In this study, we found that the drugs whose activities were most affected by the inoculum effect were SAM (35% of the strains) and Cz (25% of the strains). Although the GM MICs of all cephalosporins tested were significantly lower for the higher inoculum than for the standard inoculum (p < 0.001), no InE was detected against cefuroxime or cefotaxime, and only a limited percentage (3.8% of the strains) of InE was detected against ceftriaxone. The increase in the GM MIC was also greatest for SAM, which was 2.94 times greater at high inoculum rates, followed by Cz (2.20 times), cefuroxime (1.27 times), ceftriaxone (1.21) and cefotaxime (1.20 times). Similar to our findings, in a recent Korean study including 302 MSSA isolates of bacteremia, InE was found to be more common (43% and 66% for piperacillin-tazobactam (TZP) and SAM, respectively) and prominent (MICs were nearly 24 and 9 times higher for TZP and SAM, respectively, at higher inoculum rates) in β-lactam-β-lactamase inhibitor (BL/BLI) combinations (7). Additionally, pronounced SAMInE (9.6%) was observed more frequently than pronounced CzInE (5.8%), and no significant increase in MICs at high MSSA concentrations was observed with cefotaxime or ceftriaxone in an in vitro study of 52 MSSA isolates from blood cultures (8).
There have not been any previous reports about the proportion of CzInE or SAMInE from Türkiye. The rate of InE against Cz varied significantly across the studies. The proportion of CzInE in MSSA strains has been reported to be a median of 14.4% and ranges from 0–54.5% and varies significantly from country to country, with the highest proportion being in South American countries (36.0–54.5%), followed by Asian countries (5.8–21.8%), North American countries (0–18.7%) and European countries (2.5–11.0%) (9). The different proportions of InE could be related to the definition of InE, the prevalence and type of β-lactamase among MSSA strains, the hydrolyzing capacity of the prevalent MSSA β-lactamases in the region, β-lactam consumption rates and antimicrobial stewardship efforts in the community, the presence of resistance-carrying genes on mobile genetic elements, and preventive infection control efforts to decrease the spread of resistance genes in health care facilities (10).
In some studies such as ours, CzInE defined a 4-fold or greater increase in MIC values at high inoculum concentrations and pronounced InE as an MIC of ≥ 16 µg/ml with a high inoculum (2, 11, 12), whereas others defined it as an MIC of ≥ 16 µg/ml with a high inoculum (6). In one study from South Korea that used the same definitions, the proportions of CzInE and pronounced CzInE were reported to be 57.5% and 20%, respectively (12); in another study, the proportions of CzInE and pronounced CzInE were reported to be 20% and 4% CzInE (11), respectively, which are quite similar to our rates of 25% and 5.7%, respectively. The definition of InE that we used is more sensitive than the latter definition is. Assessment of the clinical implications of InE should include a more sensitive definition of InE to avoid the risk of falsely rejecting possible associations by using a less specific metric (13).
Some studies have suggested that CzInE is related to either type A or type C β-lactamases of MSSA and that TZP InE is related to type C β-lactamases (7, 14). In our study, all of the tested MSSA strains were found to carry type A β-lactamases. While previous studies from various countries generally reported that all four enzyme types were present, albeit in different proportions, type C was reported to be present in 94% of the MSSA strains with β-lactamases in Japan. As a plasmid-mediated and antibiotic-inducible β-lactamase, the spread of the same clones carrying the same enzyme is quite logical and possible, especially in communities with high antibiotic consumption, such as Türkiye. Widespread dissemination of β-lactamase-associated resistance genes between strains of bacteria has occurred several times in our country, such as the spread of OXA-48 among Klebsiella pneumoniae strains (15) and the spread of CTX-M among E. coli strains (16).
Although all of the strains in our study were found to carry type A β-lactamases, which have repeatedly been shown to be associated with a relatively high rate of CzInE, CzInE accounted for only 25% of our strains, suggesting that the type of β-lactamase is not the only reason for CzInE. Consistent with this, a recent study from Latin America including 690 bloodstream MSSA isolates with whole-genome sequencing revealed that the allotype rather than the type of β-lactamase could be a more accurate tool for identifying strains with a likelihood of exhibiting CzInE, and the authors reported that particular amino acid residues (E112A and G145E substitutions) were highly associated with allotypes that exhibited CzInE (17). Another study reported that single nucleotide polymorphisms of the type A blaZ gene at codons 226 and 229 (Ser226Pro and Cys229Tyr) were closely associated with CzInE (18). However, either those specific allotypes or those specific SNPs could also be determined in strains without CzInE. Clearly, CzInE is a multifactorial phenomenon, and further studies analyzing the mechanism of CzIE are still needed.
