Abstract
Trends of rising rates of resistance in Pseudomonas aeruginosa make selection of appropriate empirical therapy increasingly difficult, but whether multidrug-resistant (MDR) P. aeruginosa is associated with worse clinical outcomes is not well established. The objective of this study was to determine the impact of MDR (resistance to three or more classes of antipseudomonal agents) P. aeruginosa bacteremia on patient outcomes. We performed a retrospective cohort study of adult patients with P. aeruginosa bacteremia from 2005 to 2008. Patients were identified by the microbiology laboratory database, and pertinent clinical data were collected. Logistic regression was used to explore independent risk factors for 30-day mortality. Classification and regression tree analysis was used to determine threshold breakpoints for continuous variables. Kaplan-Meier survival analysis was used to compare time to mortality, after normalization of the patients' underlying risks by propensity scoring. A total of 109 bacteremia episodes were identified; 25 episodes (22.9%) were caused by MDR P. aeruginosa. Patients with MDR P. aeruginosa bacteremia were more likely to receive inappropriate empirical therapy (44.0% and 6.0%, respectively; P < 0.001) and had longer prior hospital stays (32.6 ± 37.3 and 14.4 ± 43.6 days, respectively; P = 0.046). Multivariate regression revealed that 30-day mortality was associated with multidrug resistance (odds ratio [OR], 6.8; 95% confidence interval [CI], 1.9 to 24.0), immunosuppression (OR, 5.0; 95% CI, 1.4 to 17.5), and an APACHE II score of ≥22 (OR, 29.0; 95% CI, 5.0 to 168.2). Time to mortality was also shorter in the MDR cohort (P = 0.011). Multidrug resistance is a significant risk factor for 30-day mortality in patients with P. aeruginosa bacteremia; efforts to curb the spread of MDR P. aeruginosa could be beneficial.
Bacterial bloodstream infections are serious medical events with life-threatening consequences and heavy impacts on health care costs through increased disease acuteness and lengths of hospital stay. Over the past decade, the numbers of bloodstream infections caused by Gram-negative bacteria have risen sharply. Among the various Gram-negative bacteria, Pseudomonas aeruginosa is known to be associated with resistance to practically all known antibiotics and is particularly problematic (3). In 2004, P. aeruginosa was placed seventh in prevalence among the causative pathogens in bloodstream infections but was second only to Candida species in bloodstream infection-related mortality. The crude mortality rate exceeded 38%, and the intensive care unit-related mortality was nearly 50% (20).
Patients with P. aeruginosa bacteremia are often medically complicated. Up to 90% of these patients have preexisting severe underlying diseases or conditions, such as malignancy, diabetes, renal failure, heart failure, cirrhosis, or other immune system-impairing disease states, or have undergone a solid organ transplant (19). Recent studies have indicated that the initial site of infection, surgery, and pneumonia are important independent predictors of mortality in patients with P. aeruginosa bacteremia, while neutropenia, the presence of a central line, or the urinary tract as the initial site of infection were not highly associated with mortality (10, 19). The rapid course of the infection, combined with a rising tendency toward resistance to multiple agents typically used in empirical treatment, makes P. aeruginosa bacteremia particularly difficult to treat successfully (10).
Paramount to the effective treatment of P. aeruginosa bacteremia is the obstacle of multidrug resistance. Recent studies demonstrated the importance of appropriate empirical antibiotic therapy in achieving favorable patient outcomes (10, 14). It can be reasoned that the probability of providing appropriate empirical therapy is lower with multidrug-resistant isolates. Furthermore, there may be only limited therapeutic options for multidrug-resistant isolates, and a less preferred (e.g., more toxic) agent(s) would have to be used as a last resort. It has been proposed that as bacteria gather mutations sequentially resulting in multidrug resistance, biofitness could be lost and the bacteria would become less virulent (11). However, there are also data to suggest that multidrug-resistant P. aeruginosa may remain fully pathogenic (7). The influence of multidrug resistance on the outcomes of patients with P. aeruginosa bloodstream infections remains controversial. The purpose of the present study was to examine the impact of multidrug resistance on outcomes in patients with P. aeruginosa bloodstream infections.
MATERIALS AND METHODS
Study sites.
