Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Prevalence of Carbapenem-Resistant Gram-Negative Infections in the United States Predominated by Acinetobacter baumannii and Pseudomonas aeruginosa

Prevalence of Carbapenem-Resistant Gram-Negative Infections in the United States Predominated by... Open Forum Infectious Diseases MAJOR ARTICLE Prevalence of Carbapenem-Resistant Gram-Negative Infections in the United States Predominated by Acinetobacter baumannii and Pseudomonas aeruginosa 1 2 3 1,* 1 4 4 4 Bin Cai, Roger Echols, Glenn Magee, Juan Camilo Arjona Ferreira, Gareth Morgan, Mari Ariyasu, Takuko Sawada, and Tsutae “Den” Nagata 1 2 3 4 Shionogi Inc., Florham Park, New Jersey; ID3C, Easton, Connecticut; Premier Research Service, Inc., Charlotte, North Carolina; Shionogi & Co., Ltd., Kitaku, Osaka, Japan Background. Carbapenem-resistant (CR) Gram-negative pathogens are recognized as a major health concern. This study exam- ined the prevalence of infections due to 4 CR Gram-negative species (Acinetobacter baumannii, Pseudomonas aeruginosa, Klebsiella pneumoniae, and Escherichia coli) in the United States and assessed their impact on hospital stays and mortality. Methods. Hospitalized patients with laboratory-confirmed infection due to any of the 4 Gram-negative pathogens were iden- tified from the Premier Healthcare Database. Proportions of CR were calculated by pathogen and infection site (blood, respiratory, urinary, or other) for the United States as whole and by census regions. Crude and adjusted odds ratios for in-hospital mortality were produced using logistic regression. Results. From 2009 to 2013, 13 262 (4.5%) of 292 742 infections due to these 4 Gram-negative pathogens were CR. Of these CR infections, 82.3% were caused by A. baumannii (22%) or P. aeruginosa (60.3%), while 17.7% were caused by K. pneumoniae or E. coli. CR patients had longer hospital stays than carbapenem-susceptible (CS) patients in all pathogen-infection site cohorts, except in the A. baumannii-respiratory cohort. The crude all cause in-hospital mortality was greater for most pathogen-infection site cohorts of the CR group compared with the CS group, especially for A. baumannii infection in the blood (crude odds ratio [OR], 3.91; 95% confidence interval [CI], 2.69–5.70). This difference for the A. baumannii-blood cohort remained aer ad ft justing for the relevant covariates (adjusted OR, 2.46; 95% CI, 1.43–4.22). Conclusion. e Th majority of CR infections and disease burden in the United States was caused by nonfermenters A. baumannii and P. aeruginosa. Patients with CR infections had longer hospital stays and higher crude in-hospital mortality. Keywords. carbapenem resistance; Gram-negative; nonfermenters; CRE; length of stay; in-hospital mortality. Infections due to carbapenem-resistant (CR) Gram-negative it was not well appreciated as a source of resistance transmission pathogens have been reported from many countries with var- until 2017 [7, 12–16], when the World Health Organization rec- iable prevalence and associated morbidity and mortality [1–4]. ognized the importance of CR A. baumannii and P. aeruginosa as e dra Th matic increase and spread of CR infections over the past equal to that of the CREs [17]. These CR nonfermenters are oen ft decade are recognized as a major public health concern [2, 5–9]. multidrug resistant and associated with substantial morbidity and Recent infection control measures have focused on the identifica- mortality [18, 19]. This study was designed to determine the prev- tion and spread of CR Enterobacteriaceae (CREs), which is mainly alence of CR organisms (CROs) in the United States, specifically due to mobile carbapenemase enzymes, both serine (KPC, GES, the Enterobacteriaceae Escherichia coli and Klebsiella pneumoniae and OXA) and metallo-carbapenemases (NDM, VIM, and IMP) and the nonfermenters Pseudomonas aeruginosa and Acinetobacter [6, 10]. Resistance mechanisms involving porin channel mutations baumannii, using an electronic database from a geographically and efflux pump overproduction may also contribute to the iden- diverse network of US hospitals. Besides determining the prev- tification of CREs [6 , 10, 11]. The frequency of carbapenem resist- alence of CR infections in the United States, the study compared ance among the non-glucose fermenting Pseudomonas aeruginosa similar infections that were carbapenem susceptible (CS) in order and Acinetobacter baumannii has also increased in recent years, but to assess the mortality attributed to the CROs. METHODS Received 18 May 2017; editorial decision 8 August 2017; accepted 9 August 2017. This retrospective study of CR Gram-negative infections *Employee at the time of the study between 2009 and 2013 was conducted using microbiology Correspondence: B. Cai, MD, PhD, Global Observational Research and Epidemiology, Shionogi Inc., Florham Park, New Jersey, 07932, USA. (bin.cai@shionogi.com). data linked with patient-level in-hospital discharge data from Open Forum Infectious Diseases the Premier Healthcare Database [20]. The Premier Healthcare © The Author 2017. Published by Oxford University Press on behalf of Infectious Diseases Database is an anonymous census of inpatients and hospi- Society of America. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits tal-clinic outpatients from geographically diverse, mixed unrestricted reuse, distribution, and reproduction in any medium, provided the original work teaching and nonteaching hospitals with varied bed sizes in the is properly cited. United States. The hospital discharge data contain information DOI: 10.1093/ofid/ofx176 Carbapenem-Resistant Gram-Negative Infections in the United States Predominated by A. baumannii & P. aeruginosa • OFID • 1 on admission, discharge, and a date-stamped log of all billed ertapenem, imipenem, meropenem) from patients who received items, including procedures, medications, laboratory, diagnos- at least 1 systemic antibiotic treatment were selected. tic, and therapeutic services at the individual patient level. Statistical Analyses Microbiology data, collected from approximately 30% of Categorical variables were described with the number of participating hospitals, include the time the microbiological patients and percentages and compared between CR and CS culture was obtained, type of pathogen identified, and anti- in each pathogen-infection site cohort using a chi-square or microbial drugs used for susceptibility testing along with the Fisher’s exact test where appropriate. Continuous variables method, result, and interpretation of antibiotic susceptibility. were described with mean, standard deviation, and median. All data were determined by individual hospital laboratories, The differences between CR and CS groups were analyzed using which were Clinical Laboratory Improvement Amendments Student’s t test or the Wilcoxon rank-sum test. A P-value <.05 (CLIA) compliant using standard laboratory methods, and indicated statistically significant differences between groups. interpretations of susceptibility results were reported as resist- In-hospital mortality was compared between CR and CS for ant (R), intermediate (I), susceptible (S), or none (N). each pathogen infection site cohort with and without adjusting e p Th athogens were further identified by infection sites for covariates, such as age, gender, race, ethnicity, mechanical (blood, respiratory, urinary, or other) based on the source of the ventilation, renal impairment, geographic region, and each of culture sample. u Th s, this study had 16 distinct pathogen-in- the Charlson comorbid conditions [21]. fection site cohorts. Within each cohort, pathogens were char- All analyses were performed using SAS 9.3 (SAS Institute acterized as CR or CS. Bacteria were defined as CR if they were Inc., Cary, North Carolina). resistant to at least 1 of the carbapenems tested and as CS if the pathogen was susceptible to or intermediate for the carbap- RESULTS enems tested. If selected patients had multiple hospitalizations This study includes data collected from 206 acute care hospitals during the study period, only data from the first hospitaliza- in the United States between January 1, 2009, and December tion were used. If a microbiologic culture identified multiple 31, 2013. The total numbers of pathogen isolates were 173 200 pathogens, the patient was counted in each pathogen group. If E.  coli, 56 552 K.  pneumoniae, 56 477 P.  aeruginosa, and 6508 a patient had multiple cultures of the same pathogen, the first A. baumannii. culture was used as index date if the pathogen had the same Carbapenem Resistance Rate susceptibility test results for all cultures. Otherwise, the first CR The overall CR rate among the 4 selected pathogens was 4.5% culture was used as the index date, and the patient was in the CR (13 262 of 292 742). A.  baumannii had the highest CR rate: group. Only cultured pathogens from patients who