Open Collections

UBC Faculty Research and Publications

Comparison of outcomes in patients with methicillin-susceptible Staphylococcus aureus (MSSA) bacteremia… Wong, Davie; Wong, Titus; Romney, Marc; Leung, Victor May 23, 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12879_2016_Article_1564.pdf [ 524.13kB ]
JSON: 52383-1.0357443.json
JSON-LD: 52383-1.0357443-ld.json
RDF/XML (Pretty): 52383-1.0357443-rdf.xml
RDF/JSON: 52383-1.0357443-rdf.json
Turtle: 52383-1.0357443-turtle.txt
N-Triples: 52383-1.0357443-rdf-ntriples.txt
Original Record: 52383-1.0357443-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessComparison of outcomes in patients withmethicillin-susceptible Staphylococcusaureus (MSSA) bacteremia who are treatedwith β-lactam vs vancomycin empirictherapy: a retrospective cohort studyDavie Wong1*, Titus Wong2,3, Marc Romney2,4 and Victor Leung2,4AbstractBackground: Prior studies suggested that vancomycin may be inferior to β-lactams for the empiric treatment ofmethicillin-susceptible S. aureus (MSSA) bacteremia. We assessed whether empiric therapy with β-lactams comparedto vancomycin was associated with differences in clinical outcomes in patients with MSSA bacteremia.Methods: We conducted a retrospective cohort study of adult inpatients with their first episode of MSSA bacteremiaat two tertiary care hospitals in Vancouver, Canada, between 2007 and 2014. Exposure was either empiric β-lactam orvancomycin therapy. All patients received definitive treatment with cloxacillin or cefazolin. The primary outcome was28-day mortality. Secondary outcomes were 90-day mortality, recurrent infection at 6 months, duration of bacteremiaand hospital length-of-stay. Outcomes were adjusted using multivariable logistic regression.Results: Of 814 patients identified, 400 met inclusion criteria (β-lactam = 200, vancomycin = 200). Overall 28-daymortality was 8.5 % (n=34). There were more cases of infective endocarditis in the β-lactam than in the vancomycingroup [45 (22.5 %) vs 23 (11.5 %), p < 0.01]. Adjusted mortality at 28 days was similar between the two groups (OR: 1.14;95 % CI: 0.49–2.64). No differences in secondary outcomes were observed. Transition to cloxacillin or cefazolin occurredwithin a median of 67.8 h in the vancomycin group.Conclusions: Empiric therapy with β-lactams was not associated with differences in all-cause mortality, recurrentinfection, microbiological cure or hospital length-of-stay compared to vancomycin. Vancomycin monotherapymay be appropriate for the empiric treatment of MSSA bacteremia if definitive therapy with cloxacillin orcefazolin can be initiated within 3 days.Keywords: Staphylococcus aureus, Bacteremia, Empiric, Therapy, Beta-lactam, VancomycinBackgroundStaphylococcus aureus is the leading cause of bacteremiaand carries a mortality of 20–30 % in the 21st century[1, 2]. Empiric vancomycin is commonly prescribed forpatients with S. aureus bacteremia (SAB) to covermethicillin-resistant S. aureus (MRSA) as up to 50–60 %of bloodstream isolates are methicillin-resistant at somecentres [3–8]. However, vancomycin is inferior to semi-synthetic anti-Staphylococcal penicillins (e.g., cloxacillin)and first generation cephalosporins (e.g., cefazolin) forthe definitive treatment of methicillin-susceptible S.aureus (MSSA) bacteremia [9–11]. Cloxacillin and cefa-zolin are considered the optimal agents against MSSAand both are equally efficacious in treating MSSAbacteremia [10]. Vancomycin is associated with higherrates of infection-related mortality, re-infection and bacte-riologic failure compared to cloxacillin or cefazolin in the* Correspondence: wongdavi@mail.ubc.ca1PGY-V Infectious Diseases Residency Training Program, University of BritishColumbia, Vancouver General Hospital, D 452 Heather Pavilion, 2733 HeatherStreet, Vancouver, BC V5Z 1 M9, CanadaFull list of author information is available at the end of the article© 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (, which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver( applies to the data made available in this article, unless otherwise stated.Wong et al. BMC Infectious Diseases  (2016) 16:224 DOI 10.1186/s12879-016-1564-5definitive treatment of MSSA bacteremia [9, 12–15].Whether vancomycin is inferior to β-lactams for em-piric therapy remains to be fully elucidated. Early stud-ies suggested that empiric vancomycin was associatedwith worse outcomes compared to empiric β-lactamtherapy [3, 16, 17], but more recent data did not dem-onstrate any differences in outcomes [15]. Althoughcontroversial, some experts recommend the addition ofa β-lactam agent to vancomycin during empiric treat-ment to ensure optimal coverage for MSSA for patientsat the highest risk of morbidity and mortality from SAB[18]. Major limitations of previous studies were failureto control for the definitive therapy prescribed whencomparing empiric regimens and neglecting to specifythe empiric β-lactams used [3, 15–17]. We assessed ifempiric β-lactam compared to vancomycin was associatedwith differences