J Antimicrob Chemother 2018; 73: 2643–2651

doi:10.1093/jac/dky259 Advance Access publication 3 August 2018

Optimal treatment of MSSA bacteraemias: a meta-analysis of cefazolin

versus antistaphylococcal penicillins
Monique R. Bidell, Nimish Patel and J. Nicholas O’Donnell *

Department of Pharmacy Practice, Albany College of Pharmacy and Health Sciences, Albany, NY, USA

*Corresponding author. Albany College of Pharmacy and Health Sciences, 106 New Scotland Ave, Albany, NY 12208, USA. Tel: !1-518-694-7203;

Downloaded from https://academic.oup.com/jac/article/73/10/2643/5066380 by guest on 11 March 2024

Fax: !1-518-694-7382; E-mail: nick.odonnell@acphs.edu orcid.org/0000-0001-5184-1385

Received 5 February 2018; returned 3 April 2018; revised 29 May 2018; accepted 7 June 2018

Background: Bacteraemias caused by MSSA are associated with significant morbidity and mortality.
Controversy exists over the optimal treatment of severe infections caused by MSSA. This systematic review and
meta-analysis aims to identify whether differences in clinical outcomes exist between cefazolin and antistaphy-
lococcal penicillins (ASPs).
Methods: PubMed, Cochrane Library and Embase were systematically searched for publications reporting clinical
outcomes of cefazolin and ASPs for adult patients with MSSA bacteraemias throughout November 2017.
Comparative studies reporting 90 day mortality associated with each treatment were included. Random effects
models were used to evaluate the impact of directed treatment agent on the odds of 30 and 90 day mortality,
clinical failure, discontinuation due to adverse effects and infection recurrence.
Results: Five hundred and ninety-nine articles were evaluated for inclusion, of which seven met all inclusion cri-
teria. Across all studies, 1589 patients received cefazolin and 2802 received an ASP. All-cause 90 day mortality
was lower in patients who received cefazolin (OR 0.63, 95% CI 0.41–0.99; I2 " 58%). Odds of discontinuation due
to adverse events was significantly lower in patients receiving cefazolin (OR 0.25, 95% CI 0.11–0.56; I2 " 13%).
No differences in clinical failure were observed (OR 0.85, 95% CI 0.41–1.76; I2 " 74%).
Conclusions: This meta-analysis identified a significant decrease in mortality associated with cefazolin therapy
for MSSA bacteraemia compared with ASPs, though no differences in clinical failure were observed. Additionally,
cefazolin appeared to be better tolerated. These results should be interpreted with caution given the uncon-
trolled and retrospective nature of the included studies.

Introduction Optimal treatment for complicated patients with deep-seated

Staphylococcus aureus is among the most commonly identified
pathogens in patients with bacteraemias.1,2 Although MRSA is often
of principal concern, MSSA remains prevalent and is associated with
significant morbidity and mortality.1,3 First-line treatment for ser-
ious MSSA infections (e.g. bacteraemia) is b-lactam therapy; name-
ly, the antistaphylococcal penicillins (ASPs) (e.g. nafcillin, oxacillin),
and the first-generation cephalosporin, cefazolin. Clinical treatment
guidelines endorsed by the Infectious Diseases Society of America
generally favour ASPs, with cefazolin recommended as an alterna-
tive agent (in the absence of CNS involvement).4–6
Both ASPs and cefazolin are highly active in vitro against MSSA
and used regularly for these infections in clinical practice.4–7
However, identification of an optimal treatment approach is critical
given the substantial risk of bacteraemia-related complications.2
Several retrospective studies of patients with serious MSSA infections
have compared clinical outcomes between ASPs and cefazolin,
though they have largely been unable to identify differences.8–14
infections such as endocarditis and osteomyelitis is of particular con-
cern and remains controversial.15–17 Some clinicians may select an
ASP for deep-seated infections based on concerns of an inoculum
effect observed with cefazolin in vitro.18–20 MICs of cefazolin for
MSSA have been shown to increase in certain isolates producing
type A b-lactamases with higher bacterial inocula.21–23 This
phenomenon is not observed with ASPs in those same strains.
While deep-seated infections may be associated with higher
inocula, the clinical relevance of this in vitro finding is not well eluci-
dated.16,19–21,24,25 Alternatively, clinicians may prefer cefazolin in
certain scenarios given its improved tolerability, more convenient
dosing (e.g. for patients receiving intermittent haemodialysis) and
lower cost.26,27
Both ASPs and cefazolin have demonstrated effectiveness for
patients with MSSA bacteraemia, and several individual studies
have not identified a consensus option resulting in improved out-
comes. Given the morbidity and mortality associated with serious

