Authors: Simone Giuliano, Jacopo Angelini, Floriana Campanile, Paola Conti, Sarah Flammini, Alberto Pagotto, Francesco Sbrana, Luca Martini, Denise D’Elia, Mohd H. Abdul-Aziz, Menino O. Cotta, Jason A. Roberts, Robert A. Bonomo, Carlo Tascini
Categories: Article, Ampicillin, Ceftobiprole, Enterococcus faecalis, Infective endocarditis, Synergism, Therapeutic drug monitoring, PK/PD, Microbiology, Medical research
Source: Scientific Reports
Authors: Simone Giuliano, Jacopo Angelini, Floriana Campanile, Paola Conti, Sarah Flammini, Alberto Pagotto, Francesco Sbrana, Luca Martini, Denise D’Elia, Mohd H. Abdul-Aziz, Menino O. Cotta, Jason A. Roberts, Robert A. Bonomo, Carlo Tascini
Enterococcus faecalis is responsible for numerous serious infections, and treatment options often include ampicillin combined with an aminoglycoside or dual beta-lactam therapy with ampicillin and a third-generation cephalosporin. The mechanism of dual beta-lactam therapy relies on the saturation of penicillin-binding proteins (PBPs). Ceftobiprole exhibits high affinity binding to nearly all E. faecalis PBPs, thus suggesting its potential utility in the treatment of severe E. faecalis infections. The availability of therapeutic drug monitoring (TDM) for ampicillin and ceftobiprole has prompted the use of this drug combination in our hospital. Due to the time-dependent antimicrobial properties of these antibiotics, an infusion administration longer than indicated was chosen. From January to December 2020, twenty-one patients were admitted to our hospital for severe E. faecalis infections and were treated with this approach. We retrospectively analyzed their clinical characteristics and pharmacological data. Most patients achieved an aggressive PK/PD target (T > 4–8 minimum inhibitory concentration, MIC) when this alternative drug combination regimen was used. Our analysis included the study of E. faecalis biofilm production, as well as the kinetics of bacterial killing of ceftobiprole alone or in combination with ampicillin. Time-kill experiments revealed strong bactericidal activity of ceftobiprole alone at concentrations four times higher than the MIC for some enterococcal strains. In cases where a bactericidal effect of ceftobiprole alone was not evident, synergism with ampicillin and bactericidal activity were demonstrated instead. The prolonged infusion of ceftobiprole, either alone or with ampicillin, emerges as a valuable option for the treatment of severe invasive E. faecalis infections.
The online version contains supplementary material available at 10.1038/s41598-025-87512-8.
Enterococcus faecalis can cause a large variety of difficult-to-treat infections and monotherapy is generally non bactericidal. Life-threatening infections such as endocarditis and bacteremia might require combination therapy^1^. Synergism between amoxicillin and cefotaxime through partial saturation of penicillin-binding proteins (PBPs) 4 and 5 by amoxicillin and total saturation of PBPs 2 and 3 by cefotaxime is known^2^ and the combination of ampicillin and ceftriaxone was found to be safe and effective for treating IE^3^. with similar effectiveness compared to ampicillin plus gentamicin^4^, and lower rates of serious side-effects^5^. Ceftobiprole is a broad-spectrum cephalosporin with in vitro activity against methicillin-resistant Staphylococcus aureus (MRSA) and E. faecalis^6^. Ceftobiprole and amoxicillin showed a synergistic and bactericidal effect against E. faecalis^7^. Owing to the pharmacodynamic characteristics of ceftobiprole, and the availability of therapeutic drug monitoring (TDM) for both ampicillin and ceftobiprole at our institution, we have been increasingly using ampicillin plus ceftobiprole combination instead of ampicillin and ceftriaxone for the treatment of IE and other invasive infections caused by E. faecalis. A TDM-guided approach was used throughout the anti-infective pharmacologic treatment to adjust the doses of these two drugs and achieve the most aggressive antimicrobial PK/PD targets. The objective of this study is to delineate the pharmacological properties of ampicillin and ceftobiprole when used in combination to treat invasive E. faecalis infections, particularly IE. Our investigation concentrated on two main (1) attainment of PK/PD targets, (2) the antimicrobial efficacy of ceftobiprole both alone and in combination with ampicillin as demonstrated by time-kill assays. The findings are related to the clinical outcomes derived from our pharmacological evaluation strategy, as previously outlined in a case series published by our team^8^.
