- Split View
-
Views
-
Cite
Cite
Jennifer S McDanel, Mary-Claire Roghmann, Eli N Perencevich, Michael E Ohl, Michihiko Goto, Daniel J Livorsi, Makoto Jones, Justin P Albertson, Rajeshwari Nair, Amy M J O’Shea, Marin L Schweizer, Comparative Effectiveness of Cefazolin Versus Nafcillin or Oxacillin for Treatment of Methicillin-Susceptible Staphylococcus aureus Infections Complicated by Bacteremia: A Nationwide Cohort Study, Clinical Infectious Diseases, Volume 65, Issue 1, 1 July 2017, Pages 100–106, https://doi.org/10.1093/cid/cix287
- Share Icon Share
Abstract
To treat patients with methicillin-susceptible Staphylococcus aureus (MSSA) infections, β-lactams are recommended for definitive therapy; however, the comparative effectiveness of individual β-lactams is unknown. This study compared definitive therapy with cefazolin vs nafcillin or oxacillin among patients with MSSA infections complicated by bacteremia.
This retrospective study included patients admitted to 119 Veterans Affairs hospitals from 2003 to 2010. Patients were included if they had a blood culture positive for MSSA and received definitive therapy with cefazolin, nafcillin, or oxacillin. Cox proportional hazards regression and ordinal logistic regression were used to identify associations between antibiotic therapy and mortality or recurrence. A recurrent infection was defined as a MSSA blood culture between 45 and 365 days after the first MSSA blood culture.
Of 3167 patients, 1163 (37%) patients received definitive therapy with cefazolin. Patients who received cefazolin had a 37% reduction in 30-day mortality (hazard ratio [HR], 0.63; 95% confidence interval [CI], .51–.78) and a 23% reduction in 90-day mortality (HR, 0.77; 95% CI, .66–.90) compared with patients receiving nafcillin or oxacillin, after controlling for other factors. The odds of recurrence (odds ratio, 1.13; 95% CI, .94–1.36) were similar among patients who received cefazolin compared with patients who received nafcillin or oxacillin, after controlling for other factors.
In this large, multicenter study, patients who received cefazolin had a lower risk of mortality and similar odds of recurrent infections compared with nafcillin or oxacillin for MSSA infections complicated by bacteremia. Physicians might consider definitive therapy with cefazolin for these infections.
(See the Editorial Commentary by Karchmer on pages 107–9.)
Infections caused by Staphylococcus aureus are often complicated by bacteremia. For patients with methicillin-susceptible S. aureus (MSSA) infections, the Infectious Diseases Society of America recommends definitive therapy with β-lactams [1]. Penicillinase-stable penicillins such as nafcillin or oxacillin are the primary choice for treating MSSA infections complicated by bacteremia, whereas cefazolin is an alternate choice [2–4]. However, nafcillin and oxacillin cause more adverse events, such as drug-induced fever, cytopenia, rash, renal dysfunction, and liver function abnormalities compared with cefazolin [5–7]. In addition, cefazolin is more convenient than nafcillin or oxacillin for patients completing their antibiotic course following hospital discharge because cefazolin can be dosed at hemodialysis and can be administered less frequently than nafcillin or oxacillin.
However, some clinicians are reluctant to use cefazolin to treat serious MSSA infections. Treatment failure has been observed among patients receiving cefazolin for MSSA infections complicated by bacteremia, specifically in patients with high-inoculum infections such as valvular vegetation or an undrained abscess [8, 9]. Treatment failure with cefazolin has been attributed to the inoculum effect, which correlates with inactivation of cefazolin by type A β-lactamases [9]. Furthermore, cefazolin has little penetration into the cerebrospinal fluid and thus may be inferior for infections involving the central nervous system.
The optimal choice of β-lactam therapy for MSSA infections complicated by bacteremia is unclear. Previous studies have assessed outcomes such as treatment failure, adverse events, or mortality among patients receiving cefazolin compared with nafcillin or oxacillin for MSSA infections complicated by bacteremia [5, 7, 10, 11]. The results varied, and for many studies, no significant difference in outcomes was observed between the therapy groups, potentially due to the limited sample size [5, 7, 10, 11]. Large multicenter studies are needed to determine if a significant difference in outcomes exists between patients treated with nafcillin or oxacillin compared with patients treated with cefazolin. This study compared definitive therapy with cefazolin vs treatment with nafcillin or oxacillin among patients with MSSA infections complicated by bacteremia at 119 Veterans Affairs (VA) hospitals.
METHODS
Study Design and Patient Population
This retrospective cohort study included patients with medical or surgical admissions to 119 acute care VA hospitals from 2003 to 2010. Patients were included if they had at least one S. aureus–positive blood culture susceptible to either methicillin or oxacillin by antimicrobial susceptibility testing, and the patients received cefazolin, nafcillin, or oxacillin for definitive therapy of their MSSA infections complicated by bacteremia. Patients were included only once in the study based on their first admission with a positive MSSA blood culture. Patients were excluded if they were admitted to a VA hospital that had <25 cases of S. aureus infections complicated by bacteremia throughout the study period, as these hospitals do not treat patients with S. aureus bacteremia on a regular basis.