C The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.
V
For permissions, please email: journals.permissions@oup.com.
2643
Bidell et al.
MSSA infections and the potential for adverse drug effects in the
setting of prolonged treatment durations, it is prudent to identify
an optimal treatment, should one exist.7 The purpose of this study
was to compare both clinical effectiveness and safety outcomes
associated with cefazolin compared with ASPs for the treatment of
MSSA bacteraemias.
Methods
Search strategy
A systematic review of the literature was conducted within PubMed/
MEDLINE, Embase and the Cochrane Central Registry of Controlled trials for
all publications from inception up to the end of November 2017. Search
terms included ‘cefazolin’, ‘nafcillin’, ‘oxacillin’, ‘cloxacillin’, ‘methicillin-sus-
ceptible Staphylococcus aureus’, ‘bacteraemia’ and ‘bloodstream infection’.
Results were limited to full-text articles available in English. References of
retrieved articles were also reviewed and assessed for possible inclusion.
Inclusion criteria
Studies were included per the following criteria: (i) hospitalized patients
with MSSA bacteraemias received definitive treatment with either cefazolin
or an ASP; (ii) clinical trial, cohort, or case–control study design; (iii) reported
the primary endpoint for each treatment group (i.e. 90 day mortality);28
and (iv) adult population. Articles were excluded if outcomes were not
reported, outcome data were not readily extractable for each treatment
group, or if the article was not available in English. Each article was eval-
uated independently for inclusion by two reviewers (J. N. O. and N. P.), with
discrepancies resolved by a third (M. R. B.).
Data extraction
Data were extracted from included studies independently by two authors
(J. N. O. and N. P.) and were standardized using a data extraction table.
Variables of interest included number of patients included in each group,
duration of treatment, severity of illness measures, study design and source
of infection.
Outcomes
The primary outcome of this meta-analysis was 90 day mortality.
Secondary outcomes included 30 day mortality, clinical failure, discontinua-
tions due to adverse effects and infection recurrence, as defined per each
individual study.
Risk of bias
Risk of bias was evaluated independently by two reviewers (M. R. B. and N. P.)
using the Newcastle–Ottawa Assessment Scale.29 This scale was developed
for evaluating non-randomized trials and assesses potential for bias within
the categories of selection, comparability and outcome assessments. A third
reviewer (J. N. O.) adjudicated unresolved discrepancies in risk of bias
evaluations.
Statistical analysis
Meta-analyses were performed using Review Manager version 5.3
(Copenhagen, Denmark). Odds ratios and 95% CIs were calculated for di-
chotomous data. Statistical heterogeneity among studies was assessed
using the v2 statistic, with the I2 statistic assessing the extent.30 A Mantel–
Haenszel random effects model was used for all analyses, as included
studies were non-randomized and are known to be inherently
heterogeneous.31
2644
Results
Literature searches identified 704 studies (469 PubMed, 203
Embase, 32 Cochrane database) for evaluation. Exclusion of dupli-
cate articles (n " 81) and non-English publications (n " 24) left
599 articles for review. Of these, seven met the criteria for inclusion
in the meta-analysis. No additional studies were identified upon
review of reference lists from the included studies (Figure 1).
Characteristics of the seven studies are included in
Table 1.8,10–12,20,32,33 All but one20 were retrospective observa-
tional studies, and three used propensity score matching.10,12,20
In total, 1589 and 2802 patients were treated with cefazolin or
Downloaded from https://academic.oup.com/jac/article/73/10/2643/5066380 by guest on 11 March 2024
an ASP (i.e. cloxacillin, oxacillin or nafcillin), respectively, for MSSA
bacteraemia. A summary of clinically relevant patient population
characteristics in the included studies is shown in Table 2. The
percentage of patients admitted to an intensive care unit at the
time of culture ranged from 10.8% (10 of 93) to 17.6% (556 of
3167) for the four studies that provided treatment-specific infor-
mation.8,12,32,33 Reporting of illness severity was heterogeneous
across studies given the variety of available classifiers.
Source of infection and presence of metastatic complications
data were extractable for all but two studies, accounting for a total
of 922 patients (Table 2).8,10,12,20,32 Bone, joint or musculoskeletal
sources were the most common foci of infection (21.6%, 199);
other reported sources in select studies were skin/soft tissue
(17.9%, 165), catheter (13.9%, 128), respiratory (11.3%, 104), in-
fective endocarditis (8.5%, 78) and other endovascular (6.0%, 20
of 362).20,32 Source of infection was unknown in 26.0% (240) of
cases. Metastatic complications were reported in 24.6% of patients
(227) who received ASPs or cefazolin.
Five studies assessed treatment failure.8,10,20,32,33 Independent
predictors of treatment failure in individual studies included dur-
ation of bacteraemia, metastatic complications including presence
of endocarditis and severe illness,10,20,32 whereas 90 day mortality
predictors included age, liver cirrhosis and hospital-acquired
bacteraemia.8,11
Mortality
Three of the seven studies used national registry databases to ver-
ify death.11,12,33 Mortality occurred within 90 days in 18.2% (289 of
1589) of cefazolin- and 25.1% (703 of 2802) of ASP-treated
patients. A statistically significant decrease in mortality was identi-
fied between patients receiving cefazolin compared with ASPs (OR
0.63, 95% CI 0.41–0.99, P " 0.05; Figure 2). Heterogeneity across
studies was moderate (I2 " 58%). A sensitivity analysis was con-
ducted to evaluate the influence of the study by McDanel et al.,33
as it represented the majority of our total population. Removal of
this study only affected the precision of the point estimate and
associated 95% CI (OR 0.49, 95% CI 0.23–1.04, P " 0.05, I2 " 65%).
Four studies also reported 30 day mortality.10,20,32,33 Among
these studies, cefazolin was associated with a significantly
decreased odds of mortality relative to ASPs, with no heterogen-
eity across studies (OR 0.58, 95% CI 0.46–0.77, P , 0.001, I2 " 0%,)
(Figure S1 available as Supplementary data at JAC Online). The
point estimate was heavily influenced by the inclusion of results
published by McDanel et al.33 (weight 96.3%). However, sensitivity
analyses confirmed similar results without the inclusion of that
study (OR 0.30, 95% CI 0.09–0.98, P " 0.05, I2 " 0%).
Optimal MSSA bacteraemia treatment JAC
469 Studies identified 203 Studies identified 32 Studies identified
via PubMed via Embase via Cochrane Database

Duplicates (n = 81)
Non-English (n = 24)

Downloaded from https://academic.oup.com/jac/article/73/10/2643/5066380 by guest on 11 March 2024

592 Articles excluded*:
599 Articles reviewed Outcomes not reported
for inclusion (n = 402)
Preclinical study (n = 271)
Invalid comparator (n = 83)
Not bacteremia (n = 83)
Not cohort or clinical trial
(n = 67)
Not MSSA (n = 30)
Pediatrics (n = 28)
Case report (n = 25)
Outcomes not extractable
7 Articles included in (n = 11)
meta-analysis

Figure 1. Evaluation of identified studies for inclusion. *Articles could be excluded for more than one reason.