in survival, recurrent infection and micro-biological cure in patients with MSSA bacteremia who re-ceived definitive therapy with cloxacillin or cefazolin.MethodsPatientsWe performed a retrospective cohort study of adult in-patients diagnosed with their first episode of MSSAbacteremia at two tertiary care hospitals in Vancouver,Canada, between January 2007 and December 2014, in-clusive. Consecutive patients were included if they hadMSSA bacteremia and either cloxacillin or cefazolin wasprescribed for definitive therapy (penicillin was an ac-ceptable alternative if the isolate was proven to be sus-ceptible). Patients were excluded if there was missingdata for 28-day mortality, no empiric therapy wasadministered, death occurred within 24 h followingdiagnosis of bacteremia, or polymicrobial bacteremia.Patients were stratified based on empiric treatmentwith β-lactams or vancomycin. The β-lactam groupreceived one or more of cloxacillin, cefazolin, β-lactam/β-lactamase inhibitor combination, a third generationcephalosporin or a carbapenem, with or without vanco-mycin. In both groups, other antimicrobials may havebeen prescribed during empiric and definitive therapy.DefinitionsBacteremia was defined as the isolation of MSSA from oneor more blood culture bottles. Bacteremia identified within72 h of hospital admission was considered community-onset, while bacteremia diagnosed after more than 72 h ofhospital admission was deemed hospital-onset. Immuno-compromised state was present if any of the followingwere described: neutropenia (≤ 1.5 × 109/L), congenital im-mune deficiencies, or use of immunosuppressants (TNF-αinhibitors, prednisone ≥ 10 mg/day or its equivalent,cancer chemotherapy, methotrexate, cyclophosphamide,mycophenolate mofetil, calcineurin inhibitors, mTORinhibitors, azathioprine and any other drug generallyconsidered to significantly weaken the immune system).Definite infective endocarditis was diagnosed using themodified Duke criteria [19]. The source of bacteremiawas either stated explicitly or inferred as the most likelysource based on available clinical data and microbio-logical results. Metastatic complications included infec-tions that occurred distant from the presumed primarysource such as septic emboli, mycotic aneurysms,osteoarticular infections, and distant abscesses. Surgicalsource control included only procedures performed inthe operating theatre. Empiric therapy began with thefirst dose of empiric antibiotics and ended with thestart of definitive therapy. Definitive therapy beganwhen antimicrobial susceptibilities were known andone of the following treatments was prescribed: 1) clox-acillin or cefazolin 2) discontinuation of other empiricantibiotics for patients already on cloxacillin or cefazo-lin empirically, or 3) continuation of empiric cloxacillinor cefazolin. Definitive therapy ended when cloxacillinor cefazolin was stopped. Time to receipt of antibioticswas measured from the time of obtaining the first posi-tive blood culture to the time of the first dose of anti-biotic. If a patient was already on antibiotics at the timeof the first positive blood culture, the time to receipt ofantibiotics was zero. We calculated the hours of em-piric β-lactam exposure based on start and stop dates,start and stop times, and dosing frequency.OutcomesOur primary outcome was 28-day all-cause in-hospitalmortality. Secondary outcomes were 90-day all-cause in-hospital mortality, recurrent infection at 6 months, dur-ation of bacteremia and hospital length-of-stay (LOS).Time to mortality was measured from the date of thefirst positive blood culture to the date of death. Recur-rent infection occurred when MSSA bacteremia was di-agnosed again following completion of a treatmentcourse for the previous episode of MSSA bacteremia.Duration of bacteremia was the time difference be-tween the first positive blood culture and the first nega-tive blood culture. Patients without follow-up bloodcultures were excluded from the analysis for durationof bacteremia. Hospital LOS was measured from thedate of admission to the date of discharge.Data extractionPatients with MSSA bacteremia were extracted from themedical microbiology laboratory information systemsand medical records were reviewed. A single reviewercollected data on patient demographics and comorbidi-ties, blood culture results and antimicrobial therapyfrom electronic and paper databases.Wong et al. BMC Infectious Diseases  (2016) 16:224 Page 2 of 9Statistical analysisOur predicted mortality difference between the β-lactamand vancomycin group was 15 % based on a previousstudy [17]. We estimated a sample size of 200 for eachgroup to capture a 15 % difference in mortality with80 % power at a two-tailed alpha level of 0.05. Baselinecategorical variables were described as counts and per-centages, and differences between groups were assessedwith chi-square or Fisher’s exact tests. Continuous vari-ables were presented as means and standard deviations,or medians and interquartile range. Differences betweengroups were assessed