The study was conducted at St. Luke's Episcopal Hospital, a 900-bed acute-care teaching hospital located in the Texas Medical Center, Houston, TX. The study was a retrospective cohort study of patients hospitalized between January 2005 and December 2008. The study was approved by the Institutional Review Board of the hospital. In view of the retrospective nature of the study, the need for informed consent from subjects was not mandated.
Patients.
All adult patients (18 years of age or greater) with blood cultures positive for Pseudomonas aeruginosa were identified from the hospital microbiology database. The treatment for each patient was at the discretion of the attending physicians. Patients without antimicrobial susceptibility profiling results were excluded. Repeat positive cultures (≤30 days apart) for the same patient were considered the same bacteremia episode.
Clonal uniqueness of the multidrug-resistant isolates.
The clonal relatedness of the first isolates of each multidrug-resistant infection episode was assessed by repetitive element-based PCR (rep-PCR), as described previously (16, 18). The DNA fragments were separated by a model 2100 bioanalyzer (Agilent Technologies, Inc., Santa Clara, CA), and the fingerprint patterns were compared visually and with DiversiLab software using the Pearson correlation coefficient (Bacterial Barcodes, Inc., Athens, GA). The isolates were considered indistinguishable (no difference in bands on visual inspection; similarity of DNA fragment pattern, 99% or greater), related (a difference of one to two bands; similarity, 95% to 98.9%), or distinct (a difference of three or more bands; similarity, less than 95%), as recommended by the manufacturer.
Data retrieval.
Pertinent data were retrieved from the patients' electronic medical records and laboratory databases, including demographic characteristics (e.g., age, ethnicity, and gender), comorbidities (e.g., cardiovascular, respiratory, and renal conditions; diabetes mellitus; and immunosuppression), severity of illness (as assessed by the APACHE II score on the first day of positive blood culture), the microorganisms isolated, the antimicrobial susceptibility profiles, and empirical treatment. The primary endpoint of the study was 30-day (all-cause) mortality. Patients discharged prior to day 30 were deemed to be alive unless proven otherwise. Secondary endpoints of the study were hospital mortality, infection-related mortality, and length of hospital stay associated with bacteremia. Infection-related mortality was assessed by an infectious diseases physician investigator (J.-P.C.) who was blinded to the antimicrobial susceptibility profiles and treatment.
Definitions.
Multidrug resistance was defined as resistance to three or more of the following four classes of agents: antipseudomonal carbapenems, antipseudomonal beta-lactams (penicillins and cephalosporins), aminoglycosides, and fluoroquinolones. Resistance to all agents within the class tested must be demonstrated for the isolate to be considered resistant to the agent class (e.g., an isolate must be resistant to amikacin, gentamicin, and tobramycin to be classified as aminoglycoside resistant). Multidrug-susceptible isolates were defined as susceptible to all the agents in each of the four antipseudomonal classes tested. Appropriate empirical therapy was defined as therapy (with at least one agent, if more than one agent was used), given within 24 h of the time that the sample for index culture was obtained, to which the isolate was found to be susceptible on the final antimicrobial susceptibility report and the doses used were appropriate for the end organ(s) function of the patients. Length of prior hospital stay was the time from the date of hospital admission to the date of first positive blood culture, and the length of hospital stay associated with bacteremia was the time from the date of first positive blood culture to the time of hospital discharge.
Statistical analysis.
Baseline demographics and clinical outcomes were compared between those with bacteremia caused by multidrug-resistant and multidrug-susceptible P. aeruginosa isolates. The Student t test or the Kruskal-Wallis test was used for continuous variables, and Fisher's exact test was used for dichotomous variables. Threshold breakpoints for continuous variables were identified by classification and regression tree (CART) analysis. Logistic regression models were used to explore independent risk factors for 30-day mortality. Univariate analyses were performed separately for each of the risk factor variables to ascertain the odds ratio (OR) and 95% confidence interval (CI). Variables with a P value of <0.2 in the univariate analyses were included in the logistic regression model for the multivariate analysis. A backward-selection process was utilized.
