The effect of intraoperative excessive bleeding on the pharmacokinetics of ampicillin and sulbactam in recipients of living donor liver transplantation in Japan

Abstract

Objective

Guidelines recommend redosing with intravenous prophylactic antibiotics when excessive bleeding exceeds 1500 mL during surgery based on the pharmacokinetics data of cefazolin. However, the necessity for redosing of other antibiotics and the threshold volume of blood loss necessitating such supplementation remain undefined. We investigated plasma antibiotic concentrations during liver transplant surgery in patients with frequent excessive bleeding.

Methods

A single-centre, prospective, observational pharmacokinetic study was conducted. Adult liver transplant recipients who received 2 g of ampicillin and 1 g of sulbactam every 3 h during surgery were included. Blood samples were collected hourly during surgery, and intraoperative bleeding amounts were reviewed from anaesthesia records. Plasma concentrations of ampicillin and sulbactam were determined using validated liquid chromatography-tandem mass spectrometry. The probability of target attainment was set at 80% free time above the MIC (fT > MIC).

Results

Twenty participants were included. Of these, 11 participants (55%) were female. The median age, body weight, and bleeding volume were 52 years, 62.1 kg, and 11 158 mL, respectively. The intraoperative clearance of ampicillin was 80.28 mL/min, and sulbactam was 77.23 mL/min. The fT > MIC for both ampicillin and sulbactam tended to be lower with bleeding > 20 000 mL than with less bleeding. Plasma concentrations of ampicillin and sulbactam were maintained during surgery without redosing, even after bleeding exceeded 1500 mL.

Conclusions

Even with excessive bleeding, administering 3 g of ampicillin/sulbactam every 3 h maintained sufficient plasma concentration. Redosing may be unnecessary unless total bleeding exceeds 20 000 mL.

Introduction

Antibiotic prophylaxis is the cornerstone of surgical site infection (SSI) prevention.¹ However, the pharmacokinetic properties during surgery may differ from those in stable conditions due to factors such as fluid administration, bleeding, and anaesthesia.^2,3 Guidelines jointly developed by the American Society of Health-System Pharmacists (ASHP), the Infectious Diseases Society of America (IDSA), the Surgical Infection Society (SIS), and the Society for Healthcare Epidemiology of America (SHEA) recommend shorter redosing intervals during surgery, specifically advocating redosing at intervals exceeding two half-lives of the antibiotics,¹ and after excessive bleeding (>1500 mL). The guidelines cite an uncontrolled open-label study that evaluated the serum and tissue concentrations of cefazolin in 11 patients during elective spinal surgery as the basis for this recommendation.⁴ Although animal and human pharmacokinetic model studies have suggested that serum and tissue antibiotic concentrations decrease during haemorrhage,⁵ clinical data supporting this threshold are scarce. Therefore, the impact of intraoperative bleeding on plasma antibiotic concentrations and whether a threshold of 1500 mL applies to other antibiotics remains unclear in real-world settings. To fill this gap, we aimed to characterize the intraoperative pharmacokinetics (PK) of ampicillin and sulbactam during orthotopic liver transplantation, which often involves excessive bleeding exceeding one total body blood volume and requires massive transfusion.^6,7

Materials and methods

Study design and setting

This single-centre, prospective observational pharmacokinetic study was conducted between 8 August 2022 and 20 February 2023 at the University of Tokyo Hospital, Japan. Candidates, who were 20 years of age or older, for scheduled living-donor liver transplantation (LDLT), including simultaneous liver and kidney transplantations, were approached and enrolled. We excluded participants who (i) had received ampicillin/sulbactam within 72 h before liver transplantation and/or (ii) were allergic to any beta-lactams.

Demographic data and intraoperative fluid count collection

Patient demographics, including age, sex, and body weight, were obtained from the medical records. The amounts of intraoperative bleeding and fluid infusion were obtained from the anaesthesia records. No intraoperative cell salvage system was used during the study period.

