陳秀熙臺灣大學:預防醫學研究所邱愉心Chiu, Yu-HsinYu-HsinChiu2010-06-022018-06-292010-06-022018-06-292008U0001-2907200818145300http://ntur.lib.ntu.edu.tw//handle/246246/184812背景 Stenotrophomonas maltophilia (SMA) 院內感染逐漸增多，且因天生多重抗藥之特性已成為重要的病原菌,尤其是在血液腫瘤病人。雖已有研究探討過SMA移生或感染與抗生素使用壓力之關係，但是過去大部分的研究設計是使用橫斷性生態學設計(cross-sectional ecological design)。極少研究採用時間數列設計(longitudinal time series design)來說明這些相關性，特別是在血液腫瘤病人身上。此外，過去的文獻報告也缺乏以病人個體層級的研究設計來探討不同種類的抗生素使用對SMA移生或感染的影響。而且也缺乏使用兩種或三種不同抗生素是否會產生協同作用之研究。人與方法 本研究採用二種時間數列的研究設計。第一種是生態學的時間數列研究設計(ecological time series design)。利用一醫學中心2002年至2006年血液腫瘤病房前瞻性院內感染監測資料，來評估醫院推行手部衛生運動和抗生素的使用對SMA移生或感染發生率之間的相關性。第二種是以病人個體為基礎的交叉性時間數列研究設計(case-based only cross-over time series design)，一病人住院的月份若有SMA移生或感染，則該月份被定義為事件。關於過去曾使用過抗生素，定義為在培養出SMA前14 天內有使用全身性抗生素。果 在血液病房，SMA的發生率在每1000人日是3.9±2.1次，比腫瘤病房的發生率高3.27倍(95%CI:1.96-5.44; p<0.001)。手部衛生運動之推行對SMA發生率沒有影響。在排除病房及個別病人的效應之後，在培養前14天曾使用過單一的抗生素，包括第4代頭芽孢素(HR:1.80; 95%CI:1.34-2.42)，對抗綠膿桿菌的盤林西林 (HR:1.60; 95%CI:1.19-2.16)，carbepenem (HR:1.45; 95%CI: 1.09-1.93)，或在培養前14天曾使用過兩種抗生素，例如：carbapenem和第4代頭芽孢素(HR:1.57; 95%CI:1.15-2.15)，或carbapenem和對抗綠膿桿菌的盤林西林 (HR:1.54; 95%CI:1.07-2.22)，都與SMA移生或感染有統計相關。而磺胺類抗生素(sulfamethoxazole/trimethoprim)的使用則會減少48%產生SMA移生或感染的機會，具有保護作用(HR:0.52; 95%CI:0.30-0.90)。Carbapenem和第4代頭芽孢菌素或對抗綠膿桿菌的盤林西林彼此不具有協同作用。論 藉由分析在病房層級與病人層級與SMA移生或感染發生率相關的因素顯示在血液病房發生率比腫瘤病房高。歸因於多種廣效性抗生素選擇壓力，而加強手部衛生以減少交叉傳播對SMA此多重抗藥細菌的移生或感染發生率並沒有影響。另外，我們的研究也顯示出，諸多全身性抗生素中，第4代頭芽孢素、對抗綠膿桿菌的盤林西林，和carbapenem對SMA的移生或感染有顯著影響。如預期的，SMA治療的首選藥物，sulfamethoxazole/trimethoprim對SMA是保護因素。研究結果建議，應審慎使用廣效性抗生素，以降低SMA的移生或感染。本研究建立的時間數列分析方法可應用於多重抗藥細菌監測資料之分析。Backgroundosocomial infections caused by Stenotrophomonas maltophilia (SMA), an emerging pathogen attributed to inherited multi-drug resistance, has increased and frequently seen in patients with hematologic malignancies. Although previous studies have reported the relationships of antibiotic uses to colonization of SMA, there are several drawbacks in previous studies. The majority of study designs are based on a cross-sectional ecological design. Very few studies applied a longitudinal time series ecological design to elucidate these associations, especially in patients with hematologic malignancies. It is also lacking of a study design at individual level to throw light on the effect of the combination of different kinds of antibiotic uses on the risk for SMA. Whether the combined use of two or three kinds of drugs in hematologic malignancy patients is synergistic is hardly addressed.atients and Methodshere are two kinds of time series study design used in our study. The first is an ecological time series design that is tailored for estimating the burden of SMA, assessing the impact of hospital-wide hand hygiene program, and studying the correlation between the consumption of antibiotic uses and incidence density of SMA at ward level by using five-year period data from 2002 to 2006. The second design is a case-based only cross-over time series design which is regarded as an atomic level. In this design, the episode of having SMA colonization or infection during a month is defined as event. A series of time epochs over five-year period was constructed for elucidating the association between antibiotic use and SMA infection at individual level. Prior therapy of antibiotic uses was defined as 14 days before culture for the ascertainment of SMA. esultshe incidence density of SMA colonization or infection in patients with malignancies was 3.9±2.1 SMA isolates per 1000 patient-days, which is higher than oncologic wards. Patients at hematologic ward were 3.27-fold (95%CI:1.96-5.44; p <0.001) risk for SMA colonization or infection compared with patients at oncologic wards, but hand hygiene promotion program to avoid exogenous transmission was lacking of significant effect on SMA colonization or infection. After excluding the effects of ward and patients, the use of 4th