Cefazolin high-inoculum effect in methicillin-susceptible Staphylococcus aureus from South American hospitals

Abstract

Objectives

Clinical failures with cefazolin have been described in high-inoculum infections caused by methicillin-susceptible Staphylococcus aureus (MSSA) producing type A β-lactamase. We investigated the prevalence of the cefazolin inoculum effect (InE) in MSSA from South American hospitals, since cefazolin is used routinely against MSSA due to concerns about the in vivo efficacy of isoxazolyl penicillins.

Methods

MSSA isolates were recovered from bloodstream (n = 296) and osteomyelitis (n = 68) infections in two different multicentre surveillance studies performed in 2001–02 and 2006–08 in South American hospitals. We determined standard-inoculum (10⁵cfu/mL) and high-inoculum (10⁷ cfu/mL) cefazolin MICs. PFGE was performed on all isolates that exhibited a cefazolin InE. Multilocus sequence typing (MLST) and sequencing of part of blaZ were performed on representative isolates.

Results

The overall prevalence of the cefazolin InE was 36% (131 isolates). A high proportion (50%) of MSSA isolates recovered from osteomyelitis infections exhibited the InE, whereas it was observed in 33% of MSSA recovered from bloodstream infections. Interestingly, Ecuador had the highest prevalence of the InE (45%). Strikingly, 63% of MSSA isolates recovered from osteomyelitis infections in Colombia exhibited the InE. MLST revealed that MSSA isolates exhibiting the InE belonged to diverse genetic backgrounds, including ST5, ST8, ST30 and ST45, which correlated with the prevalent methicillin-resistant S. aureus clones circulating in South America. Types A (66%) and C (31%) were the most prevalent β-lactamases.

Conclusions

Our results show a high prevalence of the cefazolin InE associated with type A β-lactamase in MSSA isolates from Colombia and Ecuador, suggesting that treatment of deep-seated infections with cefazolin in those countries may be compromised.

Introduction

Staphylococcus aureus is a leading cause of hospital- and community-acquired infections.¹ The agents of choice for severe infections caused by methicillin-susceptible S. aureus (MSSA) are the isoxazolyl penicillins (such as oxacillin, nafcillin and flucloxacillin). These compounds are widely used in the hospital setting for the treatment of severe MSSA infections. However, due to the need for frequent administration, clinicians tend to switch to cephalosporins such as cefazolin in order to continue therapy in an outpatient setting.^2,3 However, the use of cefazolin in the treatment of deep-seated MSSA infections in which high bacterial inocula are present may be compromised⁴ by the production of certain types of staphylococcal β-lactamases that are capable of degrading cefazolin. Four different types (A–D) of staphylococcal β-lactamase enzymes have been characterized based on their substrate specificity and amino acid sequence,^5,6 some of which are able to hydrolyse cefazolin at significant rates.⁵ Under inducing conditions (e.g. the presence of a β-lactam antibiotic), S. aureus isolates can produce large quantities of β-lactamase, a phenomenon that may compromise the effectiveness of cefazolin⁷ in infections where a high inoculum is present (e.g. endocarditis, osteomyelitis and undrained abscesses).^8,9 Under these circumstances, cefazolin degradation may occur (depending on the substrate specificity of the enzyme), leading to decreased concentrations of the antibiotic at the site of infection, with the potential to produce therapeutic failures.^8,10,11 Indeed, cefazolin degradation and inactivation has been associated with clinical failures in high-inoculum infections with S. aureus strains producing type A β-lactamase.^4,11,12