Clinical failure associated with significantly fewer discontinuations due to ad-

Five of the seven studies evaluated clinical treatment
failure.8,10,20,32,33 Definitions varied by study, but included outcomes
such as persistent bacteraemia, change of therapy due to lack of ef-
fectiveness, evidence of metastatic complications, and relapse or re-
currence (Table 1). Overall, 4.8% (70 of 1447) and 5.5% (136 of
2491) of patients treated with cefazolin or an ASP, respectively,
experienced clinical failure. No significant decrease in the odds of
clinical failure was observed in patients treated with cefazolin com-
pared those treated with an ASP (OR 0.85, 95% CI 0.41–1.76,
P " 0.66; Figure 3). Moderate heterogeneity was noted (I2 " 74%).
verse effects with no heterogeneity (OR 0.25, 95% CI 0.11–0.56,
P , 0.01, I2 " 13%; Figure 4).
Study risk of bias
Interrater agreement was 100% (M. R. B. and N. P.). The median
Newcastle–Ottawa score for included studies was 8 (range 8–9)
(Table S1). The most common risk for bias was lack of adequate
follow-up, as only three studies used national registries to verify
mortality.11,12,33

Infectious recurrence Discussion

Five studies assessed reinfection, recurrence or relapse with cefa-
zolin or ASPs within 90 days.8,10,20,32,33 All included studies consid-
ered these outcomes to be a new MSSA infection following
treatment of the index event, despite differences in terminology
(Table 1). For simplicity we selected ‘recurrence’ as the singular
term for these outcomes. No differences in rates of recurrence
were observed between treatment groups with no heterogeneity
(OR 1.28, 95% CI 0.68–2.40, P " 0.45, I2 " 16%; Figure S2).
Discontinuation due to adverse effects
Few studies evaluated tolerability of either treatment.10,20,32 In
the three studies that reported tolerability data, cefazolin was
Clinical controversy exists surrounding the optimal agent for treat-
ment of serious MSSA bacteraemias.15–17,19 In this meta-analysis
of patients with MSSA bacteraemia, a statistically significant differ-
ence in 90 day mortality was identified between patients treated
with cefazolin versus an ASP (95% CI 0.41–0.99). However, it is im-
portant to note that precision may be an issue as the upper bound
of the CI approached the null value. Therefore, interpretation of
this result should account for this. There was no difference be-
tween treatments in risk of recurrence within 90 days (P " 0.45).
Significantly fewer patients discontinued cefazolin due to adverse
effects (P , 0.01), though these data were infrequently
reported.10,17,20

2645
 Table 1. Characteristics and reported outcomes of included studies

Study design, Duration of treat- Definition of Discontinuations due

Study location, period Agents ment (days) clinical failure Failure Mortality to adverse events

2646
Bai et al. 20158 retrospective cohort; cefazolin (n " 105); median cefazolin 17 relapse (new MSSA cefazolin 6% (6/105); cloxa- 90 days: cefazolin 20% NA
Toronto; 2007–10 cloxacillin (IQR 13–31), infection within cillin 2% (4/249) (21/105); cloxacillin
Bidell et al.

(n " 249) cloxacillin 19 90 days) 30% (75/249)

(IQR 13–34)
Lee et al. 201110 retrospective, case– cefazolin (n " 49); median cefazolin 17 switching of antibiot- switch: (4 weeks): cefazolin 4 weeks: cefazolin 0% cefazolin 0% (0/41);
control; South nafcillin (n " 84) (IQR 10–28), ics or relapse (new 10% (4/41), nafcillin 0% (0/41); nafcillin 10% nafcillin 17%
Korea; 2004–09 nafcillin 15 MSSA infection relapse: (4 weeks): cefazolin (4/41) (7/41)
(IQR 10–25) within 12 weeks) 0% (0/41), nafcillin 0% 12 weeks: cefazolin 2%
(0/41) (1/41); nafcillin 12%
(5/41)
switch: (12 weeks): cefazolin
10% (4/41), nafcillin 0%
(0/41)
relapse: (12 weeks): cefazo-
lin 2% (1/41), nafcillin 2%
(1/41)
Lee et al. 201820 prospective, observa- cefazolin (n " 79); NA switching of antibiot- switch: cefazolin 16.5% 30 days: cefazolin 2.5% cefazolin 12.7%
tional cohort study; nafcillin (n " 163) ics or persistent (13/79); nafcillin 16.6% (2/79); nafcillin 8% (10/79); nafcillin
Korea; 9/2013–3/ BSI or recurrence (27/163) (13/163) 29.4% (48/163)
2015 (new MSSA infec- recurrence: cefazolin 2.5% (2/ 90 days: cefazolin 2.5%
tion within 79); nafcillin 3.7% (6/163) (2/79); nafcillin 14.7%
3 months) persistent BSI: cefazolin (24/163)
6.3% (5/79); nafcillin
13.9% (11/79)
Li et al. 201432 retrospective cohort cefazolin (n " 59); median: oxacillin 31 persistent BSI or pro- persistent BSI: cefazolin 90 days: oxacillin 6% oxacillin 21% (7/34);
study; 2008–12; oxacillin (n " 34) (IQR 21–42); gression of infec- 10.2% (6/59), oxacillin (1/16); cefazolin 0% cefazolin 3% (2/
Texas cefazolin 39 tion on therapy or 20.6% (7/34) (0/14) 59)
(IQR 28–44) relapse of infec- relapse: cefazolin 6.8%
tion within 90 days (4/59), oxacillin 8.8%
(3/34)
progression: cefazolin 6.8%
(4/59); oxacillin 14.7%
(5/34)
McDanel et al. retrospective cohort cefazolin (n " 1163); NA recurrence (new cefazolin 1.7% (20/1163); 30 days: cefazolin 10% NA
201733 study; 2003–10; nafcillin/oxacillin MSSA infection nafcillin/oxacillin 1% (113/1163); nafcillin/
United States (n " 2004) 45–90 days) (28/2004) oxacillin 15% (307/
2004)
90 days: cefazolin 20%
(231/1163); nafcillin/
oxacillin 25% (502/
2004)
Paul et al. 201111 retrospective cohort cefazolin (n " 72); 14–28 persistent BSI or NA 90 days: cefazolin 40.3% NA
study; Israel; 1988– cloxacillin relapse (29/72); cloxacillin
94, 1999–2007 (n " 281) 32.4% (91/281)
Pollett et al. retrospective, cohort cefazolin (n " 70); median: cefazolin 20 NA NA 90 days: cefazolin 7.1% NA
201612 study; 2008–13; nafcillin (n " 30) (IQR 11–20); (5/70); nafcillin 16.7%
California nafcillin 12 (5/30)
(IQR 8–29)