were receiv-     blood 40.1%, respiratory 50.4%, urine 42.0%, and other 42%. ing antibiotic treatment were included in the final data set. The CR rate for P.  aeruginosa ranked second highest and was e p Th rimary outcome variables were the rates of CR in each more variable by infection site: 10.3% in blood, 19.4% in res- pathogen-infection site cohort for all hospitals and separately piratory, 11.7% in urinary, and 12.6% in other (Figure 1). More for each US census region. Other outcome variables included than 80% of all CR infections were caused by A. baumannii or in-hospital mortality, total length of stay (LOS) in the hospital, P.  aeruginosa. This was consistent across 9 US census regions, LOS aer p ft ositive culture (iLOS), total LOS in the intensive care except for the Middle Atlantic region, where 67% (Table  1) of unit (ICU) for the CR group and CS group, and survival status within each pathogen-infection site cohort. e Th use of antibiotics with Gram-negative activity was Carbapenem Resistantce Rate assessed by class (aminoglycoside, carbapenem, cephalosporin, colistin, monobactam, penicillin, polymyxin B, quinolone, tetra- 50.4% 50.4% cycline, and other Gram-negative antibiotics) and the time when 42.0% 42.0% 40.1% 40.1% the patient received them during the hospitalization. Antibiotics 60.0% administered in the days prior to the date of the index culture 50.0% were considered “prior” antibiotics, those received from the date 40.0% 19.4% 19.4% of the index culture through 3 days aer t ft he index culture were 12.6% 12.6% 11.7% 11.7% considered “empiric,” and those received aer 3  ft days from the 10.3% 10.3% 30.0% A. baumannii index culture were considered “definitive” treatment. 20.0% 4.9% 4.9% 3.6% 3.6% 4.0% 4.0% 3.1% 3.1% P. aeruginosa Study Population 10.0% K. pneumoniae 0.2% 0.2% 0.4% 0.4% 0.2% 0.2% 0.3% 0.3% E. coli All hospitalized patients with microbiological cultures identify- 0.0% Blood Respiratory Urinary Others ing A. baumannii, P. aeruginosa, K. pneumoniae, or E.  coli from appropriate clinical specimens and with associated interpreta- Figure 1. Carbapenem resistance rate by pathogen-infection site cohort. tions of susceptibility to carbapenem antibiotics (doripenem, 2 • OFID • Cai et al CR infections were caused by the nonfermenters. For most regions, CR P.  aeruginosa outnumbered CR A.  baumannii by 3:1, except for the New England region, where 85% of CR infec- tions were due to P. aeruginosa and only 6.4% were from A. bau- mannii (Table 2). Demographic and Clinical Characteristics Compared with patients with CS A. baumannii, patients with CR isolates tended to be older when the source was blood stream infection (mean±SD, 60.5 ± 16.6 years vs 53.6 ± 21.8 years), respiratory infection (mean±SD, 62.5 ± 16.8 vs 55.8 ± 23.5 years), or other infection sites (mean±SD, 60.7 ± 16.5 vs 57.0 ± 20.0 years), but of similar age when the source was urinary tract infection (mean±SD, 61.8 ± 17.5 vs 61.5 ± 19.8 years). More than two-thirds of patients with A. baumannii infections were white; however, the proportions of nonwhite patients in the CR groups were higher than in the CS groups for all infection sites. Slightly more male patients than female patients were identified in both the CR and CS groups. Nearly half of all patients with A. baumannii infections came to the hospitals from their homes, though the proportion was slightly higher in the CS group than in the CR group. Nearly 70% of these patients were admitted via emergency rooms. Patients with CR infections were more likely to be discharged to skilled nursing facilities or rehabilitation or long-term care facilities than patients with CS infections (43% vs 24% for blood, 51% vs 42% for respiratory, 58% vs 33% for urinary, and 51% vs 31% for other). Patients with CR infections also had a higher Charlson comorbidity index score for all infection sites, a longer total LOS for infections in the blood, urinary, and other (median days of CR vs CS, 15 vs 9 for blood, 12 vs 7 for urinary, and 12 vs 8 for other), and longer total ICU stays for blood, urinary, and other (median days of CR vs CS, 8 vs 5, 8 vs 4, and 9 vs 8, respectively). Similarly, a longer LOS after positive culture was seen in the CR group than the CS group among all patients and those who survived for each infection site except respiratory (Table 3). Among nonsurvivors with blood stream infection, the median LOS after positive cul- ture was 2 days after the culture was obtained in the CR group and 4 days in the CS group. Patients with CR infections were slightly more likely to be on dialysis or have renal impairment during their hospital stays. Similar patterns in admission sources, discharge status, hemodialysis, and renal impairment were observed for patients infected with CR P.  aeruginosa, CR K.  pneumoniae, and CR E. coli compared with their CS counterparts. However, patients with CR P. aeruginosa isolates were younger than patients with CS isolates for all infection sites (mean ± SD, 60.6  ±  17.3 vs 65.2  ±  19.1 for blood, 60.6  ±  19.1 vs 63.2  ±  20.3 for respira- tory, 67.1 ± 17.1 vs 71.5 ± 17.1 for urinary, and 60.6 ± 18.5 vs 61.1  ±  20.2 for other), but they had similar Charlson comor- bidity index scores. LOS, ICU LOS, and LOS aer ft positive Carbapenem-Resistant Gram-Negative Infections in the United States Predominated by A. baumannii & P. aeruginosa • OFID • 3 Table 1. Carbapenem-resistant infections caused by A. baumannii, P. aeruginosa, K. pneumoniae, and E. coli by US census region All 4 Pathogens A. baumannii + P. aeruginosa K. pneumoniae + E. coli CR Cases CR Cases CR Cases Number of Total Total CR % of Total Total % of Total % of Total CR Cases for 4 Total % of Total % of Total CR Cases for 4 US Census Regions Hospitals Infections Cases Infections Infections CR Cases Infections Pathogens Infections CR Cases Infections Pathogens East North Central 43 47 702 3164 6.6 10 381 2542 24.5 80.3 37 321 622 1.7 19.7       East South Central 9 21 440 770 3.6 4293 703 16.4 91.3 17 147 67 0.4 8.7     Middle Atlantic 19 37 582 2499 6.6 10 449 1672 16.0 66.9 27 133 827 3.0 33.1       Mountain 15 8582 765 8.9 2260 710 31.4 92.8 6322 55 0.9 7.2 New England 9 11 964 346 2.9 2417 316 13.1 91.3 9547 30 0.3 8.7 Pacific 27 39 045 1237 3.2 6447 1070 16.6 86.5 32 598 167 0.5 13.5     South Atlantic 40 62 217 2385 3.8 14 652 2017 13.8 84.6 47 565 368 0.8 15.4       West North Central 18 17 854 403 2.3 3130 329 10.5 81.6 14 724 74 0.5 18.4     West South Central 26 46 356 1693 3.7 8955 1550 17.3 91.6 37 401 143 0.4 8.4     Total 206 292 742 13 262 4.5 62 984 10 909 17.3 82.3 229 758 2353 1.0 17.7         Abbreviations: CR, carbapenem resistant. culture were longer for CR infections than CS infections for all infection sites. The average age of patients with K. pneumoniae was approximately 60 years. There was no significant age differ- ence between patients with CR and CS infections in blood, urine, or respiratory. In other infection sites, CR patho- gen–infected patients were approximately 3 years older than those in the CS group. er Th e were more male patients than female patients in all study cohorts, except for those in the urinary cohort, where significantly higher numbers of female patients were identified. Due to the small number of patients in the CR-urinary cohort, there was no statistical significance in the observed differences in age (average of 60 years), race, or Charlson comorbidity index score. In-Hospital Mortality Patients with CR infections consistently had numerically higher crude mortality rates than patients with CS infections for each of the 16 pathogen-infection site cohorts. The crude odds ratio (OR) for CR vs CS ranged from 1.31 for E. coli in respiratory to 3.91 for A. baumannii in blood (Table 4). After adjusting for age, gender, race, ethnicity, Charlson comorbid conditions, mechanical ventilation, renal impairment, and geographic cen- sus regions, the odds of dying from CR A. baumannii in blood (adjusted OR, 2.46) and in respiratory (adjusted OR, 1.27) and from P. aeruginosa in other (adjusted OR, 1.20) remained statis- tically significant (Table 4). Patients with CR pathogens were more likely to receive more than 1 systemic antibiotic than patients with CS pathogens. Also, patients with pathogens identified from the blood, res- piratory, and other were more likely to receive 2 or more anti- biotics than if the source was urinary. The number and types of antibiotics used were highly variable. During the definitive treatment period, more than 100 combinations (defined as dif- ferent antibiotic classes regardless of the initiation time, dose, or duration) were used for each pathogen, but fewer than 20 of these combinations were used by more than 3% of patients. Fewer different antibiotic combinations were used to treat patients with CS infections than patients with CR infections. DISCUSSION This study showed that in the United States the overall