using parametric t-tests or non-parametric Mann–Whitney-U tests, as appropriate. Lo-gistic regression methods were used to model the oddsratio of death and recurrent infection in the β-lactamcompared to vancomycin group. Linear regressionmodel was conducted for hospital LOS and duration ofbacteremia. The two outcomes were log-transformed inthe analysis to improve normality of the distribution ofresiduals. All models were adjusted for pre-specifiedconfounding variables including age, sex, age-adjustedCharlson-comorbidity index [20], Pitt bacteremia score,infectious diseases consultation, infective endocarditisand time to receipt of empiric antibiotics. These factorshave been shown to affect mortality in patients withSAB [1]. The duration of bacteremia was further ad-justed for surgical source control. All analyses wereperformed using the SAS 9.4 software.ResultsWe identified 814 patients with MSSA bacteremia be-tween January 2007 and December 2014, inclusive (Fig. 1).We excluded 414 patients primarily because 60.4 % didnot receive cloxacillin or cefazolin for definitive therapy.These patients either remained on broad-spectrum anti-microbials or received vancomycin for definitive therapydue to suspected or confirmed penicillin allergy. Another22.9 % were not started on empiric therapy. Our cohortconsisted of 64.5 % males and 82.8 % of patients hadcommunity-onset bacteremia (Table 1). Infectious diseasesconsultation was obtained in most cases (70.5 %), but washigher in the β-lactam than in the vancomycin group(75 % vs 66 %, p=0.05). The most common sources ofbacteremia were unknown (26 %), injection drug use(21.8 %), peripheral or central venous catheters (15.3 %),and skin and soft tissue infections (15 %). Infective endo-carditis was diagnosed more frequently in the β-lactamthan in the vancomycin group (22.5 % vs 11.5 %, p < 0.01).The prevalence of infective endocarditis was 17.0 %. Theβ-lactam group experienced more metastatic complica-tions (36.5 % vs 26.5 %, p=0.03) and underwent surgicalsource control more frequently (21.5 % vs 13 %, p=0.02)compared to the vancomycin group.The most common empiric antimicrobials prescribedin the β-lactam group were vancomycin (76.5 %), cloxa-cillin or cefazolin (69 %), 3rd generation cephalosporins(40 %) and piperacillin-tazobactam (34.5 %). The use ofmultiple β-lactam antibiotics reflects changes made dur-ing empiric therapy. Among the subgroup of patientswho received combination therapy with β-lactam plusvancomycin (153/200), cloxacillin or cefazolin (62.1 %),3rd generation cephalosporins (47.7 %) and piperacillin-tazobactam (42.5 %) were the most common empiric β-lactams prescribed (Table 2). Cloxacillin or cefazolin(91.5 %) was the predominant empiric β-lactam used inthe monotherapy subgroup. Initiation of cloxacillin orcefazolin was delayed in the combination subgroup com-pared to the β-lactam monotherapy subgroup (medianFig. 1 Patient enrollment process. Abbreviations: MSSA methicillin-susceptible S. aureus, BSI bloodstream infectionWong et al. BMC Infectious Diseases  (2016) 16:224 Page 3 of 9Table 1 Baseline characteristics and clinical outcomes of patients with methicillin-susceptible S. aureus bacteremia. Patients receivedempiric antimicrobial therapy with either β-lactams or vancomycinPatient characteristics β-lactama (n=200) Vancomycina (n=200) P-valueAgeb 53.0 ± 16.9 57.9 ± 18.4 0.01Males 126 (63.0) 132 (66.0) 0.53Community-onset 173 (86.5) 158 (79.0) 0.05Hospital-onset 27 (13.5) 42 (21.0) 0.05HIV infection 20 (10.0) 16 (8.0) 0.48Hepatitis C infection 61 (30.5) 55 (27.5) 0.51Immunocompromised 18 (9.0) 16 (8.0) 0.72Alcohol or illicit drug abuse 82 (41.0) 71 (35.5) 0.26Intravenous drug use 64 (32.0) 56 (28.0) 0.38Charlson comorbidity indexc 3 (1.0–6.0) 4 (1.0–7.0) 0.01Pitt bacteremia scorec 1 (0–2) 1 (0–2) 0.18Infectious diseases consultation 150 (75.0) 132 (66.0) 0.05Source of bacteremiaCentral or peripheral line 24 (12.0) 37 (18.5) 0.09Skin and soft tissue 36 (18.0) 24 (12.0) 0.12Intravenous drug use 48 (24.0) 39 (19.5) 0.33Bone or joint infection 23 (11.5) 13 (6.5) 0.11Lung 6 (3.0) 7 (3.5) 1.00Other 19 (9.5) 20 (10.0) 1.00Unknown 44 (22.0) 60 (30.0) 0.09Infective endocarditis 45 (22.5) 23 (11.5) < 0.01Metastatic complications 73 (36.5) 53 (26.5) 0.03Surgical source control 43 (21.5) 26 (13.0) 0.02Empiric antimicrobialsβ-lactam 200 (100) 75 (37.5) < 0.0001Cloxacillin or cefazolin 138 (69.0) 10 (5.0) < 0.00013rd generation cephalosporin 80 (40.0) 35 (17.5) < 0.0001Piperacillin-tazobactam 69 (34.5) 35 (17.5) < 0.001Ticarcillin-clavulanic acid 3 (1.5) 1 (0.5) 0.62Carbapenem 8 (4.0) 4 (2.0) 0.24Vancomycin 153 (76.5) 197 (98.5) < 0.0001Daptomycin 2 (1.0) 1 (0.5) 1.00Linezolid 2 (1.0) 2 (1.0) 1.00Otherd 82 (41.0) 97 (48.5) 0.13Blood culture time to positivitye 20.2 (16.6–25.5) 18.5 (16.2–23.3) 0.02Duration of empiric therapye 55.4 (44.2–72.5) 52.1 (39.2–75.7) 0.55Duration of definitive therapyf 28 (13.0–42.0) 26.5 (11.0–42.0) 0.14Time to receipt of empiric therapye 1.92 (0.1–6.9) 10.8 (1.4–24.1) < 0.0001Time to