Hae Suk Cheong

Clinical Significance and Outcome of Nosocomial Acquisition of Spontaneous Bacterial Peritonitis in Patients with Liver Cirrhosis

BACKGROUND There have been few reports on the causes and treatment outcomes for nosocomial spontaneous bacterial peritonitis (SBP) in patients with liver cirrhosis. METHODS We performed a retrospective cohort study to compare the microbiological and clinical characteristics in nosocomial versus community-acquired SBP. All patients with SBP, for whom culture was proven to be positive for SBP at Samsung Medical Center (Seoul, Republic of Korea) from 1 January 2000 through 31 June 2007, were included. Medical records and laboratory data were reviewed. Nosocomial SBP was defined as SBP diagnosed after 72 h of hospitalization. RESULTS A total of 236 patients with SBP were enrolled (mean age +/- SD age, 56.6 +/- 10.7 years); 166 patients were women, and 70 were men. Nosocomial and community-acquired SBP occurred in 126 and 110 patients, respectively. Escherichia coli accounted for 102 (43.2%) of 236 isolates, Klebsiella species accounted for 33 isolates (14.0%), and Streptococcus species accounted for 23 isolates (9.8%). The overall 30-day mortality rate for nosocomial SBP was higher than that for community-acquired SBP (58.7% vs. 37.3%; P = .001). Nosocomial isolates of gram-negative organisms were significantly more resistant to third-generation cephalosporins (41% vs. 10.0%; P = .001) and quinolones (50.0% vs. 30.9%; P = .003), compared with community-acquired isolates. Multivariate analysis revealed that nosocomial infection, concomitant hepatocellular carcinoma, presentation with acute renal failure or shock, and resistance to third-generation cephalosporins were significant risk factors for 30-day mortality associated with SBP. CONCLUSIONS Nosocomial SBP has a poorer outcome than community-acquired SBP. The resistance to third-generation cephalosporins for gram-negative organisms, which are more common in nosocomial cases of SBP than in community-acquired cases of SBP, adversely affects the outcome of SBP in patients with liver cirrhosis.

Independent emergence of colistin-resistant Acinetobacter spp. isolates from Korea

High colistin resistance rates in Acinetobacter spp. were recently reported in Korean hospitals (J. Antimicrob. Chemother 2007;60:1163). In this study, we investigated if colistin-resistant Acinetobacter spp. isolates from Korean hospitals disseminated clonally or emerged independently. Multilocus sequence typing (MLST) analysis was performed for 58 colistin-resistant Acinetobacter spp. isolates: 8 isolates of the Acinetobacter baumannii subgroup A, 16 isolates of the A. baumannii subgroup B, and 34 isolates of the genomic species 13TU. A phylogenetic tree inferred from concatenated sequences of 7 MLST loci showed a clear distinction among the 3 Acinetobacter groups. In the MLST analysis, most colistin-resistant Acinetobacter spp. isolates showed different allele profiles at the 7 loci; that is, they belonged to different clones. Despite the clear distinction between the 3 Acinetobacter groups, interrelationships among the 3 groups were not consistent within the gene trees. In addition, some isolates showed clustering incongruent with their species or group identities in some gene trees. MLST analysis indicated that most colistin-resistant Acinetobacter spp. isolates from Korean hospitals arose independently. Considering the increasing use of colistin and the high recombination rate of Acinetobacter spp., independent but frequent emergence of colistin resistance in Acinetobacter spp. isolates is of great concern.

Extreme Drug Resistance in Acinetobacter baumannii Infections in Intensive Care Units, South Korea