Data were extracted from the VA electronic health record using the VA Corporate Data Warehouse and Patient Care Services databases and accessed through the VA Informatics and Computing Infrastructure. These data were last updated in February 2014 and validated by the VA Healthcare System through the VA Information Resource Center [12]. The institutional review board of the University of Iowa and the Research and Development Committee of the Iowa City VA Medical Center approved this study. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) criteria were followed when preparing the manuscript [13].
Variable Definitions
The primary outcomes examined were 30-day all-cause mortality and 90-day all-cause mortality. Thirty-day all-cause mortality was defined as death occurring within 30 days after the collection of the first positive MSSA blood culture, and 90-day all-cause mortality was defined as death occurring within 90 days after the collection of the first positive MSSA blood culture. Deaths that occurred outside of the hospital or on a subsequent admission were included as the VA uses multiple agencies (VA data, Social Security Administration Death Master File, Medicare Vital Status, and Beneficiary Identification Records Locator Subsystem Death File) to collect information regarding patient death. The VA-National Death Index Mortality Data Merge Project reported a 98% agreement between the combined sources and the National Death Index [14]. Recurrent MSSA infections complicated by bacteremia were defined as MSSA-positive blood cultures between 45 and 365 days after the first MSSA-positive blood culture. A cutoff of 45 days was selected because the recommended antibiotic treatment for some MSSA invasive infections such as osteomyelitis is up to 42 days (6 weeks).
Definitive therapy was defined as starting a definitive antimicrobial or continuing an empiric antimicrobial between days 4 and 14 after the first positive blood culture was collected [15]. Patients were excluded if they received definitive therapy with both cefazolin and nafcillin or cefazolin and oxacillin. Comorbid conditions were evaluated using the Charlson comorbidity index, which is based on the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes [16]. The modified Acute Physiology and Chronic Health Evaluation (APACHE) III score was used to measure patients’ severity of illness on hospital admission [17, 18]. A modified APACHE III score was used as not all of the included patients were admitted to the intensive care unit (ICU) and, therefore, not all tests were performed on the patients such as the Glasgow Coma Scale. ICD-9-CM codes were used to identify concurrent infections including skin and soft tissue infections, endocarditis, and osteomyelitis; chronic conditions such as hepatitis C; the stages of renal disease; and intravenous drug use [19–21]. APACHE III score was dichotomized on the median whereas age was categorized on quartiles.
Dialysis and/or end-stage renal disease (ESRD) was defined as receipt of outpatient dialysis, receipt of dialysis during the index hospital admission, or having ESRD [15]. Facilities were categorized based on complexity scores provided by the VA Healthcare Analysis and Information Group [15]. Empiric vancomycin therapy was defined as vancomycin started between 2 days before the first MSSA-positive blood culture was collected through 4 days after the blood culture was collected [15]. Hospital-onset MSSA infections complicated by bacteremia were defined as collection of the first MSSA-positive blood cultures from a patient >48 hours after hospital admission. Postinfection length of stay was defined as the number of days from collection of the first positive MSSA culture until hospital discharge or in-hospital death. Prior hospitalization was defined as admission to a VA hospital within the 12 months before the index admission.
Statistical Analysis
Bivariate analyses examined associations between patient characteristics and therapy, mortality, or recurrence using the χ2 or Fisher exact test for categorical variables and Student t test or the Wilcoxon rank-sum test for continuous variables.
Cox proportional hazards regression was used to examine the association between antibiotic therapy (cefazolin vs nafcillin or oxacillin) and mortality, where patients were censored if death occurred within 30 days or 90 days after the first positive blood culture was collected. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated based on multivariate modeling. Modified APACHE III score, Charlson comorbidity index, and additional concurrent infections (skin and soft tissue infections, endocarditis, and osteomyelitis) were incorporated in all models as these variables are strong predictors of mortality and may impact therapy selection. Additionally, the variables included in Table 1 were examined for inclusion in the multivariate models and selected based on a manual stepwise selection process. This selection method entered variables individually into the model if P < .05 in the bivariate analysis. Variables were then retained in the model if P < .05 even after subsequent variables were added to the model. A variable was a confounder if the regression coefficient for therapy was changed by >20%. Effect modification was examined by assessing the interaction of therapy with the potential effect modifier. The proportional hazards assumption was evaluated by assessing the interaction of therapy with the log of survival time, and the final model was examined using multiple time points to identify any significant changes in effect over time.