This retrospective investigation encompassed adult patients treated for E. faecalis IE or bacteremia with the combination of ampicillin plus ceftobiprole between January 2020 to December 2020. The research took place at a tertiary university hospital, the Azienda Sanitaria Universitaria Friuli Centrale of Udine, Italy. The study was approved by the local Institutional Review Board and all research was performed in accordance with relevant guidelines/regulations. Due to the retrospective and observational nature of the analysis, the patient’s signed written informed consent was waived.
Ampicillin was administered via continuous intravenous infusion, starting with a standard dose of 16 g per day. The dosage was adjusted based on renal reduced to 12 g per day for an estimated creatinine clearance (eCLCR) between 30 mL/min and 50 mL/min, and further reduced to 8 g per day when eCLCR was < 30 mL/min. Ceftobiprole was initially administered utilizing the dosing regimens recommended for distinct classes of renal function. These dosages included 500 mg q8h over 3-h extended infusions (EI) for patients with eCLCR ≥50 mL/min, 500 mg q12h over 3 h-EI for those with eCLCR of 30–50 mL/min, and 250 mg q12h over 3 h-EI for those with eCLCR <30 mL/min. Following a minimum of two days of treatment initiation, patients underwent TDM of both drugs. The objective of the TDM was to make dose adjustments that align with the optimal PK/PD target, represented by a steady-state free plasma trough concentration (fCtrough) to MIC ratio of 4–8 for the entire dosing interval (corresponding to a 100%fT > 4-8xMIC)^9^. This target was devised to achieve both clinical and microbiological cures, while concurrently mitigating the risk of overexposure and associated adverse events. We extracted demographic and clinical data from the medical records of each patient, including age, gender, weight, height, type of infection, bacterial clinical isolate and corresponding MICs of ampicillin and ceftobiprole, serum creatinine, eCLCR (calculated using the Cockcroft-Gault equation), ampicillin and ceftobiprole daily dose, and any concurrent medications.
A single blood sample was collected during continuous ampicillin infusion. Because of the 8-hourly intermittent 3-hour duration intravenous infusion of ceftobiprole, we usually performed three venous blood samples to characterize the PK profile of before dosing, at the end of the drug infusion, and four hours after dosing. The ceftobiprole dosing interval Area Under the concentration-time curve (AUC0 − t), the half-life (T1/2), the volume of distribution (Vss) and the clearance (CL) were estimated by non-compartmental model.
A high-performance liquid chromatography (HPLC) with ultraviolet (UV) detection method was used to quantify ceftobiprole in plasma. In brief, one milliliter of plasma or 0.5 mL of sample was spiked with cefepime as an internal standard (IS) and then subjected to solid-phase extraction (SPE) before HPLC analysis. A reverse phase chromatography separation was performed using Shimadzu Nexera-i equipped with UV detector (Shimadzu, Kyoto, Japan) and with an InfinityLab Poroshell 120 EC -C18 column (3.0 × 50 mm with 2.7-micron particle size) (Agilent, Santa Clara, United States). The mobile phase, consisting of a mixture of acetonitrile and 30 mM phosphate buffer (pH 3.15), was used in a gradient mode and detection was performed at a wavelength of 296 nm. The method was linear in the concentration range of 1.0–30.0 microg/mL in plasma. Only total plasma concentrations were measured for both ampicillin and ceftobiprole, whereas the free fraction was estimated. According to literature, we considered a 20% and a 16% of binding protein for ampicillin and ceftobiprole, respectively^10,11^.
E. faecalis ATCC 29,212 (vancomycin-susceptible, VSE) was used as control strain in this study. All strains were further analyzed for the in vitro antibacterial activity. Ceftobiprole was provided by Basilea Pharmaceutica International Ltd. (Basel, Switzerland); ceftaroline, linezolid and tigecycline by Pfizer Inc. (New York, NY, USA); daptomycin by Novartis (Basel, Switzerland). Penicillin, ampicillin, amoxicillin, imipenem, vancomycin, teicoplanin, gentamicin and streptomycin were purchased commercially (Sigma Chemical Co., ST. Louis, MO, USA). MICs were determined by broth microdilution and interpreted according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) clinical breakpoints^12^. In the absence of EUCAST clinical breakpoints, those of the Clinical and Laboratory Standards Institute were applied^13^.