To the Editor: Acinetobacter spp. have emerged as a cause of nosocomial infections, especially in intensive care units (ICUs). In South Korea, Acinetobacter spp. was ranked as the third most frequently found pathogen in ICUs (1). With the emergence of multidrug-resistant (MDR) or pandrug-resistant (PDR) isolates, few drugs are now available to treat MDR or PDR Acinetobacter infections; polymyxins are the only therapeutic option in many cases (2). Current polymyxin resistance rates among Acinetobacter isolates are low worldwide (3). We report the emergence of extreme drug resistance (XDR) in A. baumannii isolates from patients in ICUs of Samsung Medical Center in Seoul, South Korea. These isolates were resistant to all tested antimicrobial drugs, including polymyxin B and colistin, to which PDR isolates are normally susceptible. Sixty-three nonduplicate Acinetobacter spp. isolates were collected from the ICUs from April through November 2007. Species identification was performed based on partial RNA polymerase β-subunit gene sequences, amplified rDNA restriction analysis, and the gyrase B gene–based multiplex PCR method (3). Forty-four isolates were identified as A. baumannii: 9 as genomic species 3, six as genomic species 13TU, 2 as A. baumannii-like species, and 1 each as A. junnii and genomic species 10. In vitro susceptibility testing was performed and interpreted by using the broth microdilution method according to the Clinical and Laboratory Standards Institute guidelines (4). Colistin and polymyxin B resistances were defined as MIC >4 mg/L (4). MDR was defined as characterized by resistance to >3 classes of antimicrobial drugs, and PDR was defined as characterized by resistance to all antimicrobial drugs, regardless of colistin and polymyxin B susceptibility. XDR was defined as resistance to all antimicrobial drugs. Multilocus sequence typing (MLST) and pulsed-field gel electrophoresis (PFGE) were performed for all PDR isolates according to previously described methods (5,6). Genes encoding oxacillinases, such as those classified as OXA-23-like, OXA-24/40-like, OXA-51-like, and OXA-58-like, were detected as previously described (7). PCR and sequence analyses were performed to detect and characterize the other antimicrobial resistance genes, according to methods reported (8). Of 63 Acinetobacter isolates, 31.7% and 34.9% were resistant to imipenem and meropenem, respectively. Of the 63 isolates, 27.0% and 30.2% were resistant to polymyxin B and colistin, respectively. For the other antimicrobial drugs, Acinetobacter spp. isolates showed antimicrobial resistance rates >50%. Nineteen isolates (30.2%), all belonging to A. baumannii, were PDR. Most of these PDR isolates (16/19, 84.2%) were collected from endotracheal aspirate, and others were from peritoneal fluid and sputum. When characterized by PFGE and MLST, all PDR isolates belonged to a single clone, ST22, and all contained the blaOXA-23 and blaOXA-66 genes. ISAba1 was detected upstream of blaOXA-23 and blaOXA-66 in all PDR isolates. In addition, most PDR isolates contained blaTEM-116, blaPER-1, and blaADC-29 genes. TEM-116 is a point mutant derivative of TEM-1, Val84→Ile. All β-lactamase genes were located on a plasmid. Also, ISAba1 was located at the upstream of all the blaADC, which was shown by PCR. However, none of the isolates had blaCTX-M, blaVEB, blaIMP, blaVIM, or blaGIM. Of the PDR isolates, 8 were resistant even to colistin and polymyxin B. These 8 isolates also showed resistance to tigecycline (MICs 4 mg/L). Thus, they were resistant to all antimicrobial drugs tested in this study and were considered to have XDR. The underlying diseases of the patients whose isolates were examined varied (Table). Although 2 isolates with XDR were colonizers, 6 caused infections. All but 1 patient was treated with mechanical ventilation before isolation of the pathogen. Number of hospital days before isolation of A. baumannii was 13–256 days, and the number of ICU days before isolation was 2–38 days. Four patients were immunocompromised, and 3 had bacteremia. Among the patients with infections characterized by XDR, the overall 30-day mortality rate was 66.7%, and the infection-related 30-day mortality rate was 50.0%. All 8 isolates with XDR showed common characteristics: ST22 containing OXA-23, OXA-66, TEM-116, PER-1, and ADC-29. Table Clinical characteristics of 8 patients infected with extremely drug-resistant Acinetobacter baumannii isolates, South Korea* We report the emergence of XDR in PDR A. baumannii isolates in South Korea. Characteristics of PDR A. baumannii isolates suggest that they spread from a single clone. A single A. baumannii strain with XDR might evolve from the prevailing PDR A. baumannii and could disseminate in the ICU, probably after contamination of the hospital environment and by nosocomial transmission. In South Korea, a high resistance rate to imipenem and meropenem in Acinetobacter spp. isolates may lead to extensive use of polymyxins (3). Thus, we can hypothesize that the most prevalent carbapenem-resistant, or MDR A. baumannii clone, became PDR and then evolved into clones with XDR by acquisition of polymyxin resistance caused by antimicrobial pressure. Our investigation showed a simultaneous emergence of resistance to all antimicrobial agents available, including colistin, polymyxin B, and tigecycline. XDR poses serious problems in the treatment of patients with A. baumannii infections, especially given the slow development of new antimicrobial agents.

Clinical significance and predictors of community-onset Pseudomonas aeruginosa bacteremia.