Characteristic . | Cefazolin (n = 1163) . | Nafcillin or Oxacillin (n = 2004) . | P Value . |
---|---|---|---|
Male sex | 1133 (97) | 1974 (99) | .031 |
Age, y | |||
≤55 | 296 (25) | 474 (24) | .715 |
56–61 | 267 (23) | 479 (24) | … |
62–73 | 312 (27) | 546 (27) | … |
≥74 | 288 (25) | 505 (25) | … |
APACHE III score ≥34 | 651 (56) | 1040 (52) | .027 |
Charlson comorbidity index score, median (IQR) | 2 (1–3) | 2 (1–3) | .013 |
Other infections | |||
Skin and soft tissue infections | 287 (25) | 460 (23) | .271 |
Osteomyelitis | 138 (12) | 267 (13) | .236 |
Endocarditis | 52 (4) | 145 (7) | .002 |
Dialysis and/or ESRD | 176 (15) | 168 (8) | <.001 |
Hospital-onset infectionb | 241 (21) | 489 (24) | .018 |
Hospitalization in the prior year | 670 (58) | 1138 (57) | .652 |
Admission to intensive care unit | 178 (15) | 378 (19) | .011 |
Received empiric vancomycin | 909 (78) | 1581 (79) | .628 |
Renal stage | |||
I | 1 (<1) | 0 (0) | .367 |
II | 4 (<1) | 2 (<1) | .201 |
III | 20 (2) | 22 (1) | .140 |
IV | 14 (1) | 12 (1) | .045 |
V | 7 (1) | 14 (1) | .747 |
Chronic kidney disease, unspecified | 58 (5) | 106 (5) | .711 |
HIV infection | 18 (2) | 34 (2) | .751 |
Hepatitis C infection | 119 (10) | 218 (11) | .570 |
Intravenous drug user | 110 (9) | 221 (11) | .164 |
Diabetes | 465 (40) | 668 (33) | <.001 |
Admission year | |||
2003 | 163 (14) | 255 (13) | .316 |
2004 | 149 (13) | 285 (14) | |
2005 | 153 (13) | 256 (13) | |
2006 | 130 (11) | 238 (12) | |
2007 | 123 (11) | 259 (13) | |
2008 | 156 (13) | 250 (12) | |
2009 | 136 (12) | 233 (12) | |
2010 | 153 (13) | 228 (11) | |
Facility type | |||
1a (most complex) | 523 (47) | 1043 (54) | <.001 |
1b | 240 (21) | 413 (22) | |
1c | 204 (18) | 232 (12) | |
2 | 125 (11) | 197 (10) | |
3 (least complex) | 27 (2) | 30 (2) | |
Postinfection length of stay in the hospitalc, d, median (IQR) | 12 (8–18) | 13 (9–22) | <.001 |
30-d mortality | 113 (10) | 307 (15) | <.001 |
45-d mortality | 164 (14) | 393 (20) | <.001 |
90-d mortality | 231 (20) | 502 (25) | .001 |
MSSA recurrence | |||
45–90 d | 20 (2) | 28 (1) | .474 |
91–365 d | 38 (3) | 44 (2) | .067 |
Characteristic . | Cefazolin (n = 1163) . | Nafcillin or Oxacillin (n = 2004) . | P Value . |
---|---|---|---|
Male sex | 1133 (97) | 1974 (99) | .031 |
Age, y | |||
≤55 | 296 (25) | 474 (24) | .715 |
56–61 | 267 (23) | 479 (24) | … |
62–73 | 312 (27) | 546 (27) | … |
≥74 | 288 (25) | 505 (25) | … |
APACHE III score ≥34 | 651 (56) | 1040 (52) | .027 |
Charlson comorbidity index score, median (IQR) | 2 (1–3) | 2 (1–3) | .013 |
Other infections | |||
Skin and soft tissue infections | 287 (25) | 460 (23) | .271 |
Osteomyelitis | 138 (12) | 267 (13) | .236 |
Endocarditis | 52 (4) | 145 (7) | .002 |
Dialysis and/or ESRD | 176 (15) | 168 (8) | <.001 |
Hospital-onset infectionb | 241 (21) | 489 (24) | .018 |
Hospitalization in the prior year | 670 (58) | 1138 (57) | .652 |
Admission to intensive care unit | 178 (15) | 378 (19) | .011 |
Received empiric vancomycin | 909 (78) | 1581 (79) | .628 |
Renal stage | |||
I | 1 (<1) | 0 (0) | .367 |
II | 4 (<1) | 2 (<1) | .201 |
III | 20 (2) | 22 (1) | .140 |
IV | 14 (1) | 12 (1) | .045 |
V | 7 (1) | 14 (1) | .747 |
Chronic kidney disease, unspecified | 58 (5) | 106 (5) | .711 |
HIV infection | 18 (2) | 34 (2) | .751 |
Hepatitis C infection | 119 (10) | 218 (11) | .570 |
Intravenous drug user | 110 (9) | 221 (11) | .164 |
Diabetes | 465 (40) | 668 (33) | <.001 |
Admission year | |||
2003 | 163 (14) | 255 (13) | .316 |
2004 | 149 (13) | 285 (14) | |
2005 | 153 (13) | 256 (13) | |
2006 | 130 (11) | 238 (12) | |
2007 | 123 (11) | 259 (13) | |
2008 | 156 (13) | 250 (12) | |
2009 | 136 (12) | 233 (12) | |
2010 | 153 (13) | 228 (11) | |
Facility type | |||
1a (most complex) | 523 (47) | 1043 (54) | <.001 |
1b | 240 (21) | 413 (22) | |
1c | 204 (18) | 232 (12) | |
2 | 125 (11) | 197 (10) | |
3 (least complex) | 27 (2) | 30 (2) | |
Postinfection length of stay in the hospitalc, d, median (IQR) | 12 (8–18) | 13 (9–22) | <.001 |
30-d mortality | 113 (10) | 307 (15) | <.001 |
45-d mortality | 164 (14) | 393 (20) | <.001 |
90-d mortality | 231 (20) | 502 (25) | .001 |
MSSA recurrence | |||
45–90 d | 20 (2) | 28 (1) | .474 |
91–365 d | 38 (3) | 44 (2) | .067 |
Data are presented as No. (%) unless otherwise indicated.