BSI, bloodstream infection; NA, not applicable. Downloaded from https://academic.oup.com/jac/article/73/10/2643/5066380 by guest on 11 March 2024
 Table 2. Patient characteristics of included studies

First author, year, Metastatic

citation Age (years) Male gender ICU at time of culture Severity of illness Source of infection complication Source control
8
Bai 2015 (cefazolin .65: cefazolin 57% cefazolin 55% admitting service: hypotensive shock: catheter: cefazolin 12% (13); IE: cefazolin 2% (2); catheter removed:
n " 105, cloxacillin (60); cloxacillin (58); cloxacillin cefazolin 10% cefazolin 30% (32); cloxacillin 14% (34); cloxacillin 3% (8) cefazolin 54%
n " 249) 53% (132) 67% (166) (10); cloxacillin cloxacillin 24% (60) skin: cefazolin 25% (26); bone/joint: cefazo- (7/13); cloxacillin
18% (45) cloxacillin 17% (42) lin 8% (8); cloxa- 65% (22/34)
within 72 h of culture: respiratory: cefazolin 16% cillin 8% (20) bone/joint: cefazolin
cefazolin 13% (17); cloxacillin 18% (44) other: cefazolin 73% (11/15); clox-
(14); cloxacillin IE: cefazolin 2% (2); cloxacillin 18% (19); cloxa- acillin 46% (13/28)
22% (55) 12% (30) cillin 21% (53) abscess: cefazolin
bone/joint: cefazolin 14% (15); 50% (1/2); cloxa-
cloxacillin 11% (28) cillin 63% (12/19)
other: cefazolin 16% (17);
cloxacillin 33% (81)
Optimal MSSA bacteraemia treatment

unknown: cefazolin 33% (35);

cloxacillin 34% (85)

Lee 201110 (cefazolin mean: cefazolin cefazolin 59% NA McCabe classification, catheter: cefazolin 22% (11); cefazolin 16% (8); eradicated foci:
n " 49, nafcillin 55 + 20; nafcillin (29); nafcillin non-fatal: cefazolin nafcillin 20% (17) nafcillin 27% cefazolin 29%
n " 84) 52 + 17 58% (49) 39% (19); nafcillin osteomyelitis: cefazolin 20% (23) (14); nafcillin 26%
29% (24) (10); nafcillin 13% (11) (22)
McCabe classification, IE: cefazolin 2% (1); nafcillin
ultimately fatal: 16% (13);
cefazolin 43% (21); respiratory: cefazolin 8% (4);
nafcillin 48% (40) nafcillin 13% (11);
McCabe classification, soft tissue: cefazolin 20% (10);
rapidly fatal: cefazolin nafcillin 12% (10)
18% (9); nafcillin 24% other: cefazolin 10% (5);
(20) nafcillin 21% (18)
Lee 201820 (cefazolin 65: cefazolin 41.8% cefazolin 60.8% NA SOFA 2: cefazolin catheter: cefazolin 12.7% cefazolin 12.7% NA
n " 79, nafcillin (33); nafcillin (48); nafcillin 45.6% (36); nafcillin (10); nafcillin 12.3% (20) (10); nafcillin
n " 163) 55.2% (90) 65% (106) 68.1% (111) IE: cefazolin 1.3% (1); nafcillin 14.1% (23)
septic shock: cefazolin 6.7% (11)
1.3% (1); nafcillin endovascular: cefazolin 3.8%
7.4% (12) (3); nafcillin 8.6% (14)
bone/joint: cefazolin 35.4%
(28); nafcillin 37.4% (61)
skin: cefazolin 34.2% (27);
nafcillin 23.3% (38)
respiratory: cefazolin 3.8% (3);
nafcillin 9.8% (16)
other: cefazolin 3.8% (3);
nafcillin 7.4% (12)
unknown: cefazolin 15.2 (12);
nafcillin 13.5% (22)

Continued
JAC

2647
Downloaded from https://academic.oup.com/jac/article/73/10/2643/5066380 by guest on 11 March 2024
 Table 2. Continued

2648
First author, year, Metastatic
citation Age (years) Male gender ICU at time of culture Severity of illness Source of infection complication Source control
Bidell et al.