The inoculum effect (InE) has been defined as a significant rise in the cefazolin MIC when the bacterial inoculum size is increased to 10⁷cfu/mL (instead of the standard 10⁵cfu/mL).^8,9,13 Previous studies in the USA have reported that the prevalence of MSSA isolates exhibiting the cefazolin InE ranged from 19% to 27%.^11,14 Cefazolin (instead of oxacillin or its derivatives) is frequently used in South America as the ‘therapy of choice’ for the treatment of severe MSSA infections, since the isoxazolyl penicillins are not commercially available in certain countries, generic derivatives are thought to be clinically inferior to the innovator compound¹⁵ or because of a more favourable dosing schedule as the availability of efficient outpatient systems for intravenous administration is limited in some Latin American countries. However, the prevalence of the cefazolin InE among MSSA isolates recovered from invasive infections in South America is unknown. In this study, we sought to characterize the frequency of the cefazolin InE among MSSA isolates recovered from invasive infections (bacteraemia and osteomyelitis) in two previous prospective multicentre surveillance studies performed in South American hospitals. Additionally, we performed molecular characterization of representative MSSA isolates and typing of β-lactamase enzymes of MSSA isolates exhibiting such a phenotype. A high prevalence of InE was found in MSSA recovered from Ecuadorian and Colombian hospitals.

Materials and methods

Bacterial isolates and species identification

A total of 364 MSSA clinical isolates were included in this study. The organisms were recovered from two previous prospective multicentre clinical studies.^16,17 The first study was performed in 2001–02 and encompassed 15 tertiary care centres in five Colombian cities.¹⁶ The second study was carried out in 2006–08 and included 32 tertiary hospitals from four South American countries, including Colombia (22 hospitals), Ecuador (5 hospitals), Peru (3 hospitals) and Venezuela (2 hospitals).¹⁷ The isolates were collected from individual patients presenting with bloodstream infections (n = 296) or osteomyelitis (n = 68) (cultures directly from bone tissue after biopsy). Confirmation of species and presence of the mecA gene in all isolates was performed by a multiplex PCR assay.¹⁸

Susceptibility testing

Cefazolin (Sigma Chemicals, St Louis, MO, USA) MICs were determined using a broth microdilution method with cation-adjusted Mueller–Hinton broth II (Becton Dickinson) at standard inoculum (10⁵cfu/mL) following the CLSI recommendations.¹³ High-inoculum MICs using 10⁷cfu/mL were evaluated simultaneously in order to determine whether there was a cefazolin InE, as previously described.¹¹ MICs were determined at 24 h by two different researchers. The inocula were confirmed by performing colony counts. The high-inoculum effect was defined as a cefazolin MIC ≥16 mg/L at high inoculum and ≤8 mg/L (susceptible) at standard inoculum.¹¹ The S. aureus strains included as controls were as follows: TX0117, a high-level producer of type A β-lactamase,⁴ which exhibits the high-inoculum effect; S. aureus ATCC 29213, a producer of small amounts of type A β-lactamase; and S. aureus ATCC 25923, a β-lactamase-negative strain.^4,11

Molecular typing

Pulsed-field gel electrophoresis (PFGE) was performed on all isolates that exhibited a cefazolin InE, as previously described.^17,19 S. aureus NCTC 8325 was used as the control for molecular size fragments. Methicillin-resistant S. aureus (MRSA) circulating in South American hospitals and strains from pandemic clones were included as comparators for PFGE banding analysis and involved representatives of the community-associated MRSA USA300 Latin American variant clone (ComA), USA300, Chilean, Brazilian, Iberian, Pediatric (NRSA 387) and New York/Japan (NRSA 382) clones. Banding patterns were analysed using Gel ComparII software version 4.01 (Applied Maths, Belgium). The Dice coefficient was calculated using a tolerance of 0.5 and dendrograms were constructed by the unweighted pair group method. The percentage of similarity to define a pulsotype was ≥75% and letters were used to designate them. Multilocus sequence typing (MLST) was performed for selected isolates from the most prevalent pulsotypes and highest MICs, as previously described.²⁰

Detection of β-lactamase gene (blaZ) and sequencing

The β-lactamase gene blaZ was detected by PCR in all isolates exhibiting the InE, following the methodology previously described by Martineau et al.¹⁸ Sequencing of part of blaZ (a 355 bp fragment inside of the blaZ gene in order to identify putative amino acid differences at residues 128–216)²¹ was performed using representative isolates of the most prevalent PFGE types from blood and osteomyelitis, according to a methodology described previously.¹¹ Analysis and sequence alignments of blaZ sequences were performed using the BLAST tool from NCBI (http://blast.ncbi.nlm.nih.gov/Blast.cgi).