Time to mortality was analyzed using Kaplan-Meier survival analysis and the log-rank test, stratifying the data on the basis of the multidrug resistance status. Right censoring was used if mortality was not directly observed at the end of hospitalization. A P value of ≤0.05 was considered significant unless stated otherwise. In view of the differences in the baseline characteristics between the two cohorts, a propensity scoring analysis was used to adjust for the underlying risk factors for multidrug resistance. Logistic regression (univariate and multivariate) was used to explore the various independent risk factors associated with multidrug resistance, as detailed above. On the basis of the final parameter estimates in the multivariate model, the propensity score (an estimated probability of being infected with multidrug-resistant isolates) was assessed for each patient. Subsequently, each patient in the multidrug-resistant cohort was matched to a patient in the multidrug-susceptible group (1:1 match) with the closest propensity score. A maximal difference of 5% in the likelihood of multidrug-resistant P. aeruginosa bacteremia was allowed in the matching process. If there was more than one possible match with an identical propensity score, patients with similar sources of bacteremia (first secondary matching variable) and similar lengths of prior hospital stay (backup secondary matching variable) were given priority in the matching process. Matched patients were reanalyzed as described above. All statistical analyses were performed using the Systat program, version 12.0 (Systat Software, Inc., Point Richmond, CA).
RESULTS
Patients and bacteria.
A total of 109 adult patients with Pseudomonas aeruginosa bacteremia were identified: 84 with multidrug-susceptible P. aeruginosa bacteremia and 25 with multidrug-resistant P. aeruginosa bacteremia. The demographic and clinical characteristics of the patients are shown in Table 1. Patients in the multidrug-resistant group had a longer prior hospital stay before the first day of positive blood culture and higher baseline APACHE II scores. Overall, there was no predominant prescribing pattern (with respect to the agents selected), but patients in the multidrug-resistant group were more likely to be given more than one antipseudomonal agent empirically. No continuous or prolonged infusion of beta-lactams was used in the present study. The prevalence rate of multidrug resistance among P. aeruginosa bloodstream isolates in our institution over the study period was 13% (range, 9% to 17% annually); all of them were resistant to the antipseudomonal carbapenems and fluoroquinolones. Among the 23 multidrug-resistant isolates available, 17 unique clones were identified (Fig. 1).
TABLE 1.
Baseline characteristics of all patients
| Variable | MDSd | MDRe | P value |
|---|---|---|---|
| No. of patients | 84 | 25 | |
| Age (yr)a | 61.6 ± 14.8 | 57.4 ± 10.7 | 0.121 |
| No. (%) of patients of male gender | 49 (58.3) | 15 (60.0) | 1.000 |
| Baseline APACHE II scorea | 12.2 ± 6.0 | 14.9 ± 5.3 | 0.022f |
| Serum creatinine concn (mg/dl)a | 1.8 ± 1.6 | 1.7 ± 1.4 | 0.765 |
| Length of prior hospital stay (days)a | 14.4 ± 43.6 | 32.6 ± 37.3 | 0.046 |
| No. (%) of patients of the following ethnicity: | |||
| Caucasian | 49 (58.3) | 13 (52.0) | 0.684 |
| African American | 18 (21.4) | 8 (32.0) | 0.293 |
| Hispanic | 14 (16.7) | 4 (16.0) | 1.000 |
| Other | 3 (3.6) | 0 (0.0) | NAg |
| No. (%) of patients with the following comorbiditiesb: | |||
| Cardiovascular conditions | 66 (78.6) | 22 (88.0) | 0.393 |
| Respiratory conditions | 9 (10.7) | 8 (32.0) | 0.023 |
| Central nervous system conditions | 12 (14.3) | 3 (12.0) | 1.000 |
| Renal conditions | 25 (29.8) | 14 (56.0) | 0.031 |
| Diabetes mellitus | 29 (34.5) | 15 (60.0) | 0.035 |
| Immunnosuppression | 27 (32.1) | 9 (36.0) | 0.810 |
| No. (%) of patients with the following source of bacteremiac: | |||
| Line | 11 (13.1) | 2 (8.0) | 0.728 |
| Lung | 16 (19.0) | 11 (44.0) | 0.017 |
| Urine | 18 (21.4) | 2 (8.0) | 0.153 |
| Wound | 16 (19.0) | 3 (12.0) | 0.554 |
| Abdomen | 10 (11.9) | 5 (20.0) | 0.328 |
| Unknown | 20 (23.8) | 3 (12.0) | 0.270 |
| No. (%) of patients in whom more than one agent was used empirically | 22 (26.2) | 14 (56.0) | 0.008 |
| No. (%) of patients for whom a repeat blood culture was ordered | 53 (63.1) | 23 (92.0) | 0.006 |
Values are presented as means ± standard deviations.