BACKGROUND Pseudomonas aeruginosa bacteremia is a serious and possibly fatal condition. It is important to determine the likelihood of P. aeruginosa bacteremia when Gram-negative sepsis is suspected in community-onset infection. METHODS We performed a retrospective cohort study to identify the risk factors for P. aeruginosa infection in community-onset Gram-negative bacteremia. RESULTS A total of 106 patients with P. aeruginosa bacteremia and a total 508 patients with E. coli bacteremia were included in this study. Factors associated with P. aeruginosa bacteremia in the multivariate analysis included presentation with neutropenia, presentation with septic shock, indwelling central venous catheter, and health-care-associated infection (all P <.05). The 30-day mortality rate was 26.4% in patients with P. aeruginosa and 13.6% in those with E. coli bacteremia (P <.001). Multivariate analysis demonstrated that risk factors for mortality included a P. aeruginosa bacteremia, inappropriate initial antimicrobial therapy, a higher Charlsons weighted index of comorbidity, and a higher Pitt bacteremia score (all P <.05). In addition, urinary tract infection and benign pancreatobiliary disease were found to be protective factors for mortality based on multivariate analysis (all P <.05). CONCLUSIONS Our data suggest that initial empirical antimicrobial coverage of P. aeruginosa should be seriously considered in patients with neutropenia, presentation with septic shock, indwelling central venous catheter, or health-care-associated infection, when Gram-negative sepsis is suspected in community-onset infection.

Impact of de-escalation therapy on clinical outcomes for intensive care unit-acquired pneumonia

IntroductionDe-escalation therapy is a strategy currently used for the management of nosocomial pneumonia. In this study, we evaluated clinical outcomes and risk factors related to de-escalation therapy in patients with intensive care unit (ICU)-acquired pneumonia.MethodsThis was a retrospective observational cohort study of ICU patients who developed pneumonia more than 48 hours after admission to the ICU at Samsung Medical Center from September 2004 to December 2007.ResultsThe 137 patients comprised 44 (32.1%) who received de-escalation therapy and 93 in the non-de-escalation group. The de-escalation group showed a lower pneumonia-related mortality rate than the non-de-escalation group by day 14 (2.3% vs. 10.8%, respectively; P = 0.08) and by day 30 (2.3% vs. 14%, respectively; P = 0.03) after the diagnosis of pneumonia. The variables independently associated with ICU-acquired pneumonia-related mortality included the Acute Physiology and Chronic Health Evaluation II (APACHE II) score and the modified Clinical Pulmonary Infection Score (CPIS) after 5 days with pneumonia. The non-de-escalation group had significantly higher APACHE II score and modified CPIS after 5 days with ICU-acquired pneumonia compared to the de-escalation group. Among all patients, 20.4% (28 of 137) had negative cultures for pathogens, and 42.9% (12 of 28) received de-escalation therapy. The latter 12 patients received de-escalation therapy and survived 30 days after the diagnosis of pneumonia.ConclusionsPatients in the de-escalation group showed a significantly lower mortality rate compared to patients in the non-de-escalation group. De-escalation therapy can be safely provided to patients with ICU-acquired pneumonia if they are clinically stable by day 5, even in those whose respiratory specimen cultures yield no specific pathogens.

Impact of inappropriate antimicrobial therapy on outcome in patients with hospital-acquired pneumonia caused by Acinetobacter baumannii.

OBJECTIVES The purpose of this study was to evaluate the impact of inappropriate antimicrobial therapy on the outcome of patients with hospital-acquired pneumonia (HAP) caused by Acinetobacter baumannii. METHODS All cases of HAP caused by A. baumannii from January 2000 to March 2006 at the Samsung Medical Center (Seoul, Korea) were analyzed retrospectively. RESULTS A total of 116 patients with clinically significant Acinetobacter HAP were enrolled. Among the A. baumannii isolates, 60.3% showed multi-drug resistance (MDR), 16.4% were found to have imipenem resistance, and 15.5% had pan-drug resistance (PDR). The mean APACHE II score of the patients was 22.3 +/- 7.9. The overall in-hospital and pneumonia-related mortality rates were 47.4% and 37.9%, respectively. The univariate analysis showed that the factors associated with pneumonia-related mortality were: MDR, PDR, high APACHE II score, inappropriate empirical antimicrobial therapy, and inappropriate definitive antimicrobial treatment (All p < 0.05). Among these, a high APACHE II score and inappropriate definitive antimicrobial therapy were found to be independent factors associated with a high mortality, after adjustment for other variables. CONCLUSIONS The appropriate definitive antimicrobial therapy should be provided in patients with HAP caused by A. baumannii.