receipt of β-lactame 2.92 (0.3–13.8) 50.5 (4.8–75.5) < 0.0001Time to receipt of cloxacillin or cefazoline 31.0 (13.8–50.8) 67.8 (50.3–88.0) < 0.0001Empiric β-lactam exposure 52.3 (39.8–71.9)e60.6 ± 39.9g0 (0–16.2)e9.5 ± 38.5g< 0.0001Proportional empiric β-lactam exposure 100 (100–100)h94.5 ± 46.9i0 (0–24.5)h14.0 ± 44.8i< 0.0001Wong et al. BMC Infectious Diseases  (2016) 16:224 Page 4 of 934.8 vs 13.0 h, p=0.00). The combination subgroup hada higher Pitt bacteremia score (median 1 vs 0, p < 0.01),received more infectious diseases consultations (78.4 %vs 63.8 %, p=0.05), and experienced more metastaticcomplications (43.1 % vs 14.9 %, p < 0.001) than the β-lactam monotherapy subgroup. Both the duration ofbacteremia (median 84.8 vs 63.4 h, p=0.03) and hospitalLOS (median 26 vs 15 days, p < 0.01) were longer in thecombination subgroup.Almost half of patients in the vancomycin group re-ceived additional antimicrobials during empiric or defini-tive therapy (Table 1). Rifampin and aminoglycosides wereadded for synergy in patients with prosthetic valve infect-ive endocarditis or prosthetic joint infections. Fluoroqui-nolones, macrolides, trimethoprim-sulfamethoxazole, andclindamycin were used for either treatment of non-bacteremic co-infections or prophylaxis for other med-ical conditions. Three patients in the vancomycin groupreceived an incomplete dose of vancomycin and werecounted as not having received it.There was a greater delay in receipt of empiric anti-microbials in the vancomycin group compared to theβ-lactam group (median 10.8 vs 1.9 h, p < 0.0001)(Table 1). Seventy five (37.5 %) patients in the vanco-mycin group were briefly exposed to β-lactams duringempiric therapy. However, exposure time (median 0 vs52.3 h, p < 0.0001) and proportional time of exposure ofthe empiric period (median 0 vs 100 %, p < 0.0001) wereminiscule compared to the β-lactam group. There wereno differences in clinical outcomes between the twogroups (Table 3). The overall 28-day and 90-day mor-tality was 34 (8.5 %) and 57 (14.3 %) respectively.Among patients with infective endocarditis, 28-day and90-day mortality was 4 (8.89 %) and 7 (15.6 %) in the β-lactam group and 0 and 3 (13.0 %) in the vancomycingroup. In the vancomycin group, the subset of patientswho received brief exposure to empiric β-lactam expe-rienced faster clearance of bacteremia compared tothose who did not have any empiric β-lactam exposure(median 78.9 vs 96.3 h, p=0.04). In comparison with theβ-lactam group, patients in the vancomycin group whodid not have any empiric β-lactam exposure had slightlyhigher 90-day mortality [22 (17.6 %) vs 25 (12.5 %), p=0.01]and longer duration of bacteremia (median 96.3 vs 74.4 h,p=0.03), while those who were briefly exposed to empiricβ-lactams exhibited no difference in clinical outcomes.DiscussionThe goal of our study was to assess if empiric β-lactamscompared to vancomycin was associated with differencesin outcomes in patients with MSSA bacteremia. Wefound no differences in all-cause mortality at 28 and90 days, recurrent infection at 6 months, duration ofbacteremia or hospital LOS between patients treatedwith empiric β-lactam or vancomycin therapy. However,in the vancomycin group, the subset of patients whowere not exposed to any empiric β-lactams had higher90-day mortality and longer duration of bacteremia,while those who had even minimal exposure to empiricβ-lactams did not have worse outcomes. Patients in thevancomycin group were older, had more medical comor-bidities, were less likely to be assessed by an infectiousdiseases consultant, underwent fewer source controlprocedures, and experienced a greater delay in receipt ofempiric antibiotics compared to the β-lactam group.Table 1 Baseline characteristics and clinical outcomes of patients with methicillin-susceptible S. aureus bacteremia. Patients receivedempiric antimicrobial therapy with either β-lactams or vancomycin (Continued)Primary outcome28-day mortality 16 (8.0) 18 (9.0) 0.72Secondary outcomes90-day mortality 25 (12.5) 32 (16.0) 0.32Recurrent infection at 6 months 7 (3.5) 8 (4.0) 0.79Duration of bacteremiae,j 74.4 (48.3–130) 89.7 (56.7–132) 0.20≥ 3 daysj 98 (53.8) 111 (60.0) 0.25Hospital length of stayf 22.5 (12.5–43.0) 22 (13.0–45.0) 0.59aVariables are displayed as counts and percentages in parentheses unless otherwise specifiedbAge is represented as a mean ± standard deviation in yearscVariables are expressed as a median with interquartile range in parenthesesdOther antimicrobials used during empiric and definitive therapy included rifampin, aminoglycosides, fluoroquinolones, macrolides, trimethoprim-sulfamethoxazole,and clindamycineVariables are expressed as median hours with interquartile range in parenthesesfVariables are expressed as median days with interquartile range in parenthesesgVariables are expressed a mean ± standard deviation in hourshVariables are expressed as median percentages with interquartile range in parenthesesiVariables are expressed a mean percentage ± standard deviationjData missing for 18 and 15 