Abstract
Rationale & Objective:
Staphylococcus aureus (S. aureus) bacteremia (SAB) is associated with morbidity and mortality in patients receiving maintenance hemodialysis (HD). We evaluated changes in clinical and bacterial characteristics, and their associations with clinical outcomes following SAB in this population over a 21-year period.
Study Design:
Prospective cohort study.
Setting & Participants:
453 hospitalized, non-neutropenic, adults receiving maintenance HD who developed monomicrobial SAB between 1995 and 2015.
Exposures:
Clinical characteristics and bacterial genotype.
Outcomes:
All-cause and SAB-attributable mortality, persistent bacteremia, and metastatic infection complications.
Analytical Approach:
Proportions of participants experiencing each outcome were calculated overall and by calendar year. Secular trends were estimated using binomial risk regression, a generalized linear model with the log link function for a binomial outcome. Associations with outcomes were estimated using logistic regression.
Results:
Over the 21-year study period, patients receiving maintenance HD experienced significant increases in age- and diabetes-adjusted SAB-attributable mortality (0.45% per year, 95% confidence interval [CI] 0.36, 0.46), persistent bacteremia (0.86% per year, 95% CI 0.14, 1.55), metastatic infection complications (0.84% per year, 95% CI 0.11, 1.56), and infection with the virulent S. aureus clone USA300 (1.47% per year, 95% CI 0.33, 2.52). Over time, the suspected source of SAB was less likely to be a central venous catheter (−1.32% per year, 95% CI −2.05, −0.56) or arteriovenous (AV) graft (−1.08% per year, 95% CI −1.54, −0.56), but more likely to be a non-vascular access source (1.89% per year, 95% CI 1.29, 2.43). Patients with a non-vascular access suspected source of infection were more likely to die as a result of their S. aureus infection (Odds Ratio [OR] =3.20, 95% CI 1.36, 7.55). The increase in USA300 infections may have contributed to the observed increase in persistent bacteremia (OR=2.96, CI 1.12, 7.83) but did not explain the observed increases in SAB-attributable mortality (OR=0.83, CI 0.19, 3.61) or metastatic complications (OR=1.34, CI 0.53, 3.41).
Limitations:
Single-center, inpatient cohort.
Conclusions:
The clinical and molecular epidemiology of SAB in HD-dependent patients has changed over time, with an increase in SAB-attributable mortality and morbidity despite a decline in catheter-related infections.
Keywords: Staphylococcus aureus, bacteremia, hemodialysis, metastatic complications, MRSA
Graphical Abstract
Plain-Language Summary
Staphylococcus aureus bacteremia (SAB) is a leading cause of morbidity and mortality in patients with kidney failure on hemodialysis. Clinical characteristics and characteristics of the Staphylococcus aureus isolates have changed over time, but it is unknown how these changes have affected patient outcomes. In this study, we examined these changing characteristics and their associations with clinical outcomes among 453 hemodialysis-dependent patients hospitalized with SAB over a period of 21 years. The most common suspected sources of infection changed over time. Additionally, virulent strains of Staphylococcus aureus became more prevalent. These changes were associated with worse outcomes, including SAB-attributable mortality, persistent bacteremia, and tissue infections, all of which increased over time.
INTRODUCTION:
Infection is the second leading cause of mortality in patients with kidney failure on hemodialysis (HD), with bacteremia accounting for the majority of infections.1-3 Compared to the general population, infection-related mortality is up to 100-times greater in patients who receive dialysis.4, 5 Staphylococcus aureus (S. aureus) is the most common cause of bacteremia in HD-dependent patients, accounting for >30% of all bloodstream infections.6, 7 S. aureus bacteremia (SAB) can lead to severe complications, including endocarditis, osteomyelitis, septic arthritis, and abscesses.1, 8-10 Importantly, bacterial genotype plays a critical role in pathogenesis, as methicillin-resistant S. aureus (MRSA) infections are associated with significantly higher mortality than methicillin-susceptible S. aureus (MSSA) infections.11, 12 MRSA was first isolated in the 1960s13 and over the past 50 years, nosocomial and community-acquired MRSA infection rates among HD-dependent patients have increased14, accounting for >40% of all S. aureus samples isolated from outpatient HD centers in 2014.7
The primary objective of this study is to describe the clinical and molecular epidemiology of SAB among HD-dependent patients over a two-decade period at our institution. To accomplish this objective, we evaluated the HD-dependent subgroup of our recently published manuscript involving over 2,300 prospectively enrolled patients with SAB.15 A secondary objective of this study was to describe differences in characteristics and outcomes between HD-dependent and non-HD-dependent patients in the cohort.
METHODS
Database & Study Population
The S. aureus Bacteremia Group Prospective Cohort Study (SABG-PCS) is an ongoing, prospective cohort at Duke University Medical Center (DUMC). The overall cohort, study design, and ascertainment strategies have been described previously.15 SABG-PCS and the present study were approved by the Duke Institutional Review Board. Patients or their legal representatives provided written informed consent. If patients expired prior to notification of their blood culture results, they were included in the SABG using a Notification of Decedent Research.
Adults hospitalized with monomicrobial SAB at DUMC were eligible for enrollment in the SABG-PCS if they did not meet any of the following exclusion criteria: <18 years of age; neutropenic (absolute neutrophil count ≤1 × 109/L); had no signs or symptoms of infection; had second clinically significant bacterial pathogen other than S. aureus isolated from their blood culture; non-English speaking; or if they declined to participate.
Inclusion and Exclusion Criteria
Only patients whose initial bloodstream S. aureus isolate was available for further testing were included in the current analysis, and only the index hospitalization was considered. In our analyses comparing HD- and non-HD-dependent patients, all eligible SABG-PCS patients were included. HD-dependent patients were defined as requiring maintenance HD prior to confirmation of bacteremia. For the remaining analyses, only HD-dependent patients were included.