Abbreviations: APACHE, Acute Physiology and Chronic Health Evaluation; ESRD, end-stage renal disease; HIV, human immunodeficiency virus; IQR, interquartile range; MSSA, methicillin-susceptible Staphylococcus aureus.
aDefinitive therapy is defined as starting an antimicrobial or continuing on an empiric antimicrobial between 4 and 14 days after the first MSSA-positive blood culture was collected.
bPositive MSSA culture was collected >48 hours after admission.
cDefined as the day the first MSSA blood culture was collected until the patient was either discharged from the hospital or died.
Characteristic . | Cefazolin (n = 1163) . | Nafcillin or Oxacillin (n = 2004) . | P Value . |
---|---|---|---|
Male sex | 1133 (97) | 1974 (99) | .031 |
Age, y | |||
≤55 | 296 (25) | 474 (24) | .715 |
56–61 | 267 (23) | 479 (24) | … |
62–73 | 312 (27) | 546 (27) | … |
≥74 | 288 (25) | 505 (25) | … |
APACHE III score ≥34 | 651 (56) | 1040 (52) | .027 |
Charlson comorbidity index score, median (IQR) | 2 (1–3) | 2 (1–3) | .013 |
Other infections | |||
Skin and soft tissue infections | 287 (25) | 460 (23) | .271 |
Osteomyelitis | 138 (12) | 267 (13) | .236 |
Endocarditis | 52 (4) | 145 (7) | .002 |
Dialysis and/or ESRD | 176 (15) | 168 (8) | <.001 |
Hospital-onset infectionb | 241 (21) | 489 (24) | .018 |
Hospitalization in the prior year | 670 (58) | 1138 (57) | .652 |
Admission to intensive care unit | 178 (15) | 378 (19) | .011 |
Received empiric vancomycin | 909 (78) | 1581 (79) | .628 |
Renal stage | |||
I | 1 (<1) | 0 (0) | .367 |
II | 4 (<1) | 2 (<1) | .201 |
III | 20 (2) | 22 (1) | .140 |
IV | 14 (1) | 12 (1) | .045 |
V | 7 (1) | 14 (1) | .747 |
Chronic kidney disease, unspecified | 58 (5) | 106 (5) | .711 |
HIV infection | 18 (2) | 34 (2) | .751 |
Hepatitis C infection | 119 (10) | 218 (11) | .570 |
Intravenous drug user | 110 (9) | 221 (11) | .164 |
Diabetes | 465 (40) | 668 (33) | <.001 |
Admission year | |||
2003 | 163 (14) | 255 (13) | .316 |
2004 | 149 (13) | 285 (14) | |
2005 | 153 (13) | 256 (13) | |
2006 | 130 (11) | 238 (12) | |
2007 | 123 (11) | 259 (13) | |
2008 | 156 (13) | 250 (12) | |
2009 | 136 (12) | 233 (12) | |
2010 | 153 (13) | 228 (11) | |
Facility type | |||
1a (most complex) | 523 (47) | 1043 (54) | <.001 |
1b | 240 (21) | 413 (22) | |
1c | 204 (18) | 232 (12) | |
2 | 125 (11) | 197 (10) | |
3 (least complex) | 27 (2) | 30 (2) | |
Postinfection length of stay in the hospitalc, d, median (IQR) | 12 (8–18) | 13 (9–22) | <.001 |
30-d mortality | 113 (10) | 307 (15) | <.001 |
45-d mortality | 164 (14) | 393 (20) | <.001 |
90-d mortality | 231 (20) | 502 (25) | .001 |
MSSA recurrence | |||
45–90 d | 20 (2) | 28 (1) | .474 |
91–365 d | 38 (3) | 44 (2) | .067 |
Characteristic . | Cefazolin (n = 1163) . | Nafcillin or Oxacillin (n = 2004) . | P Value . |
---|---|---|---|
Male sex | 1133 (97) | 1974 (99) | .031 |
Age, y | |||
≤55 | 296 (25) | 474 (24) | .715 |
56–61 | 267 (23) | 479 (24) | … |
62–73 | 312 (27) | 546 (27) | … |
≥74 | 288 (25) | 505 (25) | … |
APACHE III score ≥34 | 651 (56) | 1040 (52) | .027 |
Charlson comorbidity index score, median (IQR) | 2 (1–3) | 2 (1–3) | .013 |
Other infections | |||
Skin and soft tissue infections | 287 (25) | 460 (23) | .271 |
Osteomyelitis | 138 (12) | 267 (13) | .236 |
Endocarditis | 52 (4) | 145 (7) | .002 |
Dialysis and/or ESRD | 176 (15) | 168 (8) | <.001 |
Hospital-onset infectionb | 241 (21) | 489 (24) | .018 |
Hospitalization in the prior year | 670 (58) | 1138 (57) | .652 |