Statistical analysis

The MIC means, ranges, MIC₅₀ and MIC₉₀ at standard and high inocula were calculated. Comparison between InEs among countries was performed using the Z-test for difference between proportions. ANOVA and non-parametric statistics (Wilcoxon rank-sum test and Mann–Whitney U-test) were used to compare the cefazolin standard-inoculum versus high-inoculum MIC by clinical source and country. The Z-test for differences between proportions was used to compare frequencies of the cefazolin high-inoculum effect in Colombia from the two multicentre studies. The χ² test of independence or homogeneity or Fisher exact tests were used to compare the frequency of the InE between countries. The tests were considered statistically significant if the P value was <0.05. Statistical analysis was performed using Stata software (version 10, Stata Corporation, College Station, TX, USA).

Results

Prevalence of cefazolin InE in MSSA clinical isolates from South America

A total of 364 MSSA recovered from bloodstream infections and osteomyelitis in two multicentre surveillance studies in South American hospitals were evaluated for the cefazolin InE. All isolates were susceptible to cefazolin at standard inoculum (MICs < 8 mg/L, with the highest MIC value of 2 mg/L). A marked and statistically significant increase in the MICs was observed in 131 (36%) isolates when a high inoculum was used (97 and 34 recovered from blood and bone, respectively) (Table 1). Of note, 50% (34 out of 68) of MSSA isolates recovered from osteomyelitis exhibited the InE, whereas it was documented in 33% of bloodstream isolates (97 out of 296). Among these 131 isolates, the MIC range at the standard inoculum was 0.25 to ≤2 mg/L and increased to 16 to ≤512 mg/L at the high inoculum (Table 1). Of note, 28 MSSA isolates showed very high cefazolin MICs at the high inoculum (≥128 mg/L) (18 from blood and 10 from osteomyelitis).

Table 1.

Cefazolin MIC₅₀, MIC₉₀, range and IQR (mg/L) at standard inoculum and high inoculum among 131 MSSA isolates demonstrating the InE by clinical source

Clinical source	Parameter	Cefazolin MICs (mg/L)		P valuea
		SI	HI
Bloodstream (n = 97)	mean	0.65	87.75	<0.001
	95% CI	0.59–0.70	64.1–111.3
	range	0.25–2	16–512
	IQR	0.5–1.0	16–64
	MIC50	0.5	64
	MIC90	1.0	64
Osteomyelitis (n = 34)	mean	0.80	120.94	<0.001
	95% CI	0.65–0.95	70.0–121.8
	range	0.5–2	16–512
	IQR	0.5–1	32–256
	MIC50	0.5	64
	MIC90	1.0	256
Total of isolates (n = 131)	mean	0.69	96.36	<0.001
	95% CI	0.63–0.74	74.4–118.29
	range	0.25–2	16–512
	IQR	0.5–1.0	16–64
	MIC50	0.5	64
	MIC90	1.0	256

SI, standard inoculum (∼10⁵ cfu/mL); HI, high inoculum (∼10⁷ cfu/mL); MIC₅₀, cefazolin concentration at which 50% of S. aureus isolates were inhibited; MIC₉₀, cefazolin concentration at which 90% of S. aureus isolates were inhibited.

^aStatistically significant differences were calculated by Mann–Whitney U-test comparing mean MICs at HI and SI.