Comorbidities included hypertension, congestive heart failure, coronary artery disease, and history of myocardial infarction for cardiovascular conditions; asthma and chronic obstructive pulmonary disease for respiratory conditions; stroke and cerebrospinal fluid leak for central nervous system conditions; chronic renal insufficiency for renal conditions; and organ transplantation, chronic steroid therapy (>10 mg daily of prednisone or equivalent for >1 month), neutropenia (absolute neutrophil count of <1,000 cells per mm3) postchemotherapy, autoimmune disease, and human immunodeficiency viral infection for immunosuppression.
The data may not add up to 100% if patients had more than one source of bacteremia.
MDS, multidrug susceptible.
MDR, multidrug resistant.
Analyzed by the Kruskal-Wallis test.
NA, not applicable.
FIG. 1.
Clonal relatedness of multidrug-resistant P. aeruginosa isolates (n = 23).
Clinical outcomes.
The key outcomes examined are as shown in Table 2. The overall 30-day mortality was 18.3%; it was observed in 11.9% and 40.0% of the patients in the multidrug-susceptible and multidrug-resistant cohorts, respectively (P = 0.003). In addition, overall hospital mortality and infection-related mortality were found to be 25.7% and 19.3%, respectively. Empirical therapy was more likely to be inappropriate in patients infected with multidrug-resistant isolates (P < 0.001), despite the finding that empirical combination therapy was more commonly used in this cohort.
TABLE 2.
Various (unadjusted) outcome endpoints
| Outcome endpoint | % of patients |
P valuec | ||
|---|---|---|---|---|
| Overall (n = 109) | MDSa (n = 84) | MDRb (n = 25) | ||
| 30-day mortality | 18.3 | 11.9 | 40.0 | 0.003 |
| Hospital (all-cause) mortality | 25.7 | 16.7 | 56.0 | <0.001 |
| Infection-related mortality | 19.3 | 9.5 | 52.0 | <0.001 |
| Inappropriate empirical therapy | 14.7 | 6.0 | 44.0 | <0.001 |
| Mean length of hospital stay associated with bacteremia (days) ± SD | 18.7 ± 25.0 | 16.5 ± 23.6 | 26.4 ± 28.3 | 0.120 |
MDS, multidrug susceptible.
MDR, multidrug resistant.
By comparison of the multidrug-susceptible and multidrug-resistant cohorts.
Risk factors for 30-day mortality.
The factors independently associated with 30-day mortality are shown in Table 3. In the final multivariate analysis, multidrug resistance (OR, 6.829; 95% CI, 1.945 to 23.984; P = 0.003), immunosuppression (OR, 5.001; 95% CI, 1.430 to 17.495; P = 0.012), and an APACHE II score of ≥22 (OR, 29.034; 95% CI, 5.012 to 168.1; P < 0.001) were found to be independent risk factors, after adjustment for other confounding variables.
TABLE 3.
Logistic regression analysis of the risk factors for 30-day mortalitya
| Variable | Univariate analysis |
Multivariate analysis |
||
|---|---|---|---|---|
| OR (95% CI) | P value | OR (95% CI) | P value | |
| Age | 1.022 (0.984-1.060) | 0.261 | ||
| Male | 1.820 (0.641-5.170) | 0.261 | ||
| Ethnic background | ||||
| Caucasian | 1.170 (0.436-3.142) | 0.755 | ||
| African American | 0.506 (0.136-1.888) | 0.311 | ||
| Hispanic | 0.871 (0.226-3.348) | 0.840 | ||
| Baseline serum creatinine concn | 0.997 (0.722-1.376) | 0.985 | ||
| Length of prior hospital stay | 1.009 (0.998-1.021) | 0.116 | ||
| Multidrug resistance | 4.933 (1.748-13.922) | 0.003 | 6.829 (1.945-23.984) | 0.003 |
| Appropriate empirical therapy | 0.295 (0.093-0.943) | 0.039 | ||
| Baseline APACHE II score | 1.157 (1.062-1.260) | 0.001 | ||
| Baseline APACHE II score ≥ 22 | 15.436 (3.539-67.329) | <0.001 | 29.034 (5.012-168.196) | <0.001 |
| Comorbidities | ||||
| Cardiovascular conditions | 0.658 (0.209-2.072) | 0.474 | ||
| Respiratory conditions | 3.039 (0.966-9.558) | 0.057 | ||
| Central nervous system conditions | 1.132 (0.288-4.455) | 0.859 | ||
| Renal conditions | 5.973 (2.065-17.280) | 0.001 | ||
| Diabetes mellitus | 0.981 (0.365-2.641) | 0.970 | ||
| Immunosuppression | 3.129 (1.157-8.462) | 0.025 | 5.001 (1.430-17.495) | 0.012 |
| Source of bacteremia | ||||
| Line | 0.338 (0.041-2.760) | 0.311 | ||
| Lung | 4.235 (1.522-11.787) | 0.006 | ||
| Urine | 0.194 (0.024-1.543) | 0.121 | ||
| Wound | 0.805 (0.211-3.097) | 0.751 | ||
| Abdomen | 1.773 (0.501-6.278) | 0.375 | ||
Receiver operating characteristic value of the final model, 0.820.