Definitions
SAB was determined to be either hospital- or community-acquired, according to criteria previously defined.16 Community-acquired SAB was then further categorized as either community-acquired, non-healthcare-associated (CA-NHCA) or community-acquired, healthcare–associated (CA-HCA).16 HD-dependent patients with kidney transplants were confirmed to be on HD and have non-functioning kidney transplants at the time of enrollment. Additional types of organ transplants included lung, heart, pancreas, liver, and bone marrow. Suspected sources of infection included central venous catheter (CVC), arteriovenous graft (AVG), arteriovenous fistula (AVF), or non-vascular access, with the latter defined as an infection that was determined not to originate in a CVC, AVG, or AVF (see footnote of Table 1 for more information on non-vascular access sources). Suspected sources of infection were determined based on the best judgment of the individuals who collected the data in the case report form, in addition to two independent reviewers who adjudicated the data prior to it being imported into the database. One of the independent reviewers was an infectious disease physician. These judgments were informed based on clinical progress notes from the providers who took care of the patients (including nephrologists and infectious disease physicians), and was often supported by imaging findings. At our center, only patients undergoing home HD were cannulated via the buttonhole technique, while patients undergoing in-center HD were always cannulated via the rope-ladder technique. Among patients who had a suspected AVF source of infection, only one was on home HD and had their AVF cannulation performed via the buttonhole technique. Designated clinical outcomes were all-cause mortality, SAB-attributable mortality, persistent bacteremia, and metastatic complications. SAB-attributable mortality was defined as death due to SAB and included all patients who died with persistent signs or symptoms of systemic infection, positive blood culture results, or a persistent focus of infection in the absence of another explanation for death. Outcomes were tracked for 90 days after enrollment into the SABG-PCS. A metastatic complication was defined as a S. aureus infection resulting from the spread from an initial site, either via direct extension or bloodstream seeding, and included any of the following conditions: a metastatic abscess or another deep tissue abscess (e.g. psoas or epidural abscess), infective endocarditis 17, vertebral osteomyelitis, septic arthritis, septic emboli, or septic thrombophlebitis. Persistent bacteremia was defined as ≥5 days of positive blood cultures after the initiation of an appropriate treatment regimen.15 Antibiotic therapy was considered to be appropriate if it included at least one antibiotic with in vitro activity against the patient’s bloodstream S. aureus isolate.
Table 1:
Demographics and Clinical Characteristics in Hemodialysis (HD)-Dependent vs Non-HD-Dependent Patients
| HD- Dependent |
Non-HD- Dependent |
P- value |
|
|---|---|---|---|
| n=453 | n=1862 | ||
| Age, median (Q1-Q3) a | 57 (47,67) | 60 (47,71) | 0.01 |
| Female (%) | 212 (46.8) | 784 (42.1) | 0.08 |
| Race (%) | <0.001 | ||
| Black | 347 (76.9) | 484 (26.2) | |
| White | 94 (20.8) | 1311 (70.9) | |
| Other | 10 (2.2) | 54 (2.9) | |
| Route (%) | <0.001 | ||
| HAb | 41 (9.1) | 672 (36.1) | |
| CA-HCAc | 412 (90.9) | 964 (51.8) | |
| CA-NHCAd | 0 (0) | 225 (12.1) | |
| DM (%) | 258 (57.0) | 631 (33.9) | <0.001 |
| Neoplasm (%) | 26 (5.8) | 434 (23.3) | <0.001 |
| Corticosteroid Use (%) | 45 (9.9) | 400 (21.5) | <0.001 |
| Transplante (%) | 34 (7.5) | 124 (6.7) | 0.5 |
| HIV (%) | 11 (2.4) | 55 (3.0) | 0.6 |
| Surgery w/in 30 daysf (%) | 73 (16.2) | 563 (30.4) | <0.001 |
| Previous Endocarditis (%) | 22 (4.9) | 49 (2.6) | 0.02 |
| MRSA (%) | 206 (45.5) | 898 (48.2) | 0.3 |
| USA300 (%) | 24 (5.3) | 143 (7.7) | 0.09 |
| Foreign Body Cardiac Device (%) | 44 (9.7) | 260 (14.0) | 0.02 |
| Foreign Body Central Venous Catheter (%) | 280 (61.8) | 265 (14.2) | <0.001 |
| Source of Infection Central Venous Catheterg (%) | 248 (54.9) | 331 (17.9) | <0.001 |
| Source of Infection Arteriovenous Graftg (%) | 85 (18.8) | 7 (0.4) | <0.001 |
| Source of Infection Arteriovenous Fistulag (%) | 21 (4.6) | 1 (0.1) | <0.001 |
| Source of Infection Non-Vascular Accessg,h (%) | 98 (21.7) | 1515 (81.7) | <0.001 |
P-values marked in bold are significant.
Q = quartiles
Hospital-acquired
Community-acquired, healthcare-associated
Community-acquired, non-healthcare-associated
27 non-functioning kidney transplants in HD-dependent cohort and 25 functioning kidney transplants in non-HD-dependent cohort
Within 30 days of index positive blood culture
Suspected source of infection
Includes: abscess, septic arthritis, biliary tract, burn, arterial catheter, intra-aortic balloon pump, peripheral catheter, other non-central venous catheter, cellulitis, decubitus ulcer, dermatitis, empyema, foot/leg ulcer, furuncle, gangrene, intravenous drug use, mediastinitis, percutaneous nephrostomy, peritoneal dialysis catheter, pneumonia, sinusitis, wound, and unknown/unspecified source of infection.
Susceptibility Testing & Molecular Typing
All S. aureus isolates underwent methicillin susceptibility testing and spa typing, as previously described.15 MRSA isolates were identified as USA300 if they exhibited a spa-CC08 (multilocus sequence type [MLST] CC008) genotype and were positive by PCR for Arginine Catabolic Mobile Element (ACME) and Panton-Valentine Leukocidin (PVL)-encoding genes.14, 18-21
Statistical Analysis
Patient and bacterial characteristics and clinical outcomes were summarized using medians/quartiles or counts/percentages. For comparisons between groups, statistical significance was evaluated with Mann-Whitney-U or Fisher’s Exact test. For the primary objective, the proportion of HD-dependent participants with each clinical or bacterial characteristic or experiencing each outcome was calculated overall and by calendar year. Secular trends in proportions were estimated with binomial outcome risk regression,22 fit with an independent variable calendar year as a linear, which estimates average percent change in proportion over the entire study period. An assumption was made that the denominator for all outcome proportions was equal, disregarding the competing risk of death and early discharge not attributable to SAB. Lowess curves for each of the characteristics and outcomes examined have been constructed. For two characteristics, suspected source of infection AVF and USA300 infection, secular trends were summarized only for the period 2004-2015 as there were no occurrences prior to 2004. In sensitivity analyses, we examined secular trends in suspected source of infection, bacterial characteristics, and outcomes over this same truncated time period. Univariable and multivariable logistic regression models were used to assess adjusted associations of clinical characteristics and S. aureus genotypes with mortality (both all-cause and SAB-attributable), persistent bacteremia, and metastatic complications. Analyses were performed with SAS 9.4 (SAS Institute, Cary, NC).