Ten E. faecalis strains were available for in vitro time-kill experiments. Ceftobiprole exposure was tested at concentrations of one time MIC (1X MIC), two times MIC (2X MIC), and four times MIC (4X MIC). Ceftobiprole and ampicillin antibiotic combinations were tested for four strains. The experiments were performed in duplicate in 20 ml tubes containing cation adjusted Mueller-Hinton broth (CA-MHB). (Difco, Detroit, MI) using a starting inoculum of 10^5^ -10^6^ CFU/mL with ceftobiprole (1 X MIC, 2X MIC and 4X MIC) alone or in combination with ampicillin at 1X MIC). Data points are averages from duplicate CFU/mL determination within an experiment. Bactericidal activity was defined as a ≥ 3 log10 decrease in bacterial count at 24 h. Synergy was measured at 24 h and was defined as a ≥ 2 log10 decrease in CFU/ml by the combination compared with its most active constituent and a ≥ 2 log10 decrease in the CFU/ml below the starting inocula^14^.
The ability of E. faecalis isolates to form biofilms was estimated using the 96-microtiter plate method, as described previously with some minor modifications^15^. Briefly, 150 µl of the broth cultures in trypticase soy broth (TSB; Oxoid, Basingstoke, UK) with 0.25% glucose (OD600 ~ 0.4) were inoculated in the microplates and incubated at 37° C under agitation for 24 h. After that, the wells were washed 3 times with 0.9% w/v NaCl and the remaining attached bacteria heat-fixed by incubating the plate at 37 °C for 1 h. The adherent biofilms were stained with 150 µl of 0.4% crystal violet for 30 min at room temperature. The plates were washed at least 3 times and dried at 37° C for 1 h. The crystal violet was solubilized in 150 µL of 33% acetic acid and a microplate reader (GloMax^®^ multidetection system) was used to read the optical density at a wavelength of 595 nm. Uninoculated blank wells containing tryptone soya broth (TSB; Oxoid, Basingstoke, UK) were used to determine background OD. The average from a triplicate OD value were calculated for all tested strains and negative controls. For the interpretation of the results, S. epidermidis ATCC 35,984 and E. faecalis ATCC 29,212 were used as positive (strong biofilm producer) and negative (no biofilm producer) controls, respectively. Based on the comparison between their OD values with those of the controls, strains were divided into the following no biofilm producer (0), weak biofilm producer (+ or 1), moderate biofilm producer ( + + or 2) and strong biofilm producer (+++ or 3).
A descriptive statistical analysis was used to describe the characteristics of the treated population and measurements of plasma antimicrobial concentrations. For continuous variables, means, standard deviations, medians and interquartile ranges were calculated. For discrete variables, the frequency of distribution was presented. The normality of the distribution of continuous variables was tested by the Shapiro-Wilk test. The software used for data analysis was GraphPad Prism (version 10.1.1, November 2023).
A total of 21 patients with E. faecalis invasive infections were admitted to our hospital from January 2020 to December 2020 as previously reported in Giuliano et al. where population is described^8^.
The descriptive analyses of ampicillin and ceftobiprole plasmatic concentrations are reported in Figs. 1 and 2; Tables 1 and 2.
Fig. 1Ceftobiprole (BPR) and ampicillin serum concentrations. efC BPR 50%T: estimated free serum concentration at 50% of the dosing interval; efC BPR 100%T: estimated free trough serum concentration of ceftobiprole; ef Css estimated free serum concentration of ampicillin at the steady state during a continuous infusion. The y-axis is represented on a log2 scale.
Fig. 2Spaghetti plot reporting the estimated free serum concentrations of ceftobiprole from the 21 analyzed patients measured at different time-points.
