Invited Commentary
Infectious Diseases
December 20, 2024
The Inoculum Effect and Staphylococcus aureus Infective Endocarditis—Time to Reconsider Treatment?
Sara E. Cosgrove, Loren G. Miller
JAMA Netw Open. 2024;7(12):e2451300. doi:10.1001/jamanetworkopen.2024.51300
Infective endocarditis (IE) caused by Staphylococcus aureus remains a deadly disease, with up to one-third of patients dying within 30 days of diagnosis.1 Risk factors for early mortality include older age, sepsis, new heart failure, and major neurologic events at the time of presentation.2 Jean et al3 described an additional potential risk factor for death in patients with methicillin-susceptible S aureus (MSSA) left-sided IE, namely the presence of the inoculum effect to the 2 antibiotics most commonly used to treat these infections—cefazolin and oxacillin.
The inoculum effect refers to the finding that the in vitro susceptibility of an antibiotic is affected by the amount of bacteria present.4 If the minimum inhibitory concentration (MIC) of an antibiotic at a standard inoculum (105 colony-forming units [CFU]/mL) is in the susceptible range but the MIC of an antibiotic at a high inoculum (107 CFU/mL) is in the nonsusceptible range or increased by 4-fold or greater, the inoculum effect is considered to be present. In MSSA isolates, the inoculum effect has been best described with cefazolin and is associated with the presence of staphylococcal β-lactamases.
S aureus can produce 4 types of β-lactamases that are encoded by blaZ genes and classified as type A, B, C, and D, with type A being the most prevalent.5 An inoculum effect associated with the antistaphylococcal penicillins, nafcillin, and oxacillin, which are considered to be stable against S aureus β-lactamases, has not been demonstrated robustly. Six in vitro investigations of nafcillin failed to find an inoculum effect, and only 1 study from 1975 found an inoculum effect with oxacillin.4,6 Data regarding the association between the cefazolin inoculum effect in patients with S aureus bacteremia treated with cefazolin and untoward outcomes, such as persistent bacteremia, treatment failure, or mortality, are equivocal with most human studies having only a small proportion of patients with a high inoculum infection, such as IE.7,8
In the current study, Jean et al3 evaluate the association between 30-day mortality and the presence of the blaZgene and the inoculum effect to cefazolin and oxacillin in blaZ gene-positive MSSA isolates in patients with MSSA left-sided IE. The primary strength of this study is the evaluation of patients with a true high organism burden infection, IE, where one might expect the presence of blaZ and the associated inoculum effect to impact treatment outcomes. Using data from the Jean et al article,3 we calculated that in the left-side IE cohort, blaZ was detected in 129 of 180 isolates (72%) and the inoculum effect to cefazolin or oxacillin was detected in 62 of those 129 isolates (48%), or 34% of the total left-sided IE cohort, with 15 of the 62 having inoculum effect associated with cefazolin in cefazolin-treated patients and 47 of the 62 having inoculum effect associated with oxacillin in oxacillin-treated patients. While there was no observed difference in mortality between cefazolin-treated patients and oxacillin-treated patients, there was greater 30-day all-cause mortality in patients with (1) blaZ-positive MSSA isolates compared with blaZ-negative MSSA isolates (29.5% vs 11.8%, respectively) and (2) microbiologic evidence of inoculum effect to the antibiotic used for treatment when the patients treated with cefazolin or oxacillin were combined compared with those without inoculum effect to the treatment antibiotic (40.3% vs 19.4%, respectively). In a multivariable analysis, the presence of inoculum effect with the treatment antibiotic was associated with an increased 30-day mortality (HR, 2.84; 95% CI, 1.28-6.30).
When assessed separately in an unadjusted analysis, there was a significant difference in 30-day mortality among patients treated with cefazolin who also had a cefazolin inoculum effect compared with those who did not (8 of 15 [53%] vs 6 of 30 [20%], respectively; P = .01), but not a significant difference in 30-day mortality in patients treated with oxacillin who had an oxacillin inoculum effect compared with those who did not (17 of 47 [36%] vs 7 of 37 [19%], respectively; P = .06).
While intriguing, the study has important limitations. The majority of deaths occurred early—within the first 10 days of the infection course based on the survival curves. This finding may indicate that the patients who died had baseline risk factors for mortality unrelated to the antibiotic they received. For example, data regarding neurologic complications are presented as outcomes and the frequency was high, occurring in 40% of the overall IE cohort without information on the relative proportions in the inoculum effect and no inoculum effect groups. If many of these events were present on admission, this finding would incur an independent risk of death that likely is not modified by choice of antibiotic.
The authors do not report data on antibiotic dosing or time to appropriate antibiotic initiation. Given that inadequate empirical antibiotic therapy occurs in more than one-fourth of patients with S aureus bacteremia,9 inadequate or delayed doses could be associated with failure risk in high inoculum infections. Furthermore, additional data about the presence and timing of source control and duration of bacteremia would be helpful to understand if the inoculum effect was driving clinical failure and mortality. If patients failed treatment because antibiotics were not effective due to the inoculum effect, longer duration of bacteremia would be expected, particularly if there were uncontrolled infection sources, which in turn could potentiate an inoculum effect.
Finally, the proportion of S aureus strains in this investigation that produced β-lactamases and demonstrated an inoculum effect is higher than in other studies, which could indicate that the circulating clone in the geographic area of study is unique.5 Further investigation into the S aureus strains is needed to assess for other virulence factors that may be associated with blaZ and evidence of inoculum effect and risk of adverse outcomes. Given the unique population, it is conceivable that the presence of blaZ in infecting MSSA isolates represents an epiphenomenon,10 ie, blaZ is not part of the causal pathway but instead could be associated with patient-level factors, treatment-associated differences, and/or non–blaZ-related virulence factors.
Overall, the study by Jean et al3 raises more questions than it provides answers. Is there a risk of treatment failure among isolates demonstrating an inoculum effect in other populations and in other locales where circulating MSSA strains differ? Given that there are no compelling in vitro data supporting the role of inoculum effect when oxacillin treatment is used, do the observed findings with oxacillin represent epiphenomenon? If an association between the inoculum effect and clinical outcome can be demonstrated in prospective studies, can testing for the inoculum effect be developed and employed in clinical microbiology laboratories? And even if tests can be used in a clinical setting, can integration of these tests into clinical practice improve outcomes? In the meantime, the study provides modest additional support to current data that the inoculum effect may be a risk factor for poor outcomes in patients with deep-seated MSSA infections treated with cefazolin. We believe it is prudent to treat left-sided MSSA IE with high doses of antistaphylococcal penicillins initially, with transition to cefazolin considered after blood cultures clear and source control is obtained given cefazolin’s more favorable adverse effect profile in patients requiring prolonged treatment. The story of the impact of the inoculum effect on clinical outcomes of S aureus bacteremia and IE is not new but still in its infancy.