RESULTS
Study Population
Between January 1, 1995 and December 31, 2015, a total of 2,423 unique patients were enrolled in the SABG-PCS. A total of 53 patients were excluded as their initial bloodstream isolates could not be retrieved, another 22 were excluded due to inconsistent data regarding clinical outcomes, and 7 were excluded due to inconsistent data entry for key covariates. For analyses focused on HD-dependent patients, we excluded an additional 26 participants who required kidney replacement therapy during the index hospitalization but who were not on maintenance HD prior to confirmation of bacteremia. Of the 2,315 patients included in the current analysis, 453 (19.6%) were found to be HD-dependent.
Baseline characteristics of included SABG-PCS participants are summarized in Table 1. Among HD-dependent patients, the suspected source of bacteremia was the vascular access in 78.3% (CVCs 54.9%, AVG 18.8%, and AVF 4.6%). Over the truncated 2004-2015 time period, the suspected source of bacteremia in HD-dependent patients was the vascular access in 68.6% (Table S1).
Clinical Outcomes
HD-dependent patients were more likely than non-HD-dependent patients to have persistent bacteremia (30.9% vs 20.0%, P<0.001) (Table 2A). This finding was consistent when comparing HD- and non-HD-dependent patients with MRSA bacteremia and when examining the truncated 2004-2015 time period (Table 2B, Tables S2A and S2B).
Table 2A:
Clinical Outcomes in Hemodialysis (HD)-Dependent vs Non-HD-Dependent Patients
| HD-Dependent | Non-HD-Dependent | P-value | |
|---|---|---|---|
| n=453 | n=1862 | ||
| All-Cause Mortality (%) | 106 (23.4) | 505 (27.1) | 0.1 |
| SAB a -Attributable Mortality (%) | 60 (13.2) | 253 (13.6) | 0.9 |
| Persistent Bacteremia (%) | 140 (30.9) | 373 (20.0) | <0.001 |
| Metastatic Complications (Overall) (%) | 160 (35.4) | 692 (37.2) | 0.5 |
| Metastatic Abscess (%) | 31 (6.8) | 164 (8.8) | 0.2 |
| Psoas Abscess (%) | 5 (1.1) | 41 (2.2) | 0.2 |
| Epidural Abscess (%) | 5 (1.1) | 64 (3.5) | 0.01 |
| Endocarditis (%) | 78 (17.2) | 237 (12.8) | 0.02 |
| Vertebral Osteomyelitis (%) | 16 (3.5) | 89 (4.8) | 0.3 |
| Septic Arthritis (%) | 25 (5.5) | 137 (7.4) | 0.2 |
| Septic Emboli (%) | 32 (7.1) | 105 (5.7) | 0.3 |
| Septic Thrombophlebitis (%) | 28 (6.2) | 67 (3.6) | 0.02 |
P-values marked in bold are significant.
Staphylococcus aureus-bacteremia
Table 2B:
Clinical Outcomes in Hemodialysis (HD)-Dependent vs Non-HD-Dependent Patients with MRSA Bacteremia
| HD-Dependent | Non-HD-Dependent | P-value | |
|---|---|---|---|
| n=206 | n=898 | ||
| All-Cause Mortality (%) | 65 (31.6) | 288 (32.1) | 0.9 |
| SAB a -Attributable Mortality (%) | 35 (17.0) | 163 (18.2) | 0.8 |
| Persistent Bacteremia (%) | 91 (44.2) | 235 (26.2) | <0.001 |
| Metastatic Complications (Overall) (%) | 84 (40.8) | 327 (36.5) | 0.3 |
| Metastatic Abscess (%) | 16 (7.8) | 86 (9.6) | 0.5 |
| Psoas Abscess (%) | 4 (1.9) | 19 (2.1) | 0.9 |
| Epidural Abscess (%) | 4 (1.9) | 25 (2.8) | 0.6 |
| Endocarditis (%) | 41 (19.9) | 114 (12.7) | 0.01 |
| Vertebral Osteomyelitis (%) | 9 (4.4) | 36 (4.0) | 0.9 |
| Septic Arthritis (%) | 9 (4.4) | 55 (6.1) | 0.4 |
| Septic Emboli (%) | 15 (7.3) | 56 (6.3) | 0.6 |
| Septic Thrombophlebitis (%) | 18 (8.7) | 34 (3.8) | 0.01 |
P-values marked in bold are significant.
Staphylococcus aureus-bacteremia
Changes in HD-Dependent Population over Time
Over the period from 1995 to 2015, the suspected source of SAB changed, with significant reductions in suspected CVC (annual reduction −1.32%; 95% CI −2.05, −0.56) or AVG infection (annual reduction −1.08%; 95% CI −1.54, −0.56) and significant increases in non-vascular access sources (annual increase 1.89%; 95% CI 1.29, 2.43) (Figure 1). Lowess curves for other clinical characteristics are presented in Figure S1. In sensitivity analysis examining suspected source of SAB, considering only the period from 2004 to 2015, when suspected AVF infections and USA300 infections were first observed, these trends were qualitatively similar to the overall time period, although no longer statistically significant (Figure S2).
Figure 1.
(A-D): Secular Trends in Suspected Source of Infection in Hemodialysis (HD)-Dependent Patients
While all-cause mortality did not change significantly during the study period, the prevalence of SAB-attributable mortality increased significantly (age and diabetes-adjusted annual increase 0.45%; 95% CI 0.36, 0.46). The adjusted frequency of persistent bacteremia (adjusted annual increase 0.86%; 95% CI 0.14, 1.55) and metastatic complications (adjusted annual increase 0.84%; 95% CI 0.11, 1.56) also increased during the study period (Figure 2). When outcomes were examined among only patients with MRSA, all-cause mortality decreased over time (adjusted annual decrease −1.22%; 95% CI −1.35, −0.22), while SAB-attributable mortality increased (adjusted annual increase 0.04%; 95% CI 0.01, 1.21 (Figure 3). Sensitivity analyses examining these outcomes for the 2004-2015 period are shown in Figures S3 and S4. Frequencies and proportions of outcomes by year are shown in Table S3.