disease burden and impact on survival was greatest among the non-glu- cose fermenting A. baumannii and P. aeruginosa. A. baumannii had the highest rate of CR at each infection site, followed by P. aeruginosa and K. pneumoniae. E. coli consistently had the lowest rate of CR. The combined number of CR A. bauman- nii and CR P. aeruginosa infections accounted for more than 80% of all CR infections. Though this study included both com- munity-onset and hospital-acquired infections, the pattern of CR from the selected 4 pathogens was comparable with the US 4 • OFID • Cai et al Table 2. A. baumannii and P. aeruginosa CR Infection by Region All 4 Pathogens Acinetobacter baumannii Pseudomonas aeruginosa CR Cases CR Cases CR Cases Number of % of Total % of Total % of Total CR Cases for 4 % of Total % of Total CR Cases for 4 Regions Hospitals Total Infections Cases Infections Total Infections Cases Infections Pathogens Total Infections Cases Infections Pathogens East North Central 43 47 702 3164 6.6 1360 863 63.5 27.3 9021 1679 18.6 53.1 East South Central 9 21 440 770 3.6 498 223 44.8 29.0 3795 480 12.6 62.3 Middle Atlantic 19 37 582 2499 6.6 914 288 31.5 11.5 9535 1384 14.5 55.4 Mountain 15 8582 765 8.9 326 203 62.3 26.5 1934 507 26.2 66.3 New England 9 11 964 346 2.9 183 22 12.0 6.4 2234 294 13.2 85.0 Pacific 27 39 045 1237 3.2 739 327 44.2 26.4 5708 743 13.0 60.1 South Atlantic 40 62 217 2385 3.8 1322 488 36.9 20.5 13 330 1529 11.5 64.1     West North Central 18 17 854 403 2.3 256 77 30.1 19.1 2874 252 8.8 62.5 West South Central 26 46 356 1693 3.7 909 424 46.6 25.0 8046 1126 14.0 66.5 Total 206 292 742 13 262 4.5 6507 2915 44.8 22.0 56 477 7994 14.2 60.3     Abbreviation: CR, carbapenem resistant. Table 3. Median Days Stayed in the Hospital After Positive Culture by Pathogen, Infection Site, and Survival Status Blood Respiratory Urinary Other Pathogen Length of Stay After Positive Culture CR CS CR CS CR CS CR CS Acinetobacter baumannii iLOS 8 6 8 10 8 4 7 5 iLOS among survived 11 6 9 10 8 4 7 5 iLOS among deceased 2 4 6 7 6 6 7 7 Pseudomonas aeruginosa iLOS 5 4 20 7 5 3 8 5 iLOS among survived 6 4 20 8 5 3 6 5 iLOS among deceased 3 2 18.5 4 7 4 33.5 6 Klebsiella pneumoniae iLOS 10 6 10 7 6 4 9 6 iLOS among survived 10 6 11 8 6 4 9 6 iLOS among deceased 6 3 6 4 8 4 9 6 Escherichia coli iLOS 10 6 10 6 7 4 10 5 iLOS among survived 10 7 10 6 6 4 9 5 iLOS among deceased 9 1 11 5 9 5 14 6 Abbreviations: CR, carbapenem resistant; CS, carbapenem susceptible; iLOS, length of hospital stay after positive culture. Centers for Disease Control and Prevention’s (CDC’s) National CR infection for K.  pneumoniae (48.7%) and E.  coli (57.3%). Healthcare Safety Network (NHSN) results, which focused on Differences in the frequency of CR infection by infection site health care–acquired infections via central lines, urinary cathe- have also been reported in other studies [9, 23, 24]. ters, and mechanical ventilators [9, 22]. The relative importance e CR ra Th tes varied by the geographic regions. The Middle of CR A. baumannii has also been recently emphasized by the Atlantic region had the highest proportion of CR infections European Centre for Disease Prevention and Control (ECDC), due to K.  pneumoniae and E.  coli (33.1%), while the New which reported that CR A. baumannii (resistant through the England region had a higher proportion of CR infections due production of carbapenemases) might be more widely dissemi- to P.  aeruginosa (85%). Other studies also observed the varia- nated in Europe than CREs [7, 12]. tions in frequency of CROs by hospitals and by region [22, 24, This study also showed that distribution of CR pathogens 25]. Although actual incidence cannot be estimated from this differed by infection sites. Respiratory was the most common study, the regional differences observed in this study support site of CR infection for A.  baumannii (40.3%) and P.  aerug- the understanding that infection control of resistant pathogens inosa (40.1%), while urinary was the most common site of needs to be based on local epidemiology. Table 4. In-Hospital Mortality: Crude and Adjusted Odds Ratio (CR vs CS) In-Hospital Mortality by Each Site, No. (%) Crude Adjusted Pathogen Site CR CS Odds Ratio Lower 95% CI Upper 95% CI Odds Ratio Lower 95% CI Upper 95% CI Acinetobacter Blood 103 (38.2) 55 (13.6) 3.91 2.69 5.70 2.46 1.43 4.22 baumannii Respiratory 305 (26.0) 230 (19.9) 1.41 1.16 1.71 1.27 1.01 1.58 Urinary 35 (9.6) 31 (6.2) 1.62 0.98 2.69 0.62 0.32 1.19 Other 173 (15.6) 117 (7.7) 2.24 1.74 2.87 1.19 0.88 1.61 Pseudomonas Blood 112 (33.3) 590 (20.1) 2.02 1.58 2.58 1.19 0.88 1.63 aeruginosa Respiratory 661 (20.6) 1977 (14.8) 1.50 1.36 1.65 1.11 1.00 1.24 Urinary 208 (9.8) 1031 (6.3) 1.58 1.35 1.85 1.05 0.88 1.26 Other 389 (16.8) 1139 (7.1) 2.64 2.33 2.99 1.20 1.03 1.39 Klebsiella Blood 63 (27.3) 819 (13.2) 2.46 1.83 3.32 1.10 0.75 1.61 pneumoniae Respiratory 98 (25.4) 1547 (20.6) 1.32 1.04 1.67 0.97 0.76 1.25 Urinary 90 (9.1) 1616 (5.2) 1.81 1.45 2.27 1.07 0.83 1.37 Other 76 (18.1) 863 (8.7) 2.33 1.80 3.01 1.02 0.76 1.37 Escherichia coli Blood 5 (16.1) 1709 (8.7) 2.03 0.78 5.29 1.91 0.60 6.10 Respiratory 8 (26.7) 1469 (21.7) 1.31 0.58 2.95 0.74 0.29 1.85 Urinary 13 (6.9) 4699 (3.9) 1.84 1.05 3.23 1.55 0.80 3.00 Other 8 (10.1) 1466 (5.8) 1.84 0.89 3.84 0.78 0.33 1.84 Abbreviations: CI, confidence interval; CR, carbapenem resistant; CS, carbapenem susceptible. Adjusted for age, gender, race, ethnicity, various comorbid conditions, mechanical ventilation, renal impairment, and geographic regions. Carbapenem-Resistant Gram-Negative Infections in the United States Predominated by A. baumannii & P. aeruginosa • OFID • 5 As shown in this study, a majority of the CR pathogens were was not possible to estimate CR infection rates in smaller geo- treated with multiple antibiotics, especially for A.  baumannii graphic areas, even though infections could vary from city to and P.  aeruginosa or for infections in the blood or respiratory city and hospital to hospital. This study does identify regional infection sites. Even with combination treatment, patients with differences in prevalence rates; however, it does not assess CR infections still had longer total hospital (total and aer p ft os- whether these findings are driven by individual institutions or itive culture) and ICU stays, except among nonsurvivors with episodic outbreaks within institutions. A. baumannii in blood and respiratory, where CR patients had This study only examined 2 Enterobacteriaceae, E.  coli and shorter LOS aer p ft ositive culture compared with CS patients K. pneumoniae, and 2 nonfermenters, A. baumannii and P. aerug- and the median for CR in blood was 2 days aer t ft he culture was inosa. Other nonfermenters, for example, Stenotrophomas obtained. This further highlights the treatment challenges faced maltophilia, which are highly carbapenem resistant [26, 27], by clinicians today and the need for more effective antibiotics may represent another significant source of disease caused by and more rapid information on antibiotic susceptibility. CR pathogens. A remarkably high number of different antibiotic combi- CONCLUSION nations were used to treat these infections, especially for CR pathogens. This indicates that there is no generally accepted This study showed that CR non-glucose fermenting Gram- standard of care regimen for CR infections, which underlies the negative bacteria, such as A. baumannii and P. aeruginosa, fact that there are few available antibiotics to treat CR infec- contribute to the greater disease and mortality burden when tions, and those that may be available are associated with sig- compared with CREs, such as K. pneumoniae and E. coli. This nificant toxicity. has been observed for the study as whole and for each census e a Th uthors believe that the key strength of this study is the region. These nonfermenters contributed more than 80% of large, geographically diverse data set that includes both detailed CR cases in our analysis of data over a period of 5 years. CR microbiologic data as well as patient-level information and rates differed from region to region, but the predominance of outcomes. Unlike the CDC and other data sets that focus on CR nonfermenters over CREs was consistent. The geographic hospital-acquired infections, the Premier Healthcare database variability was greater among K. pneumoniae and E. coli than includes both hospital-acquired and community-onset infec- the nonfermenters. However, because the authors cannot deter- tions, many of which are healthcare related, thus better repre- mine specific population numbers by hospital, the incidence senting the full disease burden related to CR infections in US of CR infections cannot be estimated or compared. These data hospitals. for the United States are in sharp contrast to most reports from This study selected subjects based on microbiology results, Europe and Latin America, where CREs