Table 2 shows the distribution of the cefazolin InE by country and surveillance time. In the first surveillance study, performed in 2001–02, which included 15 Colombian hospitals, a high proportion of isolates exhibited the cefazolin InE (46%) (45% of bloodstream MSSA isolates and 57% of osteomyelitis isolates). In the second and more comprehensive surveillance study, performed between 2006–08, which included hospitals in Colombia, Venezuela, Peru and Ecuador, the overall prevalence of the cefazolin InE in MSSA recovered from deep-seated infections was 34%. When the data were analysed by country of origin, Ecuador had the highest prevalence of MSSA isolates exhibiting the cefazolin InE, with almost half (45%) of the MSSA isolates showing this phenotype (40% and 56% of bloodstream and osteomyelitis isolates, respectively). The most striking finding was that >60% of MSSA isolates recovered from osteomyelitis in Colombia during this surveillance study exhibited the cefazolin InE. The lowest prevalence of the cefazolin InE was observed in Peru and Venezuela (26% and 29%, respectively).

Table 2.

Frequency of cefazolin InE in MSSA from South American hospitals

Multicentre study	Country	Frequency of InE
		blood		osteomyelitis		total isolates evaluated/MSSA with InE (%)
		no. isolates evaluated	no. (%) isolates with InE	no. isolates evaluated	no. (%) isolates with InE
2001–02	Colombia	49	22 (45)	7	4 (57)	56/26 (46)
2006–08	Colombia	150	46 (31)	16	10 (63)	166/56 (34)
	Ecuador	40	16 (40)	18	10 (56)	58/26 (45)
	Peru	39	10 (26)	0	0	39/10 (26)
	Venezuela	18	3 (17)	27	10 (37)	45/13 (29)
	total	247	75 (30)	61	30 (49)	308/105 (34)

InE, inoculum effect.

Molecular characterization of MSSA isolates exhibiting the cefazolin InE

PFGE was performed in all 131 MSSA isolates exhibiting the InE. We were able to differentiate 39 different pulsotypes among bloodstream MSSA, whereas 16 pulsotypes were identified in isolates recovered from osteomyelitis. In MSSA isolates from blood, two pulsotypes were the most frequent (designated Q and O), encompassing 20% and 10% of isolates, respectively. On the other hand, 33% of MSSA isolates recovered from osteomyelitis belonged to two major pulsotypes (designated OB and OF, with 18% and 15% of isolates, respectively). Of note, no genetic relationship with representatives of the most common MRSA pulsotypes circulating in South America (e.g. Chilean and USA300 Latin American variant) was detected by PFGE. Additionally, we were not able to detect the predominance of any country-specific pulsotype, suggesting a genetically heterogeneous population of MSSA isolates exhibiting the cefazolin InE in South America. Furthermore, we performed MLST analysis in representative isolates of the most prevalent pulsotypes from blood and osteomyelitis with the highest cefazolin MICs in each country. MLST typing revealed that the majority of MSSA tested (70%, n = 7) belonged to ST5, ST8, ST30 and ST45. Our MLST data confirm the results of PFGE, indicating that MSSA isolates exhibiting the InE in South America harbour a variety of genetic backgrounds without the predominance of a single clone or clonal cluster.

β-Lactamase typing

We performed sequencing of the blaZ gene region that encodes residues 128–216 in order to determine the type of β-lactamase present in MSSA isolates from South America. This portion of the blaZ gene has been used previously to classify the staphylococcal penicillinase.²¹ We used the same criteria indicated above for MLST in order to select isolates for typing. Among 29 isolates that were selected for β-lactamase typing, we found that the most prevalent type of β-lactamase was type A (66%), followed by type C (31%) and D (3%). Type A β-lactamase was overwhelmingly the most prevalent enzyme in bloodstream and osteomyelitis infections (67% and 63%, respectively), followed by type C (28.5% and 37.5% of isolates, respectively). Table 3 shows the cefazolin MIC distribution by type of β-lactamase enzyme. Isolates harbouring type A β-lactamase had higher mean MIC values at high inoculum than those carrying type C β-lactamase (P = 0.013).

Table 3.