Figure 2.
(A-D): Secular Trends in All-Cause Mortality, Staphylococcus aureus bacteremia (SAB)-Attributable Mortality, Persistent Bacteremia, and Metastatic Complications in Hemodialysis (HD)-Dependent Patients
Figure 3.
(A-D): Secular Trends in All-Cause Mortality, Staphylococcus aureus bacteremia (SAB)-Attributable Mortality, Persistent Bacteremia, and Metastatic Complications in Hemodialysis (HD)-Dependent Patients with Methicillin Resistant Staphylococcus aureus (MRSA) Bacteremia
Molecular Genotyping in HD-Dependent Patients
Previously, we found that a specific bacterial genotype, the epidemic community-associated MRSA clone termed USA300, was significantly associated with increasing rates of persistent bacteremia and metastatic complications.15 Thus, in the current study we tested the hypothesis that increasing rates of complications among HD-dependent patients might be due to the emergence and establishment of the virulent USA300 clone of MRSA in that population. From 1995-2015, the genotypes of S. aureus changed significantly among HD-dependent patients in the cohort. The prevalence of MRSA increased significantly in the HD-dependent population over the 21-year study period (1.44% annual increase; 95% CI 0.67, 2.17) (Figure 4A), although it remained consistent from 2004-2015 (Figure S5A). While the USA300 genotype only emerged as a cause of SAB in the HD-dependent cohort in 2004, the prevalence of USA300 increased significantly in the HD-dependent population from 2004-2015 (1.47% annual increase; 95% CI 0.33, 2.52) (Figure 4B, S5B). Despite the increase in prevalence, only 24 (5.3%) of the 453 HD-dependent patients had SAB due to USA300 infection during the study period (Table 1).
Figure 4.
(A-B): Secular Trends in Bacterial Characteristics in Hemodialysis (HD)-Dependent Patients
In univariable analysis comparing outcomes in HD-dependent patients with USA300 vs. non-USA300 bacteremia, USA300 infection was not found to be significantly associated with any of the outcomes (Table 3). In multivariable analysis, USA300 infection was found to be associated with persistent bacteremia when compared to MSSA infection (odds ratio [OR] 2.96; 95% CI 1.12, 7.83). No significant associations between USA300 infection and mortality or metastatic complications were identified in multivariable analysis. Non-USA300 MRSA was found to be associated with all-cause mortality (OR 2.27; 95% CI 1.36, 3.80), persistent bacteremia (OR 3.55; 95% CI 2.21, 5.69), and metastatic complications (OR 1.56; 95% CI 1.01, 2.43) (Table 4).
Table 3:
Univariable Analysis of Associations between Genotypes (USA300 vs non-USA300) and Outcomes in Hemodialysis (HD)-Dependent Patients with Staphylococcus aureus Bacteremia
| Genotype | USA300 | Non-USA300 | P-Value |
|---|---|---|---|
| n (%) with outcome | |||
| All-Cause Mortality | 3 (12.5) | 103 (24.0) | 0.32 |
| SABa-Attributable Mortality | 3 (12.5) | 57 (13.3) | 0.9 |
| Persistent Bacteremia | 10 (41.7) | 130 (30.3) | 0.26 |
| Metastatic Complications (Overall) | 11 (45.8) | 149 (34.8) | 0.28 |
P-values were calculated based on Fisher’s exact test.
Staphylococcus aureus bacteremia
Table 4:
Multivariable Analysis of Characteristics Associated with Mortality, Persistent Bacteremia, and Metastatic Complications in Hemodialysis (HD)-Dependent Patients
| All-Cause Mortalitya |
SABb-Attributable Mortalitya |
Persistent Bacteremiaa |
Metastatic Complicationsa |
|
|---|---|---|---|---|
| # of Events (%) | 106 (23.4) | 60 (13.2) | 140 (30.9) | 160 (35.3) |
| Variable | Odds Ratio (95% Confidence Interval) | |||
| Genotype | ||||
| MSSA | Reference | |||
| non-USA300 MRSA | 2.27 (1.36, 3.80) | 1.78 (0.95, 3.36) | 3.55 (2.21, 5.69) | 1.56 (1.01, 2.43) |
| USA300 | 0.44 (0.10, 1.88) | 0.83 (0.19, 3.61) | 2.96 (1.12, 7.83) | 1.34 (0.53, 3.41) |
| Age | ||||
| <45 Years | Reference | |||
| 45-54 Years | 1.46 (0.59, 3.63) | 1.49 (0.44, 5.14) | 1.23 (0.60, 2.54) | 1.10 (0.57, 2.11) |
| 55-64 Years | 2.30 (0.99, 5.22) | 2.41 (0.80, 7.23) | 1.34 (0.66, 2.70) | 0.95 (0.50, 1.82) |
| 65+ Years | 2.65 (1.18, 5.95) | 3.42 (1.17, 10.03) | 1.74 (0.88, 3.43) | 1.12 (0.60, 2.08) |
| Gender | ||||
| Male | Reference | |||
| Female | 1.04 (0.63, 1.72) | 0.94 (0.51, 1.74) | 1.01 (0.65, 1.58) | 0.77 (0.50, 1.16) |
| Race | ||||
| Non-Hispanic White | Reference | |||
| Black | 0.71 (0.40, 1.28) | 0.49 (0.25, 0.94) | 1.58 (0.90, 2.80) | 0.82 (0.49, 1.37) |
| Other | 2.04 (0.43, 9.76) | 2.33 (0.43, 12.49) | 6.62 (1.52, 28.91) | 2.53 (0.64, 9.96) |
| Comorbidities | ||||
| Diabetes Mellitus | 0.99 (0.59, 1.68) | 0.86 (0.46, 1.61) | 1.08 (0.68, 1.73) | 1.44 (0.92, 2.26) |
| Neoplasm | 1.91 (0.77, 4.78) | 0.64 (0.17, 2.45) | 0.44 (0.15, 1.27) | 0.22 (0.07, 0.73) |
| Transplantation | 0.67 (0.20, 2.17) | 0.54 (0.11, 2.76) | 0.91 (0.35, 2.35) | 0.46 (0.18, 1.16) |
| Corticosteroid use | 0.72 (0.27, 1.95) | 0.61 (0.16, 2.37) | 2.39 (1.11, 5.18) | 3.66 (1.72, 7.80) |
| Foreign Body Central Venous Catheter | 1.65 (0.75, 3.60) | 1.27 (0.52, 3.12) | 0.93 (0.45, 1.92) | 0.84 (0.42, 1.66) |
| Site of acquisition | ||||
| HAc | Reference | |||
| CA-HCAd | 0.25 (0.12, 0.53) | 0.39 (0.17, 0.88) | 0.89 (0.42, 1.87) | 1.82 (0.85, 3.93) |
| Source of Infection e | ||||
| Central Venous Catheter | Reference | |||
| Arteriovenous graft | 1.03 (0.38, 2.80) | 1.69 (0.54, 5.30) | 1.20 (0.50, 2.87) | 0.70 (0.31, 1.62) |
| Arteriovenous fistula | 1.36 (0.28, 6.66) | 2.65 (0.50, 13.90) | 0.30 (0.07, 1.35) | 1.51 (0.48, 4.73) |
| Non-Vascular Accessf | 3.34 (1.60, 6.97) | 3.20 (1.36, 7.55) | 2.11 (1.05, 4.22) | 1.51 (0.79, 2.91) |
Results for each variable are adjusted for all other variables shown. Odds ratios and confidence intervals marked in bold are significant
Staphylococcus aureus bacteremia
Hospital-acquired