are predominant. not based on the discharge diagnosis, avoiding potential biases Because patients with CR infections had longer hospital and in the discharge diagnoses due to various reasons, including ICU stays, more antibiotic treatments, and a higher mortality coding errors or reporting bias. The results of the microbiol- rate, the CR infections likely represent a higher cost burden ogy and drug susceptibility testing, as well as the documented to the health care system. The development of new antibi- antibiotic treatments, provide a clear picture of the pattern of otics should not only address CREs, but more importantly, it infections in acute care hospitals. should address CR non-fermenters such as P. aeruginosa and However, there are also several limitations in this study. First, A. baumannii. case definitions were based on microbiological culture results Acknowledgment that may underestimate or overestimate the number of infec- Potential coni fl cts of interest. Dr. Cai, Dr. Morgan, Dr. Arjona Ferreira, tions. Although some culture results may represent coloniza- Ms. Ariyasu, Ms. Sawada, and Dr. Nagata are full-time employees of tion rather than true infection, this study only analyzed patients Shionogi. Dr. Echols reports personal fees from Shionogi, Inc., during the for whom the cultures had susceptibility testing performed and conduct of the study. All authors have submitted the ICMJE Form for Disclosure of Potential where their physician prescribed systemic antimicrobial ther- Conflicts of Interest. Conflicts that the editors consider relevant to the con- apy. In addition, focusing only on the 4 most common Gram- tent of the manuscript have been disclosed. negative pathogens underestimates the total disease burden caused by CR infections. References Second, as microbiologic data were not collected from all 1. Kim YJ, Jun YH, Kim YR, et  al. Risk factors for mortality in patients with Pseudomonas aeruginosa bacteremia; retrospective study of impact of combin- hospitals in the Premier Network (teaching hospitals were more ation antimicrobial therapy. BMC Infect Dis 2014; 14:161. likely to provide microbiology data), there may be a reporting 2. Magiorakos AP, Suetens C, Monnet DL, et al. The rise of carbapenem resistance in Europe: just the tip of the iceberg? Antimicrob Resist Infect Control 2013; 2:6. bias favoring a more complex patient population. 3. Ny P, Nieberg P, Wong-Beringer A. Impact of carbapenem resistance on epidemi- Due to the privacy concern by Premier, participating hospi- ology and outcomes of nonbacteremic Klebsiella pneumoniae infections. Am J tals can only be identified by US census region. Therefore, it Infect Control 2015; 43:1076–80. 6 • OFID • Cai et al 4. Xu A, Zheng B, Xu YC, et al. National epidemiology of carbapenem-resistant and 15. Lesho E, Chukwuma U, Sparks M, et  al. Anatomic, geographic, and taxon-spe- extensively drug-resistant Gram-negative bacteria isolated from blood samples in cific relative risks of carbapenem resistance in the health care system of the U.S. China in 2013. Clin Microbiol Infect 2016; 22(Suppl 1):S1–8. Department of Defense. J Clin Microbiol 2016; 54:1546–51. 5. US Department of Health and Human Services, Centers for Disease Control and 16. Zilberberg MD, Kollef MH, Shorr AF. Secular trends in Acinetobacter baumannii Prevention. Antibiotic resistance threats in the United States, 2013. Available resistance in respiratory and blood stream specimens in the United States, 2003 to at: https://www.cdc.gov/drugresistance/pdf/ar-threats-2013–508.pdf. Accessed 2012: a survey study. J Hosp Med 2016; 11:21–6. January 10, 2017. 17. World Health Organization. Global priority list of antibiotic-resistant bacteria 6. European Centre for Disease Prevention and Control. Evidence brief: update on to guide research, discovery, and development of new antibiotics. Published the spread of carbapenemase-producing Enterobacteriaceae in Europe – sum- February 27, 2017. Available at: http://www.who.int/medicines/publications/ mary of the May 2015 expert assessment. Available at: http://ecdc.europa.eu/ global-priority-list-antibiotic-resistant-bacteria/en/. Accessed June 16, 2017. en/eaad/antibiotics-news/Documents/antimicrobial-resistance-EuSCAPE-Evi- 18. Vincent JL, Rello J, Marshall J, et  al. International study of the prevalence and dence-Brief.pdf. Accessed January 31, 2017. outcomes of infection in intensive care units. JAMA 2009; 302:2323–9. 7. European Centre for Disease Prevention and Control. Antimicrobial resistance 19. Micek ST, Wunderink RG, Kollef MH, et  al. An international multicenter ret- surveillance in Europe 2015. Annual Report of the European Antimicrobial rospective study of Pseudomonas aeruginosa nosocomial pneumonia: impact of Resistance Surveillance Network (EARS-Net). Stockholm: ECDC; 2017. Available multidrug resistance. Crit Care 2015; 19:219. at: http://ecdc.europa.eu/en/publications/Publications/antimicrobial-resist- 20. Premier Inc. Premier Healthcare Database: data that informs and performs. ance-europe-2015.pdf. Accessed January 31, 2017. Published January 31, 2017. Available at: https://learn.premierinc.com/pharma- 8. Sisto A, D’Ancona F, Meledandri M, et al. Carbapenem non-susceptible Klebsiella cy-and-research/premier-healthcare-database-whitepaper. Accessed June 16, pneumoniae from Micronet network hospitals, Italy, 2009 to 2012. Euro Surveill 2017. 2012; 17: pii: 20247. 21. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying 9. Sievert DM, Ricks P, Edwards JR, et  al. Antimicrobial-resistant patho- prognostic comorbidity in longitudinal studies: development and validation. J gens associated with healthcare-associated infections: summary of data r Chronic Dis 1987; 40:373–83. eported to the National Healthcare Safety Network at the Centers for Disease 22. US Centers for Disease Control and Prevention, National Center for Emerging Control and Prevention, 2009-2010. Infect Control Hosp Epidemiol 2013; and Zoonotic Infectious Disease, Division of Healthcare Quality Promotion. 34:1–14. Antibiotic Resistance Patient Safety Atlas summary of results. Available at: 10. US Centers for Disease Control and Prevention, National Center for Emerging https://www.cdc.gov/hai/pdfs/patient-safety-atlas/ar-patient-safety-atlas-sum- and Zoonotic Infectious Disease, Division of Healthcare Quality Promotion. mary-of-results.pdf. Accessed January 10, 2017. Facility guidance for control of carbapenem-resistant Enterobacteriaceae (CRE) – 23. Huijskens EG, Rossen JW, Kluytmans JA, et  al. Evaluation of yield of currently November 2015 update CRE Toolkit. Available at: https://www.cdc.gov/hai/pdfs/ available diagnostics by sample type to optimize detection of respiratory patho- cre/cre-guidance-508.pdf. Accessed January 10, 2017. gens in patients with a community-acquired pneumonia. Influenza Other Respir 11. Livermore DM. Current epidemiology and growing resistance of gram-negative Viruses 2014; 8:243–9. pathogens. Korean J Intern Med 2012; 27:128–42. 24. Kleinfauf N, Hausemann A, Kempf VAJ, Gottschalk R, Heudorf U. Burden of 12. European Centre for Disease Prevention and Control. Carbapenem-resistant carpabenem-resistant organisms in the Frankfurt/Main Metropolitan Area in Acinetobacter baumannii in healthcare settings – 8 December 2016. Stockholm: Germany 2012/2013  – first results and experiences after the introduction of ECDC; 2016. Available at: http://ecdc.europa.eu/en/publications/Publications/8- legally mandated reporting. BMC Infect Dis 2014; 12:446. Dec-2016-RRA-Acinetobacter%20baumannii-Europe.pdf. Accessed January 31, 25. Freeman R, Moore LS, Charlett A, et al. Exploring the epidemiology of carbape- 2017. nem-resistant Gram-negative bacteria in west London and the utility of routinely 13. Gniadek TJ, Carroll KC, Simner PJ. Carbapenem-resistant non-glucose-ferment- collected hospital microbiology data. J Antimicrob Chemother 2015; 70:1212–8. ing gram-negative bacilli: the missing piece to the puzzle. J Clin Microbiol 2016; 26. Brooke JS. New strategies against Stenotrophomonas maltophilia: a serious world- 54:1700–10. wide intrinsically drug-resistant opportunistic pathogen. Expert Rev Anti Infect 14. Labarca JA, Salles MJ, Seas C, Guzmán-Blanco M. Carbapenem resistance in Ther 2014; 12:1–4. Pseudomonas aeruginosa and Acinetobacter baumannii in the nosocomial setting 27. Sánchez MB. Antibiotic resistance in the opportunistic pathogen in Latin America. Crit Rev Microbiol 2016; 42:276–92. Stenotrophomonas maltophilia. Front Microbiol 2015; 6:658. Carbapenem-Resistant Gram-Negative Infections in the United States Predominated by A. baumannii & P. aeruginosa • OFID • 7 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Open Forum Infectious Diseases Oxford University Press

Prevalence of Carbapenem-Resistant Gram-Negative Infections in the United States Predominated by Acinetobacter baumannii and Pseudomonas aeruginosa

Loading next page...
