https://orcid.org/0000-0001-9154-0923
Abstract
The use of extracorporeal membrane oxygenation (ECMO) may alter blood levels of several drugs, including antibiotics, leading to under dosing of these drugs and thus to potential treatment failure. No data exist on pharmacokinetics of new antimicrobial, in particular ceftazidime/avibactam. We therefore perform this study to evaluate ceftazidime/avibactam blood levels in ECMO patients and find factors associated with underdosing.
Retrospective observational study of patients on ECMO having received ceftazidime/avibactam and in whom trough blood levels of ceftazidime and avibactam were available. Main outcome measurement was the number of patients with ceftazidime and avibactam blood levels above predefined cut-off values, derived from the European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoints for Enterobacteriaceae and Pseudomonas aeruginosa, namely 8 mg/L for ceftazidime and 4 mg/L for avibactam, and explored factors associated with underdosing.
Twenty-three ceftazidime/avibactam trough levels were available in 14 ECMO patients, all of them having received veno-venous ECMO for SARS-CoV-2-associated pneumonia. Although ceftazidime levels were above 8 mg/L in all except one patient, nine (39%) of the avibactam dosages were below 4 mg/L. Increased renal clearance (creatinine clearance > 130 mL/min) was the main factor associated with under dosing, since 7 out of the 10 dosages below the predefined cut-offs were measured in patients with this condition.
In ECMO patients receiving ceftazidime/avibactam, ceftazidime and avibactam serum levels are above EUCAST breakpoints in most cases, justifying the use of normal dosing in ECMO patients. Increased renal clearance may lead to ceftazidime and avibactam under dosing.
Introduction
Ceftazidime/avibactam is a broad-spectrum antibiotic active against difficult-to-treat pathogens, such as Enterobacteriaceae producing Ambler class A and D carbapenemases, and multidrug resistant Pseudomonas aeruginosa.1,2 Ceftazidime/avibactam may also be given in combination with aztreonam in patients with Enterobacteriaceae producing Ambler class B carbapenemase, or in patients with multidrug resistant Stenotrophomonas maltophilia infection. These pathogens might be responsible for infections developing in the most severe patients, including those on extracorporeal membrane oxygenation (ECMO).3
Whereas recent data showed that pharmacokinetic (PK) of antibiotics is generally altered on ECMO, with difficulty to reach the target for several antimicrobials including some beta-lactams, some studies reported few modifications for cephalosporins.4,5 However, no data exist on ceftazidime/avibactam PK in ECMO patients.
We therefore perform this retrospective study to evaluate serum ceftazidime/avibactam blood levels in ECMO patients and find factors associated with decreased blood levels.
Methods
Study design, participants
This was a single-centre, observational, retrospective study. Data from 14 patients treated with ceftazidime/avibactam during ECMO support between October 2020 and October 2021, and who had ceftazidime/avibactam serum level measurements, were collected and retrospectively analysed. Twenty-three samples of ceftazidime/avibactam serum levels were available in these 14 patients. All serum measurements were performed just before the next infusion and are hereafter referred to as minimal concentration (Cmin).
Ceftazidime/avibactam use
Ceftazidime/avibactam was used at a dosing of 2000/500 mg q8h over a 2-hour infusion in all except one patient (one dosage), with adjustment in case of renal failure, according to the following the manufacturer’s recommendations: for patients with creatinine clearance between 30 and 50 mL/min, dosing was 1000/250 mg q8h; for patients with creatinine clearance between 16 and 30 mL/min, dosing was 750/1 875 mg q12h; for patients with creatinine clearance between 6 and 15 mL/min or haemodialysis, dosing was 750/1 875 mg q24h; and for patients with creatinine clearance below 6 mL/min or haemodialysis, dosing was 750/1 875 mg q48h, with dose on dialysis days. In patients on continuous renal replacement therapy (CRRT), we used the dosing for patients with normal renal function.