Propensity score matching and time to mortality.
Factors independently associated with multidrug resistance were age and baseline APACHE II scores (data not shown). Forty-two patients were matched on the basis of the propensity scores for the Kaplan-Meier survival analysis. All the patients were matched with less than 1% difference in their propensity scores. The demographic and clinical characteristics of the matched patients are shown in Table 4. The time to mortality was shorter in patients with bacteremia due to multidrug-resistant isolates (Fig. 2).
TABLE 4.
Baseline characteristics of all matched patients
| Variable | MDSd | MDRe | P value |
|---|---|---|---|
| No. of patients | 21 | 21 | |
| Age (yr)a | 58.1 ± 14.8 | 58.9 ± 8.5 | 0.829 |
| No. (%) of patients of male gender | 11 (52.4) | 13 (61.9) | 0.756 |
| Baseline APACHE II scorea | 14.7 ± 5.8 | 15.0 ± 5.6 | 0.890f |
| Serum creatinine concn (mg/dl)a | 1.6 ± 1.5 | 1.5 ± 1.2 | 0.826 |
| Length of prior hospital stay (days)a | 4.4 ± 7.4 | 36.1 ± 39.7 | 0.002 |
| No. (%) of patients of the following ethnicity: | |||
| Caucasian | 12 (57.1) | 12 (57.1) | 1.000 |
| African American | 5 (23.8) | 6 (28.6) | 1.000 |
| Hispanic | 3 (14.3) | 3 (14.3) | 1.000 |
| Other | 1 (4.8) | 0 (0.0) | NAg |
| No. (%) of patients with the following comorbiditiesb: | |||
| Cardiovascular conditions | 17 (81.0) | 19 (90.5) | 0.663 |
| Respiratory conditions | 2 (9.5) | 7 (33.3) | 0.130 |
| Central nervous system conditions | 1 (4.8) | 2 (9.5) | 1.000 |
| Renal conditions | 7 (33.3) | 12 (57.1) | 0.215 |
| Diabetes mellitus | 7 (33.3) | 13 (61.9) | 0.121 |
| Immunnosuppression | 9 (42.9) | 9 (42.9) | 1.000 |
| No. (%) of patients with the following source of bacteremiac: | |||
| Line | 5 (23.8) | 2 (9.5) | 0.410 |
| Lung | 5 (23.8) | 8 (38.1) | 0.505 |
| Urine | 3 (14.3) | 2 (9.5) | 1.000 |
| Wound | 6 (28.6) | 3 (14.3) | 0.454 |
| Abdomen | 0 (0.0) | 4 (19.0) | 0.107 |
| Unknown | 3 (14.3) | 3 (14.3) | 1.000 |
| No. (%) of patients in whom more than one agent was used empirically | 6 (28.6) | 11 (52.4) | 0.208 |
| No. (%) of patients for whom a repeat blood culture was ordered | 12 (57.1) | 19 (90.5) | 0.033 |
Values presented as mean ± standard deviation.