Cefazolin MICs at high inoculum among types A and C representatives of MSSA isolates

Type of β-lactamase	Susceptibility values (mg/L)
	mean MIC	IQR	MIC50	MIC90	range
Type A isolates (n = 19)	164.21a	32–256	64	512	16–512
Type C Isolates (n = 9)	39.11	16–32	32	128	16–128

^aP = 0.0129 calculated by Wilcoxon rank-sum test to compare mean MICs at high inoculum between isolates with type A β-lactamase versus type C.

Discussion

Cefazolin is widely used in the treatment of S. aureus infections and is administered routinely as antimicrobial prophylaxis in surgery.^2,3 Cefazolin is commonly prescribed in deep-seated S. aureus infections when transitioning to outpatient therapy in patients who receive an initial course of isoxazolyl penicillins (which need to be administered at short intervals due to their short half-life). Cefazolin, with a longer half-life (1.8 h), is usually administered at 8 h intervals, which is more convenient for outpatient administration. Lee et al.²² demonstrated that cefazolin exhibited clinical outcomes similar to nafcillin for the treatment of MSSA bacteraemia and it was better tolerated (fewer side effects) than nafcillin in a retrospective, propensity score-matched, case–control study performed in Korea. Moreover, in a retrospective cohort study by Schweizer et al.²³ that included patients with MSSA bacteraemia, administration of cefazolin was independently associated with a lower adjusted rate of mortality (adjusted hazard ratio: 0.21; 95% CI: 0.09–0.47) compared with vancomycin. Similarly, Stryjewski et al.²⁴ compared the outcomes of 123 patients on haemodialysis who received vancomycin or cefazolin. Treatment failure was more common in patients receiving vancomycin (31.2%) than cefazolin (13%; P = 0.02). When the authors performed a multivariable analysis, independent factors associated with therapeutic failure included the use of vancomycin and retention of the intravenous access.²⁴

Despite the clinical efficacy and good in vitro activity of cefazolin against MSSA, failures in the treatment of endocarditis have been documented since 1973.²⁵ Indeed, Quinn et al.²⁵ were the first to demonstrate that the MIC of cefazolin increased >10-fold when tested at high inoculum (10⁷cfu/mL) against an MSSA isolate recovered from a patient who failed cefazolin therapy for endocarditis. Subsequently, cefazolin therapeutic failures and relapses in the treatment of MSSA strains that demonstrate the cefazolin InE have been confirmed.^4,26 Nonetheless, the overall clinical impact of the cefazolin InE in the treatment of MSSA infections has not been systematically investigated and only two previous studies^11,14 have attempted to identify the prevalence of the cefazolin InE among MSSA isolates. The first one¹¹ included 98 isolates recovered in several countries throughout the world from patients with a variety of S. aureus infections (complicated skin and soft tissue infections, hospital-acquired pneumonia, endocarditis and bacteraemia). In that study,¹¹ the overall prevalence of the cefazolin InE (MICs 16 to ≥128 mg/L) was only 19% (19 out of 98 isolates). In a more recent study by Livorsi et al.¹⁴ that included isolates from bloodstream infections recovered from five hospitals affiliated with the same university in the USA, the cefazolin InE was found in 27% of MSSA isolates (MICs 1–32 mg/L, with a definition of InE as a ≥4-fold increase in the cefazolin MIC at high inoculum).¹⁴ In South America, the prevalence of the InE effect was unknown, despite the fact that it may have an important clinical impact, since, at least in Colombia, cefazolin is often used as first-line therapy for severe infections caused by MSSA due to the possibility that generic oxacillin derivatives may lack in vivo therapeutic equivalence with that of the innovator molecule;¹⁵ however, clinical data supporting this approach are lacking. In other countries such as Argentina, oxacillin and derivatives are not even available in the market, making cefazolin the ‘drug of choice’ for MSSA infections. It is important to note that a range of high-inoculum cefazolin MICs were observed among MSSA, suggesting that isolates with higher MICs may be more prone to therapeutic failure, although this correlation has not been tested clinically.

