Development of machine learning models for the surveillance of colon surgical site infections.

BACKGROUND: Conventional surgical site infection (SSI) surveillance is labour-intensive. We aimed to develop machine learning (ML) models for the surveillance of SSIs for colon surgery and to assess whether the ML could improve surveillance process efficiency. METHODS: This study included cases who underwent colon surgery at a tertiary center between 2013 and 2014. Logistic regression and four ML algorithms including random forest (RF), gradient boosting (GB), and neural networks (NNs) with or without recursive feature elimination (RFE) were first trained on the entire cohort, and then re-trained on cases selected based on a previous rule-based algorithm. We assessed model performance based on the area under the curve (AUC), sensitivity, and positive predictive value (PPV). The estimated proportion of reduction in workload for chart review based on the ML models was evaluated and compared with the conventional method. RESULTS: At a sensitivity of 95%, the NN with RFE using 29 variables had the best performance with an AUC of 0.963 and PPV of 21.1%. When combining both the rule-based algorithm and ML algorithms, the NN with RFE using 19 variables had a higher PPV (28.9%) than with the ML algorithm alone, which could decrease the number of cases requiring chart review by 83.9% compared with the conventional method. CONCLUSION: We demonstrated that ML can improve the efficiency of SSI surveillance for colon surgery by decreasing the burden of chart review while providing high sensitivity. In particular, the hybrid approach of ML with a rule-based algorithm showed the best performance in terms of PPV.

Fluoroquinolones versus trimethoprim-sulfamethoxazole for the treatment of Stenotrophomonas maltophilia infections: a systematic review and meta-analysis.

BACKGROUND: Fluoroquinolones are a popular alternative to trimethoprim-sulfamethoxazole for Stenotrophomonas maltophilia infections. OBJECTIVES: To compare the effects of fluoroquinolones and trimethoprim-sulfamethoxazole on mortality of S. maltophilia infections. DATA SOURCES: PubMed and EMBASE. STUDY ELIGIBILITY CRITERIA: Clinical studies reporting mortality outcomes of S. maltophilia infections. PARTICIPANTS: Patients with clinical infections caused by S. maltophilia. INTERVENTIONS: Fluoroquinolone monotherapy in comparison with trimethoprim-sulfamethoxazole monotherapy. METHODS: Systematic review with meta-analysis technique. RESULTS: Seven retrospective cohort and seven case-control studies were included. Three cohort studies were designed to compare the two drugs, whereas others had other purposes. A total of 663 patients were identified, 332 of which were treated with trimethoprim-sulfamethoxazole (50.1%) and 331 with fluoroquinolones (49.9%). Three cohort studies were designed to compare the effect of the two drugs, whereas the others had other purposes. Levofloxacin was most frequently used among fluoroquinolones (187/331, 56.5%), followed by ciprofloxacin (114/331, 34.4%). The overall mortality rate was 29.6%. Using pooled ORs for the mortality of each study, fluoroquinolone treatment (OR 0.62, 95% CI 0.39-0.99) was associated with survival benefit over trimethoprim-sulfamethoxazole treatment, with low heterogeneity (I2 = 18%). Specific fluoroquinolones such as ciprofloxacin (OR 0.44, 95% CI 0.17-1.12) and levofloxacin (OR 0.78, 95% CI 0.48-1.26) did not show a significant difference in comparison with trimethoprim-sulfamethoxazole. In the sub-group analyses of adult and bacteraemic patients, significant differences in mortality were not observed between fluoroquinolones and trimethoprim-sulfamethoxazole. CONCLUSIONS: Based on a meta-analysis of non-randomized studies, fluoroquinolones demonstrated comparable effects on mortality of S. maltophilia infection to trimethoprim-sulfamethoxazole, supporting the use of fluoroquinolones in clinical S. maltophilia infections. Although the pooled analysis of overall studies favoured fluoroquinolones over trimethoprim-sulfamethoxazole, the studies included were observational, and sub-group analyses of certain fluoroquinolone agents did not show statistical differences with trimethoprim-sulfamethoxazole. Randomized clinical studies are needed to address these issues.

Clinical predictors of methicillin-resistance and their impact on mortality associated with Staphylococcus aureus bacteraemia.