Li 201432 (cefazolin mean: cefazolin cefazolin 75% cefazolin 7% (4); Pitt bacteraemia (me- catheter: cefazolin 10% (6), cefazolin 34% (20); catheter: cefazolin
n " 59, oxacillin 51 + 10; oxacillin (44); oxacillin oxacillin 18% (6) dian): cefazolin 0 (IQR oxacillin 3% (1) oxacillin 35% 91% (10/11);
n " 34) 51 + 14 82% (28) 0–1); oxacillin 0 (IQR IE: cefazolin 18.3% (17), (12) oxacillin 100%
0–1) oxacillin 8.8% (3) (1/1)
endovascular: cefazolin 2.2% foreign material:
(2); oxacillin 2.9% (1) cefazolin 31%
osteoarticular: cefazolin 31% (5/16); oxacillin
(18); oxacillin 59% (20) 33% (1/3)
respiratory: cefazolin 4% (2);
oxacillin 6% (2)
soft tissue: cefazolin 14% (8);
oxacillin 3% (1)
other: cefazolin 3% (2);
oxacillin 12% (4)
unknown: cefazolin 14% (8);
oxacillin 6% (2)
McDanel 201733 75: cefazolin 25% cefazolin 97% cefazolin 15% (178); modified APACHE III NA NA NA
(cefazolin n " 1163, (288); nafcillin/ (1133); nafcil- nafcillin/oxacillin 34: cefazolin 56%
nafcillin/oxacillin oxacillin 25% lin/oxacillin 19% (378) (651); nafcillin/oxacil-
n " 2004) (505) 99% (1974) lin 52% (1040)
Paul 201111 (cefazolin mean: 69 + 16.8 55.1% (298/541) 5.2% (28/541) shock at onset: 12.2% catheter: 11.6% (63/541) NA NA
n " 72, cloxacillin (66) IE: 6.5% (25/541)
n " 281) bone/joint: 5% (27/541)
skin/soft tissue: 15.3% (83/
541)
respiratory: 10% (54/541)
other: 21.3% (115/541)
unknown: 22.6% (122/541)
Pollett 201612 mean: cefazolin cefazolin 31% cefazolin 13% (9); met SIRS criteria: catheter: cefazolin 19% (13), IE: cefazolin 14% NA
(cefazolin n " 70, 53 + 15.1), (22); nafcillin nafcillin 27% (8) cefazolin 69% (48), nafcillin 10% (3) (10); nafcillin
nafcillin n " 30) nafcillin 23% (7) nafcillin 67% (20) MSK/bone: cefazolin 7.1% (5); 17% (5)
50.4 + 12.5 nafcillin 10% (3) vertebral osteo-
respiratory: cefazolin 1.4% (1); myelitis: cefazo-
nafcillin 13.3% (4) lin 6% (4);
cellulitis: cefazolin 4% (3); nafcillin 7% (2)
nafcillin 0 (0)
other: cefazolin 21.4% (15);
nafcillin 10% (3)
unknown: cefazolin 40% (28);
nafcillin 53% (16)

IE, infective endocarditis; MSK, musculoskeletal; SIRS, systemic inflammatory response syndrome.

Downloaded from https://academic.oup.com/jac/article/73/10/2643/5066380 by guest on 11 March 2024

Optimal MSSA bacteraemia treatment JAC
Cefazolin ASPs Odds Ratio Odds Ratio
Study or Subgroup Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI
Bai et al. 2015 21 105 75 249 22.7% 0.58 [0.33, 1.00]
Lee et al. 2011 1 41 5 41 3.7% 0.18 [0.02, 1.61]
Lee et al. 2018 2 79 24 163 7.3% 0.15 [0.03, 0.65]
Li et al. 2014 0 59 1 34 1.8% 0.19 [0.01, 4.74]
McDanel et al. 2017 231 1163 502 2004 32.8% 0.74 [0.62, 0.88]
Paul et al. 2011 29 72 91 281 23.1% 1.41 [0.83, 2.40]
Pollett et al. 2016 5 70 5 30 8.6% 0.38 [0.10, 1.44]

Total (95% CI) 1589 2802 100.0% 0.63 [0.41, 0.99]

Total events 289 703

Downloaded from https://academic.oup.com/jac/article/73/10/2643/5066380 by guest on 11 March 2024

2 2 2
Heterogeneity: Tau = 0.15; Chi = 14.15, df = 6 (P = 0.03; I = 58%)
Test for overall effect Z = 2.00 (P = 0.05) 0.01 0.1 1 10 100
Favours Cefazolin Favours ASPs

Figure 2. Forest plot of ORs for 90 day mortality.

Cefazolin ASPs Odds Ratio Odds Ratio

Study or Subgroup Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI
Bai et al. 2015 6 105 4 249 15.2% 3.71 [1.03, 13.44]
Lee et al. 2011 6 41 6 41 15.9% 1.00 [0.29, 3.40]
Lee et al. 2018 24 79 82 163 24.5% 0.43 [0.24, 0.76]
Li et al. 2014 14 59 16 34 20.0% 0.35 [0.14, 0.86]
McDanel et al. 2017 20 1163 28 2004 24.4% 1.23 [0.69, 2.20]

Total (95% CI) 1447 2491 100.0% 0.85 [0.41, 1.76]

Total events 70 136
2 2 2
Heterogeneity: Tau = 0.48; Chi = 15.38, df = 4 (P = 0.004); I = 74%
0.01 0.1 1 10 100
Test for overall effect: Z = 0.44 (P = 0.66)
Favours Cefazolin Favours ASPs

Figure 3. Forest plot of ORs for clinical failure.

