Abstract
Accessory gene regulator (agr) dysfunction in Staphylococcus aureus has been associated with a longer duration of bacteremia. We aimed to assess the independent association between agr dysfunction in S. aureus bacteremia and 30-day in-hospital mortality. This retrospective cohort study included all adult inpatients with S. aureus bacteremia admitted between 1 January 2003 and 30 June 2007. Severity of illness prior to culture collection was measured using the modified acute physiology score (APS). agr dysfunction in S. aureus was identified semiquantitatively by using a δ-hemolysin production assay. Cox proportional hazard models were used to measure the association between agr dysfunction and 30-day in-hospital mortality, statistically adjusting for patient and pathogen characteristics. Among 814 patient admissions complicated by S. aureus bacteremia, 181 (22%) patients were infected with S. aureus isolates with agr dysfunction. Overall, 18% of patients with agr dysfunction in S. aureus died, compared to 12% of those with functional agr in S. aureus (P = 0.03). There was a trend toward higher mortality among patients with S. aureus with agr dysfunction (adjusted hazard ratio [HR], 1.34; 95% confidence interval [CI], 0.87 to 2.06). Among patients with the highest APS (scores of >28), agr dysfunction in S. aureus was significantly associated with mortality (adjusted HR, 1.82; 95% CI, 1.03 to 3.21). This is the first study to demonstrate an independent association between agr dysfunction and mortality among severely ill patients. The δ-hemolysin assay examining agr function may be a simple and inexpensive approach to predicting patient outcomes and potentially optimizing antibiotic therapy.
Staphylococcus aureus is the second most common pathogen in health care-associated infections, responsible for 15% of all health care-associated infections reported to the National Healthcare Safety Network from 2006 to 2007 (13). S. aureus bacteremia is associated with high patient morbidity and mortality and confers a considerable financial burden to an already overwhelmed health care system (4, 5). Currently, many clinical laboratories devote significant resources to the rapid detection of the S. aureus mecA gene, which confers resistance to beta-lactam antibiotics, in an attempt to improve patient outcomes. However, other genetic elements of S. aureus may have similar clinical and epidemiologic importance.
The accessory gene regulator (agr) locus of S. aureus is a quorum-sensing, global regulon that coordinates the expression of dozens of housekeeping genes as well as the expression of secreted and cell-associated virulence factors (7, 37). Mutations causing agr dysfunction in the agr locus confer changes in the expression of autolysins and hemolysins and have global effects on bacterial phenotypes, including pathogenicity (8, 26, 34, 38).
S. aureus agr dysfunction has been associated with a longer duration of bacteremia and with attenuated vancomycin bactericidal activity (7, 25, 27). The development of the vancomycin-intermediate S. aureus (VISA) phenotype has been accompanied by a marked reduction or complete loss of agr function, both through loss-of-function mutations in agr and through upstream suppression of agr expression (23, 25, 27, 34). agr knockout mutants have also demonstrated an increased proclivity to develop the VISA phenotype (7, 27, 37). In vitro studies have demonstrated that inactivation of agr attenuates the bactericidal activity of vancomycin, and certain clones of S. aureus with dysfunctional agr may have a survival advantage under vancomycin selective pressure, possibly related to decreased autolysis (26, 27). Studies of nosocomial bacteremia have shown that infection with S. aureus without agr function was more common among patients with persistent or prolonged S. aureus bacteremia than among patients with resolving S. aureus bacteremia (7, 21).
To our knowledge, no study has measured the independent association between agr dysfunction and mortality among S. aureus-infected patients. The purpose of this study was to assess the association between agr dysfunction and mortality in a large cohort of patients with S. aureus bacteremia while statistically adjusting for important patient characteristics, such as comorbid conditions and severity of illness, and pathogen phenotypes and genotypes.
(This work was presented in part at the joint 48th Annual Interscience Conference on Antimicrobial Agents and Chemotherapy-46th Annual Infectious Disease Society of America Scientific Meeting, Washington, DC, 27 October 2008.)
MATERIALS AND METHODS
Study design and patient population.
This retrospective cohort study included all adult admissions to the University of Maryland Medical Center between 1 January 2003 and 30 June 2007 with a positive blood culture for S. aureus. Each admission was treated as an independent event, and therefore patients may have been included in the study more than once. The first positive S. aureus blood culture from each admission was saved by the clinical microbiology laboratory for research purposes and used for this study. Eligible patients were identified using a relational database containing medical, pharmaceutical, and microbiologic data. This database has been validated in previous studies and has positive and negative predictive values in excess of 99% compared to paper medical records (9, 10, 18, 19, 22). The relational database was compared to paper medical records for 236 (29%) patients in our study cohort and was found to have positive predictive values of at least 97% and negative predictive values of at least 98%.
Additional variables that were not available in the relational database (e.g., components of the modified acute physiology score [APS]) were collected by a research nurse via chart review. This study was approved by the institutional review board of the University of Maryland, Baltimore, MD. A waiver of consent was granted given the retrospective nature of the project.
Laboratory methods. (i) Clinical strains.
All S. aureus-positive blood cultures during the study period were initially identified by the clinical microbiology laboratory and were stored at −80°C. Identification of S. aureus was determined by Gram stain, colony morphology, and catalase- and coagulase-positive reactions.