Table 1Pharmacokinetic parameters of ceftobiprole.Number of investigated 12efCmin(mg/L)Dose-normalized efCmin(mg/L)efCmax(mg/L)Dose-normalized efCmax(mg/L)efAUC0 − t(mg/L.h)Dose-normalized efAUC0 − t(mg/L.h)Half-life(h)Vd(L)Vd/kg (L/kg)CL(L/h)Mean7.4986.52819.514.226123.427.545.1323.5830.22242.596Std. Deviation3.3152.7924.7128.94342.1911.082.3111.6090.063591.333Coefficient of variation (%)44.2142.7724.1521.1634.1840.2245.0344.9028.5951.36Minimum2.2262.22310.9926.1762.8312.572.4861.6860.15051.203Maximum12.6711.5726.0256.21216.449.8310.16.6850.33685.591efCmin: estimated free trough serum concentration; efCmax: estimated free peak serum concentration; efAUC0-t: estimated free area under the time-concentration curve throughout the dosing; volume of distribution; Vd/ volume of distribution per kg; CL: drug clearance. All the blood samples were drawn when ceftobiprole serum concentrations were at steady state.
Table 2Antibiotic data for achievement of Pharmacokinetic/Pharmacodynamic (PK/PD) target.PK/PD parameterAntibioticCeftobiprole (100%; n = 21)Ampicillin (100%; n = 21)50% T~> MICachieved90.5%; 19-50% T> 4−MICachieved71.0%; 15-100% T> MICachieved85.5%; 1881.0%; 17100% T> 4−MIC~achieved62.0%; 1376.0%; 16MIC - minimum inhibitory concentration; 50% T > MIC - at least, an half of the duration of a dosing interval the antibiotic concentration remains above the MIC of the known or suspected pathogen; 50% T > 4xMIC - at least, an half of the duration of a dosing interval for which the antibiotic concentration remains above a concentration that is four-fold the MIC of the known or suspected pathogen (maximal activity); 100% T > xMIC - the entire duration of a dosing interval for which the antibiotic concentration remains above the MIC of the known or suspected pathogen; 100% T > 4xMIC - the entire duration of a dosing interval for which the antibiotic concentration remains above a concentration that is four-fold the MIC of the known or suspected pathogen (maximal activity).
The MIC50 and the MIC90 of the isolated E. faecalis strains for ceftobiprole was 0.5 mg/L and 1 mg/L, respectively. The MIC50 and the MIC90 of the isolated E. faecalis strains for ampicillin was 1 mg/L and 2 mg/L, respectively. We calculated the PK/PD index for both ceftobiprole and ampicillin (Fig. 3). We found that the free serum concentrations of ceftobiprole were 18-fold the MIC (median 17.9; IQR 5.8–27.2)], whereas the free serum concentrations of ampicillin were 46-fold the MIC (median 46.1; IQR 29.7–77.0). The coefficient of variation of the PK/PD index of ceftobiprole and ampicillin were 70.5% and 67.9%, respectively. For both antimicrobials, at least 80% of patients achieved the PK/PD target of 100% T > MIC, and at least 60% the optimal ones represented by 100% T > 4MIC, as detailed in Table 2. A descriptive overview of clinical, microbiological, and pharmacological data of each case is reported in Table 3.
Fig. 3Ceftobiprole (BPR) and ampicillin PK/PD index. The observed Pharmacokinetic/Pharmacodynamic (PK/PD) ratios at 100% of dosing interval (100% fT > MIC). A ratio of 1 represents the minimum PK/PD therapeutic target, whereas the ratio of 4 (dotted red line) represents a high antimicrobial activity. The reported MICs derive from pathogen isolates or surrogates from EUCAST when unknown. The PK/PD index was calculated referring to estimated free serum concentrations of antibiotics. The y-axis is represented on a log2 scale. Legend. MIC - minimum inhibitory concentration; EUCAST - European Committee on Antimicrobial Susceptibility Testing.