 
/lp/oxford-university-press/prevalence-of-carbapenem-resistant-gram-negative-infections-in-the-fpeohJ1Ygn

References (55)

Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press on behalf of Infectious Diseases Society of America.
eISSN
2328-8957
DOI
10.1093/ofid/ofx176
Publisher site
See Article on Publisher Site

Abstract

Open Forum Infectious Diseases MAJOR ARTICLE Prevalence of Carbapenem-Resistant Gram-Negative Infections in the United States Predominated by Acinetobacter baumannii and Pseudomonas aeruginosa 1 2 3 1,* 1 4 4 4 Bin Cai, Roger Echols, Glenn Magee, Juan Camilo Arjona Ferreira, Gareth Morgan, Mari Ariyasu, Takuko Sawada, and Tsutae “Den” Nagata 1 2 3 4 Shionogi Inc., Florham Park, New Jersey; ID3C, Easton, Connecticut; Premier Research Service, Inc., Charlotte, North Carolina; Shionogi & Co., Ltd., Kitaku, Osaka, Japan Background. Carbapenem-resistant (CR) Gram-negative pathogens are recognized as a major health concern. This study exam- ined the prevalence of infections due to 4 CR Gram-negative species (Acinetobacter baumannii, Pseudomonas aeruginosa, Klebsiella pneumoniae, and Escherichia coli) in the United States and assessed their impact on hospital stays and mortality. Methods. Hospitalized patients with laboratory-confirmed infection due to any of the 4 Gram-negative pathogens were iden- tified from the Premier Healthcare Database. Proportions of CR were calculated by pathogen and infection site (blood, respiratory, urinary, or other) for the United States as whole and by census regions. Crude and adjusted odds ratios for in-hospital mortality were produced using logistic regression. Results. From 2009 to 2013, 13 262 (4.5%) of 292 742 infections due to these 4 Gram-negative pathogens were CR. Of these CR infections, 82.3% were caused by A. baumannii (22%) or P. aeruginosa (60.3%), while 17.7% were caused by K. pneumoniae or E. coli. CR patients had longer hospital stays than carbapenem-susceptible (CS) patients in all pathogen-infection site cohorts, except in the A. baumannii-respiratory cohort. The crude all cause in-hospital mortality was greater for most pathogen-infection site cohorts of the CR group compared with the CS group, especially for A. baumannii infection in the blood (crude odds ratio [OR], 3.91; 95% confidence interval [CI], 2.69–5.70). This difference for the A. baumannii-blood cohort remained aer ad ft justing for the relevant covariates (adjusted OR, 2.46; 95% CI, 1.43–4.22). Conclusion. e Th majority of CR infections and disease burden in the United States was caused by nonfermenters A. baumannii and P. aeruginosa. Patients with CR infections had longer hospital stays and higher crude in-hospital mortality. Keywords. carbapenem resistance; Gram-negative; nonfermenters; CRE; length of stay; in-hospital mortality. Infections due to carbapenem-resistant (CR) Gram-negative it was not well appreciated as a source of resistance transmission pathogens have been reported from many countries with var- until 2017 [7, 12–16], when the World Health Organization rec- iable prevalence and associated morbidity and mortality [1–4]. ognized the importance of CR A. baumannii and P. aeruginosa as e dra Th matic increase and spread of CR infections over the past equal to that of the CREs [17]. These CR nonfermenters are oen ft decade are recognized as a major public health concern [2, 5–9]. multidrug resistant and associated with substantial morbidity and Recent infection control measures have focused on the identifica- mortality [18, 19]. This study was designed to determine the prev- tion and spread of CR Enterobacteriaceae (CREs), which is mainly alence of CR organisms (CROs) in the United States, specifically due to mobile carbapenemase enzymes, both serine (KPC, GES, the Enterobacteriaceae Escherichia coli and Klebsiella pneumoniae and OXA) and metallo-carbapenemases (NDM, VIM, and IMP) and the nonfermenters Pseudomonas aeruginosa and Acinetobacter [6, 10]. Resistance mechanisms involving porin channel mutations baumannii, using an electronic database from a geographically and efflux pump overproduction may also contribute to the iden- diverse network of US hospitals. Besides determining the prev- tification of CREs [6 , 10, 11]. The frequency of carbapenem resist- alence of CR infections in the United States, the study compared ance among the non-glucose fermenting Pseudomonas aeruginosa similar infections that were carbapenem susceptible (CS) in order and Acinetobacter baumannii has also increased in recent years, but to assess the mortality attributed to the CROs. METHODS Received 18 May 2017; editorial decision 8 August 2017; accepted 9 August 2017. This retrospective study of CR Gram-negative infections *Employee at the time of the study between 2009 and 2013 was conducted using microbiology Correspondence: B. Cai, MD, PhD, Global Observational Research and Epidemiology, Shionogi Inc., Florham Park, New Jersey, 07932, USA. (bin.cai@shionogi.com). data linked with patient-level in-hospital discharge data from Open Forum Infectious Diseases the Premier Healthcare Database [20]. The Premier Healthcare © The Author 2017. Published by Oxford University Press on behalf of Infectious Diseases Database is an anonymous census of inpatients and hospi- Society of America. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits tal-clinic outpatients from geographically diverse, mixed unrestricted reuse, distribution, and reproduction in any medium, provided the original work teaching and nonteaching hospitals with varied bed sizes in the is properly cited. United States. The hospital discharge data contain information DOI: 10.1093/ofid/ofx176 Carbapenem-Resistant Gram-Negative Infections in the United States Predominated by A. baumannii & P. aeruginosa • OFID • 1 on admission, discharge, and a date-stamped log of all billed ertapenem, imipenem, meropenem) from patients who received items, including procedures, medications, laboratory, diagnos- at least 1 systemic antibiotic treatment were selected. tic, and therapeutic services at the individual patient level. Statistical Analyses Microbiology data, collected from approximately 30% of Categorical variables were described with the number of participating hospitals, include the time the microbiological patients and percentages and compared between CR and CS culture was obtained, type of pathogen identified, and anti- in each pathogen-infection site cohort using a chi-square or microbial drugs used for susceptibility testing along with the Fisher’s exact test where appropriate. Continuous variables method, result, and interpretation of antibiotic susceptibility. were described with mean, standard deviation, and median. All data were determined by individual hospital laboratories, The differences between CR and CS groups were analyzed using which were Clinical Laboratory Improvement Amendments Student’s t test or the Wilcoxon rank-sum test. A P-value <.05 (CLIA) compliant using standard laboratory methods, and indicated statistically significant differences between groups. interpretations of susceptibility results were reported as resist- In-hospital mortality was compared between CR and CS for ant (R), intermediate (I), susceptible (S), or none (N). each pathogen infection site cohort with and without adjusting e p Th athogens were further identified by infection sites for covariates, such as age, gender, race, ethnicity, mechanical (blood, respiratory, urinary, or other) based on the source of the ventilation, renal impairment, geographic region, and each of culture sample. u Th s, this study had 16 distinct pathogen-in- the Charlson comorbid conditions [21]. fection site cohorts. Within each cohort, pathogens were char- All analyses were performed using SAS 9.3 (SAS Institute acterized as CR or CS. Bacteria were defined as CR if they were Inc., Cary, North Carolina). resistant to at least 1 of the carbapenems tested and as CS if the pathogen was susceptible to or intermediate for the carbap- RESULTS enems tested. If selected patients had multiple hospitalizations This study includes data collected from 206 acute care hospitals during the study period, only data from the first hospitaliza- in the United States between January 1, 2009, and December tion were used. If a microbiologic culture identified multiple 31, 2013. The total numbers of pathogen isolates were 173 200 pathogens, the patient was counted in each pathogen group. If E.  coli, 56 552 K.  pneumoniae, 56 477 P.  aeruginosa, and 6508 a patient had multiple cultures of the same pathogen, the first A. baumannii. culture was used as index date if the pathogen had the same Carbapenem Resistance Rate susceptibility test results for all cultures. Otherwise, the first CR The overall CR rate among the 4 selected pathogens was 4.5% culture was used as the index date, and the patient was in the CR (13 262 of 292 742). A.  baumannii had the