(ii) agr phenotype and genotype studies.
We operationalized agr dysfunction by using the level of δ-hemolysin production. This is the standard approach for the measurement of agr function because RNAIII, the effector of agr regulation, encodes δ-hemolysin (34). δ-Hemolysin production was measured by streaking the S. aureus isolate adjacent to a β-hemolysin disk (Remel, Lenexa, KS) on a tryptic soy agar plate with 5% sheep blood, incubating it at 37°C overnight, and evaluating it for synergistic hemolysis within the β-hemolysin zone produced by the disk containing bacterial growth. The presence of synergistic hemolysis within the β-hemolysin zone indicates the production of δ-hemolysin by the test organism and, therefore, a functional agr locus (27). agr dysfunction was defined as the complete absence of δ-hemolysin within the β-hemolysin zone, as evidenced by the lack of synergistic hemolysis (Fig. 1). Multiplex PCR was used to determine the agr group genotype as described previously, with appropriate control strains for agr groups I, II, III, and IV (11).
FIG. 1.
Measurement of agr dysfunction by δ-hemolysin assay. Isolates were streaked near a β-hemolysin disk in order to test for δ-hemolysin production.
(iii) Antimicrobial susceptibility testing.
Vancomycin, daptomycin, and linezolid MICs were measured using the microdilution Etest (AB Biodisk, Solna, Sweden; bioMérieux, Durham, NC) according to the manufacturers' instructions and following Clinical and Laboratory Standards Institute guidelines (1, 3). The VISA phenotype was defined as a vancomycin MIC of ≥4 μg/ml (33). The Etest macrodilution method was used to determine if the patient was infected with heterogeneous VISA (hVISA) as described previously (40).
(iv) Additional genotype studies.
PCR was used to determine the presence of the Panton-Valentine leukocidin (PVL) and arginine catabolic mobile element (ACME) genes (6, 17). The polymorphic X region of the staphylococcal protein A (spa) gene was sequenced and typed according to previously described procedures (12, 31). The USA300 clone was defined as any S. aureus isolate that carried spa type motif MBQBLO and was PVL positive and ACME positive, as previously validated (14).
Clinical and outcome variable definitions.
The primary outcome of interest was 30-day in-hospital mortality, defined as death occurring in the hospital during the period from culture collection to 30 days after culture collection. Severity of illness was measured 24 h before the time the culture was obtained, using the modified APS. If the blood culture was obtained within 24 h of hospital admission, the APS at the time of admission was calculated. We chose to measure severity of illness as close to exactly 24 h before culture collection as possible. This was done in order to address the underlying severity of illness before infection. Measurement of the severity of illness after culture collection would measure the severity of the bacteremia, which should not be controlled for since it is in the causal pathway (20, 29). The modified APS is based on the acute physiology and chronic health evaluation (APACHE) III score (16). Since the APACHE III was designed for use in intensive care unit (ICU) patients, the score has been modified by excluding variables that are not applicable to a more general inpatient population, as in several previously published studies (16, 22, 32). Aggregate comorbidity was measured using the chronic disease score. Medications prescribed in the first 24 h of admission were given a weighted value which was used to calculate the chronic disease score for each admission (39). The chronic disease score has been validated for use in studies of antibiotic resistance, including studies of methicillin-resistant S. aureus (MRSA), and is used in a similar fashion to the Charlson comorbidity index but may be more predictive (18). Blood cultures collected more than 48 h after admission were designated hospital-onset infections, and blood cultures collected within 48 h of admission were designated community-onset infections. Polymicrobial infection was defined as having more than one microorganism present in the same blood culture. Patients who received any antibiotics from the University of Maryland Medical Center within a year of the index admission were categorized as having received antibiotics in the past year.
Statistical analysis.
Bivariable associations were assessed using the chi-square test or Fisher's exact test for categorical variables and the Student t test or the Wilcoxon rank sum test for continuous variables. Stratification was used to test for potential effect modification. Unadjusted relative risks (RRs) were calculated to assess the associations between categorical variables and 30-day in-hospital mortality. Cox proportional hazard models were fit to measure the hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between agr dysfunction and 30-day in-hospital mortality. Patients were censored upon hospital discharge. The proportional hazards assumption that the effect of a variable is constant over time was tested for each variable in the final model by assessing the interaction of the variable with a function of time. All variables that were significantly (α ≤ 0.1) associated with agr dysfunction or mortality in the bivariate analysis, as well as variables considered to be biologically important (e.g., methicillin resistance), were included in the initial (full) multivariable Cox proportional hazard model. Variables that were not significantly associated with the outcome (α > 0.05) were removed from the full multivariable model in succession. Each of the removed variables was then reinserted into the model to assess whether it was a confounder, which was considered the case if the variable altered the regression coefficient of the primary exposure variable by >20%. If so, that variable was included in the model.
Since other studies have shown that the risk of mortality among bacteremic patients changes depending on underlying severity of illness, the cohort was divided into quartiles by initial severity-of-illness scores, and separate analyses were performed (15). To assess effect modification with severity of illness, we also examined the significance of an interaction term representing the interaction between a high severity-of-illness score and agr function. All analyses were performed using SAS software (SAS Institute, Cary, NC), version 9.1.