Cefazolin ASPs Odds Ratio Odds Ratio

Study or Subgroup Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI
Lee et al. 2011 0 41 7 41 7.5% 0.06 [0.00, 1.01]
Lee et al. 2018 10 79 48 163 71.0% 0.35 [0.17, 0.73]
Li et al. 2014 2 59 7 34 21.5% 0.14 [0.03, 0.70]

Total (95% CI) 179 238 100.0% 0.25 [0.11, 0.56]

Total events 12 62
2 2 2
Heterogeneity: Tau = 0.10; Chi = 2.30, df = 2 (P = 0.32); I = 13%
Test for overall effect: Z = 3.38 (P = 0.0007) 0.01 0.1 1 10 100
Favours Cefazolin Favours ASPs

Figure 4. Forest plot of ORs for discontinuation due to adverse effects.

It is important to consider our findings in the context of poten-
tial predictors of the outcomes of interest. With respect to foci of
infection, there was a higher proportion of patients who received
cefazolin who had a catheter source [16.0% (58 of 362) versus
9.6% (54 of 560)] and fewer had a respiratory source [7.5% (27 of
362) versus 13.8% (77 of 560)] compared with those who received
an ASP. One study8 found neither catheter nor respiratory source
to be significantly associated with 90 day mortality upon multivari-
ate analysis; however, Lee et al.10 found pneumonia to be signifi-
cantly associated with treatment failure. Predictors of 90 day
mortality in the included studies included age, liver cirrhosis and
hospital-acquired bacteraemia.8,11 With respect to treatment
failure, deep-seated infection including infective endocarditis and
metastatic infection were significant predictors in multiple stud-
ies.10,20 Notably, these characteristics appeared similar between
treatment groups across the meta-analysis population.
Source control is critical to effective treatment, particularly in
severe infections such as MSSA bacteraemia. Treatment group-
specific source control data were only reported for three of the
seven studies in this meta-analysis.8,10,32 Pooling of these three
studies revealed comparable rates of source control for cefazolin
[41.4% (48 of 116)] compared with ASPs [42.0% (71 of 169)]; how-
ever, data were unavailable for the majority of the meta-analysis
population, limiting interpretation.