generation cephalosporins (HR:1.80; 95%CI:1.34-2.42), anti-Pseudomonas penicillins (HR:1.60; 95%CI:1.19-2.16), carbepenem (HR:1.45; 95%CI: 1.09-1.93), and exposure to carbapenem and 4th generation cephalosporins (HR:1.57; 95%CI:1.15-2.15), or carbapenem and anti-Pseudomonas penicillins (HR:1.54; 95%CI:1.07-2.22) were responsible for SMA colonization or infection . Sulfamethoxazole/trimethoprim (HR:0.52; 95%CI:0.30-0.90) decreased the chance of 48% of SMA colonization or infections and had protective effect. Carbapenem and 4th generation cephalosporins or anti-Pseudomonas penicillins made independent contribution to SMA colonization or infection. However, carbapenem in combination with 4th generation cephalosporins or anti-Pseudomonas penicillins had no synergistic effects. onclusionsy using data at ward level and patient level from a tertiary hospital in Taiwan, we found the incidence density of SMA colonization or infection was higher at hematological ward than oncology ward. In patients with malignancies, the high burden of such endogenous SMA infection may be attributed to the exposure to more kinds of broad-spectrum antibiotics. Our study demonstrated that 4th generation cephalosporins, anti-Pseudomonas penicillins, and carbapenem make significant contribution to SMA colonization or infection independent of the frequent use of carbapenem. We also found that sulfamethoxazole/trimethoprim was a protective factor for SMA. These findings suggest that to avoid broad-spectrum antibiotic use may prohibit patients from colonization or infection of SMA in the patients with hematologic malignancies. Our time-series design ca be applied to the surveillance of multi-drug resistance.Contents試委員審定書.………………………………………………………………………..i文摘要…………………………………………………………………….……….….iibstract……………………………………………………………………………….….vontents………………………………………………………………………………....ixntroduction …………………………………………………………………………….1iterature Review………………………………………………………….……………3atients and Methods…………………………………………………………………...7 Study Population……………………………………………………………………..7 Study Design…………………………………………………………………………8 Ethic Approval……………………………………………………………………….9 Data Collection……………………………………………………………………….9 Clinical Data……………………………………………………………………….9 Clinical Isolates…………………………………………………………….………9ntibiotics Consumption………………………………………………………….10 Antimicrobial Susceptibility……………………………………………………….11 Hospital-wide Hand Hygiene Program………………………………………...…….12tatistical Analysis…………………………………………………………………...13 Multi-level Model…………………………………………………………………15esults………………………………………………………………………….....…....17 Prevalence Density…………………………………………………………………..17 Prevalence Density and Consumption of Antibiotics…………………………….….18 Demographic Characteristics of Study Patients…………………………….….........18 Site-specific Distributions…………………………………………………………...19 Antimicrobial Susceptibility…………………………………………………………20 Ecological Analysis…………………………………………………………………..20 Atomic Analysis……………………………………………………………………...22 Patients at Hamatologic Ward………………………………………………………..24 Patients at Oncologic wards …..……………………………...………………….......25 Multi-level Analysis……………………………………………………………….....26iscussion………………………………………………………………………………27 Clinical Epidemiology……………………………………………………………….27cological Analysis…………………………………………………………………..28 Atomic Analysis……………………………………………………………………...29trengths……………………………………………………………………………...31imitations…………………………………………………………………………...33onclusions…………………………………………………………………………….34ables and Figures……………………………………………………………………...35eferences……………………………………………………………………………....66ontents of Tablesable 1 Summary of literature review for Stenotrophomonas maltophilia……….……37able 2 Annual prevalence densities of Stenotrophomonas maltophilia from study patients at hematologic ward and oncologic wards during 2002~2006………………………………………………………………...……43able 3 Demographic characteristics in 375 cancer patients with Stenotrophomonasaltophilia colonization or infection during 2002-2006………………..