Admission to intensive care unit | 178 (15) | 378 (19) | .011 |
Received empiric vancomycin | 909 (78) | 1581 (79) | .628 |
Renal stage | |||
I | 1 (<1) | 0 (0) | .367 |
II | 4 (<1) | 2 (<1) | .201 |
III | 20 (2) | 22 (1) | .140 |
IV | 14 (1) | 12 (1) | .045 |
V | 7 (1) | 14 (1) | .747 |
Chronic kidney disease, unspecified | 58 (5) | 106 (5) | .711 |
HIV infection | 18 (2) | 34 (2) | .751 |
Hepatitis C infection | 119 (10) | 218 (11) | .570 |
Intravenous drug user | 110 (9) | 221 (11) | .164 |
Diabetes | 465 (40) | 668 (33) | <.001 |
Admission year | |||
2003 | 163 (14) | 255 (13) | .316 |
2004 | 149 (13) | 285 (14) | |
2005 | 153 (13) | 256 (13) | |
2006 | 130 (11) | 238 (12) | |
2007 | 123 (11) | 259 (13) | |
2008 | 156 (13) | 250 (12) | |
2009 | 136 (12) | 233 (12) | |
2010 | 153 (13) | 228 (11) | |
Facility type | |||
1a (most complex) | 523 (47) | 1043 (54) | <.001 |
1b | 240 (21) | 413 (22) | |
1c | 204 (18) | 232 (12) | |
2 | 125 (11) | 197 (10) | |
3 (least complex) | 27 (2) | 30 (2) | |
Postinfection length of stay in the hospitalc, d, median (IQR) | 12 (8–18) | 13 (9–22) | <.001 |
30-d mortality | 113 (10) | 307 (15) | <.001 |
45-d mortality | 164 (14) | 393 (20) | <.001 |
90-d mortality | 231 (20) | 502 (25) | .001 |
MSSA recurrence | |||
45–90 d | 20 (2) | 28 (1) | .474 |
91–365 d | 38 (3) | 44 (2) | .067 |
Data are presented as No. (%) unless otherwise indicated.
Abbreviations: APACHE, Acute Physiology and Chronic Health Evaluation; ESRD, end-stage renal disease; HIV, human immunodeficiency virus; IQR, interquartile range; MSSA, methicillin-susceptible Staphylococcus aureus.
aDefinitive therapy is defined as starting an antimicrobial or continuing on an empiric antimicrobial between 4 and 14 days after the first MSSA-positive blood culture was collected.
bPositive MSSA culture was collected >48 hours after admission.
cDefined as the day the first MSSA blood culture was collected until the patient was either discharged from the hospital or died.
Ordinal logistic regression was used to examine the association between therapy and recurrence of MSSA infections complicated by bacteremia. Odds ratios (ORs) and 95% CIs were calculated. Patients were categorized by outcome: no recurrent MSSA infections complicated by bacteremia, 45–90 days until a MSSA infection complicated by bacteremia, 91–365 days until a MSSA infection complicated by bacteremia, and death occurred within 45–365 days after the first MSSA-positive blood culture but no MSSA infections complicated by bacteremia during this time period. The reference group for the analysis was no recurrent MSSA infections complicated by bacteremia. Patients were excluded from the analysis if they died within 45 days from the day the first MSSA-positive blood cultures were collected. The same set of model variables were assessed, entered, and included using the same stepwise method and criteria described above. The modified APACHE III score and Charlson comorbidity index were retained in all models regardless of statistical significance, as these variables are strong predictors of therapy type and recurrent MSSA infections complicated by bacteremia. The proportional odds assumption was examined for treatment type. All data analysis was completed using SAS version 9.4 software (SAS Institute, Cary, North Carolina). A subanalysis was performed that examined dialysis and/or ESRD, as these patients may be more likely to receive cefazolin due to the ease of dosing cefazolin for dialysis.