Outcomes
Main outcome measurement was the number of patients with Cmin above predefined cut-off values, derived from the EUCAST breakpoints.6 The EUCAST breakpoint of ceftazidime/avibactam for Enterobacteriaceae and Pseudomonas is 8 mg/L, and the avibactam concentration for susceptibility testing purpose is 4 mg/L.6 Therefore, we arbitrarily set the cut-offs at 8 mg/L for ceftazidime and 4 mg/L for avibactam to define patients whose dosages were or not above Cmin.7 Secondary outcomes included proportion of patients with ceftazidime Cmin ≥ 4×MIC of pathogen, proportion of patients with optimal joint PK/PD target (defined as both Cmin/MIC ratio for ceftazidime ≥ 4 and Cmin/CT ratio for avibactam > 1),8 and parameters that may influence ceftazidime/avibactam concentrations.
Estimated glomerular filtration rate (eGFR) was calculated using the UxV/P formula, where U is urinary creatinine, V the 24-hour urinary output and P serum creatinine. As defined in a previous study exploring beta-lactam PK in ECMO patients, complete renal failure was defined as no urine output, renal failure as an eGFR between 10 and 60 mL/min, normal renal function as an eGFR between 61 and 130 mL/min and increased renal clearance as an eGFR > 130 mL/min.9 These eGFR were obtained using the 24-hour urine collection before the antibiotic assay.
Statistical analysis
Data are expressed as median [IQR] for continuous variables. The data distribution was described with the use of boxplots expressing median, interquartile range and extreme values.
Ethical considerations
In accordance with French law, informed written consent was not obtained because this observational study did not modify existing diagnostic or therapeutic strategies. Patients and/or relatives were informed about the study and could decline inclusion. The study was approved by the Société de Réanimation de Langue Française ethics committee (CE SRLF 19-70). The database is registered with the Commission Nationale de l’Informatique et des Libertés (CNIL, registration no. 1950673).
Results
A total of 14 patients on veno-venous ECMO (all for SARS-CoV-2-associated pneumonia) received ceftazidime/avibactam for ventilator-associated pneumonia and were included in this study. All canulations were at the femoro-jugular site. Each patient’s characteristics and pathogens responsible for infection are described in Table 1. Five patients had acute kidney injury (AKI) according to KDIGO (Kidney Disease Improving Global Outcomes) classification, at the time of ceftazidime/avibactam measurement, among whom four needed CRRT.
Baseline characteristics and ICU course of the patients
| Parameter | Patients N = 14 |
|---|---|
| Age, years | 49 [41–56] |
| Male sex | 11 (78.5) |
| Body mass index, kg/m2 | 32.6 [29.0–37.9] |
| Overweighta | 4 (28.5) |
| Obesityb | 9 (64) |
| Severity score at ICU admission | |
| SOFA | 11.5 [9.0–12.0] |
| SAPS II | 57 [54–59] |
| AKI (KDIGO range 1 to 3 during ATB) | 5 (35.7) |
| Renal replacement therapy at the time of dosage | 4 (28.5) |
| Intermittent haemodialysis | 1 (7) |
| CRRT | 3 (21.4) |
| Creatinine clearance, mL/minc | 81 [12–136] |
| Creatinine clearance <10 mL/min | 4 (29) |
| Creatinine clearance 10–60 mL/min | 1 (7) |
| Creatinine clearance 61–130 mL/min | 4 (29) |
| Creatinine clearance >130 mL/min | 5 (35) |
| Albumin serum level within 48 h before dosage, g/L | 23 [16–25] |
| Time between ECMO start and ATB, days | 26 [22–33] |
| Time between ECMO start and TDM, days | 32 [28–39] |
| Time between ATB start and TDM, days | 4.5 [2–5] |
| Pathogen typed | |
| EBSL Enterobacteriaceae | 2 (14.2) |