RESULTS
There were 814 episodes of S. aureus bacteremia in 756 unique patients who had S. aureus isolated from the blood during the study period. Of these, 86% had received vancomycin on the index admission and 42% had received antibiotics in the past year. One hundred nine (13.4%) patients died in the hospital within 30 days of culture collection.
Among the S. aureus isolates collected, 63% were considered community-onset infections (culture collection within 48 hours of admission). Sixty percent of the isolates were MRSA, 2.1% were nonsusceptible to daptomycin, and all were susceptible to linezolid. The vancomycin MIC distribution of the isolates was as follows: <1 μg/ml, 4.3%; 1 μg/ml, 19.6%; 1.5 μg/ml, 52.5%; 2 μg/ml, 21.3%; and >2 μg/ml, 2.3%. Two (0.2%) isolates were VISA, and 4.3% were hVISA. Genotyping analyses at the agr locus found that 52% of isolates were of agr group I, 37% were of agr group II, 9% were of agr group III, <1% were of agr group IV, and 2% were unable to be typed. Twenty-two percent of the isolates (n = 181) had agr dysfunction as measured by the δ-hemolysin assay.
Thirty-seven percent (n = 300) of the isolates had a spa type associated with the ST8 clone, and among these, 63% were classified as USA300 (i.e., MBQBLO motif positive, PVL positive, and ACME positive). Thirty percent (n = 245) of the isolates had spa types associated with the ST5 clone, and 5% of the isolates had spa types associated with the ST30 clone. More details about the S. aureus clones in this cohort can be found in Table S1 in the supplemental material.
S. aureus isolates with dysfunctional agr were more likely to carry the MDMGMK spa type motif and to be of agr group II and had lower vancomycin MICs than isolates with functional agr (P < 0.05). Patients infected with S. aureus with dysfunctional agr were more likely to have a central venous catheter before culture collection, were more likely to be admitted in an earlier year, and were less likely to be injection drug users (P < 0.01) than patients infected with S. aureus with functional agr (Table 1).
TABLE 1.
Pathogen and patient characteristics stratified by agr function
| Characteristic | No. (%) of infections with characteristic (unless indicated otherwise) |
P value | ||
|---|---|---|---|---|
| Strains with functional agr (n = 633 [77.8% of infections]) | Strains with dysfunctional agr (n = 181 [22.2% of infections]) | All strains (n = 814) | ||
| Pathogen characteristics | ||||
| Methicillin resistance | 372 (58.8) | 116 (64.1) | 488 (60.0) | 0.20 |
| agr groupa | <0.01 | |||
| I | 363 (58.3) | 60 (34.1) | 423 (52.0) | |
| II | 221 (35.4) | 79 (44.9) | 300 (36.8) | |
| III | 36 (5.8) | 35 (19.9) | 71 (8.7) | |
| IV | 3 (0.5) | 2 (1.1) | 5 (0.6) | |
| Vancomycin MIC (μg/ml) | <0.01 | |||
| 0.5 | 1 (0.2) | 3 (1.7) | 4 (0.5) | |
| 0.75 | 18 (2.8) | 14 (7.7) | 32 (3.9) | |
| 1 | 113 (17.8) | 45 (24.9) | 158 (19.4) | |
| 1.5 | 339 (53.6) | 89 (49.2) | 427 (52.5) | |
| 2 | 149 (23.5) | 24 (13.3) | 173 (21.3) | |
| >2 | 13 (2.1) | 6 (3.3) | 19 (2.3) | |
| hVISA phenotype | 27 (4.3) | 8 (4.4) | 35 (4.3) | 0.93 |
| Hospital-onset infection | 223 (35.2) | 72 (39.8) | 295 (36.2) | 0.26 |
| USA300 clone | 171 (27.0) | 19 (10.5) | 190 (23.3) | <0.01 |
| spa type motif MDMGMK | 156 (24.6) | 78 (43.1) | 234 (28.8) | <0.01 |
| Yr of isolation | 0.01 | |||
| 2003 | 158 (25.0) | 51 (28.2) | 209 (25.7) | |
| 2004 | 147 (23.2) | 57 (31.5) | 204 (25.1) | |
| 2005 | 151 (23.8) | 34 (18.8) | 185 (22.7) | |
| 2006 | 117 (18.5) | 33 (18.2) | 150 (18.4) | |
| 2007b | 60 (9.5) | 6 (3.3) | 66 (8.1) | |
| Patient characteristics | ||||
| Age (yr) (mean ± SD) | 48 ± 16 | 52 ± 17 | 49 ± 16 | <0.01 |
| Female sex | 240 (37.9) | 70 (38.7) | 310 (38.1) | 0.85 |
| Admitted to ICU prior to culture collection | 186 (29.4) | 61 (33.7) | 247 (30.3) | 0.27 |
| Endocarditis | 110 (17.4) | 28 (15.5) | 138 (17.0) | 0.55 |
| Median (IQR) severity-of-illness score | 16 (8-28) | 19 (11-28) | 17 (9-28) | 0.08 |
| Median (IQR) chronic disease score | 1.0 (0-2.1) | 1.0 (1.0-2.2) | 1.0 (0-2.1) | 0.14 |
| Received vancomycin on index admission | 540 (85.3) | 160 (88.4) | 700 (86.0) | 0.29 |
| Receipt of antibiotics during previous yr | 263 (41.6) | 75 (41.4) | 338 (41.5) | 0.98 |
| Injection drug use | 219 (34.6) | 32 (17.7) | 251 (30.8) | <0.01 |
| Presence of central venous catheter before culture collection | 322 (63.0) | 114 (50.9) | 436 (53.6) | <0.01 |
Fifteen isolates were unable to be typed for agr group.