RESULTS
A total of 11154 patients with MSSA infections complicated by bacteremia were identified. Of these patients, 7718 patients were excluded for not receiving definitive therapy with cefazolin, nafcillin, or oxacillin between 4 and 14 days after the first MSSA-positive blood culture was collected (Figure 1). Of the 3436 patients who received definitive therapy with cefazolin, nafcillin, or oxacillin, 269 patients were excluded for receiving definitive therapy with both cefazolin and penicillinase-stable penicillins. APACHE III score (≥34; 55% vs 53%, P = .693), age (≥62 years; 50% vs 52%, P = .40), and mortality rate (10% vs 13%, P = .182) were similar among patients who received both therapies compared with the patients who received only one therapy. However, patients who received both therapies were more likely to have received dialysis and/or have ESRD (17% vs 11%, P = .004). During the definitive period, 187 (70%) patients switched from nafcillin or oxacillin to cefazolin whereas 59 (22%) patients switched from cefazolin to nafcillin or oxacillin. Twenty-three (9%) patients started both therapies on the same day.
Of the 3167 patients receiving definitive therapy with cefazolin, nafcillin, or oxacillin, 37% were treated with cefazolin (Table 1). Compared with the patients who received nafcillin or oxacillin, patients treated with cefazolin were more likely to have a higher APACHE III score (≥34; 56% vs 52%, P = .027), receive dialysis or have ESRD (15% vs 8%, P < .001), and have diabetes (40% vs 33%, P < .001). Furthermore, the patients treated with cefazolin were less likely to have endocarditis (4% vs 7%, P = .002), have a hospital-onset infection (21% vs 24%, P = .018), be hospitalized in a more complex VA facility (P < .001), and have a slightly shorter postinfection length of stay in the hospital (12 days vs 13 days, P < .001). Mortality significantly differed between the 2 therapy groups (Table 1 and Figure 2). Twenty-five percent of patients receiving nafcillin or oxacillin died at 90 days, whereas only 20% of patients who received cefazolin died (P = .001).
In the multivariate models, patients treated with cefazolin for their MSSA infections complicated by bacteremia had a 37% reduction in 30-day mortality risk compared with patients treated with nafcillin or oxacillin (HR, 0.63; 95% CI, .51–.78; Table 2; Supplementary Figure 1) after adjusting for skin and soft tissue infections, endocarditis, osteomyelitis, ICU admission, and diabetes, as well as APACHE III and Charlson comorbidity index scores. Similarly, the adjusted risk of 90-day mortality was 23% lower for patients receiving cefazolin compared with patients receiving nafcillin or oxacillin (HR, 0.77; 95% CI, .66–.90; Table 2; Supplementary Figure 2). The adjusted odds of having a recurrent MSSA infection complicated by bacteremia did not significantly differ between the cefazolin and the nafcillin or oxacillin treatment groups (OR, 1.13; 95% CI, .94–1.36; Table 2; Supplementary Figure 3). Only 2% of patients treated with cefazolin (n = 27) and 3% of patients treated with nafcillin or oxacillin (n = 65; P = .136) had intracranial and intraspinal abscesses; there were no statistically significant differences in mortality (30-day: 4% vs 2%, P = .503; 90-day: 11% vs 6%, P = .415) or recurrence (4% vs 0%, P = .176) comparing the treatment groups in that patient population. Eighty-nine percent of the patients had information available on whether they received an infectious disease (ID) consult between the collection of the first positive MSSA blood culture until hospital discharge or in-hospital death. ID consult was not available for all patients as this information was collected for another study. When only the patients who had information on ID consult were examined, risk estimates for 30-day mortality, 90-day mortality, and recurrent MSSA infections complicated by bacteremia were similar to the risk estimates in the adjusted models that did not include ID consult (data not shown). In a subanalysis excluding patients who received dialysis or had ESRD, risk estimates were similar to the full cohort for 30-day mortality (HR, 0.66; 95% CI, .52–.82), 90-day mortality (HR, 0.80; 95% CI, .68–.94), and recurrent MSSA infection complicated by bacteremia (OR, 1.13; 95% CI, .92–1.38). When only patients with dialysis and/or ESRD were examined, the risk estimates did not substantially change compared with the full cohort for 30-day mortality (HR, 0.48; 95% CI, .24–.97), 90-day mortality (HR, 0.60; 95% CI, .37–.98), and recurrent MSSA infection complicated by bacteremia (OR, 1.03; 95% CI, .64–1.68).