highest CR rate: group. Only cultured pathogens from patients who were receiv-     blood 40.1%, respiratory 50.4%, urine 42.0%, and other 42%. ing antibiotic treatment were included in the final data set. The CR rate for P.  aeruginosa ranked second highest and was e p Th rimary outcome variables were the rates of CR in each more variable by infection site: 10.3% in blood, 19.4% in res- pathogen-infection site cohort for all hospitals and separately piratory, 11.7% in urinary, and 12.6% in other (Figure 1). More for each US census region. Other outcome variables included than 80% of all CR infections were caused by A. baumannii or in-hospital mortality, total length of stay (LOS) in the hospital, P.  aeruginosa. This was consistent across 9 US census regions, LOS aer p ft ositive culture (iLOS), total LOS in the intensive care except for the Middle Atlantic region, where 67% (Table  1) of unit (ICU) for the CR group and CS group, and survival status within each pathogen-infection site cohort. e Th use of antibiotics with Gram-negative activity was Carbapenem Resistantce Rate assessed by class (aminoglycoside, carbapenem, cephalosporin, colistin, monobactam, penicillin, polymyxin B, quinolone, tetra- 50.4% 50.4% cycline, and other Gram-negative antibiotics) and the time when 42.0% 42.0% 40.1% 40.1% the patient received them during the hospitalization. Antibiotics 60.0% administered in the days prior to the date of the index culture 50.0% were considered “prior” antibiotics, those received from the date 40.0% 19.4% 19.4% of the index culture through 3 days aer t ft he index culture were 12.6% 12.6% 11.7% 11.7% considered “empiric,” and those received aer 3  ft days from the 10.3% 10.3% 30.0% A. baumannii index culture were considered “definitive” treatment. 20.0% 4.9% 4.9% 3.6% 3.6% 4.0% 4.0% 3.1% 3.1% P. aeruginosa Study Population 10.0% K. pneumoniae 0.2% 0.2% 0.4% 0.4% 0.2% 0.2% 0.3% 0.3% E. coli All hospitalized patients with microbiological cultures identify- 0.0% Blood Respiratory Urinary Others ing A. baumannii, P. aeruginosa, K. pneumoniae, or E.  coli from appropriate clinical specimens and with associated interpreta- Figure 1. Carbapenem resistance rate by pathogen-infection site cohort. tions of susceptibility to carbapenem antibiotics (doripenem, 2 • OFID • Cai et al CR infections were caused by the nonfermenters. For most regions, CR P.  aeruginosa outnumbered CR A.  baumannii by 3:1, except for the New England region, where 85% of CR infec- tions were due to P. aeruginosa and only 6.4% were from A. bau- mannii (Table 2). Demographic and Clinical Characteristics Compared with patients with CS A. baumannii, patients with CR isolates tended to be older when the source was blood stream infection (mean±SD, 60.5 ± 16.6 years vs 53.6 ± 21.8 years), respiratory infection (mean±SD, 62.5 ± 16.8 vs 55.8 ± 23.5 years), or other infection sites (mean±SD, 60.7 ± 16.5 vs 57.0 ± 20.0 years), but of similar age when the source was urinary tract infection (mean±SD, 61.8 ± 17.5 vs 61.5 ± 19.8 years). More than two-thirds of patients with A. baumannii infections were white; however, the proportions of nonwhite patients in the CR groups were higher than in the CS groups for all infection sites. Slightly more male patients than female patients were identified in both the CR and CS groups. Nearly half of all patients with A. baumannii infections came to the hospitals from their homes, though the proportion was slightly higher in the CS group than in the CR group. Nearly 70% of these patients were admitted via emergency rooms. Patients with CR infections were more likely to be discharged to skilled nursing facilities or rehabilitation or long-term care facilities than patients with CS infections (43% vs 24% for blood, 51% vs 42% for respiratory, 58% vs 33% for urinary, and 51% vs 31% for other). Patients with CR infections also had a higher Charlson comorbidity index score for all infection sites, a longer total LOS for infections in the blood, urinary, and other (median days of CR vs CS, 15 vs 9 for blood, 12 vs 7 for urinary, and 12 vs 8 for other), and longer total ICU stays for blood, urinary, and other (median days of CR vs CS, 8 vs 5, 8 vs 4, and 9 vs 8, respectively). Similarly, a longer LOS after positive culture was seen in the CR group than the CS group among all patients and those who survived for each infection site except respiratory (Table 3). Among nonsurvivors with blood stream infection, the median LOS after positive cul- ture was 2 days after the culture was obtained in the CR group and 4 days in the CS group. Patients with CR infections were slightly more likely to be on dialysis or have renal impairment during their hospital stays. Similar patterns in admission sources, discharge status, hemodialysis, and renal impairment were observed for patients infected with CR P.  aeruginosa, CR K.  pneumoniae, and CR E. coli compared with their CS counterparts. However, patients with CR P. aeruginosa isolates were younger than patients with CS isolates for all infection sites (mean ± SD, 60.6  ±  17.3 vs 65.2  ±  19.1 for blood, 60.6  ±  19.1 vs 63.2  ±  20.3 for respira- tory, 67.1 ± 17.1 vs 71.5 ± 17.1 for urinary, and 60.6 ± 18.5 vs 61.1  ±  20.2 for other), but they had similar Charlson comor- bidity index scores. LOS, ICU LOS, and LOS aer ft positive Carbapenem-Resistant Gram-Negative Infections in the United States Predominated by A. baumannii & P. aeruginosa • OFID • 3 Table 1. Carbapenem-resistant infections caused by A. baumannii, P. aeruginosa, K. pneumoniae, and E. coli by US census region All 4 Pathogens A. baumannii + P. aeruginosa K. pneumoniae + E. coli CR Cases CR Cases CR Cases Number of Total Total CR % of Total Total % of Total % of Total CR Cases for 4 Total % of Total % of Total CR Cases for 4 US Census Regions Hospitals Infections Cases Infections Infections CR Cases Infections Pathogens Infections CR Cases Infections Pathogens East North Central 43 47 702 3164 6.6 10 381 2542 24.5 80.3 37 321 622 1.7 19.7       East South Central 9 21 440 770 3.6 4293 703 16.4 91.3 17 147 67 0.4 8.7     Middle Atlantic 19 37 582 2499 6.6 10 449 1672 16.0 66.9 27 133 827 3.0 33.1       Mountain 15 8582 765 8.9 2260 710 31.4 92.8 6322 55 0.9 7.2 New England 9 11 964 346 2.9 2417 316 13.1 91.3 9547 30 0.3 8.7 Pacific 27 39 045 1237 3.2 6447 1070 16.6 86.5 32 598 167 0.5 13.5     South Atlantic 40 62 217 2385 3.8 14 652 2017 13.8 84.6 47 565 368 0.8 15.4       West North Central 18 17 854 403 2.3 3130 329 10.5 81.6 14 724 74 0.5 18.4     West South Central 26 46 356 1693 3.7 8955 1550 17.3 91.6 37 401 143 0.4 8.4     Total 206 292 742 13 262 4.5 62 984 10 909 17.3 82.3 229 758 2353 1.0 17.7         Abbreviations: CR, carbapenem resistant. culture were longer for CR infections than CS infections for all infection sites. The average age of patients with K. pneumoniae was approximately 60 years. There was no significant age differ- ence between patients with CR and CS infections in blood, urine, or respiratory. In other infection sites, CR patho- gen–infected patients were approximately 3 years older than those in the CS group. er Th e were more male patients than female patients in all study cohorts, except for those in the urinary cohort, where significantly higher numbers of female patients were identified. Due to the small number of patients in the CR-urinary cohort, there was no statistical significance in the observed differences in age (average of 60 years), race, or Charlson comorbidity index score. In-Hospital Mortality Patients with CR infections consistently had numerically higher crude mortality rates than patients with CS infections for each of the 16 pathogen-infection site cohorts. The crude odds ratio (OR) for CR vs CS ranged from 1.31 for E. coli in respiratory to 3.91 for A. baumannii in blood (Table 4). After adjusting for age, gender, race, ethnicity, Charlson comorbid conditions, mechanical ventilation, renal impairment, and geographic cen- sus regions, the odds of dying from CR A. baumannii in blood (adjusted OR, 2.46) and in respiratory (adjusted OR, 1.27) and from P. aeruginosa in other (adjusted OR, 1.20) remained statis- tically significant (Table 4). Patients with CR pathogens were more likely to receive more than 1 systemic antibiotic than patients with CS pathogens. Also, patients with pathogens identified from the blood, res- piratory, and other were more likely to receive 2 or more anti- biotics than if the source was urinary. The number and types of antibiotics used were highly variable. During the definitive treatment period, more than 100 combinations (defined as dif- ferent antibiotic classes regardless of the initiation time, dose, or duration) were used for each pathogen, but fewer than 20 of these combinations were used by more than 3% of patients. Fewer different antibiotic combinations were used to treat patients with CS infections than patients