2649
Bidell et al.
A limitation of many studies conducted previously is the small
sample size, which may prevent detection of outcome differences
between groups. The largest study in this meta-analysis was by
McDanel et al.,33 comprising 72% (3436 of 4749) of the pooled
population. This was a retrospective study comprised of patients
from 119 Veterans Affairs hospitals between 2003 and 2010. After
adjusting for confounding variables (i.e. diabetes, APACHE III 34,
admission to ICU, and diagnosis of infective endocarditis), cefazo-
lin was associated with a 23% decrease in 90 day mortality
(adjusted HR 0.77, 95% CI 0.66–0.90) and 37% decrease in 30 day
mortality compared with an ASP (adjusted HR 0.63, 95% CI 0.51–
0.78). This is consistent with the findings of our meta-analysis, and
sensitivity analyses showed that removal of this study resulted in
minimal changes in point estimates for 30 and 90 day mortality.
Unfortunately, data on source control and antibiotic dosing were
not reported for the study population.
Our findings demonstrate a significant difference in 30 and
90 day mortality between cefazolin and ASPs for serious MSSA infec-
tions. These data contrast concerns regarding an inoculum effect
described with cefazolin in preclinical studies, leading some clinicians
to prefer ASPs for the treatment of MSSA bacteraemias. This effect
has been well described in vitro in isolates expressing a Type A blaZ
gene.21–23 However, several considerations hinder the ability to draw
definitive conclusions on the clinical relevance of this phenomenon.
Firstly, blaZ may be affected by the bacterial growth phase, and asso-
ciated b-lactamases exhibit variable affinities for cefazolin.22
Secondly, in vitro studies do not capture the impact of the human im-
mune response, nor the impact of drug re-dosing that is employed in
clinical practice.19 Thirdly, the prevalence of blaZ and the inoculum
effect appears to vary widely.24,34–36 A multicentre study from the
Chicago area found ,2% of MSSA isolates exhibited a cefazolin in-
oculum effect leading to resistant MICs,23 while a single centre study
from Atlanta showed a higher rate of cefazolin inoculum effect at
17%.21 Several studies attempting to characterize the clinical rele-
vance of this finding report inconsistent results.20,21,24,25 It may be of
value to evaluate this at an institutional level. If concerns exist
regarding use of cefazolin in the setting of high inoculum infection
with concurrent concerns for ASP toxicity, clinicians may consider se-
quential therapy, starting with ASPs followed by cefazolin after the
initial decrease in the inoculum.15 Of note, this stepwise approach
was not evaluated in the studies included in this meta-analysis.
In the absence of a clear clinical benefit with one agent, it is im-
portant to consider drug tolerability. Our findings of improved tol-
erability with cefazolin over ASPs is consistent with similar
literature not included in these analyses. Recently, Flynt et al.26
compared rates of acute kidney injury in patients receiving cefazo-
lin or nafcillin for MSSA bacteraemias. They identified a significant
increase in risk with patients receiving nafcillin, after controlling for
ICU admission and endocarditis (OR 2.74, 95% CI 1.1–6.6).
Youngster et al.27 also identified significant tolerability issues with
nafcillin compared with cefazolin in outpatients, with increased
rates of rash (13.9 versus 4.2%, P , 0.01), renal dysfunction
(11.4 versus 3.3%, P , 0.01), liver function abnormalities (8.1 ver-
sus 1.6%, P " 0.01) and early discontinuation of antibiotics (33.8
versus 6.7%, P , 0.01).
There are a number of limitations that should be considered
when interpreting the findings of this analysis. Conclusions were
derived from retrospective studies, which are inherently subject to
bias. While the majority of studies were found to have limited risk
2650
of bias using the Newcastle–Ottawa scoring system, use of ASPs in
patients deemed to be clinically ‘sicker’ or having a deep-seated in-
fection must be considered. Though we believe our findings to be
of clinical value, there is risk that these may be over-generalized,
given that important predictors of outcomes were not consistently
evaluated or reported across studies. Important missing data in-
clude MIC information, time to clearance of blood cultures, time to
appropriate therapy, antibiotic dosing strategies and prevalence of
blaZ/inoculum effect. Furthermore, severity of illness scoring
(e.g. APACHE II, SOFA, Pitt Bacteremia Score) and adjusted odds of
mortality were not consistently reported across all studies, limiting
adjustment for these effects on mortality. Source control data
Downloaded from https://academic.oup.com/jac/article/73/10/2643/5066380 by guest on 11 March 2024
were also reported inconsistently. While patients with complicated
bacteraemias and deep-seated infections were included in all
studies, the representation within each study population varied
substantially. The utilization of Mantel–Haenszel random effects
models result in a more conservative estimate of effects and a bias
towards the null, but is appropriate when included studies are un-
controlled since inherent bias exists within these study designs.31
This approach also may have low power to detect differences
when heterogeneity is high and number of included studies is low.
Only studies published in English were included, which may affect
the precision of the point estimates. Furthermore, the definition of
treatment failure varied by study (Table 1). Clinicians would benefit
from the creation of standardized definitions in future treatment
guidelines.
In conclusion, this meta-analysis identified a significant de-
crease in 90 day mortality between cefazolin and ASPs for the
treatment of MSSA bacteraemia. Additionally, cefazolin appeared
to be better tolerated than ASPs in these patients. It is important to
note that adjustment for confounders was limited by the extract-
able data reported in each study. The prevalence of MSSA exhibit-
ing a cefazolin inoculum effect is likely variable and should be
considered at the individual level. While differences in specific out-
comes (i.e. 90 day mortality and discontinuation due to adverse
events) were observed between cefazolin and ASPs in uncontrolled
studies, it is important for clinicians to consider patient-specific fac-
tors when selecting antimicrobial treatment for patients with inva-
sive MSSA infections.
Funding
This study was carried out as part of our routine work.
Transparency declarations
None to declare.
Supplementary data
Table S1 and Figures S1 and S2 appear as Supplementary data at JAC
Online.
References
1 van Hal SJ, Jensen SO, Vaska VL et al. Predictors of mortality in
Staphylococcus aureus bacteremia. Clin Microbiol Rev 2012; 25: 362–86.
Optimal MSSA bacteraemia treatment
2 Keynan Y, Rubinstein E. Staphylococcus aureus bacteremia, risk factors,
complications, and management. Crit Care Clin 2013; 29: 547–62.
3 Weiner LM, Webb AK, Limbago B et al. Antimicrobial-resistant pathogens
associated with healthcare-associated infections: summary of data reported
to the National Healthcare Safety Network at the Centers for Disease Control
and Prevention, 2011-2014. Infect Control Hosp Epidemiol 2016; 37:
1288–301.