Table 3Patient-level clinical, microbiological, and pharmacological characterization of the 21 analyzed patients who received the ampicillin and ceftobiprole combination.Case n°DiagnosisDuration ampicillin/ceftobiprole treatment (days)Microbiological eradicationClinicalcureCeftobiproleAmpicillinDose(mg/day)Cmin(mg/L)Cmax (mg/L)C1 h post infusion(mg/L)EstimatedAUC0 − t (mg*h/L)MIC(mg/L)Dose(g/day) by continuous infusionCss (mg/L)MIC(mg/L)#1Prosthetic aortic valve endocarditis8yesyes500 q 8 h6.1423.98NA110.650.125NANA0.50#2Prosthetic aortic valve endocarditis30noyes350 q 8 h10.6723.4218.07129.360.25862.701.00#3Prosthetic aortic valve endocarditis22noyes500 q 12 h10.3323.3921.27199.310.25428.800.50#4Prosthetic aortic valve endocarditis5yesno, death250 q 24 h2.6513.088.73148.301.0016145.701.00#5Native aortic and mitral valve endocarditis19yesno, relapse500 q 8 h6.0115.589.0474.800.251639.700.50#6Native mitral valve endocarditis33yesyes500 q 8 h9.1422.9819.65127.082.001688.342.00#7Native aortic valve endocarditisNAyesyes500 q 8 h6.8821.24NA105.87NA1211.67NA#8Prosthetic aortic valve endocarditis and spondylodiscitis40yesyes500 q 12 h5.8022.3517.14153.731.001665.650.50#9Bacteremia23yesyes250 q 12 h12.316.76NA173.330.50NA88.530.50#10Prosthetic aortic valve endocarditis43yesyes500 q 8 h8.9326.6620.95136.950.51641.341.00#11Native mitral valve endocarditis21yesno, death500 q 8 h15.8028.0326.1174.091.00NANA0.50#12Prosthetic aortic valve endocarditis25yesyes500 q 8 h10.35NANANA0.5NA63.222.00#13Native aortic valve endocarditis27yesyes500 q 8 h3.9716.6411.2175.200.251648.881.00#14Prosthetic aortic and native mitral valve endocarditis60nono, death500 q 8 h7.3526.5917.82123.460.251638.281.00#15Bacteremia14yesyes500 q 8 hNA25.0018.44NA1.001636.001.00#16Bacteremia46yesyes500 q 8 h13.4229.0723.30163.360.258111.021.00#17Bacteremia14yesyes500 q 12 h13.7730.9826.64257.580.508NA1.00#18Bacteremia27yesyes500 q 8 h3.6713.63NA63.900.501640.730.50#19Bacteremia12yesyes500 q 8 h10.9428.97NA152.440.5012NA0.50#20Bacteremia17yesyes500 q 8 h9.15NANANA1.001640.402.00#21Bacteremia10yesyes500 q 8 h2.88NANANA0.501635.570.50Every patient with ceftobiprole received 500 mg q8h over 3-h. microbiological eradication was defined as obtaining negative follow-up blood cultures at any point during ampicillin/ceftobiprole treatment following the initial positive blood culture or obtaining negative valve cultures in patients who underwent surgery. Cmin - total trough serum concentration; Cmax − total peak serum concentration; C1 h post infusion - total serum concentration sampled 1 h after the end of ceftobiprole infusion; AUC0 − t - total area under the time-concentration curve throughout the dosing; MIC - minimal inhibitory concentration; Css - serum concentration at steady state; NA - not available; q 8 h - antibiotic dose administered every 8 h; q 12 h - antibiotic dose administered every 12 h; q 24 h - antibiotic dose administered every 24 h.
Ten E. faecalis strains were available for time-kill experiments and biofilm production assessment. In two of these isolates (see Fig. 4and Supplementary Fig. 1), which were no-biofilm producers with a ceftobiprole MIC of 0.25 mg/L, ceftobiprole exhibited bactericidal activity at 8 and 24 h after exposure to drug concentrations four times the MIC. In another isolate of E. faecalis, characterized by weak biofilm production and a ceftobiprole MIC of 0.5 mg/L, bactericidal activity was observed at 8 and 24 h after exposure to drug concentrations four times the MIC (Fig. 5). In an additional isolate with a ceftobiprole MIC of 1 mg/L and moderate biofilm production, bactericidal activity of ceftobiprole alone was observed at four times the MIC at 4-, 8-, and 24-hours post-exposure to the drug (see Supplementary Fig. 2). For another E. faecalis strain with a ceftobiprole MIC of 0.25 mg/L and for which a very high level of biofilm production was recognized (+++), ceftobiprole was bactericidal alone at two and four times the MIC at 24-hours post-exposure; addition of ampicillin at concentration equal to its MIC did not result in synergism (Fig. 6). In Fig. 7, the time-kill experiment of an E. faecalis isolate with moderate biofilm production and a ceftobiprole MIC of 0.5 mg/L is illustrated. Ceftobiprole did not exhibit a bactericidal effect at concentrations of 1, 2, and 4 times the MIC, even after 24-hour incubation, for this strain. Additionally, the isolate showed regrowth after 8 h of drug exposure. Furthermore, the antibiotic gradient diffusion strip conducted on this strain revealed an MIC of 0.75 mg/L (Fig. 8A). Colonies displaying heteroresistance were observed within the inhibition zone, and the antibiotic gradient diffusion strip performed on these internal colonies of heteroresistance showed an MIC of 16 mg/L (Fig. 8B). When this isolate was exposed to ceftobiprole at 1X MIC, 2X MIC, and 4X MIC, along with ampicillin at a concentration equal to its MIC, a synergistic effect was observed at 8- and 24-hours post-exposure to the drugs (Fig. 9). The same pattern was observed for another E. faecalis strain, which was a strong producer of biofilm in vitro. Ceftobiprole exhibited weak activity alone; however, when combined with ampicillin at a concentration equal to its MIC, a synergistic killing effect was demonstrated (see Supplementary Fig. 3).