Only the first 6 months of 2007 were assessed.
When the bivariable statistical associations between molecular characteristics and 30-day in-hospital mortality were assessed, agr dysfunction was the only pathogen characteristic evaluated that was statistically associated with 30-day in-hospital mortality (P = 0.03). Other pathogen characteristics, including methicillin resistance, increased daptomycin, linezolid, or vancomycin MIC, hVISA phenotype, polymicrobial infection, presence of the PVL or ACME gene, and pulsed-field gel electrophoresis (PFGE) type USA300, were not associated with 30-day in-hospital mortality (P > 0.05) (Table 2). There was a trend toward increased mortality among patients infected with S. aureus with agr type II (unadjusted RR = 1.41; 95% CI, 0.99 to 2.00).
TABLE 2.
Pathogen characteristics stratified by 30-day in-hospital mortalitya
| Characteristic | No. (%) of admissions with characteristic |
P value | ||
|---|---|---|---|---|
| Nonsurvivors (n = 109 [13.4% of admissions]) | Survivors (n = 705 [86.6% of admissions]) | All admissions (n = 814) | ||
| agr dysfunction | 33 (30.3) | 149 (21.0) | 182 (22.2) | 0.03 |
| Methicillin resistance | 70 (64.2) | 418 (59.3) | 488 (60.0) | 0.33 |
| hVISA phenotype | 7 (6.4) | 28 (4.0) | 35 (4.3) | 0.24 |
| Daptomycin MIC of >1 μg/ml | 3 (2.9) | 13 (2.0) | 16 (2.1) | 0.53 |
| Vancomycin MIC (μg/ml) | 0.78 | |||
| 0.5 | 1 (0.9) | 3 (0.4) | 4 (0.5) | |
| 0.75 | 7 (6.4) | 25 (3.6) | 32 (3.8) | |
| 1 | 19 (17.6) | 140 (19.9) | 159 (19.6) | |
| 1.5 | 54 (50.0) | 373 (52.9) | 427 (52.5) | |
| 2 | 26 (24.1) | 147 (20.9) | 173 (21.3) | |
| >2 | 2 (1.9) | 17 (2.4) | 19 (2.3) | |
| Polymicrobial infection | 12 (11.0) | 77 (10.9) | 89 (10.9) | 0.98 |
| agr group II | 50 (45.9) | 250 (35.5) | 300 (36.8) | 0.05 |
| PVL | 27 (24.8) | 224 (31.8) | 251 (30.8) | 0.14 |
| ACMEb | 30 (27.5) | 187 (28.4) | 217 (28.2) | 0.73 |
| USA300 clonec | 21 (19.3) | 169 (24.0) | 190 (23.3) | 0.28 |
No isolates were nonsusceptible to linezolid.
ACME was measured for all MRSA isolates and for methicillin-susceptible S. aureus isolates that were PVL positive and had spa type motif MBQBLO.
The USA300 clone was defined as MBQBLO spa motif positive, PVL positive, and ACME positive.
In total, 18% of patients infected with S. aureus with dysfunctional agr died within 30 days of culture collection, compared to 12% of patients infected with S. aureus with functional agr. Infection with S. aureus with dysfunctional agr was significantly associated with 30-day in-hospital mortality (unadjusted RR = 1.52; 95% CI, 1.05 to 2.21). In contrast, methicillin resistance (i.e., presence of the mecA gene) was not statistically associated with 30-day in-hospital mortality (unadjusted RR = 1.20; 95% CI, 0.83 to 1.73).
We used contingency tables to calculate the sensitivity of the δ-hemolysin assay for predicting mortality. Lack of δ-hemolysin production (i.e., agr dysfunction) was able to predict mortality with a sensitivity of 30%, specificity of 79%, positive predictive value of 18%, and negative predictive value of 88%. There was a nonsignificant trend toward a higher hazard of 30-day in-hospital mortality among patients infected with S. aureus with dysfunctional agr than among patients infected with S. aureus with functional agr after statistically adjusting for severity of illness, age, aggregate comorbidity, methicillin resistance, previous admission in the past year, presence of a central venous catheter, receipt of vancomycin at any time during the index admission, and vancomycin MIC (adjusted HR = 1.34; 95% CI, 0.87 to 2.06; P = 0.18).
Eighty-six percent (700/814 admissions) of the total cohort received vancomycin during the index admission. Ninety percent of the most severely ill patients (modified APS of >28) received vancomycin empirically or definitively. Limiting the cohort to patients who received vancomycin empirically or definitively did not change the magnitude of the adjusted association between loss of agr function and 30-day in-hospital mortality.