| MDR Pseudomonas aeruginosa | 11 (78.5) |
| Class A Carbapenemase Enterobacteriaceae | 0 |
| Class B Carbapenemase Enterobacteriaceaee | 5 (35.7) |
| Class D Carbapenemase Enterobacteriaceae | 1 (7,1) |
| Outcomes | |
| Cure of infection | 3 (21) |
| Mortality | 14 (58) |
| Parameter | Patients N = 14 |
|---|---|
| Age, years | 49 [41–56] |
| Male sex | 11 (78.5) |
| Body mass index, kg/m2 | 32.6 [29.0–37.9] |
| Overweighta | 4 (28.5) |
| Obesityb | 9 (64) |
| Severity score at ICU admission | |
| SOFA | 11.5 [9.0–12.0] |
| SAPS II | 57 [54–59] |
| AKI (KDIGO range 1 to 3 during ATB) | 5 (35.7) |
| Renal replacement therapy at the time of dosage | 4 (28.5) |
| Intermittent haemodialysis | 1 (7) |
| CRRT | 3 (21.4) |
| Creatinine clearance, mL/minc | 81 [12–136] |
| Creatinine clearance <10 mL/min | 4 (29) |
| Creatinine clearance 10–60 mL/min | 1 (7) |
| Creatinine clearance 61–130 mL/min | 4 (29) |
| Creatinine clearance >130 mL/min | 5 (35) |
| Albumin serum level within 48 h before dosage, g/L | 23 [16–25] |
| Time between ECMO start and ATB, days | 26 [22–33] |
| Time between ECMO start and TDM, days | 32 [28–39] |
| Time between ATB start and TDM, days | 4.5 [2–5] |
| Pathogen typed | |
| EBSL Enterobacteriaceae | 2 (14.2) |
| MDR Pseudomonas aeruginosa | 11 (78.5) |
| Class A Carbapenemase Enterobacteriaceae | 0 |
| Class B Carbapenemase Enterobacteriaceaee | 5 (35.7) |
| Class D Carbapenemase Enterobacteriaceae | 1 (7,1) |
| Outcomes | |
| Cure of infection | 3 (21) |
| Mortality | 14 (58) |
Results are expressed as n (%) or median [IQR].
IQR, interquartile range; BMI, body mass index; SOFA, sepsis-related organ failure; SAPS II, simplified acute physiological score; ECMO, extracorporeal membrane oxygenation; AKI, acute kidney injury; KDIGO, kidney disease: improving global outcomes; CRRT, continuous renal replacement therapy; ATB, antimicrobial treatment; EBSL, extended-spectrum beta-lactamase; MDR, multidrug resistant; TDM, therapeutic drug monitoring.
aDefined by a BMI between 25–30 kg/m2.
bDefined by a BMI >30 kg/m2.
cCalculated by the U × V/P formula, where U is urinary creatinine, V the 24-h urinary output and P is serum creatinine.
dFive patients had polymicrobial infection.
eFor these patients with class B carbapenemase infection, ceftazidime/avibactam was given in combination with aztreonam. Two had monomicrobial infection with class B carbapanemase-producing Enterobacteriaceae, and three had polymicrobial infection with class B carbapanemase-producing Enterobacteriaceae and a Pseudomonas aeruginosa.
Baseline characteristics and ICU course of the patients
| Parameter | Patients N = 14 |
|---|---|
| Age, years | 49 [41–56] |
| Male sex | 11 (78.5) |
| Body mass index, kg/m2 | 32.6 [29.0–37.9] |
| Overweighta | 4 (28.5) |
| Obesityb | 9 (64) |
| Severity score at ICU admission | |
| SOFA | 11.5 [9.0–12.0] |
| SAPS II | 57 [54–59] |
| AKI (KDIGO range 1 to 3 during ATB) | 5 (35.7) |
| Renal replacement therapy at the time of dosage | 4 (28.5) |
| Intermittent haemodialysis | 1 (7) |
| CRRT | 3 (21.4) |
| Creatinine clearance, mL/minc | 81 [12–136] |
| Creatinine clearance <10 mL/min | 4 (29) |
| Creatinine clearance 10–60 mL/min | 1 (7) |
| Creatinine clearance 61–130 mL/min | 4 (29) |
| Creatinine clearance >130 mL/min | 5 (35) |
| Albumin serum level within 48 h before dosage, g/L | 23 [16–25] |
| Time between ECMO start and ATB, days | 26 [22–33] |
| Time between ECMO start and TDM, days | 32 [28–39] |
| Time between ATB start and TDM, days | 4.5 [2–5] |
| Pathogen typed | |
| EBSL Enterobacteriaceae | 2 (14.2) |
| MDR Pseudomonas aeruginosa | 11 (78.5) |