Model . | Outcome . | Regression Analysis . | Risk Ratio (95% Confidence Interval) . | P Value . |
---|---|---|---|---|
Mortality | Hazard Ratio | |||
Unadjusted | 30-d mortality | Cox proportional hazards regression | 0.51 (.49–.76) | <.001 |
Adjusteda | 30-d mortality | Cox proportional hazards regression | 0.63 (.51–.78) | <.001 |
Unadjusted | 90-d mortality | Cox proportional hazards regression | 0.75 (.65–.88) | <.001 |
Adjustedb | 90-d mortality | Cox proportional hazards regression | 0.77 (.66–.90) | .001 |
Recurrencec | Odds Ratiod | |||
Unadjusted | Recurrence | Ordinal logistic regression | 1.20 (1.00–1.44) | .047 |
Adjustede | Recurrence | Ordinal logistic regression | 1.13 (.94–1.36) | .211 |
Model . | Outcome . | Regression Analysis . | Risk Ratio (95% Confidence Interval) . | P Value . |
---|---|---|---|---|
Mortality | Hazard Ratio | |||
Unadjusted | 30-d mortality | Cox proportional hazards regression | 0.51 (.49–.76) | <.001 |
Adjusteda | 30-d mortality | Cox proportional hazards regression | 0.63 (.51–.78) | <.001 |
Unadjusted | 90-d mortality | Cox proportional hazards regression | 0.75 (.65–.88) | <.001 |
Adjustedb | 90-d mortality | Cox proportional hazards regression | 0.77 (.66–.90) | .001 |
Recurrencec | Odds Ratiod | |||
Unadjusted | Recurrence | Ordinal logistic regression | 1.20 (1.00–1.44) | .047 |
Adjustede | Recurrence | Ordinal logistic regression | 1.13 (.94–1.36) | .211 |
aAdjusted for skin and soft tissue infections, endocarditis, osteomyelitis, Acute Physiology and Chronic Health Evaluation (APACHE) III score, Charlson comorbidity index, admission to the intensive care unit (ICU), and diabetes.
bAdjusted for skin and soft tissue infections, endocarditis, osteomyelitis, APACHE III score, Charlson comorbidity index, hospital-onset infection, admission to the ICU, and diabetes.
cRecurrence is defined as a methicillin-susceptible Staphylococcus aureus (MSSA)–positive blood culture within 45–365 days after first MSSA-positive blood culture.
dOdds ratios are based on a multivariate ordinal logistic regression. As such, the odds ratio represents the cumulative odds of recurrence and is assumed to be the same across cumulative probabilities.
eAdjusted for APACHE III score, Charlson comorbidity index, and hospital-onset infection.
Model . | Outcome . | Regression Analysis . | Risk Ratio (95% Confidence Interval) . | P Value . |
---|---|---|---|---|
Mortality | Hazard Ratio | |||
Unadjusted | 30-d mortality | Cox proportional hazards regression | 0.51 (.49–.76) | <.001 |
Adjusteda | 30-d mortality | Cox proportional hazards regression | 0.63 (.51–.78) | <.001 |
Unadjusted | 90-d mortality | Cox proportional hazards regression | 0.75 (.65–.88) | <.001 |
Adjustedb | 90-d mortality | Cox proportional hazards regression | 0.77 (.66–.90) | .001 |
Recurrencec | Odds Ratiod | |||
Unadjusted | Recurrence | Ordinal logistic regression | 1.20 (1.00–1.44) | .047 |
Adjustede | Recurrence | Ordinal logistic regression | 1.13 (.94–1.36) | .211 |
Model . | Outcome . | Regression Analysis . | Risk Ratio (95% Confidence Interval) . | P Value . |
---|---|---|---|---|
Mortality | Hazard Ratio | |||
Unadjusted | 30-d mortality | Cox proportional hazards regression | 0.51 (.49–.76) | <.001 |
Adjusteda | 30-d mortality | Cox proportional hazards regression | 0.63 (.51–.78) | <.001 |
Unadjusted | 90-d mortality | Cox proportional hazards regression | 0.75 (.65–.88) | <.001 |
Adjustedb | 90-d mortality | Cox proportional hazards regression | 0.77 (.66–.90) | .001 |
Recurrencec | Odds Ratiod | |||
Unadjusted | Recurrence | Ordinal logistic regression | 1.20 (1.00–1.44) | .047 |
Adjustede | Recurrence | Ordinal logistic regression | 1.13 (.94–1.36) | .211 |
aAdjusted for skin and soft tissue infections, endocarditis, osteomyelitis, Acute Physiology and Chronic Health Evaluation (APACHE) III score, Charlson comorbidity index, admission to the intensive care unit (ICU), and diabetes.
bAdjusted for skin and soft tissue infections, endocarditis, osteomyelitis, APACHE III score, Charlson comorbidity index, hospital-onset infection, admission to the ICU, and diabetes.
cRecurrence is defined as a methicillin-susceptible Staphylococcus aureus (MSSA)–positive blood culture within 45–365 days after first MSSA-positive blood culture.
dOdds ratios are based on a multivariate ordinal logistic regression. As such, the odds ratio represents the cumulative odds of recurrence and is assumed to be the same across cumulative probabilities.
eAdjusted for APACHE III score, Charlson comorbidity index, and hospital-onset infection.