4 Baddour LM, Wilson WR, Bayer AS et al. Infective endocarditis in adults:
diagnosis, antimicrobial therapy, and management of complications: a scien-
tific statement for healthcare professionals from the American Heart
Association. Circulation 2015; 132: 1435–86.
5 Berbari EF, Kanj SS, Kowalski TJ et al. 2015 Infectious Diseases Society of
America (IDSA) clinical practice guidelines for the diagnosis and treatment of
native vertebral osteomyelitis in adults. Clin Infect Dis 2015; 61: e26–46.
6 Mermel LA, Allon M, Bouza E et al. Clinical practice guidelines for the diag-
nosis and management of intravascular catheter-related infection: 2009
Update by the Infectious Diseases Society of America. Clin Infect Dis 2009; 49:
1–45.
7 del Rio A, Cervera C, Moreno A et al. Patients at risk of complications of
Staphylococcus aureus bloodstream infection. Clin Infect Dis 2009; 48 Suppl
4: S246–53.
8 Bai AD, Showler A, Burry L et al. Comparative effectiveness of cefazolin ver-
sus cloxacillin as definitive antibiotic therapy for MSSA bacteraemia: results
from a large multicentre cohort study. J Antimicrob Chemother 2015; 70:
1539–46.
9 Chen PY, Chuang YC, Wang JT et al. Impact of prior healthcare-associated
exposure on clinical and molecular characterization of methicillin-susceptible
Staphylococcus aureus bacteremia: results from a retrospective cohort study.
Medicine (Baltimore) 2015; 94: e474.
10 Lee S, Choe PG, Song KH et al. Is cefazolin inferior to nafcillin for treatment
of methicillin-susceptible Staphylococcus aureus bacteremia? Antimicrob
Agents Chemother 2011; 55: 5122–6.
11 Paul M, Zemer-Wassercug N, Talker O et al. Are all b-lactams similarly ef-
fective in the treatment of methicillin-sensitive Staphylococcus aureus bac-
teraemia? Clin Microbiol Infect 2011; 17: 1581–6.
12 Pollett S, Baxi SM, Rutherford GW et al. Cefazolin versus nafcillin for
methicillin-sensitive Staphylococcus aureus bloodstream infection in a
California tertiary medical center. Antimicrob Agents Chemother 2016; 60:
4684–9.
13 Rao SN, Rhodes NJ, Lee BJ et al. Treatment outcomes with cefazolin ver-
sus oxacillin for deep-seated methicillin-susceptible Staphylococcus aureus
bloodstream infections. Antimicrob Agents Chemother 2015; 59: 5232–8.
14 Renaud CJ, Lin X, Subramanian S et al. High-dose cefazolin on consecu-
tive hemodialysis in anuric patients with staphylococcal bacteremia.
Hemodial Int 2011; 15: 63–8.
15 Karchmer AW. Definitive treatment for methicillin-susceptible
Staphylococcus aureus bacteremia: data versus a definitive answer? Clin
Infect Dis 2017; 65: 107–9.
16 Loubet P, Burdet C, Vindrios W et al. Cefazolin versus anti-staphylococcal
penicillins for treatment of methicillin-susceptible Staphylococcus aureus
bacteraemia: a narrative review. Clin Microbiol Infect 2018; 24: 125–32.
17 Li J, Echevarria KL, Traugott KA. b-Lactam therapy for methicillin-
susceptible Staphylococcus aureus bacteremia: a comparative review of cefa-
zolin versus antistaphylococcal penicillins. Pharmacotherapy 2017; 37: 346–60.
18 Sabath LD, Garner C, Wilcox C et al. Effect of inoculum and of
b-lactamase on the anti-staphylococcal activity of thirteen penicillins and
cephalosporins. Antimicrob Agents Chemother 1975; 8: 344–9.
19 Craig WA, Bhavnani SM, Ambrose PG. The inoculum effect: fact or arti-
fact? Diagn Microbiol Infect Dis 2004; 50: 229–30.
20 Lee S, Song KH, Jung SI et al. Comparative outcomes of cefazolin versus
nafcillin for methicillin-susceptible Staphylococcus aureus bacteraemia: a
JAC
prospective multicentre cohort study in Korea. Clin Microbiol Infect 2018; 24:
152–8.
21 Livorsi DJ, Crispell E, Satola SW et al. Prevalence of blaZ gene types and
the inoculum effect with cefazolin among bloodstream isolates of
methicillin-susceptible Staphylococcus aureus. Antimicrob Agents Chemother
2012; 56: 4474–7.
22 Lee SH, Park WB, Lee S et al. Association between type A blaZ gene
polymorphism and cefazolin inoculum effect in methicillin-susceptible
Staphylococcus aureus. Antimicrob Agents Chemother 2016; 60:
6928–32.
23 Wang SK, Sigar I, Gilchrist A et al. The prevalence of an inoculum effect
with cefazolin and the association with certain blaZ gene types among
Downloaded from https://academic.oup.com/jac/article/73/10/2643/5066380 by guest on 11 March 2024
methicillin-susceptible Staphylococcus aureus isolates from four major
Chicago medical centers. Open Forum Infect Dis 2016; 3 Suppl 1: Abstract
292.
24 Lee S, Kwon KT, Kim HI et al. Clinical implications of cefazolin inoculum ef-
fect and b-lactamase type on methicillin-susceptible Staphylococcus aureus
bacteremia. Microb Drug Resist 2014; 20: 568–74.
25 Nannini EC, Stryjewski ME, Singh KV et al. Inoculum effect with cefazolin
among clinical isolates of methicillin-susceptible Staphylococcus aureus: fre-
quency and possible cause of cefazolin treatment failure. Antimicrob Agents
Chemother 2009; 53: 3437–41.
26 Flynt LK, Kenney RM, Zervos MJ et al. The safety and economic impact of
cefazolin versus nafcillin for the treatment of methicillin-susceptible
Staphylococcus aureus bloodstream infections. Infect Dis Ther 2017; 6:
225–31.
27 Youngster I, Shenoy ES, Hooper DC et al. Comparative evaluation of the
tolerability of cefazolin and nafcillin for treatment of methicillin-susceptible
Staphylococcus aureus infections in the outpatient setting. Clin Infect Dis
2014; 59: 369–75.
28 Harris PNA, McNamara JF, Lye DC et al. Proposed primary endpoints for
use in clinical trials that compare treatment options for bloodstream infection
in adults: a consensus definition. Clin Microbiol Infect 2017; 23: 533–41.
29 Stang A. Critical evaluation of the Newcastle-Ottawa scale for the assess-
ment of the quality of nonrandomized studies in meta-analyses. Eur J
Epidemiol 2010; 25: 603–5.
30 Higgins JP, Thompson SG, Deeks JJ et al. Measuring inconsistency in
meta-analyses. BMJ 2003; 327: 557–60.
31 Falagas ME, Tansarli GS, Ikawa K et al. Clinical outcomes with extended
or continuous versus short-term intravenous infusion of carbapenems and
piperacillin/tazobactam: a systematic review and meta-analysis. Clin Infect
Dis 2013; 56: 272–82.
32 Li J, Echevarria KL, Hughes DW et al. Comparison of cefazolin versus oxa-
cillin for treatment of complicated bacteremia caused by methicillin-
susceptible Staphylococcus aureus. Antimicrob Agents Chemother 2014; 58:
5117–24.
33 McDanel JS, Roghmann MC, Perencevich EN et al. Comparative effective-
ness of cefazolin versus nafcillin or oxacillin for treatment of methicillin-
susceptible Staphylococcus aureus infections complicated by bacteremia: a
nationwide cohort study. Clin Infect Dis 2017; 65: 100–6.
34 Chong YP, Park SJ, Kim ES et al. Prevalence of blaZ gene types and the
cefazolin inoculum effect among methicillin-susceptible Staphylococcus aur-
eus blood isolates and their association with multilocus sequence types and
clinical outcome. Eur J Clin Microbiol Infect Dis 2015; 34: 349–55.
35 Song KH, Jung SI, Lee S et al. Characteristics of cefazolin inoculum effect-
positive methicillin-susceptible Staphylococcus aureus infection in a multi-
centre bacteraemia cohort. Eur J Clin Microbiol Infect Dis 2017; 36: 285–94.
36 Rincón S, Reyes J, Carvajal LP et al. Cefazolin high-inoculum effect in
methicillin-susceptible Staphylococcus aureus from South American hospi-
tals. J Antimicrob Chemother 2013; 68: 2773–8.
2651