| Class A Carbapenemase Enterobacteriaceae | 0 |
| Class B Carbapenemase Enterobacteriaceaee | 5 (35.7) |
| Class D Carbapenemase Enterobacteriaceae | 1 (7,1) |
| Outcomes | |
| Cure of infection | 3 (21) |
| Mortality | 14 (58) |
| Parameter | Patients N = 14 |
|---|---|
| Age, years | 49 [41–56] |
| Male sex | 11 (78.5) |
| Body mass index, kg/m2 | 32.6 [29.0–37.9] |
| Overweighta | 4 (28.5) |
| Obesityb | 9 (64) |
| Severity score at ICU admission | |
| SOFA | 11.5 [9.0–12.0] |
| SAPS II | 57 [54–59] |
| AKI (KDIGO range 1 to 3 during ATB) | 5 (35.7) |
| Renal replacement therapy at the time of dosage | 4 (28.5) |
| Intermittent haemodialysis | 1 (7) |
| CRRT | 3 (21.4) |
| Creatinine clearance, mL/minc | 81 [12–136] |
| Creatinine clearance <10 mL/min | 4 (29) |
| Creatinine clearance 10–60 mL/min | 1 (7) |
| Creatinine clearance 61–130 mL/min | 4 (29) |
| Creatinine clearance >130 mL/min | 5 (35) |
| Albumin serum level within 48 h before dosage, g/L | 23 [16–25] |
| Time between ECMO start and ATB, days | 26 [22–33] |
| Time between ECMO start and TDM, days | 32 [28–39] |
| Time between ATB start and TDM, days | 4.5 [2–5] |
| Pathogen typed | |
| EBSL Enterobacteriaceae | 2 (14.2) |
| MDR Pseudomonas aeruginosa | 11 (78.5) |
| Class A Carbapenemase Enterobacteriaceae | 0 |
| Class B Carbapenemase Enterobacteriaceaee | 5 (35.7) |
| Class D Carbapenemase Enterobacteriaceae | 1 (7,1) |
| Outcomes | |
| Cure of infection | 3 (21) |
| Mortality | 14 (58) |
Results are expressed as n (%) or median [IQR].
IQR, interquartile range; BMI, body mass index; SOFA, sepsis-related organ failure; SAPS II, simplified acute physiological score; ECMO, extracorporeal membrane oxygenation; AKI, acute kidney injury; KDIGO, kidney disease: improving global outcomes; CRRT, continuous renal replacement therapy; ATB, antimicrobial treatment; EBSL, extended-spectrum beta-lactamase; MDR, multidrug resistant; TDM, therapeutic drug monitoring.
aDefined by a BMI between 25–30 kg/m2.
bDefined by a BMI >30 kg/m2.
cCalculated by the U × V/P formula, where U is urinary creatinine, V the 24-h urinary output and P is serum creatinine.
dFive patients had polymicrobial infection.
eFor these patients with class B carbapenemase infection, ceftazidime/avibactam was given in combination with aztreonam. Two had monomicrobial infection with class B carbapanemase-producing Enterobacteriaceae, and three had polymicrobial infection with class B carbapanemase-producing Enterobacteriaceae and a Pseudomonas aeruginosa.
Figure 1 displays the distribution of the 23 ceftazidime/avibactam Cmin, according to eGFR or RRT. Ceftazidime Cmins were above the predefined cut-off value (i.e. 8 mg/L) in all except one patient who had increased eGFR. Avibactam Cmins were above the predefined cut-off value (i.e. 4 mg/L) for all except nine measurements: six were performed in patients with increased eGFR; one received a half-dose of ceftazidime/avibactam (subsequent dosage after increasing the dosing was above 4 mg/L) and two had low Cmin despite normal eGFR and normal dosing of ceftazidime/avibactam. Taking into account the actual MICs of pathogens responsible for infection showed same results. Moreover, results were also similar when considering aggressive PK/PD for ceftazidime: all ceftazidime Cmin were ≥4× the actual MIC of pathogen (except for two patients with class B carbapenemase-producing Enterobacteriaceae). Considering the joint PK/PD target, as previously described,8 also yielded similar results: for all except nine measurements (those with avibactam Cmin < 4 mg/L), the joint PK/PD target was optimal.