When the cohort was stratified into quartiles by severity of illness, 4.1% of patients in quartile 1 (modified APS of <9) died within 30 days of hospitalization, 4.8% of patients in quartile 2 died (modified APS of 10 to 17), 13.9% of patients in quartile 3 died (modified APS of 18 to 28), and 29.6% of patients in quartile 4 died (modified APS of ≥28). Among the patients in the quartile with the highest severity-of-illness scores, patients infected with S. aureus with agr dysfunction had a significantly higher hazard of mortality than patients infected with S. aureus with functional agr after statistically adjusting for vancomycin MIC, receipt of vancomycin, aggregate comorbidity, methicillin resistance, presence of a central venous catheter, previous admission in the past year, and age (adjusted HR = 1.82; 95% CI, 1.03 to 3.21; P = 0.04) (Table 3). Similarly, when we assessed the interaction between a high severity-of-illness score and agr dysfunction among the entire cohort, the adjusted interaction term was statistically significant (P = 0.049).
TABLE 3.
Adjusted association between agr dysfunction and mortality, stratified by severity of illness
| Quartile based on severity of illness | Mortality (%) among patients infected with S. aureus |
Association between agr dysfunction and mortality (adjusted HR [95% CI])a | |
|---|---|---|---|
| Strains with dysfunctional agr | Strains with functional agr | ||
| 4th (modified APS of >28; n = 216 [27% of patients]) | 44.0 | 25.3 | 1.82 (1.03-3.21) |
| 3rd (modified APS of >17 and ≤28; n = 195 [24% of patients]) | 14.8 | 13.5 | 1.01 (0.41-2.48) |
| 2nd (modified APS of >9 and ≤17; n = 208 [26% of patients]) | 4.2 | 5.0 | 0.43 (0.08-2.47) |
| 1st (modified APS of ≤9; n = 195 [24% of patients]) | 3.3 | 4.2 | 0.28 (0.02-3.40) |
Statistically adjusted for vancomycin MIC, receipt of vancomycin, aggregate comorbidity, methicillin resistance, presence of a central venous catheter, previous admission in the past year, and age, using Cox proportional hazard models.
Subcohort analyses were also performed. When repeat admissions were excluded from the cohort, the association between infection with S. aureus with agr dysfunction and mortality remained significant (unadjusted RR = 1.50; 95% CI, 1.03 to 2.19). Another subgroup analysis was performed among 245 patients infected with S. aureus strains with spa types that correlate with the ST5 multilocus sequence type, which is considered to be made up of health care-associated strains. In this subgroup, infection with S. aureus with dysfunctional agr was a significant risk factor for mortality (unadjusted RR = 1.70; 95% CI, 1.02 to 2.82).
DISCUSSION
Although S. aureus bloodstream infections are associated with high mortality in general, specific pathogen and host factors have been found to be predictive of inferior responses to vancomycin therapy and of poor outcomes. This is the first study to demonstrate an independent association between agr dysfunction and higher 30-day mortality in severely ill patients with S. aureus bacteremia.
In our study, agr dysfunction was significantly associated with mortality only among the most severely ill patients. While agr dysfunction was associated with a significant (91%) increase in mortality for the sickest quartile of patients, there was no association for the other quartiles (Table 3). Based on what is known about the effects of agr dysfunction on S. aureus, we hypothesize that this observation has its foundation in two components of the host-pathogen-drug triad. First, it has been determined that S. aureus with a loss of or reduced agr function is not killed as effectively in vitro by vancomycin as S. aureus with preserved agr function. Pharmacodynamic models have shown a decrease in vancomycin bactericidal activity against S. aureus with dysfunctional agr compared to S. aureus with functional agr, and vancomycin kill-curve studies of clinical MRSA isolates stratified by δ-hemolysin production demonstrated significantly less killing by vancomycin of MRSA strains that did not produce δ-hemolysin than of those strains that did produce δ-hemolysin (24, 28, 36, 37). While the mechanism of this phenomenon is not understood, it is believed to depend in part on diminished autolysis of S. aureus with dysfunctional agr. A host who is severely ill may rely on bactericidal therapy in order to survive much more so than a host who is not as sick, with potentially a more intact immune system. While the importance of bactericidal therapy in the prompt eradication of bacteremia has been demonstrated, such analyses have not been done based on severity of host illness and on actual mortality. Since this was a retrospective cohort study, we did not have the means to measure duration of bacteremia but suspect that this effect on patient mortality by strains with agr dysfunction may lie in the longer duration of bacteremia, given the ample data linking duration of bacteremia to increased mortality and agr dysfunction to prolonged bacteremia (2, 7). Testing for agr dysfunction may identify a subgroup of patients who need close vancomycin therapeutic monitoring or should receive other antistaphylococcal antibiotics.
A second component that we believe is affected by agr dysfunction is the pathogenicity of the organism in bloodstream infections. agr dysfunction has been linked to reduced killing by innate host defense cationic peptides from platelets called thrombin-induced platelet microbicidal proteins (tPMPs). Resistance to these peptides has been associated with prolonged bacteremia and an enhanced ability to establish endovascular infection such as endocarditis (7, 25). Fowler et al. compared 21 patients with persistent MRSA bacteremia to 18 patients with resolving MRSA bacteremia and found that patients with persistent bacteremia were significantly more likely to have strains with reduced susceptibility to tPMPs and defective δ-hemolysin production than patients with resolving bacteremia. Therefore, a host with a low severity of illness may be able to overcome conditions of prolonged bacteremia and endocarditis much more effectively, thus ensuring survival, than a host with a high severity of illness.