Although our data cannot be generalized to all hospitals in the region, our results are the first to show that MSSA recovered from invasive infections (bloodstream and osteomyelitis) in South American hospitals exhibit a higher frequency of the cefazolin InE than previously reported in other parts of the world. Most worrisome is the finding that almost half of MSSA isolates recovered from osteomyelitis infections (a disease that requires appropriate bactericidal therapy) exhibited such an InE. These findings may have implications for the use of cefazolin for osteoarticular infections in the region. Additionally, a high prevalence of InE in MSSA isolates recovered from bloodstream infections was found in Ecuador and Colombia (40% and 31%, respectively), suggesting that the activity of cefazolin may be compromised in at least one-third of patients presenting with MSSA bacteraemia in those countries and emphasizing the need for clinical data regarding the outcomes of patients treated with cefazolin in the region. Of note, the InE may also affect other cephalosporins and testing of this phenomenon with other compounds of the same class would also be of importance.

The clinical relevance of our findings is unclear. Although clinical failures when using cefazolin in high-inoculum infections such as infective endocarditis have been documented,^4,12,25,26 the paucity of clinical data precludes making strong recommendations at this point. It is also unclear if the cefazolin InE would have relevance in other infections with lower bacterial inocula. Our data suggest that in clinical settings with a high prevalence of MSSA isolates exhibiting the InE, it would be prudent for the laboratory to test for this phenotype in deep-seated MSSA infections, although the cost–benefit ratio of this approach needs to be carefully evaluated. Moreover, strategies to overcome the cefazolin InE (e.g. higher cefazolin dose or combinations with β-lactamase inhibitors) in serious MSSA infections would need to be considered if an important clinical impact of this phenotype is confirmed. Therefore, prospective clinical data are urgently needed to address this issue.

In agreement with previous reports,^11,14,27 type A was the most frequent β-lactamase detected in MSSA isolates exhibiting the cefazolin InE. This is not surprising, since type A β-lactamase has a higher kinetic affinity for cefazolin than other types.⁵ Nonetheless, 31% of MSSA isolates exhibiting the InE harboured type C β-lactamase, which has also been previously associated with the cefazolin InE.¹¹ It is important to note that both type A and C β-lactamases are often encoded by genes present on transmissible elements, such as plasmids and transposons,^28,29 suggesting that the high prevalence of InE observed in the region may be due to the dissemination of mobile elements among MSSA, although other factors, such as hyperproduction of the enzyme,³⁰ may also contribute to the InE. Another important conclusion from our results is that the cefazolin InE could not be explained by the dissemination of a specific clone, since a high degree of genetic diversity was found among the MSSA exhibiting the cefazolin InE, with a prevalence of four genetic lineages that belong to the most important S. aureus (including MRSA) clonal complexes (CCs) disseminated worldwide (CCs 5, 8, 45 and 30).³¹ Our results also suggest that similar genetic backgrounds are present in both MSSA and MRSA¹⁷ from South America, raising the possibility that the changing molecular epidemiology of hospital-associated (HA)-MRSA in the region that has been previously reported^17,32 may originate from a pool of genetic lineages of circulating MSSA. Indeed, the two current most successful HA-MRSA clones prevalent in South America belong to CC5 (Chilean clone) and CC8 (USA300 Latin American variant), suggesting that successful MSSA lineages within CC5 and CC8 that acquired the SCCmec by horizontal dissemination^33–35 may explain the dynamics of the population genetics of MRSA in the region.

In summary, a high prevalence of cefazolin InE was detected among MSSA isolates from northern South America and was associated with β-lactamase types A and C, suggesting that treatment of deep-seated infections with cefazolin (particularly in Colombia and Ecuador) may be compromised. Further studies are needed to determine the clinical impact of the cefazolin InE in severe S. aureus infections in the region.

Funding

This work was supported by a research award from the Universidad El Bosque, ‘Research and Development Projects Program’ Project PCI 2011-235. C. A. A. is supported by NIH grant R01-AI093749 from the National Institute of Allergy and Infectious Diseases (NIAID), part of the US National Institutes of Health.

Transparency declarations

None to declare.