Ceftazidime (white boxes) and avibactam (grey boxes) trough levels according to glomerular filtration rate, estimated by creatinine clearance. The box plots report: the internal horizontal line is the median; the lower and upper box limits are quartile 1 and quartile 3, respectively; bars represent the 95% confidence interval and circles represent outliers.
Among our 14 patients, three were cured of their infection without recurrence, whereas 11 had treatment failure (death during treatment, persistent infection or relapse with the same pathogen). Among the three patients with clinical success (all survived), two had avibactam serum level below the predefined cut-off.
Discussion
Low antibiotic serum levels may lead to the treatment failure in critically ill patients with severe infections, whereas high levels can lead to severe complications. Harmful or inadequate antibiotic concentration in patients on ECMO seems to be difficult to assess.
In this study, ceftazidime and avibactam Cmins were above predefined cut-offs derived from EUCAST breakpoints in most ECMO patients, since 22/23 dosages (96%) of ceftazidime and 14/23 (61%) of avibactam dosages were above these breakpoints. Moreover, all ceftazidime Cmins were above 4× the actual MIC of pathogens. Increased eGFR was the most frequent parameter that may lead to have ceftazidime and avibactam Cmin below these cut-offs (one out of one, and six out of nine, respectively).
Ceftazidime and avibactam share similar PK profiles: they have low protein binding (<10%), are poorly metabolized and are hydrophilic molecules, which reduces the risk of sequestration on ECMO membrane.4,9,10 Our results are in line with previous studies, which showed that cephalosporins PK is not altered in ECMO patients, their blood levels being above the MIC and predefined breakpoints, and is similar to patients without ECMO.5,9,10,11 Indeed, in a recently published study on antimicrobial PK in ECMO patients, whereas most antimicrobials did not reach the predefined target, cephalosporin (ceftriaxone and cefepime) blood levels seemed not affected by ECMO.4,12 Moreover, we also highlighted here that increased renal clearance seems to be an important factor that may influence serum concentration of beta-lactam in ECMO patients. On the other side, unlike recently published data, RRT was not associated with underdosing for ceftazidime and avibactam, perhaps because we used normal dosing in our patients requiring RRT.4,9 A pragmatic approach could be to systematically perform therapeutic drug monitoring in ECMO patients requiring antimicrobials, all the more in patients requiring RRT or in patients with increased eGFR.12–14 This strategy may help to avoid toxicity, mainly neurological in patients with too-high blood levels, and increase the dose in patients with low blood levels to avoid infection treatment failure.13,15–17
Limitations of our study include its monocentric retrospective design with small number of patients, the lack of a control group (patients without ECMO) and the lack of data on antibiotics tissue levels: indeed, blood measurement does not represent the tissue level. Additionally, we have only one assay, which prevents us from conducting more detailed PK measurements, a limitation that should be considered in future studies. Moreover, the small number of included patients renders formal conclusion on clinical outcomes hazardous. Another limitation is the interpretability of our results regarding the MIC of pathogens responsible for infection: indeed, the MIC is determined for the combination of drugs (ceftazidime/avibactam), whereas the two molecules (ceftazidime and avibactam) are measured separately, and being the target for one molecule does not necessarily predict the correct concentration for the other.18 Therefore, the relationship between MIC and Cmin is difficult to assess, and the time above MIC, which may be the best PK parameter, is impossible to give. One solution is to combine results of Cmin of both molecules to have a joint PK/PD target attainment, which has recently been showed to be associated with treatment success.8
In conclusion, ceftazidime and avibactam serum levels were above EUCAST breakpoints in most cases. Our results indicate that usual dosing could be used in ECMO patients, with attainment of the recommended target in most patients. Increased renal clearance seems to be the main factor that influence ceftazidime/avibactam serum concentration and that may lead to ceftazidime and avibactam under dosing. Our data suggest the importance to monitor ceftazidime and avibactam trough levels, particularly in patients with increased eGFR. Larger studies are warranted to confirm our results.
Funding
This study was conducted as part of our routine work.
Transparency declarations
C.-E.L. received lecture fees from Merck, Aerogen and AdvanzPharma, and grants from Eumedica and Merck, all outside the submitted work. C.-E.L. received travel grant from Pfizer, the manufacturer of ceftazidime-avibactam. The other authors have no conflicts of interest to declare in relationship to this manuscript.