patients in the β-lactam and vancomycin group, respectivelyWong et al. BMC Infectious Diseases  (2016) 16:224 Page 5 of 9Table 2 Baseline characteristics and clinical outcomes of patients with methicillin-susceptible S. aureus bacteremia. Patients receivedeither empiric combination therapy with β-lactam plus vancomycin or empiric β-lactam monotherapyPatient characteristics β-lactam plus vancomycina (n=153) β-lactam monotherapya (n=47) P-valueAgeb 51.3 ± 16.9 58.7 ± 16.1 < 0.01Males 92 (60.1) 34 (72.3) 0.17Community-onset 137 (89.5) 36 (76.6) 0.03Hospital-onset 16 (10.5) 11 (23.4) 0.03HIV infection 18 (11.8) 2 (4.26) 0.17Hepatitis C infection 50 (32.7) 11 (23.4) 0.28Immunocompromised 12 (7.84) 6 (12.8) 0.38Alcohol or illicit drug abuse 68 (44.4) 14 (29.8) 0.09Intravenous drug use 54 (35.3) 10 (21.3) 0.08Charlson comorbidity indexc 3 (1–6) 3 (1–5) 0.30Pitt bacteremia scorec 1 (0–2) 0 (0–1) < 0.01Infectious diseases consultation 120 (78.4) 30 (63.8) 0.05Source of bacteremiaCentral or peripheral line 15 (9.80) 9 (19.1) 0.12Skin and soft tissue 25 (16.3) 11 (23.4) 0.28Intravenous drug use 43 (28.1) 5 (10.6) 0.02Bone or joint infection 15 (9.80) 8 (17.0) 0.19Lung 4 (2.61) 2 (4.26) 0.63Other 14 (9.15) 5 (10.6) 0.78Unknown 37 (24.2) 7 (14.9) 0.23Infective endocarditis 40 (26.1) 5 (10.6) 0.03Metastatic complications 66 (43.1) 7 (14.9) < 0.001Surgical source control 31 (20.3) 12 (25.5) 0.43Empiric antimicrobialsCloxacillin or cefazolin 95 (62.1) 43 (91.5) < 0.00013rd generation cephalosporin 73 (47.7) 7 (14.9) < 0.0001Piperacillin-tazobactam 65 (42.5) 4 (8.51) < 0.0001Ticarcillin-clavulanic acid 2 (1.31) 1 (2.13) 0.55Carbapenem 7 (4.58) 1 (2.13) 0.68Daptomycin 0 2 (4.26) 0.05Linezolid 1 (0.65) 1 (2.13) 0.42Otherd 60 (39.2) 22 (46.8) 1.00Blood culture time to positivitye 20.3 (16.3–25.3) 19.8 (18.0–27) 0.36Duration of empiric therapye 55.5 (44.9–73.8) 54.3 (38.8–64.5) 0.08Duration of definitive therapyf 31 (14–43) 26 (12–40) 0.18Time to receipt of empiric therapye 1.6 (0.03–6.17) 3.95 (0.58–15.5) 0.08Time to receipt of β-lactame 2.77 (0.25–13.7) 3.95 (0.58–16.6) 0.74Time to receipt of cloxacillin or cefazoline 34.8 (21.2–58.4) 13.0 (1.83–23.8) 0.00Empiric β-lactam exposuree 52.3 (39.8–73.8) 51.7 (38.8–64.5) 0.42Proportional empiric β-lactam exposureg 100 (86.1–100) 100 (95.2–100) 0.01Primary outcome28-day mortality 14 (9.15) 2 (4.26) 0.37Wong et al. BMC Infectious Diseases  (2016) 16:224 Page 6 of 9More cases of infective endocarditis and metastatic com-plications were diagnosed β-lactam group. Despite thehigh prevalence of MRSA at both of our institutions(25 % and 38 %), only 76.5 % of patients in the β-lactamgroup received vancomycin empirically as well. Perhapsthe awareness of MRSA was low among some treatingclinicians or patients who did not receive empiric vanco-mycin were judged to be at low risk for MRSA infection.Although the combination of β-lactams with vanco-mycin exhibits synergistic killing against MRSA, neithersynergy nor antagonism was observed against MSSA invitro [21]. Therefore, the addition of vancomycin to β-lactams would not be expected to influence microbio-logical cure in MSSA bacteremia.Interestingly, the differential time delay in receipt ofempiric antimicrobials was unexpected. The β-lactamgroup received antimicrobial therapy earlier possibly be-cause these patients were more severely ill as reflectedin their higher rate of infective endocarditis and meta-static complications, despite similar Pitt bacteremiascores between the two groups. Infectious diseases con-sultation may have also contributed to earlier initiationof antibiotics in the β-lactam group.In the β-lactam group, patients who received empiricβ-lactam plus vancomycin were generally sicker as indi-cated by their higher Pitt bacteremia score, higher rateof infective endocarditis and metastatic complications,and longer duration of bacteremia and hospital LOS,compared to those who received empiric β-lactammonotherapy. The greater severity of illness in thiscombination subgroup may explain the initial use ofbroad-spectrum β-lactams (ceftriaxone or piperacillin-tazobactam), with subsequent de-escalation to cloxacillinor cefazolin in some patients during the empiric period bythe infectious diseases consultant when S. aureus wasidentified in the blood culture. De-escalation occurredwithin a median of 34.8 h, which follows the time to posi-tivity of the first blood culture (median 20.3 h).The 28-day and 90-day mortality in our study was lowat 34 (8.5 %) and 57 (14.3 %) respectively, but is withinthe range of 3.6 to 51.7 % reported in a meta-analysis ofpatients with MSSA bacteremia from catheter-related in-fections and infective endocarditis by Cosgrove et al.