Although the mechanisms of CzInE and SAM InE have not been fully defined, both are clearly related to the β-lactamases of MSSA. We found that while the GM MICs of cefuroxime, ceftriaxone, and cefotaxime were not affected by the presence or absence of β-lactamase, only the MICs of SAM and Cz were significantly affected by the presence or absence of β-lactamase; the GM MIC values of SAM against strains with β-lactamase were 23 times greater than those without β-lactamase, and this difference became even more pronounced at higher inocula, with the 61 and GM MIC values of Cz against the strains with β-lactamase being 1.66 and 3.86 times greater than those without β-lactamase at standard and high inocula, respectively. Additionally, the GM zone diameters of the 52 MSSA strains subjected to PG and cefazoline were significantly lower for β-lactamase-positive strains than for β-lactamase-negative strains, both at standard and high inoculum rates (< 0.001). However, the interaction of β-lactamases with InE seems different between SAMInE and Cz InE. Cz InE is repeatedly reported to be related to either hyperproduction or better performance of type A and type C BlaZ on cefazoline (2); SAMInE probably results from the decreased inhibition of the β-lactamase inhibitor sulbactam against MSSA β-lactamase (7). In accordance with that suggestion, in our study, we observed that while SAMInE was strongly related to the increase in only the minimum inhibitory concentration (MIC) of SAM (p < 0.001), CzInE was strongly related to the decrease in the PG and Cz zone diameters (p < 0.001 for both) and the increase in the Cz MIC (p < 0.001). and high (28.26 ± 2.63 versus 24.39 ± 2.33, p < 0.001) inoculum. The fact that SAMInE is specific to SAM and does not affect susceptibility to other antimicrobial agents indicates that the mechanism of SAMInE is specific to SAM and is not caused by excessive production of β-lactamase. It is suggested that if the presence of InE against one β-lactamase inhibitor is shown, then InE and decreased activity of other β-lactamase inhibitors could be foreseen. In accordance with this suggestion, TZP is frequently affected by MSSA β-lactamases (7). In an in vitro study, clavulanic acid and tazobactam were 93 and 11 times more active than sulbactam against the β-lactamases of MSSA strains (19).
The mean Cz zone diameter of the strains with CzInE was significantly lower than that of the strains without CzInE, both in standard and high inoculum, in our study (27.94 ± 3.09 versus 23.85 ± 1.99, p < 0.001). A Cz zone diameter of < 28 mm was found to be 100% sensitive for the definition of CzInE among type A β-lactamase-carrying MSSA strains in our study. In a recent study from Australia, it was shown that type A and type C β-lactamases of MSSA could be defined by calculations using zone diameters of cefazolin, cephalothin, and oxacillin, with a sensitivity and specificity of 88.6% and 96.6%, respectively (20). These results show that both the InE and β-lactamase types can be predicted via simple disk diffusion tests and that the zone diameter of Cz can be used as a screening test to define the InE of Cz and to define the necessity of further testing.
We found that the rate of β-lactamase positivity (26/26 vs 18/26, p = 0.004) of the infected MSSA strains was significantly higher and that the rates of neutropenia (4/26 vs. 1/26 p = 0.350), admission to the ICU (6/26 vs. 1/26, p = 0.99), and mortality (6/26 vs. 2/26, p = 0.248) were higher in patients infected with the strains showing InE against SAM or Cz than in patients not infected. In another study, metastatic cancer, recent close contact with a chronically ill patient, and resistance to clindamycin and erythromycin among the causative MSSA strains (7) were found to be risk factors for CzInE. As neutropenia frequently develops in patients with metastatic cancer, the incidence, underlying mechanism and clinical implications of CzInE among patients with cancer or neutropenia should be evaluated in further studies. We did not find any associations between SAM or CzInE and other comorbidities or test results of the patients or with the susceptibility or resistance of the strains to certain antibiotics.
It is unknown whether being less active against the β-lactamase of MSSA and having a more frequent InE among MSSA strains than Cz in SAM will affect the treatment effectiveness of the two agents.
In our study, the mortality rate in patients infected with MSSA strains showing SAMInE and treated with SAM was higher than that in those not treated with SAM (37.5% vs 68%, p = 0.044, OR 7.8, 95% CI 1.23–49.68), and the causative MSSA strains of deceased patients presented higher SAM InE (62.5% vs 34%, p = 0.235) than surviving patients did. We could find only one study analyzing the clinical impact of SAMInE. In that study, which included 302 patients with MSSA bacteremia, the mortality rates of the SAM InE-positive (n = 27) group, who received empirical β-lactam/β-lactamase inhibitors, were significantly greater than those of the negative (n = 23) (32.4% vs 5.6%, p = 0.04) group, and the mortality rate of the SAM InE-positive (n = 28) group, who received definitive β-lactam/β-lactamase inhibitors for the treatment of MSSA bacteremia, was also significantly greater than that of the negative (n = 14) group (26.1% vs. 8.3%, p = 0.38) (7).