Community-acquired, healthcare associated
Suspected source of infection
Includes: abscess, septic arthritis, biliary tract, burn, arterial catheter, intra-aortic balloon pump, peripheral catheter, other non-central venous catheter, cellulitis, decubitus ulcer, dermatitis, empyema, foot/leg ulcer, furuncle, gangrene, intravenous drug use, mediastinitis, percutaneous nephrostomy, peritoneal dialysis catheter, pneumonia, sinusitis, wound, and unknown/unspecified source of infection
Patient Characteristics and Clinical Outcomes among HD-Dependent Patients
Next, we considered potential associations between patient characteristics and clinical outcomes. Age >65 years was strongly associated with both all-cause (OR 2.65; 95% CI 1.18, 5.95) and SAB-attributable (OR 3.42; CI 1.17, 10.03) mortality. Black patients were less likely to experience SAB-attributable mortality compared to non-Hispanic White patients (OR 0.49; CI 0.25, 0.94). Corticosteroid use was associated with persistent bacteremia (OR 2.39; CI 1.11-5.18) and metastatic complications (OR 3.66; CI 1.72-7.80). The presence of a foreign body CVC was not found to be significantly associated with any of the outcomes. As compared to HD-dependent patients with a CVC as the suspected source of infection, HD-dependent patients with a suspected non-vascular access source were significantly more likely to suffer overall mortality (OR 3.34; 95% CI: 1.60, 6.97), to die as a result of their S. aureus infection (OR 3.20; 95% CI 1.36, 7.55), and to have persistent bacteremia (OR 2.11; 95% CI 1.05, 4.22) (Table 4).
DISCUSSION
This study demonstrates that the clinical and molecular epidemiology of SAB in HD-dependent patients has changed significantly over the past two decades. These findings have several key implications.
The suspected source of infection changed during the study period. Fewer patients had a CVC or AVG suspected source of infection, while the frequency of a non-vascular access suspected source of infection increased. The decrease in suspected CVC-associated infections may reflect the impact of the Fistula First Breakthrough Initiative (FFBI), which sought to increase the prevalence of AVFs in HD-dependent patients, while reducing CVC and AVG use.23 This initiative came about in part because of the greatly increased risk of infection in patients who dialyze via a CVC.24 Interestingly, suspected AVF sources of infection only emerged in our cohort in the year 2004, after the implementation of the FFBI in 2003. It is possible that this could be related to process-of-care issues that occurred due to increasing AVF use after the FFBI was implemented, but more research is needed to fully elucidate this relationship.
While previous reports have noted the presence of foreign body CVC as being associated with all-cause mortality25, this trend was not seen in our study. Importantly, however, there are underlying patient factors that may predispose patients to getting a CVC placed as opposed to an AVF for primary HD access, some of which have been shown to contribute to mortality.26 We found a higher likelihood of all-cause and SAB-attributable mortality among patients whose suspected source of infection was not their HD access. This finding may be due in part to delays in source control for non-vascular access sources of infection. The increase in suspected non-vascular access sources of infection may explain in part the increased prevalence of SAB-attributable mortality over time. The cause of the increased prevalence of persistent bacteremia and metastatic complications is likely to be multifactorial and related to both changes in patient characteristics and molecular epidemiology of SAB.
Nearly three-quarters of the HD-dependent patients with SAB in our study were Black. Interestingly, among Black patients with SAB, there was a significant association with decreased SAB-attributable mortality. This finding is consistent with the well-described “survival paradox,” with older Black patients on HD actually having lower mortality than older White patients on HD27,28, and with some29 but not all30 previous reports of outcomes among patients with SAB. However, more research is needed to better clarify potential interactions between race and clinical outcome among HD-dependent patients with SAB and other bloodstream infections.31
Consistent with previous studies, we noticed a significant association between MRSA and adverse outcomes.11, 12, 32 The prevalence of bloodstream infection caused by the USA300 strain of MRSA increased in the HD-dependent cohort during the study period. Although a few studies have evaluated shifts in molecular epidemiology of S. aureus among HD-dependent patients,33-35 they were limited by a retrospective study design, small patient population, and/or short study period. Consistent with what was seen in the overall cohort15, our study demonstrated that USA300 was strongly associated with the development of persistent bacteremia in HD-dependent patients. However, among the HD-dependent patients, no significant independent associations were found between bacteremia with USA300 and metastatic complications. It is possible that our study was underpowered to account for this difference, as only 24 HD-dependent patients were infected with USA300 over the study period. It is also possible that HD-dependent patients with SAB are diagnosed sooner and receive earlier treatment and source control than non-HD-dependent patients given their thrice-weekly HD regimen and frequent use of empiric antibiotics for suspected bacteremia. Further study is needed with adequately sized cohorts to better establish the relationship between bacterial genotype and clinical outcomes among HD-dependent patients with SAB.