We investigated the clinical predictors of methicillin-resistance and their impact on mortality in 371 patients with Staphylococcus aureus bacteraemia identified from two prospective multi-centre studies. Methicillin resistant S. aureus (MRSA) accounted for 42.2% of community-onset and 74.5% of hospital-onset cases. No significant clinical difference was found between patients infected with MRSA vs. methicillin-sensitive S. aureus (MSSA), except that the former were more likely to have had hospital-onset bacteraemia and received antibiotics in the preceding 90 days. After stratifying according to the acquisition site, prior antibiotic use was the only independent predictor of having MRSA in both community-onset and hospital-onset cases. The frequency of inappropriate empirical antibiotic therapy was higher in patients with MRSA than in those with MSSA bacteraemia. However, methicillin resistance was not a predictor of mortality in patients and the clinical characteristics and outcomes of both MRSA and MSSA bacteraemia were similar. This study indicates that there are no definitive clinical or epidemiological risk factors which could distinguish MRSA from MSSA cases with the exception of the previous use of antibiotics for having MRSA bacteraemia, which emphasises the prudent use of glycopeptide treatment of patients at risk for invasive MRSA infections.

Suggested new breakpoints of anti-MERS-CoV antibody ELISA titers: performance analysis of serologic tests.

To provide optimal cut-off values of anti-Middle East respiratory syndrome coronavirus (MERS-CoV) serologic tests, we evaluated performance of ELISA IgG, ELISA IgA, IFA IgM, and IFA IgG using 138 serum samples of 49 MERS-CoV-infected patients and 219 serum samples of 219 rRT-PCR-negative MERS-CoV-exposed healthcare personnel and patients. The performance analysis was conducted for two different purposes: (1) prediction of neutralization activity in MERS-CoV-infected patients, and (2) epidemiologic surveillance of MERS-CoV infections among MERS-CoV-exposed individuals. To evaluate performance according to serum collection time, we used 'days post onset of illness (dpoi)' and 'days post exposure (dpex)' assessing neutralization activity and infection diagnosis, respectively. Performance of serologic tests improved with delayed sampling time, being maximized after a seroconversion period. In predicting neutralization activity, ELISA IgG tests showed optimal performance using sera collected after 21 dpoi at cut-off values of OD ratio 0.4 (sensitivity 100% and specificity 100%), and ELISA IgA showed optimal performance using sera collected after 14 dpoi at cut-off value of OD ratio 0.2 (sensitivity 85.2% and specificity 100%). In diagnosis of MERS-CoV infection, ELISA IgG exhibited optimal performance using sera collected after 28 dpex, at a cut-off value of OD ratio 0.2 (sensitivity 97.3% and specificity 92.9%). These new breakpoints are markedly lower than previously suggested values (ELISA IgG OD ratio 1.1, sensitivity 34.8% and specificity 100% in the present data set), and the performance data help serologic tests to be practically used in the field of MERS management.

Empirical combination of a ß-lactam to vancomycin may not improve outcomes of methicillin-susceptible Staphylococcus aureus bacteremia, compared to vancomycin monotherapy.

To evaluate effect of empirical combination of a ß-lactam to vancomycin and vancomycin monotherapy in Staphylococcus aureus bacteremia (MSSA-B), we conducted a retrospective cohort study. Electronic medical records of individuals who were diagnosed with MSSA-B between January 2005 and February 2015 at a tertiary care center were reviewed. Patients were classified into three groups according to empirical antibiotic regimen (BL group, ß-lactam; VAN group, vancomycin; BV group, combination of ß-lactam and vancomycin), and 30-day all-cause mortality of each group was compared. During the study period, 561 patients with MSSA-B were identified. After exclusion of 198 patients (36 with poly-microbial infection, 114 expired within 2 days, and 48 already received parenteral antibiotics) and a matching process, 46 patients for each group were included. Baseline characteristics were similar except for severity and comorbidity scores. The 30-day mortality for all three groups were not significantly different (BL 4.3%, VAN 6.5%, BV 8.7%; P = 0.909). In a multivariate analysis, type of empirical antibiotic regimen was not statistically associated with 30-day all-cause mortality. In comparison with the VAN group, the BV group yielded a HR of 0.579 (95% CI = 0.086-3.890, P = 0.574). Pitt bacteremia score was the only significant factor for mortality. The empirical combination of a ß-lactam to vancomycin was not associated with lower mortality in treating MSSA-B, compared to vancomycin monotherapy.