Fig. 4BPR S 58,972,655 E. faecalis. BPR MIC = 0.25 mg/L. Biofilm = 0. Time-kill experiments with BPR alone at 1X, 2X, 4X MIC.
Fig. 5BPR S 53,906,575 E. faecalis. BPR MIC = 0.5 mg/L. Biofilm = +. Time-kill experiments with BPR alone at 1X, 2X, 4X MIC.
Fig. 6BPR S 54,524,183 E. faecalis. BPR MIC = 0.25 mg/L. Biofilm = +++. Time-kill experiments with BPR at 1X, 2X, 4X MIC alone and in combination with ampicillin 1XMIC.
Fig. 7BPR S 59,245,935 E. faecalis. BPR MIC = 0.5 mg/L. Biofilm = ++. Time-kill experiments with BPR alone at 1X, 2X, 4X MIC.
Fig. 8BPR S 59,245,935 E. faecalis. BPR MIC = 0.75 mg/L. Biofilm = ++. antibiotic gradient diffusion strip (E-test) wild type (panel A) and internal colonies (Panel B).
Fig. 9BPR S 59,245,935 E. faecalis. BPR MIC = 0.5 mg/L. Biofilm = ++. Time-kill experiments with BPR at 1X, 2X, 4X MIC alone and in combination with ampicillin 1X MIC.
One of the most significant features that distinguishes Enterococcus from Streptococcus is its tolerance to penicillin, defined as an MBC to MIC ratio greater than 32^16^. Tolerance to beta-lactams justifies the need to use a combination of drugs to achieve bactericidal effect, as no antibacterial agent has proven to be bactericidal on its own against Enterococcus spp., so far^16^. Penicillin alone is bactericidal against Enterococcus only at high doses^17^; however, when combined with streptomycin, a bactericidal effect is achieved at lower penicillin doses^17^. The successful mechanism of the penicillin and aminoglycoside combination is based on the destruction of peptidoglycan by penicillin, leading to aminoglycoside accumulation inside the cell^18^. In this context, aminopenicillins are even more effective, being four times more active against susceptible Enterococcus cells than penicillin^19^. Since the second decade of the 2000s, a new treatment regimen for severe E. faecalis infections has emerged^4^. This regimen includes the combination of two beta-lactams, ampicillin and ceftriaxone^4^. This dual beta-lactam regimen has been demonstrated to be non-inferior to the combination of ampicillin and gentamicin. No significant differences in mortality during antimicrobial treatment were observed between the ampicillin-ceftriaxone combination and the ampicillin-gentamicin combination (22% vs. 21%, P = .81). Similarly, comparable outcomes were noted for mortality at the 3-month follow-up (8% vs. 7%, P = .72), treatment failure (1% vs. 2%, P = .54), and relapse rates (3% vs. 4%, P = .67)^4^. Moreover, due to a lower prevalence of nephrotoxicity (OR 0.45 [0.26–0.77], P = .0182) and drug withdrawal caused by adverse effects (OR 0.11 [0.03–0.46], P = .0160), dual beta-lactam therapy may be a preferable option compared to ampicillin plus aminoglycoside, as demonstrated by Mirna and colleagues^20^. The importance of combination therapy for infective endocarditis caused by E. faecalis has recently been further emphasized by Danneels and colleagues^21^. They highlighted that the recurrence rate of enterococcal infective endocarditis was significantly higher in patients receiving monotherapy with ampicillin compared to combination therapy with ampicillin plus gentamicin or the sequential treatment of ampicillin and gentamicin followed by ampicillin and ceftriaxone^21^. The mechanism of dual beta-lactam combination therapy in E. faecalis might be explained by the concept of the critical level of PBP binding. PBPs are proteins involved in the building of peptidoglycan chains