Recently, much emphasis has been placed on the association between other elements of S. aureus, such as mecA and vancomycin MIC, and clinical outcomes. However, in our bivariable analysis, agr dysfunction was the only characteristic of S. aureus that was associated with patient mortality. Continued primary emphasis on mecA+ S. aureus or on MRSA, in both the clinical and infection prevention senses, may not be an optimal strategy to improve patient outcomes and patient safety going forward. Future studies should assess the ability of agr function testing to predict poor patient outcomes and optimize antibiotic therapy. In addition, if these findings are validated, then infection control efforts might be directed at preventing the transmission of S. aureus with agr dysfunction instead of or in addition to strains that contain mecA.
Potential limitations in the clinical component of this study include its retrospective, observational nature and the fact that it reflects the experience of a single center. Thus, there is a risk of unmeasured confounding. However, a randomized controlled trial would not be ethical because patients cannot be randomized to infection with S. aureus with dysfunctional agr. In addition, as stated above, we did not quantify the duration of bacteremia, a critical determinant of patient mortality in S. aureus bacteremia. With respect to limitations in the laboratory phase of this study, it is possible that agr dysfunction mutations could have occurred in the laboratory rather than in vivo. However, Traber et al. demonstrated that agr mutations, which may lead to agr dysfunction, occur before or during an infection rather than during specimen handling (35). Also, posthandling mutations would be independent of patient mortality and would be expected to bias our results only toward the null hypothesis. Another laboratory limitation is that δ-hemolysin assay may not be as sensitive a marker for agr dysfunction as other laboratory tests, such as Northern blot hybridization to detect RNAIII production (41). When Shopsin et al. measured δ-hemolysin activity as well as performing Northern blotting for RNAIII and assessment of exoprotein profiles, they found that δ-hemolysin activity was a fairly specific marker for agr dysfunction (30). Measurement of δ-hemolysin activity is also easier and less expensive than Northern blotting (27).
In summary, this study demonstrates that agr dysfunction in S. aureus bacteremia isolates is associated independently with excessive mortality among severely ill patients. The δ-hemolysin assay to semiquantitatively examine agr function may be a simple and inexpensive approach for predicting patient outcomes and potentially optimizing antibiotic therapy.
Supplementary Material
Acknowledgments
We thank Colleen Reilly, Jingkun Zhu, and Kristen Schratz for database maintenance and extraction; Jacqueline Baitch and Atlisa Young for medical chart review; and Sandra McCain, Tarah Ranke, Gwen Robinson, Mary Lee, Kezia Alexander, and O. Colin Stine for their assistance with the performance and interpretation of laboratory testing. The agr group III positive control, NRS 143, was provided by the Network of Antimicrobial Resistance in S. aureus (NARSA), supported under NIAID/NIH contract HHSN272200700055C.
This study was funded in part by an investigator-initiated research grant from Pfizer, Inc. (IIR GA5951BG). A. D. Harris was supported by National Institutes of Health grants 1R01A160859-01A1 and 1K24AI079040-01A1. J. P. Furuno was supported by National Institutes of Health grant 1K01AI071015-02. J. K. Johnson was supported by National Institutes of Health grant 1K12RR02350-03. J. C. McGregor was supported by National Institutes of Health grant KL2RR024141. E. N. Perencevich was supported by U.S. Department of Veterans Affairs Health Services Research and Development grants RCD-02-026-2 and IIR-05-123-1. M. D. Shardell was supported by National Institutes of Health grant K12HD043489.
M.L.S. and E.N.P. have received investigator-initiated research support from Pfizer, Inc. M.L.S. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. G.S. is a consultant for Astellas, Cubist, Pfizer, and Ortho-McNeal, has served as a speaker for Astellas, Cubist, and Pfizer, and has received research support from Cubist. J.P.F., A.D.H., J.K.J., M.D.S., J.C.M., and K.A.T. have no conflicts of interest.
Footnotes
Published ahead of print on 20 December 2010.
Supplemental material for this article may be found at http://aac.asm.org/.
REFERENCES
- 1.Andrews, J. M. 2001. Determination of minimum inhibitory concentrations. J. Antimicrob. Chemother. 48(Suppl. 1):5-16. [DOI] [PubMed] [Google Scholar]
- 2.Chang, F. Y., et al. 2003. A prospective multicenter study of Staphylococcus aureus bacteremia: incidence of endocarditis, risk factors for mortality, and clinical impact of methicillin resistance. Medicine (Baltimore) 82:322-332. doi: 10.1097/01.md.0000091185.93122.40. [DOI] [PubMed] [Google Scholar]
- 3.Clinical and Laboratory Standards Institute. 2008. Performance standards for antimicrobial susceptibility testing. CLSI document M100-S19, 19th informational supplement. CLSI, Wayne, PA.