Our study has limitations. The number of HD-dependent patients who experienced SAB due to USA300 was small, limiting the power to detect statistically significant differences. Our data come from patients hospitalized at a single academic referral center. Thus, our data may not fully generalize to other regions of the country or to HD-dependent patients who are treated with SAB in the outpatient setting or in community hospitals. With an increase in the outpatient management of bacteremia in HD-dependent patients in recent years, it is possible that only HD-dependent patients with more severe SAB infections required hospitalization in later years of the study, which may have contributed to the worsening clinical outcomes we observed over time. While suspected source of infection was determined based on clinical notes and imaging findings reviewed by the data collector and verified by two independent clinical adjudicators, strict Infectious Diseases Society of America36 and Kidney Health Initiative37 guidelines on catheter-related blood stream infections (CRBSI) and arteriovenous graft or fistula associated blood stream infections were not met. These updated guidelines were published in 2009 and 2018, respectively, while data collection for this cohort began in 1995. Due to missing time to event data for events other than death, we were unable to censor patients in our analyses who may have died from causes other than SAB prior to experiencing SAB-attributable mortality, persistent bacteremia, or metastatic complications, or who were lost to follow-up within the 90 days post-enrollment. Therefore, in performing analyses for SAB-attributable mortality, persistent bacteremia and metastatic complications, our denominator was likely higher than the true number of patients eligible to experience these outcomes. However, we know that the number of patients who died for reasons other than SAB remained constant over time (Figure S6) and the result of an inflated denominator would be a conservative estimate. Finally, while the point estimates and 95% confidence intervals in our binomial risk regression models are intended to represent trends (e.g. annual increase or decrease) in patient and bacterial characteristics and outcomes over the study period, there may be individual years when the direction is opposite that of the overall point estimate. For this reason, we have included figures with all reported point estimates and have not reported point estimates for trends that were clearly non-linear.
Despite these limitations, our study showed that the clinical presentation, clinical outcomes, and molecular epidemiology of SAB in HD-dependent patients have changed significantly over the past two decades. Suspected source of infection shifted over the study period with non-vascular access infections becoming more prevalent. SAB-attributable mortality, persistent bacteremia, and metastatic complications increased, and patients who had a non-vascular access source of infection were more likely to die from S. aureus than those whose source was a CVC. USA300 emerged as an increasingly prevalent cause of SAB and was found to be strongly associated with persistent bacteremia. More research is needed to better clarify the clinical impact of the emergence of USA300 among the HD-dependent patients. We plan to address this in future studies by genotyping bacterial isolates in HD-dependent patients with SAB beyond the year 2015. Determining this will help to decide whether there is future clinical and therapeutic utility in genotyping S. aureus isolates among HD-dependent patients with SAB, or if we should focus exclusively on patient-specific factors to improve outcomes in this patient population.
Supplementary Material
Table S1: Demographics and Clinical Characteristics in Hemodialysis (HD)-Dependent vs Non-HD-Dependent Patients from 2004-2015
Table S2A: Clinical Outcomes in Hemodialysis (HD)-Dependent vs Non-HD-Dependent Patients 2004-2015
Table S2B: Clinical Outcomes in Hemodialysis (HD)-Dependent vs Non-HD-Dependent Patients with Methicillin Resistant Staphylococcus aureus (MRSA) Bacteremia 2004-2015
Table: S3: Frequencies and Proportions of Patients Experiencing Clinical Outcome by Year
Figure S1 (13 Panel Figure) Lowess Curves of Clinical Characteristics in Hemodialysis (HD)-Dependent Patients; Panel A-Age; Panel B-Female; Panel C-White Race; Panel D-Healthcare Associated Infection; Panel E-Diabetes Mellitus; Panel F-Neoplasm; Panel G-Corticosteroid Use; Panel H-Transplant; Panel I-HIV; Panel J-Surgery in Last 30 Days; Panel K-Previous Endocarditis; Panel L-Foreign Body Cardiac Device; Panel M-Foreign Body Central Venous Catheter
Figure S2 (4 Panel Figure): Secular Trends in Suspected Source of Infection in Hemodialysis (HD)-Dependent Patients from 2004-2015; Panel A-Suspected Source of Infection CBC; Panel B-Suspected Source of Infection AVG; Panel C-Suspected Source of Infection AVF; Panel D-Suspected Source of Infection Non-Vascular Access
Figure S3 (4 Panel Figure) Secular Trends in All-Cause Mortality, Staphylococcus aureus bacteremia (SAB)-Attributable Mortality, Persistent Bacteremia, and Metastatic Complications in Hemodialysis (HD)-Dependent Patients from 2004-2015; Panel A-All-Cause Mortality; Panel B-SAB-Attributable Mortality; Panel C-Persistent Bacteremia; Panel D-Metastatic Complications
Figure S4 (4 Panel Figure) Secular Trends in All-Cause Mortality, Staphylococcus aureus bacteremia (SAB)-Attributable Mortality, Persistent Bacteremia, and Metastatic Complications in Hemodialysis (HD)-Dependent Patients with Methicillin Resistant Staphylococcus aureus (MRSA) Bacteremia from 2004-2015; Panel A-All-Cause Mortality; Panel B-SAB-Attributable Mortality; Panel C-Persistent Bacteremia; Panel D-Metastatic Complications
Figure S5 (2 Panel Figure) Secular Trends in Bacterial Characteristics in Hemodialysis (HD)-Dependent Patients from 2004-2015; Panel A- Methicillin Resistant Staphylococcus aureus (MRSA) Bacteremia; Panel B-USA300 Infection
Figure S6 (1 Panel Figure) Secular Trend in Non-Staphylococcus aureus bacteremia (SAB)-Attributable Mortality
Support:
This research was supported in part by R01-AI068804 and T32- DK007731. VGF was supported by K24-AI093969. The funders did not have a role in study design, data collection, analysis, reporting, or the decision to submit for publication.