[22]. Definite infective endocarditis was diagnosed in 68(17 %) of our patients, which is similar to rates reportedin previous studies [10, 14, 16, 23].Table 2 Baseline characteristics and clinical outcomes of patients with methicillin-susceptible S. aureus bacteremia. Patients receivedeither empiric combination therapy with β-lactam plus vancomycin or empiric β-lactam monotherapy (Continued)Secondary outcomes90-day mortality 21 (13.7) 4 (8.51) 0.45Recurrent infection at 6 months 5 (3.27) 2 (4.26) 0.67Duration of bacteremiae,h 84.8 (52.5–136) 63.4 (30.5–114) 0.03≥ 3 daysh 81 (56.3) 17 (44.7) 0.27Hospital length of stayf 26 (13–45) 15 (10–30) < 0.01aVariables are displayed as counts and percentages in parentheses unless otherwise specifiedbAge is represented as a mean ± standard deviation in yearscVariables are expressed as a median with interquartile range in parenthesesdOther antimicrobials used during empiric and definitive therapy included rifampin, aminoglycosides, fluoroquinolones, macrolides, trimethoprim-sulfamethoxazole,and clindamycineVariables are expressed as median hours with interquartile range in parenthesesfVariables are expressed as median days with interquartile range in parenthesesgVariables are expressed as median percentages with interquartile range in parentheseshData missing for 9 patients in each subgroupTable 3 Outcome analysis comparing β-lactam versus vancomycin group. Variables were adjusted for predefined confoundingvariables, including age, sex, age-adjusted Charlson-comorbidity index, Pitt bacteremia score, infectious diseases consultation,infective endocarditis and time to receipt of empiric antibiotics. Duration of bacteremia was further adjusted for surgical source controlOutcomes Crude OR (95 % CI) P-value Adjusted OR (95 % CI) P-value28-day mortality 0.88 (0.44–1.78) 0.72 1.14 (0.49–2.64) 0.7690-day mortality 0.75 (0.43–1.32) 0.32 1.01 (0.51–2.02) 0.97Recurrent infection at 6 months 0.87 (0.31–2.45) 0.79 1.27 (0.39–4.11) 0.69Ratio of Mean (95 % CI) P-value Adjusted Ratio of Mean (95 % CI) P-valueDuration of bacteremia 0.92 (0.78–1.07) 0.27 0.94 (0.79–1.11) 0.44Hospital length-of-stay 0.96 (0.81–1.14) 0.65 0.95 (0.80–1.14) 0.60Abbreviations: OR odds ratio, CI confidence intervalWong et al. BMC Infectious Diseases  (2016) 16:224 Page 7 of 9In the vancomycin group, the median time to defini-tive treatment with cloxacillin or cefazolin was 67.8 h,which is comparable to other studies [3, 16, 17]. In theKhatib study, a delayed clearance of bacteremia (≥ 3 days)was observed in the vancomycin group compared to theβ-lactam group (57.6 % vs 37.5 %). However, there wereno differences in all-cause or attributable mortality be-tween groups. Among injection drug users with predom-inantly right-sided MSSA infective endocarditis, Lodiseet al. demonstrated that infection-related mortality waslower with empiric β-lactam than with vancomycinmonotherapy (11.4 % vs 39.3 %, p=0.005) [17]. Evenwhen patients were switched from vancomycin to asemi-synthetic penicillin within a median of 3 days,infection-related mortality remained high at 40.9 %. Theoverall mortality in this cohort was unusually high at22.2 % compared to a rate of 0–4 % described in a sys-tematic review by Yung et al. [24]. In contrast, a recentstudy by McDanel et al. found that empiric β-lactamtherapy (predominantly piperacillin-tazobactam and cef-triaxone) compared to vancomycin was not associatedwith differences in mortality in patients with MSSAbacteremia [15]. However, the McDanel study excludedpatients who received empiric vancomcyin plus β-lactams, and clinical outcomes were evaluated independ-ent of the antimicrobial prescribed for definitive therapy.At institutions where MRSA prevalence is significant,vancomycin is generally accepted as an appropriate em-piric antimicrobial for SAB, but due to conflicting resultsfrom prior studies, the addition of a β-lactam agent toempiric therapy remains controversial. β-lactam mono-therapy may be inadequate empiric treatment if thestrain is methicillin-resistant, but mortality outcomesfrom retrospective studies are mixed [3, 6, 8, 25]. Ultim-ately, the choice of empiric therapy will depend on pa-tient factors, the prevalence of MRSA in the population,and the ability of the microbiology laboratory to rapidlydifferentiate MSSA from MRSA. From the perspective ofantimicrobial stewardship, vancomycin monotherapyseems favourable as the reduction in usage of β-lactamsmay decrease the potential for drug-drug interactionsand adverse effects during empiric therapy for SAB.A particular strength of our study is the inclusion ofa large proportion of patients who received optimalanti-MSSA agents (cloxacillin and cefazolin) empiric-ally. The major limitation of previous studies was thelack of transparency regarding the empiric β-lactamsprescribed [3, 17]. This is important because not all β-lactams have the same activity against MSSA. In oneretrospective study, second and third generation cepha-losporins and β-lactam/β-lactamase inhibitor combina-tions were inferior to cloxacillin and cefazolin forempiric treatment of MSSA bacteremia [26]. We didnot perform subgroup analysis for different β-lactamsbecause antimicrobials were frequently switched duringempiric therapy.Our study has several limitations. The reason for thelack of difference in the primary outcome is likely multi-factorial. Because