and are distinguished into three classes according to their enzymatic (1) class A, dual-functional PBPs with both glycosyltransferase and transpeptidase activities; (2) class B, transpeptidases; and (3) class C, carboxypeptidases and endopeptidases^22^. Enterococci have three class A and three class B PBPs^23,24^. Several class B PBPs, specifically PBP4 and PBP5, are involved in reduced beta-lactam susceptibility in E. faecalis and E. faecium, respectively^22^. In the case of ampicillin and ceftriaxone combination, ampicillin partially saturates essential PBP4 and PBP5, resulting in the upregulation of non-essential PBP2 and PBP3, to which ceftriaxone subsequently binds, achieving total PBP saturation^2,25,26^. Fontana and coworkers^27^ highlighted three interesting points about the role of PBPs as the killing target of beta-lactams in E. faecalis. First, a direct relationship was demonstrated between bacterial growth rate and penicillin the higher the growth rate the lower the penicillin MIC was^27^. Secondly, in conditions of optimal E. faecalis growth, PBP3 was found to be the only lethal target of a certain amount of PBP3 saturation was necessary for lethality^27^. Lastly, a reduced level of PBP3 saturation was found in conditions of suboptimal growth; however, even when the saturation of PBP3 by penicillin obtained in conditions of non-optimal growth (i.e. in a situation in which a higher dose of antibiotic is required for inhibition because of reduced susceptibility) was the same of that realized in condition of optimal growth, still PBP3 alone could not be identified as the sole lethal target^27^. All these findings demonstrate that inhibition by beta-lactams is indirect and non-constant, depending on which PBP is essential for growth in a certain physiologic status and on its affinity for the antibiotic^28^. Whatever the crucial PBPs are for each growth condition, a “critical” level of PBP binding is required for cell killing. This “critical” level might be a composite of beta-lactam affinity to the PBPs and number of PBPs bound. Hence beta-lactam and, at a greater extent, beta-lactam combinations, by binding to different PBP with various levels of affinity, can increase PBP saturation to the critical level necessary for bacterial killing. Unlike ceftriaxone or cefotaxime, ceftobiprole exhibits intrinsic activity against E. faecalis, which is linked to its high affinity for almost all enterococcal PBPs. By strongly binding to PBP1, PBP2, and PBP3, and saturating PBP4 and PBP5 as well^29^, ceftobiprole might be able to achieve the “critical” level of PBP binding required to eliminate E. faecalis. In addition, while mutations causing overexpression of pbp4 may be associated with increased MICs of ceftriaxone and reduced effectiveness of the combination of ceftriaxone and ampicillin^30^, the bactericidal activity of ceftobiprole is not significantly influenced by increased pbp4 gene expression or point mutations causing alterations in the catalytic site motif of PBP4^31^. These results indicate the need to explore alternative antibiotic combinations for treating different contemporary isolates of E. faecalis causing IE^30^, and one potential combination to consider might be ampicillin plus ceftobiprole. Other potential options for treating severe E. faecalis infections include combinations such as ceftobiprole with daptomycin, gentamicin, or levofloxacin, which have shown synergistic effects in certain time-kill studies^32–34^.