with CR infections. DISCUSSION This study showed that in the United States the overall disease burden and impact on survival was greatest among the non-glu- cose fermenting A. baumannii and P. aeruginosa. A. baumannii had the highest rate of CR at each infection site, followed by P. aeruginosa and K. pneumoniae. E. coli consistently had the lowest rate of CR. The combined number of CR A. bauman- nii and CR P. aeruginosa infections accounted for more than 80% of all CR infections. Though this study included both com- munity-onset and hospital-acquired infections, the pattern of CR from the selected 4 pathogens was comparable with the US 4 • OFID • Cai et al Table 2. A. baumannii and P. aeruginosa CR Infection by Region All 4 Pathogens Acinetobacter baumannii Pseudomonas aeruginosa CR Cases CR Cases CR Cases Number of % of Total % of Total % of Total CR Cases for 4 % of Total % of Total CR Cases for 4 Regions Hospitals Total Infections Cases Infections Total Infections Cases Infections Pathogens Total Infections Cases Infections Pathogens East North Central 43 47 702 3164 6.6 1360 863 63.5 27.3 9021 1679 18.6 53.1 East South Central 9 21 440 770 3.6 498 223 44.8 29.0 3795 480 12.6 62.3 Middle Atlantic 19 37 582 2499 6.6 914 288 31.5 11.5 9535 1384 14.5 55.4 Mountain 15 8582 765 8.9 326 203 62.3 26.5 1934 507 26.2 66.3 New England 9 11 964 346 2.9 183 22 12.0 6.4 2234 294 13.2 85.0 Pacific 27 39 045 1237 3.2 739 327 44.2 26.4 5708 743 13.0 60.1 South Atlantic 40 62 217 2385 3.8 1322 488 36.9 20.5 13 330 1529 11.5 64.1     West North Central 18 17 854 403 2.3 256 77 30.1 19.1 2874 252 8.8 62.5 West South Central 26 46 356 1693 3.7 909 424 46.6 25.0 8046 1126 14.0 66.5 Total 206 292 742 13 262 4.5 6507 2915 44.8 22.0 56 477 7994 14.2 60.3     Abbreviation: CR, carbapenem resistant. Table 3. Median Days Stayed in the Hospital After Positive Culture by Pathogen, Infection Site, and Survival Status Blood Respiratory Urinary Other Pathogen Length of Stay After Positive Culture CR CS CR CS CR CS CR CS Acinetobacter baumannii iLOS 8 6 8 10 8 4 7 5 iLOS among survived 11 6 9 10 8 4 7 5 iLOS among deceased 2 4 6 7 6 6 7 7 Pseudomonas aeruginosa iLOS 5 4 20 7 5 3 8 5 iLOS among survived 6 4 20 8 5 3 6 5 iLOS among deceased 3 2 18.5 4 7 4 33.5 6 Klebsiella pneumoniae iLOS 10 6 10 7 6 4 9 6 iLOS among survived 10 6 11 8 6 4 9 6 iLOS among deceased 6 3 6 4 8 4 9 6 Escherichia coli iLOS 10 6 10 6 7 4 10 5 iLOS among survived 10 7 10 6 6 4 9 5 iLOS among deceased 9 1 11 5 9 5 14 6 Abbreviations: CR, carbapenem resistant; CS, carbapenem susceptible; iLOS, length of hospital stay after positive culture. Centers for Disease Control and Prevention’s (CDC’s) National CR infection for K.  pneumoniae (48.7%) and E.  coli (57.3%). Healthcare Safety Network (NHSN) results, which focused on Differences in the frequency of CR infection by infection site health care–acquired infections via central lines, urinary cathe- have also been reported in other studies [9, 23, 24]. ters, and mechanical ventilators [9, 22]. The relative importance e CR ra Th tes varied by the geographic regions. The Middle of CR A. baumannii has also been recently emphasized by the Atlantic region had the highest proportion of CR infections European Centre for Disease Prevention and Control (ECDC), due to K.  pneumoniae and E.  coli (33.1%), while the New which reported that CR A. baumannii (resistant through the England region had a higher proportion of CR infections due production of carbapenemases) might be more widely dissemi- to P.  aeruginosa (85%). Other studies also observed the varia- nated in Europe than CREs [7, 12]. tions in frequency of CROs by hospitals and by region [22, 24, This study also showed that distribution of CR pathogens 25]. Although actual incidence cannot be estimated from this differed by infection sites. Respiratory was the most common study, the regional differences observed in this study support site of CR infection for A.  baumannii (40.3%) and P.  aerug- the understanding that infection control of resistant pathogens inosa (40.1%), while urinary was the most common site of needs to be based on local epidemiology. Table 4. In-Hospital Mortality: Crude and Adjusted Odds Ratio (CR vs CS) In-Hospital Mortality by Each Site, No. (%) Crude Adjusted Pathogen Site CR CS Odds Ratio Lower 95% CI Upper 95% CI Odds Ratio Lower 95% CI Upper 95% CI Acinetobacter Blood 103 (38.2) 55 (13.6) 3.91 2.69 5.70 2.46 1.43 4.22 baumannii Respiratory 305 (26.0) 230 (19.9) 1.41 1.16 1.71 1.27 1.01 1.58 Urinary 35 (9.6) 31 (6.2) 1.62 0.98 2.69 0.62 0.32 1.19 Other 173 (15.6) 117 (7.7) 2.24 1.74 2.87 1.19 0.88 1.61 Pseudomonas Blood 112 (33.3) 590 (20.1) 2.02 1.58 2.58 1.19 0.88 1.63 aeruginosa Respiratory 661 (20.6) 1977 (14.8) 1.50 1.36 1.65 1.11 1.00 1.24 Urinary 208 (9.8) 1031 (6.3) 1.58 1.35 1.85 1.05 0.88 1.26 Other 389 (16.8) 1139 (7.1) 2.64 2.33 2.99 1.20 1.03 1.39 Klebsiella Blood 63 (27.3) 819 (13.2) 2.46 1.83 3.32 1.10 0.75 1.61 pneumoniae Respiratory 98 (25.4) 1547 (20.6) 1.32 1.04 1.67 0.97 0.76 1.25 Urinary 90 (9.1) 1616 (5.2) 1.81 1.45 2.27 1.07 0.83 1.37 Other 76 (18.1) 863 (8.7) 2.33 1.80 3.01 1.02 0.76 1.37 Escherichia coli Blood 5 (16.1) 1709 (8.7) 2.03 0.78 5.29 1.91 0.60 6.10 Respiratory 8 (26.7) 1469 (21.7) 1.31 0.58 2.95 0.74 0.29 1.85 Urinary 13 (6.9) 4699 (3.9) 1.84 1.05 3.23 1.55 0.80 3.00 Other 8 (10.1) 1466 (5.8) 1.84 0.89 3.84 0.78 0.33 1.84 Abbreviations: CI, confidence interval; CR, carbapenem resistant; CS, carbapenem susceptible. Adjusted for age, gender, race, ethnicity, various comorbid conditions, mechanical ventilation, renal impairment, and geographic regions. Carbapenem-Resistant Gram-Negative Infections in the United States Predominated by A. baumannii & P. aeruginosa • OFID • 5 As shown in this study, a majority of the CR pathogens were was not possible to estimate CR infection rates in smaller geo- treated with multiple antibiotics, especially for A.  baumannii graphic areas, even though infections could vary from city to and P.  aeruginosa or for infections in the blood or respiratory city and hospital to hospital. This study does identify regional infection sites. Even with combination treatment, patients with differences in prevalence rates; however, it does not assess CR infections still had longer total hospital (total and aer p ft os- whether these findings are driven by individual institutions or itive culture) and ICU stays, except among nonsurvivors with episodic outbreaks within institutions. A. baumannii in blood and respiratory, where CR patients had This study only examined 2 Enterobacteriaceae, E.  coli and shorter LOS aer p ft ositive culture compared with CS patients K. pneumoniae, and 2 nonfermenters, A. baumannii and P. aerug- and the median for CR in blood was 2 days aer t ft he culture was inosa. Other nonfermenters, for example, Stenotrophomas obtained. This further highlights the treatment challenges faced maltophilia, which are highly carbapenem resistant [26, 27], by clinicians today and the need for more effective antibiotics may represent another significant source of disease caused by and more rapid information on antibiotic susceptibility. CR pathogens. A remarkably high number of different antibiotic combi- CONCLUSION nations were used to treat these infections, especially for CR pathogens. This indicates that there is no generally accepted This study showed that CR non-glucose fermenting Gram- standard of care regimen for CR infections, which underlies the negative bacteria, such as A. baumannii and P. aeruginosa, fact that there are few available antibiotics to treat CR infec- contribute to the greater disease and mortality burden when tions, and those that may be available are associated with sig- compared with CREs, such as K. pneumoniae and E. coli. This nificant toxicity. has been observed for the study as whole and for each census e a Th uthors believe that the key strength of this study is the region. These nonfermenters contributed more than 80% of large, geographically diverse data set that includes both detailed CR cases in our analysis of data over a period of 5 years. CR microbiologic data as well as patient-level information and rates differed from region to region, but the predominance of outcomes. Unlike the CDC and other data sets that focus on CR nonfermenters over CREs was consistent. The geographic hospital-acquired infections, the Premier Healthcare database variability was greater among K. pneumoniae and E. coli than includes both hospital-acquired and community-onset infec- the nonfermenters. However, because the authors cannot deter- tions, many of which are healthcare related, thus better repre- mine specific population numbers by hospital, the incidence senting the full disease burden related to CR infections in US of CR infections cannot be estimated or compared. These data hospitals. for the United States are in sharp contrast