of the low event rate in both groups,our study was potentially underpowered to detect a sig-nificant difference in mortality. The lower than expecteddeath rate may be partly due to the exclusion of patientswho died within 24 h of the diagnosis of SAB and of pa-tients who remained on broad-spectrum β-lactams. Thisgroup may have represented a sicker population andthus, we may have selected for less critically ill patients.Due to the retrospective nature of this study, baselinecharacteristics between the two groups were significantlydifferent, although we did attempt to control for thesedifferences in the multivariable model. Future studieswill need to employ matching strategies to eliminate thisimbalance. Although there appears to be a protective ef-fect of even brief exposure to empiric β-lactams, cautionmust be exercised when interpreting this data due to thesmall sample size and major differences in baseline char-acteristics between treatment groups. Because the sourceof bacteremia was not identified in a significant propor-tion of patients in both groups, clinical outcomes mayhave been impacted by a lack of source control. As twodifferent microbiology laboratories were involved in thestudy, differences in the detection methods of MSSAmay have affected the timing of definitive therapy.Obtaining subsequent blood cultures was often delayedor sometimes not performed at all, which may have ledto an overestimation of the duration of bacteremia inboth groups. Data regarding adverse effects were notcollected due to the inherent difficulty of establishingdrug-related events in a retrospective study. We werenot able to determine if patients received appropriatedosing of antibiotics because data on antibiotic dosesand vancomycin trough levels were not collected. A ran-domized controlled trial would be needed to confirmour study findings.ConclusionsEmpiric therapy with β-lactams was not associatedwith differences in all-cause mortality, recurrent infec-tion, microbiological cure or hospital LOS comparedto vancomycin in patients with MSSA bacteremia.Vancomycin monotherapy may be appropriate for theempiric treatment of MSSA bacteremia if definitivetherapy with cloxacillin or cefazolin can be initiatedwithin 3 days.AbbreviationsBSI: bloodstream infection; LOS: length-of-stay; MRSA: methicillin-resistantStaphylococcus aureus; MSSA: methicillin-susceptible Staphylococcus aureus;SAB: Staphylococcus aureus bacteremia.Wong et al. BMC Infectious Diseases  (2016) 16:224 Page 8 of 9AcknowledgementsWe thank the members of Centre for Health Evaluation & Outcome Sciences,Dr. Joel Singer, Rachel McKay, Joseph Puyat and Hong Qian, for theirassistance with the statistical analyses.FundingThis study was not supported by external funding.Availability of data and materialsData supporting the findings are included in the manuscript.Authors’ contributionsDW conceived and designed the study, collected and analyzed the data,drafted and revised the manuscript. TW participated in study design andrevised the manuscript. MR made critical revisions to the manuscript. VLparticipated in study design, interpreted the data and revised themanuscript. All authors read and approved the final manuscript.Competing interestsThe authors declare that they have no competing interests.Consent to publishNot applicable.Ethics and consent to participateThe study was approved by the research ethics board at the University ofBritish Columbia with waiver of informed consent, and received institutionalapproval from Vancouver Coastal Health and Providence Healthcare.Author details1PGY-V Infectious Diseases Residency Training Program, University of BritishColumbia, Vancouver General Hospital, D 452 Heather Pavilion, 2733 HeatherStreet, Vancouver, BC V5Z 1 M9, Canada. 2Department of Pathology andLaboratory Medicine, University of British Columbia, Vancouver, BC, Canada.3Division of Medical Microbiology and Infection Control, Vancouver GeneralHospital, JPPN1, Medical Microbiology Laboratory, 899 W 12th Ave,Vancouver, BC V5Z 1 M9, Canada. 4Division of Medical Microbiology, St.Paul’s Hospital, Medical Microbiology Laboratory, 1081 Burrard St., Vancouver,BC V6Z 1Y6, Canada.Received: 26 November 2015 Accepted: 12 May 2016References1. Van Hal SJ, Jensen SO, Vaska VL, Espedido BJ, Paterson DL, Gosbell IB.Predictors of mortality in Staphylococcus aureus bacteremia. Clin MicrobiolRev. 2012;25:362–86.2. Thwaites GE, Edgeworth JD, Gkrania-Klotsas E, Kirby A, Tilley R, Torok ME, etal. Clinical management of Staphylococcus aureus bacteraemia. Lancet InfectDis. 2011;11:208–22.3. Khatib R, Saeed S, Sharma M, Riederer K, Fakih MG, Johnson LB. Impact of initialantibiotic choice and delayed appropriate treatment on the outcome ofStaphylococcus aureus bacteremia. Eur J Clin Microbiol Infect Dis. 2006;25:181–5.4. Paul M, Kariv G, Goldberg E, Raskin M, Shaked H, Hazzan R, et al. Importance ofappropriate empirical antibiotic therapy for methicillin-resistant Staphylococcusaureus bacteraemia. J Antimicrob Chemother. 