed., pp. 219-263). Sudbury, Mass. : Jones and Bartlett.