- 4.Cosgrove, S. E., et al. 2005. The impact of methicillin resistance in Staphylococcus aureus bacteremia on patient outcomes: mortality, length of stay, and hospital charges. Infect. Control Hosp. Epidemiol. 26:166-174. [DOI] [PubMed] [Google Scholar]
- 5.Cosgrove, S. E., et al. 2003. Comparison of mortality associated with methicillin-resistant and methicillin-susceptible Staphylococcus aureus bacteremia: a meta-analysis. Clin. Infect. Dis. 36:53-59. [DOI] [PubMed] [Google Scholar]
- 6.Diep, B. A., et al. 2006. Complete genome sequence of USA300, an epidemic clone of community-acquired methicillin-resistant Staphylococcus aureus. Lancet 367:731-739. doi: 10.1016/S0140-6736(06)68231-7. [DOI] [PubMed] [Google Scholar]
- 7.Fowler, V. G., Jr., et al. 2004. Persistent bacteremia due to methicillin-resistant Staphylococcus aureus infection is associated with agr dysfunction and low-level in vitro resistance to thrombin-induced platelet microbicidal protein. J. Infect. Dis. 190:1140-1149. doi: 10.1086/423145. [DOI] [PubMed] [Google Scholar]
- 8.Fujimoto, D. F., and K. W. Bayles. 1998. Opposing roles of the Staphylococcus aureus virulence regulators, agr and sar, in Triton X-100- and penicillin-induced autolysis. J. Bacteriol. 180:3724-3726. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Furuno, J. P., et al. 2007. Value of performing active surveillance cultures on intensive care unit discharge for detection of methicillin-resistant Staphylococcus aureus. Infect. Control Hosp. Epidemiol. 28:666-670. doi: 10.1086/518348. [DOI] [PubMed] [Google Scholar]
- 10.Furuno, J. P., et al. 2005. Methicillin-resistant Staphylococcus aureus and vancomycin-resistant enterococci co-colonization. Emerg. Infect. Dis. 11:1539-1544. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Gilot, P., G. Lina, T. Cochard, and B. Poutrel. 2002. Analysis of the genetic variability of genes encoding the RNA III-activating components agr and TRAP in a population of Staphylococcus aureus strains isolated from cows with mastitis. J. Clin. Microbiol. 40:4060-4067. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Harmsen, D., et al. 2003. Typing of methicillin-resistant Staphylococcus aureus in a university hospital setting by using novel software for spa repeat determination and database management. J. Clin. Microbiol. 41:5442-5448. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Hidron, A. I., et al. 2008. NHSN annual update: antimicrobial-resistant pathogens associated with healthcare-associated infections: annual summary of data reported to the National Healthcare Safety Network at the Centers for Disease Control and Prevention, 2006-2007. Infect. Control Hosp. Epidemiol. 29:996-1011. doi: 10.1086/591861. [DOI] [PubMed] [Google Scholar]
- 14.Johnson, J. K., et al. 2007. Skin and soft tissue infections caused by methicillin-resistant Staphylococcus aureus USA300 clone. Emerg. Infect. Dis. 13:1195-1200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kim, P. W., et al. 2005. Risk of mortality with a bloodstream infection is higher in the less severely ill at admission. Am. J. Respir. Crit. Care Med. 171:616-620. doi: 10.1164/rccm.200407-916OC. [DOI] [PubMed] [Google Scholar]
- 16.Knaus, W. A., et al. 1991. The APACHE III prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults. Chest 100:1619-1636. [DOI] [PubMed] [Google Scholar]
- 17.Lina, G., et al. 1999. Involvement of Panton-Valentine leukocidin-producing Staphylococcus aureus in primary skin infections and pneumonia. Clin. Infect. Dis. 29:1128-1132. [DOI] [PubMed] [Google Scholar]
- 18.McGregor, J. C., et al. 2005. Utility of the chronic disease score and Charlson comorbidity index as comorbidity measures for use in epidemiologic studies of antibiotic-resistant organisms. Am. J. Epidemiol. 161:483-493. doi: 10.1093/aje/kwi068. [DOI] [PubMed] [Google Scholar]
- 19.McGregor, J. C., et al. 2006. Comorbidity risk-adjustment measures were developed and validated for studies of antibiotic-resistant infections. J. Clin. Epidemiol. 59:1266-1273. doi: 10.1016/j.jclinepi.2006.01.016. [DOI] [PubMed] [Google Scholar]
- 20.McGregor, J. C., et al. 2007. A systematic review of the methods used to assess the association between appropriate antibiotic therapy and mortality in bacteremic patients. Clin. Infect. Dis. 45:329-337. doi: 10.1086/519283. [DOI] [PubMed] [Google Scholar]
- 21.Moise, P. A., G. Sakoulas, A. Forrest, and J. J. Schentag. 2007. Vancomycin in vitro bactericidal activity and its relationship to efficacy in clearance of methicillin-resistant Staphylococcus aureus bacteremia. Antimicrob. Agents Chemother. 51:2582-2586. doi: 10.1128/AAC.00939-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Osih, R. B., et al. 2007. Impact of empiric antibiotic therapy on outcomes in patients with Pseudomonas aeruginosa bacteremia. Antimicrob. Agents Chemother. 51:839-844. doi: 10.1128/AAC.00901-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Renzoni, A., et al. 2004. Modulation of fibronectin adhesins and other virulence factors in a teicoplanin-resistant derivative of methicillin-resistant Staphylococcus aureus. Antimicrob. Agents Chemother. 48:2958-2965. doi: 10.1128/AAC.48.8.2958-2965.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Rose, W. E., M. J. Rybak, B. T. Tsuji, G. W. Kaatz, and G. Sakoulas. 2007. Correlation of vancomycin and daptomycin susceptibility in Staphylococcus aureus in reference to accessory gene regulator (agr) polymorphism and function. J. Antimicrob. Chemother. 59:1190-1193. doi: 10.1093/jac/dkm091. [DOI] [PubMed] [Google Scholar]