to most reports from This study selected subjects based on microbiology results, Europe and Latin America, where CREs are predominant. not based on the discharge diagnosis, avoiding potential biases Because patients with CR infections had longer hospital and in the discharge diagnoses due to various reasons, including ICU stays, more antibiotic treatments, and a higher mortality coding errors or reporting bias. The results of the microbiol- rate, the CR infections likely represent a higher cost burden ogy and drug susceptibility testing, as well as the documented to the health care system. The development of new antibi- antibiotic treatments, provide a clear picture of the pattern of otics should not only address CREs, but more importantly, it infections in acute care hospitals. should address CR non-fermenters such as P. aeruginosa and However, there are also several limitations in this study. First, A. baumannii. case definitions were based on microbiological culture results Acknowledgment that may underestimate or overestimate the number of infec- Potential coni fl cts of interest. Dr. Cai, Dr. Morgan, Dr. Arjona Ferreira, tions. Although some culture results may represent coloniza- Ms. Ariyasu, Ms. Sawada, and Dr. Nagata are full-time employees of tion rather than true infection, this study only analyzed patients Shionogi. Dr. Echols reports personal fees from Shionogi, Inc., during the for whom the cultures had susceptibility testing performed and conduct of the study. All authors have submitted the ICMJE Form for Disclosure of Potential where their physician prescribed systemic antimicrobial ther- Conflicts of Interest. Conflicts that the editors consider relevant to the con- apy. In addition, focusing only on the 4 most common Gram- tent of the manuscript have been disclosed. negative pathogens underestimates the total disease burden caused by CR infections. References Second, as microbiologic data were not collected from all 1. Kim YJ, Jun YH, Kim YR, et  al. Risk factors for mortality in patients with Pseudomonas aeruginosa bacteremia; retrospective study of impact of combin- hospitals in the Premier Network (teaching hospitals were more ation antimicrobial therapy. BMC Infect Dis 2014; 14:161. likely to provide microbiology data), there may be a reporting 2. Magiorakos AP, Suetens C, Monnet DL, et al. The rise of carbapenem resistance in Europe: just the tip of the iceberg? Antimicrob Resist Infect Control 2013; 2:6. bias favoring a more complex patient population. 3. Ny P, Nieberg P, Wong-Beringer A. Impact of carbapenem resistance on epidemi- Due to the privacy concern by Premier, participating hospi- ology and outcomes of nonbacteremic Klebsiella pneumoniae infections. Am J tals can only be identified by US census region. Therefore, it Infect Control 2015; 43:1076–80. 6 • OFID • Cai et al 4. Xu A, Zheng B, Xu YC, et al. National epidemiology of carbapenem-resistant and 15. Lesho E, Chukwuma U, Sparks M, et  al. Anatomic, geographic, and taxon-spe- extensively drug-resistant Gram-negative bacteria isolated from blood samples in cific relative risks of carbapenem resistance in the health care system of the U.S. China in 2013. Clin Microbiol Infect 2016; 22(Suppl 1):S1–8. Department of Defense. J Clin Microbiol 2016; 54:1546–51. 5. US Department of Health and Human Services, Centers for Disease Control and 16. Zilberberg MD, Kollef MH, Shorr AF. Secular trends in Acinetobacter baumannii Prevention. Antibiotic resistance threats in the United States, 2013. Available resistance in respiratory and blood stream specimens in the United States, 2003 to at: https://www.cdc.gov/drugresistance/pdf/ar-threats-2013–508.pdf. Accessed 2012: a survey study. J Hosp Med 2016; 11:21–6. January 10, 2017. 17. World Health Organization. Global priority list of antibiotic-resistant bacteria 6. European Centre for Disease Prevention and Control. Evidence brief: update on to guide research, discovery, and development of new antibiotics. Published the spread of carbapenemase-producing Enterobacteriaceae in Europe – sum- February 27, 2017. Available at: http://www.who.int/medicines/publications/ mary of the May 2015 expert assessment. Available at: http://ecdc.europa.eu/ global-priority-list-antibiotic-resistant-bacteria/en/. Accessed June 16, 2017. en/eaad/antibiotics-news/Documents/antimicrobial-resistance-EuSCAPE-Evi- 18. Vincent JL, Rello J, Marshall J, et  al. International study of the prevalence and dence-Brief.pdf. Accessed January 31, 2017. outcomes of infection in intensive care units. JAMA 2009; 302:2323–9. 7. European Centre for Disease Prevention and Control. Antimicrobial resistance 19. Micek ST, Wunderink RG, Kollef MH, et  al. An international multicenter ret- surveillance in Europe 2015. Annual Report of the European Antimicrobial rospective study of Pseudomonas aeruginosa nosocomial pneumonia: impact of Resistance Surveillance Network (EARS-Net). Stockholm: ECDC; 2017. Available multidrug resistance. Crit Care 2015; 19:219. at: http://ecdc.europa.eu/en/publications/Publications/antimicrobial-resist- 20. Premier Inc. Premier Healthcare Database: data that informs and performs. ance-europe-2015.pdf. Accessed January 31, 2017. Published January 31, 2017. Available at: https://learn.premierinc.com/pharma- 8. Sisto A, D’Ancona F, Meledandri M, et al. Carbapenem non-susceptible Klebsiella cy-and-research/premier-healthcare-database-whitepaper. Accessed June 16, pneumoniae from Micronet network hospitals, Italy, 2009 to 2012. Euro Surveill 2017. 2012; 17: pii: 20247. 21. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying 9. Sievert DM, Ricks P, Edwards JR, et  al. Antimicrobial-resistant patho- prognostic comorbidity in longitudinal studies: development and validation. J gens associated with healthcare-associated infections: summary of data r Chronic Dis 1987; 40:373–83. eported to the National Healthcare Safety Network at the Centers for Disease 22. US Centers for Disease Control and Prevention, National Center for Emerging Control and Prevention, 2009-2010. Infect Control Hosp Epidemiol 2013; and Zoonotic Infectious Disease, Division of Healthcare Quality Promotion. 34:1–14. Antibiotic Resistance Patient Safety Atlas summary of results. Available at: 10. US Centers for Disease Control and Prevention, National Center for Emerging https://www.cdc.gov/hai/pdfs/patient-safety-atlas/ar-patient-safety-atlas-sum- and Zoonotic Infectious Disease, Division of Healthcare Quality Promotion. mary-of-results.pdf. Accessed January 10, 2017. Facility guidance for control of carbapenem-resistant Enterobacteriaceae (CRE) – 23. Huijskens EG, Rossen JW, Kluytmans JA, et  al. Evaluation of yield of currently November 2015 update CRE Toolkit. Available at: https://www.cdc.gov/hai/pdfs/ available diagnostics by sample type to optimize detection of respiratory patho- cre/cre-guidance-508.pdf. Accessed January 10, 2017. gens in patients with a community-acquired pneumonia. Influenza Other Respir 11. Livermore DM. Current epidemiology and growing resistance of gram-negative Viruses 2014; 8:243–9. pathogens. Korean J Intern Med 2012; 27:128–42. 24. Kleinfauf N, Hausemann A, Kempf VAJ, Gottschalk R, Heudorf U. Burden of 12. European Centre for Disease Prevention and Control. Carbapenem-resistant carpabenem-resistant organisms in the Frankfurt/Main Metropolitan Area in Acinetobacter baumannii in healthcare settings – 8 December 2016. Stockholm: Germany 2012/2013  – first results and experiences after the introduction of ECDC; 2016. Available at: http://ecdc.europa.eu/en/publications/Publications/8- legally mandated reporting. BMC Infect Dis 2014; 12:446. Dec-2016-RRA-Acinetobacter%20baumannii-Europe.pdf. Accessed January 31, 25. Freeman R, Moore LS, Charlett A, et al. Exploring the epidemiology of carbape- 2017. nem-resistant Gram-negative bacteria in west London and the utility of routinely 13. Gniadek TJ, Carroll KC, Simner PJ. Carbapenem-resistant non-glucose-ferment- collected hospital microbiology data. J Antimicrob Chemother 2015; 70:1212–8. ing gram-negative bacilli: the missing piece to the puzzle. J Clin Microbiol 2016; 26. Brooke JS. New strategies against Stenotrophomonas maltophilia: a serious world- 54:1700–10. wide intrinsically drug-resistant opportunistic pathogen. Expert Rev Anti Infect 14. Labarca JA, Salles MJ, Seas C, Guzmán-Blanco M. Carbapenem resistance in Ther 2014; 12:1–4. Pseudomonas aeruginosa and Acinetobacter baumannii in the nosocomial setting 27. Sánchez MB. Antibiotic resistance in the opportunistic pathogen in Latin America. Crit Rev Microbiol 2016; 42:276–92. Stenotrophomonas maltophilia. Front Microbiol 2015; 6:658. Carbapenem-Resistant Gram-Negative Infections in the United States Predominated by A. baumannii & P. aeruginosa • OFID • 7

Journal

Open Forum Infectious DiseasesOxford University Press

Published: Aug 16, 2017

There are no references for this article.