Comorbidities included hypertension, congestive heart failure, coronary artery disease, and history of myocardial infarction for cardiovascular conditions; asthma and chronic obstructive pulmonary disease for respiratory conditions; stroke and cerebrospinal fluid leak for central nervous system conditions; chronic renal insufficiency for renal conditions; organ transplantation, chronic steroid therapy (>10 mg daily of prednisone or equivalent for >1 month), neutropenia (absolute neutrophil count of <1,000 cells per mm3) postchemotherapy, autoimmune disease, and human immunodeficiency viral infection for immunosuppression.
The data may not add up to 100% if the patients had more than one source of bacteremia.
MDS, multidrug susceptible.
MDR, multidrug resistant.
Analyzed by the Kruskal-Wallis test.
NA, not applicable.
FIG. 2.
Kaplan-Meier analysis of time to (all-cause) mortality in matched patients (n = 42). All-cause mortality was observed in 2 patients (9.5%) in the multidrug-susceptible group and 13 patients (61.9%) in the multidrug-resistant group (P = 0.001).
DISCUSSION
P. aeruginosa bacteremia is problematic in acutely and critically ill hospitalized patients. The rate of mortality within the first 3 to 5 days of the onset of bacteremia is high; with as many as 50% of isolates being resistant to standard empirical antibiotic therapies, it becomes particularly difficult to select the appropriate initial antibiotic therapy. Studies have shown that appropriate empirical therapy is paramount to improve mortality outcomes; an inappropriate choice of initial antibiotics is associated with increased mortality (10, 14). However, many clinicians believe that multidrug resistance is associated with reduced virulence (2, 15) and therefore had a less detrimental impact on patient outcomes.
The economic impact of multidrug resistance in Gram-negative bacilli has been reviewed previously (5). In addition, multidrug resistance has also previously been shown to be associated with worse clinical outcomes in pneumonia caused by Gram-negative organisms (12) and infections due to P. aeruginosa (4, 6). Our findings corroborated the results reported from a previous study that patients with multidrug-resistant bacteremia were more likely to receive inappropriate empirical therapy (14). In the univariate analysis, appropriate empirical therapy was associated with decreased 30-day mortality (OR, 0.295; P = 0.039). However, the only independent variables significantly related to 30-day mortality in the final model were multidrug resistance, immunosuppression, and a baseline APACHE II score of ≥22.
In the present study, we provided evidence that the multidrug-resistant isolates were clonally diverse, a common limitation and potential confounding variable in studies of similar design. We showed that multidrug-resistant P. aeruginosa was significantly associated with an increase in the 30-day mortality. A similar trend was also observed with respect to hospital and infection-related mortality. As with the other studies, the patient cohorts had significant underlying differences. Patients infected with multidrug-resistant bacteria were more likely to have been hospitalized for a longer period of time and presented with higher APACHE II scores at the onset of bacteremia. Furthermore, a higher proportion of these patients had respiratory/renal comorbidities, diabetes, and a respiratory source of bacteremia. These underlying differences could have biased our assessment of the true impact of multidrug resistance on patient outcomes. As a result, a previously adopted (1, 8) multivariate analysis and propensity scoring system were used to minimize the influence of pertinent confounders. Subsequently, Kaplan-Meier survival analysis also revealed a significantly shorter time to mortality in the multidrug-resistant cohort, after adjustment for the underlying risk factors for multidrug resistance.
Despite our best efforts, there were several limitations associated with the retrospective design of the present study. Empirical treatment and definitive treatment were at the discretion of the respective attending medical teams rather than adherence to a prestandardized protocol. An assumption was made that all the clinicians were equally up-to-date with the latest local resistance patterns, the latest treatment guidelines, and knowledge of the optimal supportive care/antimicrobial dosing to be used and had a similar threshold for soliciting assistance from an infectious diseases consultant. The timeliness with which the results were provided by the microbiology laboratory to the clinicians was also assumed to be similar. An unintentional bias could have been introduced as more experienced infectious diseases consultants were more likely to be involved in complicated (multidrug-resistant) cases, but the bias would have been more favorable for the multidrug-resistant cohort. Assessments of the length of the prior hospital stay and the length of the hospital stay associated with bacteremia could also be biased, as these arbitrary measures might be due to various reasons.