DISCUSSION
This large observational study identified a reduced risk of mortality among patients treated with cefazolin compared with patients treated with nafcillin or oxacillin for MSSA infections complicated by bacteremia. Recurrence of MSSA infections complicated by bacteremia was similar between the 2 treatment groups. These results support current guidelines that list cefazolin as an alternative therapy in the treatment of patients with severe, invasive MSSA infections.
Previous studies evaluated outcomes associated with cefazolin or penicillinase-stable penicillin treatment for patients with MSSA infections complicated by bacteremia [5, 7, 10, 11]. These studies reported risk estimates favoring therapy with cefazolin but did not detect a significant difference between the 2 treatment groups [5, 7, 10, 11]. Li et al reported slightly lower rates of 30-day mortality, 90-day mortality, and recurrent bacteremia among patients receiving cefazolin vs oxacillin for complicated MSSA bacteremia [7]. When comparing cefazolin with cloxacillin for MSSA infections complicated by bacteremia, Bai et al found a nonsignificant 42% reduced risk of 90-day mortality for patients treated with cefazolin in a matched propensity score study whereas Paul et al reported nonsignificant decreased odds of 90-day mortality for patients receiving cefazolin [10, 11]. Furthermore, Lee et al found increased odds of treatment failure, which included both mortality and recurrent infections at 4 weeks and 12 weeks for patients receiving nafcillin vs cefazolin for MSSA infections complicated by bacteremia [5]. The previous studies were much smaller than the current study. Bai et al had the largest sample size of 354 patients whereas the present study included 3167 patients [11]. Therefore, previous studies potentially did not have statistical power to achieve significance.
Prior studies used a variety of outcome definitions [5, 7, 10, 11]. Unfortunately, no standardized definition exists for reporting mortality or recurrent MSSA infections. Even though 90-day mortality rates were examined, follow-up periods (eg, days from admission, culture collection, or initiating therapy) may have differed among studies. This occurrence may have been reflected in the varying rates of mortality (range, 1%–34%) [7, 10, 11]. Furthermore, our study captured deaths that occurred outside of hospital admission whereas previous studies may not have had access to this information.
This study had limitations. First, there was potential for confounding by indication. Patients receiving cefazolin may have differed from patients receiving penicillinase-stable penicillins in ways that were associated with mortality. We attempted to limit the potential for unadjusted confounding by including both measures of illness severity based on vital signs and laboratory values at baseline (ie, APACHE III score) and comorbidities (ie, Charlson comorbidity index). Similarly, residual confounding may impact the outcome due to the sickest patients receiving nafcillin instead of cefazolin. Physicians may be less likely to try cefazolin when the patient has a serious MSSA infection (eg, endocarditis, sepsis) or a high-inoculum infection (eg, valvular vegetation, undrained abscess). Third, information on source control (eg, intravenous catheter removal), dose of therapy, cause of death, and adverse events were not available. In addition, we only studied mortality and recurrence with bacteremia. Our results do not apply to recurrence without bacteremia or other complications of S. aureus bacteremia. Last, a large proportion of the patients who had MSSA infections complicated by bacteremia did not receive definitive therapy with nafcillin, oxacillin, or cefazolin due to receiving empiric or definitive therapy with another antibiotic, dying, being discharged before receiving therapy, or only receiving empiric but not definitive therapy with nafcillin, oxacillin, or cefazolin. Additionally, 269 patients were excluded for receiving definitive therapy with both cefazolin and penicillinase-stable penicillins. Therefore, the results of this study may not be generalizable to all patients who have MSSA infections complicated by bacteremia.
Information was not available regarding intravenous antibiotics received after hospital discharge. After hospital discharge, many veterans receive additional intravenous antibiotics at home for their MSSA infection complicated by bacteremia. The antibiotics are frequently administered by home healthcare services, which are not usually VA affiliated. Therefore, information to inspect duration of therapy and switching therapy due to adverse reactions after hospital discharge is not incorporated in the VA dataset that was used for this study.
In conclusion, definitive therapy with cefazolin may be superior to nafcillin or oxacillin when examining mortality among patients with MSSA infections complicated by bacteremia. Clinicians may want to consider prescribing cefazolin for patients with MSSA infections complicated by bacteremia not involving the central nervous system; however, more research is needed such as randomized trials comparing the efficacy of cefazolin with nafcillin or oxacillin to optimize MSSA bacteremia treatment outcomes among specific groups of patients.
Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Notes
Disclaimer. The VA Office of Research and Development had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation of the manuscript. The opinions expressed are those of the authors and not necessarily those of the Department of Veterans Affairs or the United States government.
Financial support. This work was supported by a VA Health Services Research and Development (HSR&D) Career Development Award (award number CDA 11–211; principal investigator, M. L. S.). M. L. S. was supported by a VA HSR&D Career Development Award (CDA 11–215).
Potential conflicts of interest. J. S. M. has received speaker honoraria from bioMérieux. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
References
Author notes
Presented in Part: IDWeek, San Diego, California, October 7–11, 2015. Abstract 52930.
Comments