2010;65:2658–65.5. David MZ, Daum RS. Community-associated methicillin-resistantStaphylococcus aureus: epidemiology and clinical consequences of anemerging epidemic. Clin Microbiol Rev. 2010;23:616–87.6. Lodise TP, McKinnon PS, Swiderski L, Rybak M. Outcomes analysis ofdelayed antibiotic treatment for hospital-acquired Staphylococcus aureusbacteremia. Clin Infect Dis. 2003;36:1418–23.7. Marchaim D, Kaye KS, Fowler VG, Anderson DJ, Chawla V, Golan Y. Case–control study to identify factors associated with mortality among patientswith methicillin-resistant Staphylococcus aureus bacteraemia. Clin MicrobiolInfect. 2010;16:747–52.8. Schweizer ML, Furuno JP, Harris AD, Johnson JK, Shardell MD, McGregor JC.Empiric antibiotic therapy for Staphylococcus aureus bacteremia May Not reducein-hospital mortality: a retrospective cohort study. PLoS One. 2010;5:e11432.9. Kim SH, Kim KH, Kim HB, Kim NJ, Kim EC, Oh MD, et al. Outcome ofvancomycin treatment in patients with methicillin-susceptible Staphylococcusaureus bacteremia. Antimicrob Agents Chemother. 2008;52:192–7.10. Lee S, Choe PG, Song KH, Park SW, Kim HB, Kim NJ, et al. Is cefazolin inferiorto nafcillin for treatment of methicillin-susceptible Staphylococcus aureusbacteremia? Antimicrob Agents Chemother. 2011;55:5122–6.11. Liu C, Bayer A, Cosgrove SE, Daum RS, Fridkin SK, Gorwitz RJ, et al. Clinicalpractice guidelines by the infectious diseases society of America for thetreatment of methicillin-resistant Staphylococcus aureus infections in adultsand children. Clin Infect Dis. 2011;52:1–38.12. Chang FY, Peacock JE, Musher Jr DM, Triplett P, MacDonald BB, Mylotte JM, etal. Staphylococcus aureus bacteremia recurrence and the impact of antibiotictreatment in a prospective multicenter study. Medicine. 2003;82:333–9.13. Chan KE, Warren HS, Thadhani RI, Steele DJR, Hymes JL, Maddux FW, et al.Prevalence and outcomes of antimicrobial treatment for Staphylococcus aureusbacteremia in outpatients with ESRD. J Am Soc Nephrol. 2012;23:1551–9.14. Stryjewski ME, Szczech LA, Benjamin Jr DK, Inrig JK, Kanafani ZA, EngemannJJ, et al. Use of vancomycin or first-generation cephalosporins for thetreatment of hemodialysis-dependent patients with MSSA bacteremia. ClinInfect Dis. 2007;44:190–6.15. McDanel JS, Perencevich EN, Diekema DJ, Herwaldt LA, Smith TC,Chrischilles EA, et al. Comparative effectiveness of beta-lactams versusvancomycin for treatment of methicillin-susceptible Staphylococcus aureusbloodstream infections among 122 hospitals. Clin Infect Dis. 2015;61:361–7.16. Schweizer ML, Furuno JP, Harris AD, Johnson JK, Shardell MD, McGregor JC,et al. Comparative effectiveness of nafcillin or cefazolin versus vancomycinin methicillin-susceptible Staphylococcus aureus bacteremia. BMC Infect Dis.2011;11:279.17. Lodise Jr TP, McKinnon PS, Levine DP, Rybak MJ. Impact of empirical-therapyselection on outcomes of intravenous drug users with infective endocarditiscaused by MSSA. Antimicrob Agents Chemother. 2007;51:3731–3.18. McConeghy KW, Bleasdale SC, Rodvold KA. The empirical combination ofvancomycin and a β-lactam for staphylococcal bacteremia. Clin Infect Dis.2013;57:1760–5.19. Li JS, Sexton DJ, Mick N, Nettles R, Fowler Jr VG, Ryan T, et al. Proposedmodifications to the duke criteria for the diagnosis of infective endocarditis.Clin Infect Dis. 2000;30:633–8.20. Schneeweiss S, Wang PS, Avorn J, Glynn RJ. Improved comorbidityadjustment for predicting mortality in medicare populations. Health ServRes. 2003;38:1103–20.21. Joukhadar C, Pillai S, Wennersten C, Moellering RC, Eliopoulos Jr GM. Lack ofbactericidal antagonism or synergism in vitro between oxacillin andvancomycin against methicillin-susceptible strains of Staphylococcus aureus.Antimicrob Agents Chemother. 2010;54:773–7.22. Cosgrove SE, Sakoulas G, Perencevich EN, Schwaber MJ, Karchmer AW,Carmeli Y. Comparison of mortality associated with methicillin-resistant andmethicillin-susceptible Staphylococcus aureus bacteremia: a meta-analysis.Clin Infect Dis. 2003;36:53–9.23. Bai AD, Showler A, Burry L, Steinberg M, Ricciuto DR, Fernandes T, et al.Comparative effectiveness of cefazolin versus cloxacillin as definitiveantibiotic therapy for MSSA bacteraemia: results from a large multicentrecohort study. J Antimicrob Chemother. 2015;70:1539–46.24. Yung D, Kottachchi D, Neupane B, Haider S, Loeb M. Antimicrobials forright-sided endocarditis in intravenous drug users: a systematic review.J Antimicrob Chemother. 2007;60:921–8.25. Rodriguez-Bano J, Millan AB, Dominguez MA, Borraz C, Gonzalez MP,Almirante B, et al. Impact of inappropriate empirical therapy for sepsis dueto health care-associated methicillin-resistant Staphylococcus aureus. J Infect.2009;58:131–7.26. Paul M, Zemer-Wassercug N, Talker O, Lishtzinsky Y, Lev B, Samra Z, et al. Areall beta-lactams similarly effective in the treatment of methicillin-sensitiveStaphylococcus aureus bacteraemia. Clin Microbiol Infect. 2011;17:1581–6.Wong et al. BMC Infectious Diseases  (2016) 16:224 Page 9 of 9


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items