Recent Results of In Situ Abdominal Aortic Reconstruction with Cryopreserved Arterial Allograft.

OBJECTIVE: To evaluate treatment outcomes of in situ abdominal aortic reconstruction with cryopreserved arterial allograft (CAA) for patients with abdominal aortic infection. MATERIALS AND METHODS: A retrospective review of prospectively collected data was conducted of patients who underwent in situ aortic reconstruction using CAA for primary, secondary, or prosthetic infection of the abdominal aorta between May 2006 and July 2015, at a single institution. Clinical presentation, indications for treatment, procedural details, early post-operative mortality and morbidity, late death, and graft related complications during the follow up period were investigated. Patient survival and event free survival (any death or re-operation) were calculated using the Kaplan-Meier method. RESULTS: Twenty-five patients (male, n = 20, 80%; mean age, 70.2 ± 8.7 years) underwent in situ abdominal aortic reconstruction (48% aortic, 52% aorto-bi-iliac) with vessel size and ABO matched CAA for treatment of abdominal aortic infection caused by infected abdominal aortic aneurysm (n = 15), aortic prosthesis infection (n = 7), aortic reconstruction with concomitant colon resection (n = 2), and primary suppurative aortitis (n = 1). The median follow up was 19.1 months (range 1-73 months). There were seven post-operative deaths including two (8%) early (<30 days) and five (20%) late deaths There were three (12%) graft related complications including thrombotic occlusion of the CAA, aneurysmal dilatation, and aorto-enteric fistula. Three years after CAA implantation, patient survival was 74% and the event free survival was 58%. CONCLUSIONS: It is believed that in situ abdominal aortic reconstruction with CAA is a useful option for treating primary, secondary, or prosthetic infection of the abdominal aorta.

Emergence of an extended-spectrum ß-lactamase-producing serotype K1 Klebsiella pneumoniae ST23 strain from Asian countries.

Extended-spectrum ß-lactamase (ESBL) production has been very rare in serotype K1 Klebsiella pneumoniae ST23 strains, which are well-known invasive community strains. Among 92 ESBL-producing strains identified in 218 isolates from nine Asian countries, serotype K1 K. pneumoniae strains were screened. Two ESBL-producing K. pneumoniae isolates from Singapore and Indonesia were determined to be serotype K1 and ST23. Their plasmids, which contain CTX-M-15 genes, are transferable rendering the effective transfer of ESBL resistance plasmids to other organisms.

Sequence type 72 community-associated meticillin-resistant Staphylococcus aureus emerged as a predominant clone of nasal colonization in newly admitted patients.

Current knowledge of community-associated (CA) meticillin-resistant Staphylococcus aureus (MRSA) carriage in hospitalized patients is incomplete. Genotypic characteristics of 637 nasal MRSA isolates from newly admitted patients in South Korea were investigated. Sequence type (ST) 72 accounted for 52.1%, 46.3%, and 52.8% of the isolates during the periods of 2007-2008, 2009-2010, and 2013-2014, respectively. Instead of classic MRSA clones responsible for healthcare-associated infections, including ST5 and ST239, MRSA with community genotype ST72 was the predominant strain in newly admitted patients regardless of age and home province of the patients. Active strategies are needed to prevent healthcare-associated infection by CA-MRSA.

Generic piperacillin/tazobactam is not associated with galactomannan false-positivity in adult patients with cancer: a case-control study.

Recent products of piperacillin/tazobactam (PTZ) from the original manufacturer, previously considered a major cause of galactomannan (GM) false-positivity, are reported not to be related to it. However, data regarding generic PTZ are limited and controversial. To evaluate the effect of generic PTZ on GM false-positivity in Korea, we performed a case-control study in adult patients with cancer. A case-control study was designed. Electronic medical records of cancer patients who were admitted and tested for serum GM between March and June 2014 at a tertiary care university hospital were reviewed. During the study period, a single generic PTZ (C manufacturer, Korea) was used. Patients who received PTZ within 24 h prior to serum GM testing were enrolled. Age- and GM test date-matched non-PTZ patients were selected as controls. A total of 110 patients received PTZ within 24 h prior to serum GM testing during the study period. The GM optical density index (ODI) of the PTZ group did not vary significantly from that of the control group (p = 0.251). The percentage of false-positive patients in the PTZ group was also similar to that of the control group (p = 0.538). There was no statistical relationship between GM ODI titer and time interval from PTZ administration (p = 0.095) or cumulative PTZ dose (p = 0.416). In a case-control study that evaluated 220 patients, a generic PTZ in Korea was not related to GM false-positivity.