The mechanism(s) of multidrug resistance was investigated and found to be heterogeneous; there was also a low prevalence of hypermutation among the multidrug-resistant isolates (17). Since the multidrug resistance phenotype could be due to different molecular mechanisms, results from our single-center study may not be directly applicable to other institutions where there is a predominant mechanism of multidrug resistance. In our statistical analysis, multidrug resistance was identified as an independent risk factor. In reality, the likelihood of appropriate empirical therapy is highly (and negatively) correlated to multidrug resistance. Since a delay in starting effective antimicrobial therapy has been shown to be associated with a higher rate of mortality among patients with P. aeruginosa bacteremia (9, 13, 14), it may not be feasible to clearly distinguish the relative importance of the two variables with the present study design. Investigations are ongoing to examine the virulence of and host immune response to multidrug-resistant isolates.
In conclusion, multidrug-resistant P. aeruginosa bacteremia was significantly associated with a longer prior hospital stay and was less likely than multidrug-susceptible P. aeruginosa bacteremia to be treated with appropriate empirical antibiotics. Patients with APACHE II scores of ≥22, immunosuppression, and infection with a multidrug-resistant strain were at higher risk for 30-day mortality. The time to mortality was also significantly shorter in the multidrug-resistant cohort, even after adjustment of the risk factors for multidrug resistance. Measures to combat/curb multidrug-resistant bacterial infections (e.g., vigilant infection control practices, antimicrobial stewardship, and development of new antipseudomonal agents which overcome existing resistance mechanisms) would likely benefit patient outcomes.
Acknowledgments
This study was supported by a grant from the Investigator-Sponsored Study Program of AstraZeneca.
Footnotes
Published ahead of print on 28 June 2010.
REFERENCES
- 1.Aksoy, O., D. J. Sexton, A. Wang, P. A. Pappas, W. Kourany, V. Chu, V. G. Fowler, Jr., C. W. Woods, J. J. Engemann, G. R. Corey, T. Harding, and C. H. Cabell. 2007. Early surgery in patients with infective endocarditis: a propensity score analysis. Clin. Infect. Dis. 44:364-372. [DOI] [PubMed] [Google Scholar]
- 2.Becker, J. M., L. K. Henry, W. Jiang, and Y. Koltin. 1995. Reduced virulence of Candida albicans mutants affected in multidrug resistance. Infect. Immun. 63:4515-4518. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bonomo, R. A., and D. Szabo. 2006. Mechanisms of multidrug resistance in Acinetobacter species and Pseudomonas aeruginosa. Clin. Infect. Dis. 43(Suppl. 2):S49-S56. [DOI] [PubMed] [Google Scholar]
- 4.Cao, B., H. Wang, H. Sun, Y. Zhu, and M. Chen. 2004. Risk factors and clinical outcomes of nosocomial multi-drug resistant Pseudomonas aeruginosa infections. J. Hosp. Infect. 57:112-118. [DOI] [PubMed] [Google Scholar]
- 5.Giske, C. G., D. L. Monnet, O. Cars, and Y. Carmeli. 2008. Clinical and economic impact of common multidrug-resistant gram-negative bacilli. Antimicrob. Agents Chemother. 52:813-821. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Harris, A., C. Torres-Viera, L. Venkataraman, P. DeGirolami, M. Samore, and Y. Carmeli. 1999. Epidemiology and clinical outcomes of patients with multiresistant Pseudomonas aeruginosa. Clin. Infect. Dis. 28:1128-1133. [DOI] [PubMed] [Google Scholar]
- 7.Hocquet, D., P. Berthelot, M. Roussel-Delvallez, R. Favre, K. Jeannot, O. Bajolet, N. Marty, F. Grattard, P. Mariani-Kurkdjian, E. Bingen, M. O. Husson, G. Couetdic, and P. Plesiat. 2007. Pseudomonas aeruginosa may accumulate drug resistance mechanisms without losing its ability to cause bloodstream infections. Antimicrob. Agents Chemother. 51:3531-3536. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ingram, P. R., D. C. Lye, D. A. Fisher, W. P. Goh, and V. H. Tam. 2009. Nephrotoxicity of continuous versus intermittent infusion of vancomycin in outpatient parenteral antimicrobial therapy. Int. J. Antimicrob. Agents 34:570-574. [DOI] [PubMed] [Google Scholar]