…….44able 4 Site-specific distribution of 458 isolates of Stenotrophomonas maltophilia n cancer patients during 2002-2006………………………………...………..46able 5 Antimicrobial susceptibility of Stenotrophomonas maltophilia clinicalsolates in study patients during 2002-2006……………………..……………47able 6 Antimicrobial susceptibility of Stenotrophomonas maltophilia clinical isolates rom different sites in study patients during 2002-2006………………………48able 7 Univariate ecological analysis of potential factors for Stenotrophomonas altophilia colonization or infection in cancer patients……………...............49able 8 Multivariate ecological analysis of potential factors for Stenotrophomonasaltophilia colonization or infection in cancer patients………….…………..50able 9 Univariate individual-level analysis of the impact of prior antibiotic use on Stenotrophomonas maltophilia colonization or infection in cancer patients; One category of antibiotic….……………..…………………………………51able 10 Parsimonious model (M1) of prior antibiotic use and Stenotrophomonas maltophilia colonization or infection in cancer patients; One category of antibiotic…….…………………………….……………………..…………..52able 11 Comparison between M1 and other insignificant variables about one ind of antibiotic to exclude negative confounders………………………….53able 12 Evaluation of additional effects of two categories of antibiotics on Stenotrophomonas maltophilia colonization or infection in cancer patients............................................................................................................54able 13 Univariate individual-level analysis of prior antibiotic use and Stenotrophomonas maltophilia colonization or infection in cancer patients; Two categories of antibiotics….……………………………….….……...55able 14 Parsimonious model (M2) of prior antibiotic use and Stenotrophomonas maltophilia colonization or infection in cancer patients; Two categories of antibiotics…..…………………………………………………..….………..56able 15 Comparison between M2 and other insignificant variables about two lasses of antibiotics to exclude negative confounders…..………..………...57able 16 Evaluation of synergistic effects of two categories of antibiotics n Stenotrophomonas maltophilia colonization or infection in ancer patients……………………………………………………………….58able 17 Univariate individual-level analysis of prior antibiotics use and Stenotrophomonas maltophilia colonization or infection in ancer patients; Three categories of antibiotics……………………………..59able 18 Univariate individual-level analysis of prior antibiotic use and tenotrophomonas maltophilia colonization or infection in patients with ematologic malignancies…………………………………………………...60able 19 Parsimonious model of prior antibiotic use and Stenotrophomonas altophilia colonization or infection in patients with hematologic malignancies…………………………………………………………………61able 20 Univariate individual-level analysis of prior antibiotic use for tenotrophomonas maltophilia colonization or infection in patients with olid tumors……………………………………………………………...…..62able 21 Parsimonious model of prior antibiotic use for Stenotrophomonas altophilia colonization or infection in patients with solid tumors……………………………………………………………………63able 22 Results of the multi-level model in Bayesian approach…………….……......64ontents of Figuresigure 1 Monthly prevalence density of hematologic ward and oncologic wards during 2002~2006………………………...……….35igure 2 Carbapenem use and prevalence density of Stenotrophomonas maltophilia in patients at hematologic ward during 2002~2006……………………………………………………………….......36application/pdf552423 bytesapplication/pdfen-USStenotrophomonas maltophilia抗生素血液腫瘤時間序列生態學研究個體研究多層次分析antibioticsmalignancytime seriesecological studyindividual studymulti-level analysisthesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/184812/1/ntu-97-R95846012-1.pdf