Two distinct clones of carbapenem-resistant Acinetobacter baumannii isolates from Korean hospitals

We investigated the characteristics of 48 carbapenem-resistant Acinetobacter baumannii isolates collected from 5 tertiary care hospitals in Korea by multilocus sequencing typing, pulsed-field gel electrophoresis, and polymerase chain reaction amplification of the antimicrobial resistance determinants. We identified 2 distinct main clones of carbapenem-resistant A. baumannii isolates, which showed different antimicrobial resistance profiles and are also differentiated by the kinds of oxacillinase (OXA) carbapenemases and Acinetobacter-derived cephalosporinase (ADC) beta-lactamases. One main clone, ST22:A, had 27 carbapenem-resistant isolates (56.3%), showed high polymyxin B and colistin resistances (33.3% and 37.0%, respectively), and contained both bla(OXA-51-like) and bla(OXA-23-like) genes and the bla(ADC-29) or bla(ADC-30) gene. In contrast, the other main clone, ST28:B, included 15 isolates (31.3%), showed complete susceptibilities to polymyxin B and colistin, and contained only the bla(OXA-51-like) gene and bla(ADC-31) or bla(ADC-32) genes. The distribution of these main carbapenem-resistant A. baumannii clones did not relate to locality, indicating that they are widespread in Korean hospitals. In addition, we found new types of PER beta-lactamases, PER-6.

Changes of serotype and genotype in Streptococcus pneumoniae isolates from a Korean hospital in 2007

We investigated the change in clones and serotypes of Streptococcus pneumoniae isolates in a Korean tertiary-care hospital. Serotypes of S. pneumoniae isolates were determined by the capsular quellung method, and in vitro susceptibility testing was performed by broth microdilution method. Multilocus sequence typing was performed to determine the genotypes of the S. pneumoniae isolates. The erm(B) and mef(A) genes in erythromycin-resistant isolates were also detected using the duplex polymerase chain reaction method. During the 2 periods assayed (1998-2000 and 2007), 7-valent pneumococcal conjugate vaccine (PCV7) serotypes decreased significantly from 58.3% to 30.9% (P = 0.001). Especially, serotypes 19F and 23F decreased significantly from 31.7% to 8.5% (P < 0.0001) and from 20.0% to 7.4% (P = 0.021), respectively. In contrast to the other PCV7 serotypes, serotype 14 coupled with CC554 emerged in 2007, which may indicate no effect of PCV7 against serotype 14 isolates from Korea and the possibility of a different subtype. Of the non- PCV7 serotypes, serotype 19A increased from 8.3% to 14.9% (P = 0.227) and serotype 15 increased from 0% to 8.5% (P = 0.023). The increase of serotype 19A was due to the expansion of a preexisting clone with serotype 19A, ST320. However, S. pneumoniae isolates of serotype 15 showed diverse STs. Our data may provide helpful information in local vaccine serotype expansion or replacement in Korea.

Mortality of Community-Acquired Pneumonia in Korea: Assessed with the Pneumonia Severity Index and the CURB-65 Score

The pneumonia severity index (PSI) and CURB-65 are widely used tools for the prediction of community-acquired pneumonia (CAP). This study was conducted to evaluate validation of severity scoring system including the PSI and CURB-65 scores of Korean CAP patients. In the prospective CAP cohort (participated in by 14 hospitals in Korea from January 2009 to September 2011), 883 patients aged over 18 yr were studied. The 30-day mortalities of all patients were calculated with their PSI index classes and CURB scores. The overall mortality rate was 4.5% (40/883). The mortality rates per CURB-65 score were as follows: score 0, 2.3% (6/260); score 1, 4.0% (12/300); score 2, 6.0% (13/216); score 3, 5.7% (5/88); score 4, 23.5% (4/17); and score 5, 0% (0/2). Mortality rate with PSI risk class were as follows: I, 2.3% (4/174); II, 2.7% (5/182); III, 2.3% (5/213); IV, 4.5% (11/245); and V, 21.7% (15/69). The subgroup mortality rate of Korean CAP patients varies based on the severity scores and CURB-65 is more valid for the lower scores, and PSI, for the higher scores. Thus, these variations must be considered when using PSI and CURB-65 for CAP in Korean patients.

The cefazolin inoculum effect in methicillin-susceptible Staphylococcus aureus blood isolates: their association with dysfunctional accessory gene regulator (agr).

We evaluated the clinical significance of the cefazolin inoculum effect (CIE) in methicillin-susceptible Staphylococcus aureus (MSSA) isolates. In total, 146 isolates were recovered from patients with MSSA bacteremia at 9 hospitals in Korea. The CIE was observed in 16 MSSA isolates, and while type A was the only detected β-lactamase in MSSA isolates exhibiting the CIE, no strains expressing type B, C, or D β-lactamases exhibited this effect. The CIE was only observed in agr group III and I isolates and was significantly more common in isolates with agr dysfunction than in those with functional agr (P<0.001). Even among isolates producing type A β-lactamase, the CIE was also prevalent in isolates with dysfunctional agr than in isolates with functional agr (P=0.025). This study demonstrates an association between the CIE of MSSA isolates and agr dysfunction, in addition to those between the CIE and type A β-lactamase.