From a pharmacokinetic perspective, the results of this study demonstrate that intravenous infusion of ceftobiprole and ampicillin, tailored to optimize the time-dependent properties of both drugs by the extension of the time of administration, achieved plasma concentrations that were well above the MICs. Given the innovative administration method for ceftobiprole and ampicillin, performing TDM ensured adequate exposure to the antibacterial therapy, enabling the attainment and maintenance of concentrations suitable for achieving aggressive PK/PD targets compared to intermittent infusion. Considering the potential synergistic effect of ampicillin and ceftobiprole, we speculate that the possible resulting reduction in MICs could lead to a higher PK/PD target attainment rate than those observed in this study. This might yield two key enhanced clinical and microbiological efficacy, and a possible reduction in the required drug dosage, offering advantages in terms of both safety and Pharmacoeconomics. It is worth noting that our high value of the PK/PD indices for both ceftobiprole and ampicillin were mainly derived from the low MIC values reported for the isolated pathogens, rather than the high antimicrobial serum concentrations. Nevertheless, our data show that the use of prolonged infusion administration of ampicillin and ceftobiprole is associated with high serum concentrations, despite the initial dose being adjusted based on baseline renal function to minimize antibiotic overexposure and reduce the risk of toxicity. This may also suggest the use of a TDM-based approach to accordingly modify the dosing of antimicrobial therapy^9^. Among some E. faecalis isolates, ceftobiprole exhibited bactericidal activity at multiple time points post-exposure at various multiples of the MIC, regardless biofilm production ability of the strain. Overall, even in isolates strongly producing biofilm, a synergistic effect with ampicillin was not demonstrable when ceftobiprole exhibited high bactericidal activity by itself at concentrations two or four times the MIC. However, the consistent demonstration of the bactericidal effect of ceftobiprole alone was not observed across all isolates, particularly among those which were biofilm producers (see Fig. 7and Supplementary Fig. 3). Moreover, these isolates showed regrowth after 8 h of drug exposure and for one of them ceftobiprole heteroresistance was observed. We hypothesized that three strain-specific mechanisms underlie the loss of ceftobiprole bactericidal resistance and heteroresistance, biofilm-forming capacity, and quorum sensing system. Non-susceptibility to ceftobiprole in E. faecalis primarily arises from the overexpression of pbp4, driven by mutations in the promoter region, and alterations in the catalytic site motifs of PBP4 which might have the potential to interfere with the formation of the ceftobiprole/PBP4 complex^31^. Antimicrobial non-susceptibility is linked to a fitness cost, resulting in a trade-off in the presence of antibiotics, resistant strains outcompete susceptible bacteria, while in antibiotic-free environments, they are disadvantaged compared to susceptible strains^35^. Heteroresistance, characterized by a lower fitness cost than full resistance, is thus a more favorable phenotype; furthermore, it is unstable and reversible^36^. In biofilms, bacteria act collectively, resembling a multicellular organism, to respond to various environmental conditions^37^. This phenomenon is known as quorum sensing system^37–39^. In the biofilm context, under antibiotic pressure, subpopulations of resistant and heteroresistant bacteria gain a selective advantage over antibiotic-susceptible populations. Resistance and heteroresistance phenotypes are regulated by quorum sensing through the production of autoinducers^40–44^. This sacred triangle, comprising biofilm formation, quorum sensing, and the development of resistance and, particularly, heteroresistance, due to its advantage over full resistance, could elucidate the behavior of some of our biofilm-forming E. faecalis isolates, wherein ceftobiprole alone did not exhibit bactericidal activity. In Pseudomonas aeruginosa, ceftazidime binds mainly to PBP1a and PBP3^45^, quenching quorum sensing system and preventing the expression of specific quorum-sensing-related virulence phenotypes^46^. Furthermore, in P. aeruginosa synergism between beta-lactams against biofilm that is potentially related to inhibition of multiple PBP targets has already been hypothesized^47^. Similarly, we could hypothesize that ampicillin and ceftobiprole combination, by binding enterococcal PBPs at a critical level necessary to inhibit the quorum sensing system, might revert heteroresistance, disrupt biofilm, and lead to a synergistic effect that kills enterococci. Our study has several limitations that may affect the generalizability of our findings. These include the retrospective and monocentric nature of our analysis, the limited sample size, and the absence of a treatment control group. Additionally, only total drug concentrations were measured, with the free fractions of ampicillin and ceftobiprole being estimated.
Our study suggests the potential of ceftobiprole alone or in combination with ampicillin as a promising treatment option for severe E. faecalis infections, particularly IE. Through pharmacological characterization and antibacterial activity assessments, we have highlighted the potential efficacy of this combination therapy, which may offer advantages over traditional antibiotic regimens. By building upon previous clinical discussions and recognizing the urgency in finding alternative treatments amidst emerging resistance patterns, our findings offer a significant contribution towards identifying more effective treatment regimens for managing E. faecalis infections. Moving forward, further research, including larger prospective studies, is warranted to fully elucidate the potential benefits and optimize the use of ceftobiprole and ampicillin combination therapy in clinical practice. Overall, our study might represent an important advancement in the field of antimicrobial therapy and has the potential to improve outcomes for patients with IE due to E. faecalis.
Below is the link to the electronic supplementary material.
Supplementary Material 1