- 9.Kang, C. I., S. H. Kim, H. B. Kim, S. W. Park, Y. J. Choe, M. D. Oh, E. C. Kim, and K. W. Choe. 2003. Pseudomonas aeruginosa bacteremia: risk factors for mortality and influence of delayed receipt of effective antimicrobial therapy on clinical outcome. Clin. Infect. Dis. 37:745-751. [DOI] [PubMed] [Google Scholar]
- 10.Kang, C. I., S. H. Kim, W. B. Park, K. D. Lee, H. B. Kim, E. C. Kim, M. D. Oh, and K. W. Choe. 2005. Bloodstream infections caused by antibiotic-resistant gram-negative bacilli: risk factors for mortality and impact of inappropriate initial antimicrobial therapy on outcome. Antimicrob. Agents Chemother. 49:760-766. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kugelberg, E., S. Lofmark, B. Wretlind, and D. I. Andersson. 2005. Reduction of the fitness burden of quinolone resistance in Pseudomonas aeruginosa. J. Antimicrob. Chemother. 55:22-30. [DOI] [PubMed] [Google Scholar]
- 12.Kwa, A. L., J. G. Low, E. Lee, A. Kurup, H. L. Chee, and V. H. Tam. 2007. The impact of multidrug resistance on the outcomes of critically ill patients with Gram-negative bacterial pneumonia. Diagn. Microbiol. Infect. Dis. 58:99-104. [DOI] [PubMed] [Google Scholar]
- 13.Lodise, T. P., Jr., N. Patel, A. Kwa, J. Graves, J. P. Furuno, E. Graffunder, B. Lomaestro, and J. C. McGregor. 2007. Predictors of 30-day mortality among patients with Pseudomonas aeruginosa bloodstream infections: impact of delayed appropriate antibiotic selection. Antimicrob. Agents Chemother. 51:3510-3515. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Micek, S. T., A. E. Lloyd, D. J. Ritchie, R. M. Reichley, V. J. Fraser, and M. H. Kollef. 2005. Pseudomonas aeruginosa bloodstream infection: importance of appropriate initial antimicrobial treatment. Antimicrob. Agents Chemother. 49:1306-1311. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Ravizzola, G., F. Pirali, A. Paolucci, L. Terlenghi, L. Peroni, A. Colombi, and A. Turano. 1987. Reduced virulence in ciprofloxacin-resistant variants of Pseudomonas aeruginosa strains. J. Antimicrob. Chemother. 20:825-829. [DOI] [PubMed] [Google Scholar]
- 16.Syrmis, M. W., M. R. O'Carroll, T. P. Sloots, C. Coulter, C. E. Wainwright, S. C. Bell, and M. D. Nissen. 2004. Rapid genotyping of Pseudomonas aeruginosa isolates harboured by adult and paediatric patients with cystic fibrosis using repetitive-element-based PCR assays. J. Med. Microbiol. 53:1089-1096. [DOI] [PubMed] [Google Scholar]
- 17.Tam, V. H., K. T. Chang, K. Abdelraouf, C. G. Brioso, M. Ameka, L. A. McCaskey, J. S. Weston, J. P. Caeiro, and K. W. Garey. 2010. Prevalence, resistance mechanisms, and susceptibility of multidrug-resistant bloodstream isolates of Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 54:1160-1164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Tam, V. H., K. T. Chang, M. T. LaRocco, A. N. Schilling, S. K. McCauley, K. Poole, and K. W. Garey. 2007. Prevalence, mechanisms, and risk factors of carbapenem resistance in bloodstream isolates of Pseudomonas aeruginosa. Diagn. Microbiol. Infect. Dis. 58:309-314. [DOI] [PubMed] [Google Scholar]
- 19.Vidal, F., J. Mensa, M. Almela, J. A. Martinez, F. Marco, C. Casals, J. M. Gatell, E. Soriano, and M. T. Jimenez de Anta. 1996. Epidemiology and outcome of Pseudomonas aeruginosa bacteremia, with special emphasis on the influence of antibiotic treatment. Analysis of 189 episodes. Arch. Intern. Med. 156:2121-2126. [PubMed] [Google Scholar]
- 20.Wisplinghoff, H., T. Bischoff, S. M. Tallent, H. Seifert, R. P. Wenzel, and M. B. Edmond. 2004. Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin. Infect. Dis. 39:309-317. [DOI] [PubMed] [Google Scholar]