Risk factors for cytomegalovirus gastrointestinal diseases in adult patients with cancer.

Cytomegalovirus (CMV) gastrointestinal (GI) disease has been noticed frequently in cancer patients, causing abdominal pain, diarrhea, and GI bleeding. However, little is known about its actual incidence, clinical presentation, and the risk factors for its development among cancer patients. To answer these questions, we analyzed all cases that occurred during an 18-year period at our center. A case-control study was performed to identify risk factors for CMV GI disease. Electronic medical records were reviewed from individuals who were admitted and diagnosed with CMV GI disease during the period of January 1995 through March 2013 at a tertiary care center. Two CMV disease-free cancer patients were matched as controls. A total of 98 episodes of CMV GI disease were included in this study, and the overall incidence rate was 52.5 per 100,000 cancer patients, with an increasing trend throughout the study period. According to multivariate analysis, male sex, low body mass index, lymphopenia, hematological malignancy, and steroid use and red blood cell transfusion within 1 month prior to the CMV disease were identified to be independent risk factors. Among these factors, RBC transfusion showed the highest odds ratio (OR = 5.09). Male sex, low body mass index, lymphopenia, hematological malignancy, steroid use, and red blood cell transfusion within 1 month prior to the CMV disease diagnosis were independent risk factors for the development of CMV GI disease in adult patients with cancer.

Predictors of uropathogens other than Escherichia coli in patients with community-onset acute pyelonephritis.

AIM: A constant reduction in the incidence of community-onset acute pyelonephritis (CO-APN) caused by Escherichia coli has been shown with a parallel increase incidence caused by other organisms. Therefore, we evaluated the risk factors and outcome of non-E. coli as uropathogens in patients with community-onset APN. METHODS: As a part of a nationwide multicentre surveillance study conducted in Korea, a total of 416 patients with CO-APN were collected with their epidemiological, antibiotic treatment and outcome data. RESULTS: The risk factors and outcomes of non-E. coli as uropathogens were evaluated in a total of 416 patients with culture-confirmed CO-APN. Non-E. coli caused 127 cases (30.5%) of CO-APN. CO-APN caused by non-E. coli resulted in higher inappropriate empirical therapy (38.6% vs. 20.1%, p < 0.001), longer hospital stay (12.6 days vs. 6.7 days, p = 0.005) and higher 30-day mortality (9.4% vs. 3.8% p = 0.020) compared with CO-APN caused by E. coli. Multivariate analyses showed that male gender (OR, 3.48; CI, 2.13-5.67; p < 0.001), underlying haematological disease (OR, 5.32; CI, 1.17-24.254; p = 0.031), underlying benign prostate hyperplasia (OR, 2.61; CI, 1.02-6.74; p = 0.046), chronic indwelling urethral catheter (OR, 6.34; CI, 1.26-31.84; p = 0.025) and admission history in the previous 6 months (OR, 2.12; CI, 1.23-3.58; p = 0.005) were predictors for CO-APN caused by a non-E. coli isolate. CONCLUSIONS: Community-onset APN caused by non-E. coli represents a distinct subset of urinary tract infections with worse outcomes. The defined risk factors related with non-E. coli should be taken into consideration when empirical antibiotic therapy is prescribed in patients with community-onset APN.

Risk factors for levofloxacin-nonsusceptible Streptococcus pneumoniae in community-acquired pneumococcal pneumonia: a nested case-control study.