Financial Disclosure:
VGF reports Grant/Research Support: MedImmune, Cerexa/Forest/Actavis/Allergan, Pfizer, Advanced Liquid Logics, Theravance, Novartis, Cubist/Merck; Medical Biosurfaces; Locus; Affinergy; Contrafect; Karius; Genentech, Regeneron, Basilea. Paid Consultant: Pfizer, Novartis, Galderma, Novadigm, Durata, Debiopharm, Genentech, Achaogen, Affinium, Medicines Co., Cerexa, Tetraphase, Trius, MedImmune, Bayer, Theravance, Cubist, Basilea, Affinergy, Janssen, xBiotech, Contrafect, Regeneron, Basilea, Destiny. Membership: Merck Co-Chair V710 Vaccine. Educational fees: Green Cross, Cubist, Cerexa, Durata, Theravance; Debiopharm. Royalties: UpToDate. Patent pending: sepsis diagnostics. The remaining authors declare that they have no relevant financial interests.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES:
- 1.Inrig JK, Reed SD, Szczech LA, et al. Relationship between clinical outcomes and vascular access type among hemodialysis patients with Staphylococcus aureus bacteremia. Clin J Am Soc Nephrol. 2006;1(3): 518–524. [DOI] [PubMed] [Google Scholar]
- 2.Jaber BL. Bacterial infections in hemodialysis patients: pathogenesis and prevention. Kidney Int. 2005;67(6): 2508–2519. [DOI] [PubMed] [Google Scholar]
- 3.Saran R, Robinson B, Abbott KC, et al. US Renal Data System 2018 Annual Data Report: epidemiology of kidney disease in the United States. Am J Kidney Dis. 2019;73(3). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Sarnak MJ, Jaber BL. Mortality caused by sepsis in patients with end-stage renal disease compared with the general population. Kidney Int. 2000;58(4): 1758–1764. [DOI] [PubMed] [Google Scholar]
- 5.Vogelzang JL, van Stralen KJ, Noordzij M, et al. Mortality from infections and malignancies in patients treated with renal replacement therapy: data from the ERA-EDTA registry. Nephrol Dial Transplant. 2015;30(6): 1028–1037. [DOI] [PubMed] [Google Scholar]
- 6.Tokars JI, Miller ER, Stein G. New national surveillance system for hemodialysis-associated infections: initial results. Am J Infect Control. 2002;30(5): 288–295. [DOI] [PubMed] [Google Scholar]
- 7.Nguyen DB, Shugart A, Lines C, et al. National Healthcare Safety Network (NHSN) Dialysis Event Surveillance Report for 2014. Clin J Am Soc Nephrol. 2017; 12(7): 1139–1146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Marr KA, Sexton DJ, Conlon PJ, Corey GR, Schwab SJ, Kirkland KB. Catheter-related bacteremia and outcome of attempted catheter salvage in patients undergoing hemodialysis. Ann Intern Med. 1997;127(4): 275–280. [DOI] [PubMed] [Google Scholar]
- 9.Marr KA, Kong L, Fowler VG, et al. Incidence and outcome of Staphylococcus aureus bacteremia in hemodialysis patients. Kidney Int. 1998;54(5): 1684–1689. [DOI] [PubMed] [Google Scholar]
- 10.Kovalik EC, Raymond JR, Albers FJ, et al. A clustering of epidural abscesses in chronic hemodialysis patients: risks of salvaging access catheters in cases of infection. J Am Soc Nephrol. 1996;7(10): 2264–2267. [DOI] [PubMed] [Google Scholar]
- 11.Cosgrove SE, Sakoulas G, Perencevich EN, Schwaber MJ, Karchmer AW, Carmeli Y. Comparison of mortality associated with methicillin-resistant and methicillin-susceptible Staphylococcus aureus bacteremia: a meta-analysis. Clin Infect Dis. 2003;36(1): 53–59. [DOI] [PubMed] [Google Scholar]
- 12.del Rio A, Cervera C, Moreno A, Moreillon P, Miró JM. Patients at Risk of Complications of Staphylococcus aureus Bloodstream Infection. Clinical Infectious Diseases. 2009;48(Supplement_4): S246–S253. [DOI] [PubMed] [Google Scholar]
- 13.Barber M Methicillin-resistant staphylococci. J Clin Pathol. 1961;14: 385–393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Carrel M, Perencevich EN, David MZ. USA300 Methicillin-Resistant Staphylococcus aureus, United States, 2000-2013. Emerg Infect Dis. 2015;21(11): 1973–1980. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Souli M, Ruffin F, Choi SH, et al. Changing Characteristics of Staphylococcus aureus Bacteremia: Results From a 21-Year, Prospective, Longitudinal Study. Clin Infect Dis. 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Friedman ND, Kaye KS, Stout JE, et al. Health care--associated bloodstream infections in adults: a reason to change the accepted definition of community-acquired infections. Ann Intern Med. 2002;137(10): 791–797. [DOI] [PubMed] [Google Scholar]
- 17.Li JS, Sexton DJ, Mick N, et al. Proposed modifications to the Duke criteria for the diagnosis of infective endocarditis. Clin Infect Dis. 2000;30(4): 633–638. [DOI] [PubMed] [Google Scholar]
- 18.Berlon NR, Qi R, Sharma-Kuinkel BK, et al. Clinical MRSA isolates from skin and soft tissue infections show increased in vitro production of phenol soluble modulins. J Infect. 2015;71(4): 447–457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.David MZ, Daum RS, Bayer AS, et al. Staphylococcus aureus bacteremia at 5 US academic medical centers, 2008-2011: significant geographic variation in community-onset infections. Clin Infect Dis. 2014;59(6): 798–807. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.O'Hara FP, Suaya JA, Ray GT, et al. spa Typing and Multilocus Sequence Typing Show Comparable Performance in a Macroepidemiologic Study of Staphylococcus aureus in the United States. Microb Drug Resist. 2016;22(1): 88–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Diep BA, Gill SR, Chang RF, et al. Complete genome sequence of USA300, an epidemic clone of community-acquired meticillin-resistant Staphylococcus aureus. Lancet. 2006;367(9512): 731–739. [DOI] [PubMed] [Google Scholar]
- 22.Wacholder S Binomial regression in GLIM: estimating risk ratios and risk differences. Am J Epidemiol. 1986;123(1): 174–184. [DOI] [PubMed] [Google Scholar]