There are also several studies comparing the effectiveness of SAM or other BL/BLI combinations and other first-choice regimens, including either Cz or ASPs. In a retrospective cohort study of 478 patients with MSSA bacteremia, mortality was found to be similar between patients treated with cloxacillin and cefazoline, but it was nearly two times (OR 2.68, p = 0.08) greater among patients treated with BL/BLI combinations, including TZP (n = 32), ampicillin-clavulanate (n = 28) and SAM (n = 1) (21). In a retrospective cohort study of our group including 127 MSSA bacteremia patients, the mortality rate of patients treated with Cz (2/30, 6.6%) was lower than that of patients treated with SAM (9/47, 19%); however, the difference did not reach statistical significance (p = 0.082) (22).
In the study of Uda et al., another BL/BLI in combination with TZP for definitive therapy of MSSA bacteremia was found to be associated with treatment failure for MSSA bacteremia (OR = 17, p = 0.003) (23).
In a USA study including more than 400 MSSA bacteremia cases, while no difference in mortality was observed between ASP and cefazolin or fluoroquinolones, higher mortality was observed with TZP than with ASP/cefazolin (HR, 0.10; 95% CI, 0.01–0.78), suggesting that TZP may not be as effective as monotherapy for MSSA bacteremia (24).
Finally, in a recent retrospective observational study from Japan comparing the clinical efficacy of SAM (41 patients) with that of Cz (30 patients) in patients with MSSA bacteremia, the mortality rate did not differ between the groups (25).
All of these data suggest that, owing to the lower effectiveness of sulbactam against the β-lactamases MSSA and SAMInE, SAM treatment of MSSA bacteremia and other higher inoculum infections could be less effective than other options, including Cz. SAM is a frequent replacement for first-line antimicrobials for MSSA bacteremia, including ASP and Cz, especially in the case of the unavailability of those agents, such as those used here in Türkiye, Argentina, and previously in Japan, or if those agents could not be used owing to adverse effects. Therefore, additional studies are urgently needed on this subject, and the limited evidence along with our findings support the hypothesis that the BL/BLI of SAM or TZP could have decreased activity against some β-lactamases of MSSA, which could lead to decreased activity of those agents during treatment.
In our study, the mortality rate in patients infected with MSSA strains showing CzInE and treated with Cz was greater than that in those not treated with Cz (12.5% vs. 6.8%, p = 0.514); additionally, the mortality rate of patients infected with the strains showing CzInE (4/13, 30%) was greater than the mortality rate of patients infected with the strains not showing CzInE (4/39, 10%) (p = 0.096), but the differences were not statistically significant in either of the findings, and the numbers were insufficient to reach a definite conclusion. Cz is the most studied drug for the presence and clinical consequences of InE in MSSA strains that cause bacteremia, but the results of these studies are conflicting. In a recent meta-analysis of 23 observational studies, CzInE was defined in 0%-55% of the cases, and the mortality rate of serious infections caused by MSSA did not differ significantly between the strains with and without CIE. However, the quality of the included studies was low (26). A well-designed study without limitations is urgently needed to answer the question about the clinical impact of CzInE.
As cefuroxime was found to be highly effective in vitro against MSSA strains, with a GM MIC similar to that of Cz and was shown to be unaffected by InE in our study and other studies (27, 28), it could be evaluated as an alternative agent to SAM or cefazoline in the case of InE against those antimicrobials. In a recent study of 268 patients with MSSA bacteremia who were empirically treated with a mean of 3 days of either ASP flucloxaciline, cefuroxime or ceftriaxone, the duration of bacteremia or SAB-related mortality did not differ between the groups (29). Therefore, it is important to conduct comparative studies with Cz and cefuroxime in such MSSA infections. However, as the β-lactamases of MSSA are not constitutive but inducible with exposure to β-lactam antibiotics, close monitoring of the susceptibility of all antibiotics used along with rational antimicrobial usage efforts is of utmost importance (30).
Our study has several limitations, including its retrospective and observational design and the small number of patients in each group for some comparisons. However, our study contributes to the current limited knowledge concerning the incidence and clinical consequences of SAMInE in patients with MSSA bacteraemia. Additionally, for the first time, we found a disk diffusion test zone diameter breakpoint for the screening of CzInE.
In conclusion, InE is more frequently encountered against SAM than Cz is among MSSA strains causing bacteremia, probably because of the decreased activity of sulbactam against some of the type A β-lactamases of the MSSA strains. SAM treatment of patients infected with MSSA strains harboring SAMInE may increase mortality. Additional studies that provide stronger evidence are needed concerning the incidence and clinical consequences of InE among MSSA strains of deep-seated infections caused by β-lactam agents, including not only Cz but also BL/BLI. A cefazolin zone diameter of < 28 mm could be used as a screening method to define Cz InE.