This study was performed to evaluate the clinical features of community-onset levofloxacin-nonsusceptible pneumococcal pneumonia and to identify risk factors for levofloxacin resistance. Using the database of a surveillance study of community-acquired pneumococcal infections in Asian countries, we conducted a nested case-control study to identify risk factors for levofloxacin-nonsusceptible S. pneumoniae in community-acquired pneumonia in adults. Of 981 patients with pneumococcal pneumonia, 46 (4.7 %) had levofloxacin-nonsusceptible S. pneumoniae, of whom 39 evaluable cases were included in the analysis. All cases were from Korea, Taiwan, and Hong Kong. Among patients with levofloxacin-susceptible S. pneumoniae, 490 controls were selected based on patient country. Of the 39 cases of levofloxacin-nonsusceptible pneumococcal pneumonia, 23 (59.0 %) were classified as healthcare-associated, while 164 (33.5 %) of the 490 controls of levofloxacin-susceptible S. pneumoniae (P = 0.001) were classified as healthcare-associated. Multivariate analysis showed that previous treatment with fluoroquinolones, cerebrovascular disease, and healthcare-associated infection were significantly associated with levofloxacin-nonsusceptible pneumococcal pneumonia (all P < 0.05). Levofloxacin-nonsusceptible pneumococci pose an important new public health threat in our region, and more information on the emergence and spread of these resistant strains will be necessary to prevent spread throughout the population.

Impact of a computerized alert system for bacteremia notification on the appropriate antibiotic treatment of Staphylococcus aureus bloodstream infections.

A computerized alert system (CAS) has been introduced to notify bacteremia in real time. We evaluated the impact of the CAS on the administration of appropriate antibiotics in patients with Staphylococcus aureus bloodstream infections (BSIs). We retrospectively reviewed the medical records of patients with S. aureus BSI for each 1-year control and intervention periods, before and after the implementation of the CAS. The proportions of appropriate antibiotic treatment were compared between the control and intervention periods. The 30-day mortality of S. aureus bacteremia was also assessed in the study population. A total of 313 patients were included in the study. Appropriate antibiotics were initiated 7 h earlier in the intervention period (mean time, 13.5 h vs. 20.0 h; p = 0.136). The administration of appropriate antibiotics within the 24 h after blood acquisition was similar between the two periods, but this significantly increased from 3.3% in the control period to 10.6% in the intervention during the 24-36 h interval (p = 0.012). In the subgroup analysis, similar trends were observed in patients with methicillin-resistant isolates (6.7% vs. 18.2%; p = 0.032) and hospital-onset infection (3.5% vs. 17.1 %; p = 0.004). The independent risk factors for 30-day mortality of S. aureus bacteremia were age, a high Pitt bacteremia score, an increased Charlson's weighted index of comorbidity, and hospital-onset infection, although the appropriateness of antibiotic therapy within 36 h and the CAS were not identified as predictors. The CAS increased the proportion of appropriate antimicrobial therapy during the 24-36 h interval after bacteremia onset in patients with S. aureus BSIs.

Association of levofloxacin resistance with mortality in adult patients with invasive pneumococcal diseases: a post hoc analysis of a prospective cohort.

OBJECTIVE: This study was conducted to identify risk factors for mortality and to evaluate the impact of antimicrobial resistance on outcome in adult patients with invasive pneumococcal disease (IPD). METHODS: A post hoc analysis of an observational cohort study on community-acquired pneumococcal infections was conducted and a total of 136 adult patients with IPD were analyzed in this study. RESULTS: Pneumonia was the most common type of infection (n = 84, 61.8 %), followed by primary bacteremia (n = 15, 11.0 %) and meningitis (n = 15, 11.0 %). One hundred and three patients (75.7 %) had concomitant pneumococcal bacteremia. The overall 30-day mortality rate was 26.5 % (36/136), and factors associated with 30-day mortality were corticosteroid use, presentation with septic shock, and development of acute respiratory distress syndrome (ARDS) (all P < 0.05). While penicillin and erythromycin resistance were associated with a lower mortality, an association between levofloxacin resistance and increased mortality was found in the univariate analysis; however, statistical significance was not reached (P = 0.083). Multivariable analysis showed that presentation with septic shock, corticosteroid use, development of ARDS, and levofloxacin resistance were independent factors associated with 30-day mortality. Of the five patients with IPD caused by levofloxacin-resistant Streptococcus pneumoniae, three (60 %) died within 30 days of diagnosis. CONCLUSION: Levofloxacin resistance was associated with increased mortality, along with septic shock, prior use of corticosteroids, and development of ARDS, in adult patients with IPD. Our data suggest that the emergence of levofloxacin resistance among invasive pneumococcal isolates is now becoming a challenge for clinicians managing community-acquired bacterial infections.