- 23.Lee T Fistula First Initiative: Historical Impact on Vascular Access Practice Patterns and Influence on Future Vascular Access Care. Cardiovasc Eng Technol. 2017;8(3): 244–254. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Lok CE, Foley R. Vascular access morbidity and mortality: trends of the last decade. Clin J Am Soc Nephrol. 2013;8(7): 1213–1219. [DOI] [PubMed] [Google Scholar]
- 25.Ravani P, Palmer SC, Oliver MJ, et al. Associations between hemodialysis access type and clinical outcomes: a systematic review. J Am Soc Nephrol. 2013;24(3): 465–473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Brown RS, Patibandla BK, Goldfarb-Rumyantzev AS. The Survival Benefit of "Fistula First, Catheter Last" in Hemodialysis Is Primarily Due to Patient Factors. J Am Soc Nephrol. 2017;28(2): 645–652. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Kalantar-Zadeh K, Kovesdy CP, Norris KC. Racial survival paradox of dialysis patients: robust and resilient. Am J Kidney Dis. 2012;60(2): 182–185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Kucirka LM, Grams ME, Lessler J, et al. Association of race and age with survival among patients undergoing dialysis. JAMA. 2011;306(6): 620–626. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Roghmann MC. Predicting methicillin resistance and the effect of inadequate empiric therapy on survival in patients with Staphylococcus aureus bacteremia. Arch Intern Med. 2000;160(7): 1001–1004. [DOI] [PubMed] [Google Scholar]
- 30.Klevens RM, Morrison MA, Nadle J, et al. Invasive methicillin-resistant Staphylococcus aureus infections in the United States. JAMA. 2007;298(15): 1763–1771. [DOI] [PubMed] [Google Scholar]
- 31.van Hal SJ, Jensen SO, Vaska VL, Espedido BA, Paterson DL, Gosbell IB. Predictors of mortality in Staphylococcus aureus Bacteremia. Clin Microbiol Rev. 2012;25(2): 362–386. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Chong YP, Park S-J, Kim HS, et al. Persistent Staphylococcus aureus bacteremia: a prospective analysis of risk factors, outcomes, and microbiologic and genotypic characteristics of isolates. Medicine. 2013;92(2): 98–108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Hidalgo M, Carvajal LP, Rincon S, et al. Methicillin-Resistant Staphylococcus aureus USA300 Latin American Variant in Patients Undergoing Hemodialysis and HIV Infected in a Hospital in Bogota, Colombia. PLoS One. 2015;10(10): e0140748. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Patel G, Jenkins SG, Mediavilla JR, et al. Clinical and molecular epidemiology of methicillin-resistant Staphylococcus aureus among patients in an ambulatory hemodialysis center. Infect Control Hosp Epidemiol. 2011;32(9): 881–888. [DOI] [PubMed] [Google Scholar]
- 35.Johnson LB, Venugopal AA, Pawlak J, Saravolatz LD. Emergence of community-associated methicillin-resistant Staphylococcus aureus infection among patients with end-stage renal disease. Infect Control Hosp Epidemiol. 2006;27(10): 1057–1062. [DOI] [PubMed] [Google Scholar]
- 36.Mermel LA, Allon M, Bouza E, et al. Clinical practice guidelines for the diagnosis and management of intravascular catheter-related infection: 2009 Update by the Infectious Diseases Society of America. Clin Infect Dis. 2009;49(1): 1–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Allon M, Brouwer-Maier DJ, Abreo K, et al. Recommended Clinical Trial End Points for Dialysis Catheters. Clin J Am Soc Nephrol. 2018;13(3): 495–500. [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.
Supplementary Materials
Table S1: Demographics and Clinical Characteristics in Hemodialysis (HD)-Dependent vs Non-HD-Dependent Patients from 2004-2015
Table S2A: Clinical Outcomes in Hemodialysis (HD)-Dependent vs Non-HD-Dependent Patients 2004-2015
Table S2B: Clinical Outcomes in Hemodialysis (HD)-Dependent vs Non-HD-Dependent Patients with Methicillin Resistant Staphylococcus aureus (MRSA) Bacteremia 2004-2015
Table: S3: Frequencies and Proportions of Patients Experiencing Clinical Outcome by Year
Figure S1 (13 Panel Figure) Lowess Curves of Clinical Characteristics in Hemodialysis (HD)-Dependent Patients; Panel A-Age; Panel B-Female; Panel C-White Race; Panel D-Healthcare Associated Infection; Panel E-Diabetes Mellitus; Panel F-Neoplasm; Panel G-Corticosteroid Use; Panel H-Transplant; Panel I-HIV; Panel J-Surgery in Last 30 Days; Panel K-Previous Endocarditis; Panel L-Foreign Body Cardiac Device; Panel M-Foreign Body Central Venous Catheter
Figure S2 (4 Panel Figure): Secular Trends in Suspected Source of Infection in Hemodialysis (HD)-Dependent Patients from 2004-2015; Panel A-Suspected Source of Infection CBC; Panel B-Suspected Source of Infection AVG; Panel C-Suspected Source of Infection AVF; Panel D-Suspected Source of Infection Non-Vascular Access
Figure S3 (4 Panel Figure) Secular Trends in All-Cause Mortality, Staphylococcus aureus bacteremia (SAB)-Attributable Mortality, Persistent Bacteremia, and Metastatic Complications in Hemodialysis (HD)-Dependent Patients from 2004-2015; Panel A-All-Cause Mortality; Panel B-SAB-Attributable Mortality; Panel C-Persistent Bacteremia; Panel D-Metastatic Complications
Figure S4 (4 Panel Figure) Secular Trends in All-Cause Mortality, Staphylococcus aureus bacteremia (SAB)-Attributable Mortality, Persistent Bacteremia, and Metastatic Complications in Hemodialysis (HD)-Dependent Patients with Methicillin Resistant Staphylococcus aureus (MRSA) Bacteremia from 2004-2015; Panel A-All-Cause Mortality; Panel B-SAB-Attributable Mortality; Panel C-Persistent Bacteremia; Panel D-Metastatic Complications
Figure S5 (2 Panel Figure) Secular Trends in Bacterial Characteristics in Hemodialysis (HD)-Dependent Patients from 2004-2015; Panel A- Methicillin Resistant Staphylococcus aureus (MRSA) Bacteremia; Panel B-USA300 Infection
Figure S6 (1 Panel Figure) Secular Trend in Non-Staphylococcus aureus bacteremia (SAB)-Attributable Mortality