- 25.Sakoulas, G., et al. 2005. Reduced susceptibility of Staphylococcus aureus to vancomycin and platelet microbicidal protein correlates with defective autolysis and loss of accessory gene regulator (agr) function. Antimicrob. Agents Chemother. 49:2687-2692. doi: 10.1128/AAC.49.7.2687-2692.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Sakoulas, G., et al. 2003. Staphylococcus aureus accessory gene regulator (agr) group II: is there a relationship to the development of intermediate-level glycopeptide resistance? J. Infect. Dis. 187:929-938. [DOI] [PubMed] [Google Scholar]
- 27.Sakoulas, G., et al. 2002. Accessory gene regulator (agr) locus in geographically diverse Staphylococcus aureus isolates with reduced susceptibility to vancomycin. Antimicrob. Agents Chemother. 46:1492-1502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Sakoulas, G., et al. 2006. Effects of prolonged vancomycin administration on methicillin-resistant Staphylococcus aureus (MRSA) in a patient with recurrent bacteraemia. J. Antimicrob. Chemother. 57:699-704. doi: 10.1093/jac/dkl030. [DOI] [PubMed] [Google Scholar]
- 29.Schweizer, M. L., et al. 2 July 2010. Empiric therapy for Staphylococcus aureus bacteremia may not reduce in-hospital mortality. PLoS One 5:e11432. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Shopsin, B., et al. 2008. Prevalence of agr dysfunction among colonizing Staphylococcus aureus strains. J. Infect. Dis. 198:1171-1174. doi: 10.1086/592051. [DOI] [PubMed] [Google Scholar]
- 31.Shopsin, B., et al. 1999. Evaluation of protein A gene polymorphic region DNA sequencing for typing of Staphylococcus aureus strains. J. Clin. Microbiol. 37:3556-3563. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Sunenshine, R. H., et al. 2007. Multidrug-resistant Acinetobacter infection mortality rate and length of hospitalization. Emerg. Infect. Dis. 13:97-103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Tenover, F. C., and R. C. Moellering, Jr. 2007. The rationale for revising the Clinical and Laboratory Standards Institute vancomycin minimal inhibitory concentration interpretive criteria for Staphylococcus aureus. Clin. Infect. Dis. 44:1208-1215. doi: 10.1086/513203. [DOI] [PubMed] [Google Scholar]
- 34.Traber, K., and R. Novick. 2006. A slipped-mispairing mutation in AgrA of laboratory strains and clinical isolates results in delayed activation of agr and failure to translate delta- and alpha-haemolysins. Mol. Microbiol. 59:1519-1530. doi: 10.1111/j.1365-2958.2006.04986.x. [DOI] [PubMed] [Google Scholar]
- 35.Traber, K. E., et al. 2008. agr function in clinical Staphylococcus aureus isolates. Microbiology 154:2265-2274. doi: 10.1099/mic.0.2007/011874-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Tsuji, B. T., Y. Harigaya, A. J. Lesse, G. Sakoulas, and J. M. Mylotte. 2009. Loss of vancomycin bactericidal activity against accessory gene regulator (agr) dysfunctional Staphylococcus aureus under conditions of high bacterial density. Diagn. Microbiol. Infect. Dis. 64:220-224. doi: 10.1016/j.diagmicrobio.2009.01.028. [DOI] [PubMed] [Google Scholar]
- 37.Tsuji, B. T., M. J. Rybak, K. L. Lau, and G. Sakoulas. 2007. Evaluation of accessory gene regulator (agr) group and function in the proclivity towards vancomycin intermediate resistance in Staphylococcus aureus. Antimicrob. Agents Chemother. 51:1089-1091. doi: 10.1128/AAC.00671-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Villaruz, A. E., et al. 2009. A point mutation in the agr locus rather than expression of the Panton-Valentine leukocidin caused previously reported phenotypes in Staphylococcus aureus pneumonia and gene regulation. J. Infect. Dis. 200:724-734. doi: 10.1086/604728. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Von Korff, M., E. H. Wagner, and K. Saunders. 1992. A chronic disease score from automated pharmacy data. J. Clin. Epidemiol. 45:197-203. [DOI] [PubMed] [Google Scholar]
- 40.Walsh, T. R., et al. 2001. Evaluation of current methods for detection of staphylococci with reduced susceptibility to glycopeptides. J. Clin. Microbiol. 39:2439-2444. doi: 10.1128/JCM.39.7.2439-2444.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Wright, J. S., III, et al. 2005. The agr radiation: an early event in the evolution of staphylococci. J. Bacteriol. 187:5585-5594. doi: 10.1128/JB.187.16.5585-5594.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.