Community-associated Panton-Valentine leukocidin-negative meticillin-resistant Staphylococcus aureus clone (ST72-MRSA-IV) causing healthcare-associated pneumonia and surgical site infection in Korea.

BACKGROUND: Community-associated meticillin-resistant Staphylococcus aureus (CA-MRSA) has emerged as an important pathogen worldwide in a continent-specific manner. Clinical characteristics of infections caused by CA-MRSA other than USA300, especially in healthcare settings, have not been well established. AIM: To conduct a retrospective cohort study to determine the clinical characteristics of infections caused by Panton-Valentine leukocidin (PVL)-negative, multilocus sequence type (ST) 72 staphylococcal cassette chromosome mec (SCCmec) type IV, a major CA-MRSA clone in Korea. METHODS: ST72-IV isolates, which were susceptible to fluoroquinolones, gentamicin, rifampicin, and cotrimoxazole, were presumptively identified among 4667 MRSA isolates and then confirmed by SCCmec typing and multilocus sequence typing. A total of 124 cases of ST72-IV infections were analysed. FINDINGS: The annual incidence of infections by ST72-IV per 100,000 admissions increased from 45.5 to 66.3 cases during 2007-2009. The most frequently occurring type of infection was skin and soft tissue infection (SSTI) (46.0%), followed by pneumonia (27.4%) and bone and joint infection (9.7%). Surgical site infection accounted for 22.6% and 32.5% of community-onset (CO) healthcare-associated infection and hospital-onset (HO) infection, respectively. Pneumonia was most frequent (45.0%) among HO infection. Multivariate analysis showed that pneumonia increased the odds of all-cause mortality (odds ratio: 18.8; 95% confidence interval: 2.6-133.9) compared with other types of infection. CONCLUSIONS: Increasing trends were observed in annual incidence of CO and HO infections by ST72-IV in Korea. Pneumonia was the most frequent among HO infection and was associated with higher mortality. These findings pose important implications for successful antibiotic therapy and infection control in the era of CA-MRSA.

Tacrolimus as a risk factor for tuberculosis and outcome of treatment with rifampicin in solid organ transplant recipients.

BACKGROUND: The purpose of this study was to investigate the incidence, risk factors, and treatment outcome of tuberculosis (TB) in solid organ transplant (SOT) recipients treated with rifampicin. METHODS: The incidence density of TB was calculated by a retrospective cohort study. Risk factors for TB were analyzed by a nested case-control study. Treatment outcome and effects of anti-TB drugs on immunosuppressants and allograft were compared between patients whose initial 2-month intensive regimen included rifampicin and those whose intensive regimen did not. RESULTS: Among the 2144 SOT recipients over 16 years, 40 cases of TB were found (1.7%). The incidence density was 372 cases per 10(5) patient years (95% confidence interval [CI], 270-503), which was 4 times higher than for the general Korean population (90 cases per 10(5) person years). The median time to the development of TB was 234 days (range, 33-3940 days). The use of tacrolimus (odds ratio [OR] 4.90; 95% CI, 1.74-13.80; P = 0.003) and cytomegalovirus (CMV) infection within the prior 3 months (OR 4.62; 95% CI, 1.44-14.87; P = 0.01) were found to be risk factors for TB. Patients whose intensive regimen included rifampicin were more likely to have an increased dose of calcineurin inhibitors than patients whose intensive regimen did not include rifampicin (13/15 [86.7%] vs. 3/14 [21.4%], P = 0.001). Graft rejection and mortality did not differ between the 2 groups. CONCLUSIONS: Use of tacrolimus and CMV infection were major risk factors for TB in SOT recipients. The graft outcome and mortality did not differ whether rifampicin was used or not during the first 2-month intensive phase.

Fecal carriage of serotype K1 Klebsiella pneumoniae ST23 strains closely related to liver abscess isolates in Koreans living in Korea.

We determined the fecal carriage rate of serotype K1 Klebsiella pneumoniae in healthy Koreans and studied their genetic relationship with liver abscess isolates. We compared the carriage according to the country of residence. The stool specimens were collected through health promotion programs in Korea. K. pneumoniae strains were selected and tested for K1 by PCR. Serotype K1 isolates were characterized by multilocus sequence typing and pulsed field gel electrophoresis. A total of 248 K. pneumoniae isolates were obtained from 1,174 Koreans. Serotype K1 was identified in 57 (4.9%), of which 54 (94.7%) were ST 23 and were closely related to the liver abscess isolates. Participants aged >25 years showed a higher fecal carriage rate than those ≤ 25 (P = 0.007). The proportion of serotype K1 out of K. pneumoniae isolates in foreigners of Korean ethnicity who had lived in other countries was lower compared with those who had lived in Korea (5.6% vs 24.1%, P = 0.024). A substantial proportion of Koreans >25 years carries serotype K1 K. pneumoniae ST23 strains, which are closely related to liver abscess isolates. Differences in carriage rates by country of residence suggests that environmental factors might play an important role in the carriage of this strain.

Impact of inappropriate empiric antimicrobial therapy on outcome in Pseudomonas aeruginosa bacteraemia: a stratified analysis according to sites of infection.

BACKGROUND: The purpose of this study was to evaluate the impact of inappropriate empiric antimicrobial therapy on the outcome of Pseudomonas aeruginosa bacteraemia according to the primary infection site. METHODS: A retrospective cohort study including 202 patients with P. aeruginosa bacteraemia was performed. High-risk sites of infection were defined as the lung, intra-abdominal non-hepatobiliary tract or unknown source. RESULTS: Of the 202 patients with P. aeruginosa bacteraemia, 80 (39.6%) had received inappropriate empiric antimicrobial therapy. No significant difference in the 30-day mortality rate was found between the inappropriate therapy group and the appropriate therapy group (19/80 [23.8%] vs. 32/122 [26.2%], P = 0.692). Patients with pneumonia or non-hepatobiliary tract intra-abdominal infection showed significant association with high mortality, while those with urinary tract or hepatobiliary tract infection showed negative associations with mortality. In the subgroup analysis including 98 patients with high-risk sites of infection, the mortality rate of the inappropriate therapy group was significantly higher than that of the appropriate therapy group (14/26 [53.8%] vs. 23/72 [31.9%], P = 0.035). Inappropriate empiric antimicrobial therapy was also found to be one of the independent risk factors for mortality in patients with high-risk sites of infection (odds ratio [OR] 8.69; 95% confidence interval [CI] 1.86-40.59), along with renal disease, corticosteroid use, polymicrobial infection and higher Pitt bacteraemia score. CONCLUSION: Inappropriate empiric antimicrobial therapy adversely affected the outcome of P. aeruginosa bacteraemia in patients with high-risk sites of infection. Our data suggest that the impact of inappropriate antimicrobial therapy on the outcome of P. aeruginosa bacteraemia may be dependent on the primary site of infection.

Risk factors for treatment failure in patients with prosthetic joint infections.

A retrospective, observational cohort study was conducted to describe the incidence, clinical and microbiological findings and to evaluate risk factors for treatment failure associated with prosthetic joint infections (PJIs). We retrospectively reviewed the medical records of all patients undergoing total knee or total hip prosthesis implantation in our institution between 1994 and 2008. Our institution is a 1950-bed tertiary care university hospital and referral centre. A total of 93 patients with PJIs was identified although only 68 patients had undergone prosthesis implantation at our hospital. The overall infection rate was 0.63%. The most common organisms isolated were Gram positive (76.5%), including meticillin-resistant staphylococci. Two-stage arthroplasty was performed in 48 (51.6%) patients, and debridement and retention of the prosthesis in 34 (36.5%) patients. When 43 patients followed up for more than two years after treatment were included in treatment outcome analysis, the overall treatment failure rate was 41.8% (18/43). Staphylococcus aureus infection was the only clinical variable associated with treatment failure (odds ratio: 11.9; 95% confidence interval: 1.07-133.9; P=0.044), after adjustment for the other variables. In conclusion, S. aureus was the most common pathogen isolated in patients with PJI, and an independent risk factor for treatment failure in patients with PJI.