Antibiotic administration and blood sample collection

Two grams of ampicillin and one gram of sulbactam in 50 mL of 0.9% sodium chloride or 5% glucose were intravenously administered using a syringe pump over 30 min immediately before skin incision and every 3 h thereafter throughout the surgery. No additional dose of ampicillin/sulbactam was administered, regardless of the amount of bleeding, according to the institutional protocol. After the first administration of ampicillin/sulbactam, 2 mL of blood was drawn from the radial arterial catheter every hour throughout the surgery. All blood samples were refrigerated at 4°C and centrifuged at 3000 g for 15 min within 24 h to separate the plasma.

Measurement of ampicillin and sulbactam concentration and pharmacokinetic/pharmacodynamic analysis

The total plasma concentrations of ampicillin and sulbactam were determined using validated liquid chromatography-tandem mass spectrometry. The lower limit of detection was 0.5 mg/L for both ampicillin and sulbactam. Non-compartmental analysis (NCA) was used to estimate the pharmacokinetic parameters of ampicillin and sulbactam. The AUC was calculated using the linear trapezoidal method based on the plasma concentrations. The clearances were calculated based on the total dose of each antibiotic and the AUC. The volume of distribution (Vd) was calculated by dividing the dose of each antibiotic by the plasma concentration of each antibiotic 1 h after the first administration, and the body-weight-normalized Vd was determined by dividing each participant’s individual Vd by their body weight and then averaging those values. The probability of target attainment (PTA) was defined as the estimated duration that the free concentration remains above the MIC (fT > MIC). The free (unbound) plasma concentrations were calculated by multiplying the total concentrations by previously reported free fraction (0.28 for ampicillin and 0.38 for sulbactam).⁸ The PTA was set at the probability of fT > MIC of 32/16 mg/L for ampicillin/sulbactam based on the susceptibility breakpoints proposed by the CLSI, where a MIC of 8/4 mg/L or lower was considered susceptible and 32/16 mg/L or higher resistant for the common SSI pathogens. The target of fT > MIC of 32/16 was set at 80%.^9,10

Statistical analysis

Median and interquartile range (IQR) were calculated to describe the distribution of results. Pearson’s correlation coefficient was used to assess correlations between quantitative variables. Fisher’s exact test was employed to evaluate differences in proportions of the participants with fT > MIC above 0.8, depending on intraoperative blood loss exceeding 20 000 mL or not. The level of statistical significance was set at 0.05. R version 4.3.0, along with the ggplot2 package, was used for analysis.

Patient consent statement

This study was approved by the Institutional Review Board of the University of Tokyo Hospital (approval number 2022090NI). All the participants provided written informed consent.

Results

Of the 22 LDLT candidates who consented to this study, one did not undergo LDLT because of COVID-19 in a donor, and another had been administered antibiotic prophylaxis before transplantation. The remaining 20 participants who underwent LDLT were included in the analysis. Table 1 summarizes the demographic characteristics of the participants. The median age was 52.0 years, and 11 (55%) were female. The median body weight and body mass index were 62.1 kg and 22.5, respectively. Two participants underwent renal replacement therapy before transplantation for hepatorenal syndrome. Excluding these two participants, the median calculated creatinine clearance was 99.57 mL/min. Participants had a median of 11.2 L (IQR, 6.0–14.1) of bleeding with the median bleeding rate 1105 mL/h (IQR, 746.7–1664.4).

Table 1.

Basic characteristics of the 20 recipients of living-donor liver transplantation

	N (%) or median (interquartile range)
Age, years	52.0 (44.8–56.0)
Female sex	11 (55%)
Height, cm	165.4 (160.0–170.2)
Body weight, kg	62.1 (53.5–74.5)
Body mass index, kg/m2	22.5 (19.8–25.2)
Underlying liver diseases
Alcoholic liver disease	8 (40%)
Non-alcoholic steatohepatitis	4 (20%)
Autoimmune hepatitis	2 (10%)
Primary sclerosing cholangitis	2 (10%)
Primary biliary cholangitis	1 (5%)
Congenital biliary atresia	1 (5%)
Idiopathic portal hypertension	1 (5%)
Multiple causes	1 (5%)a
MELD score	18.5 (14.5–23.5)
The ratio of graft weight to recipient standard liver volume, %	46.0 (41.3–50.0)
Renal replacement therapy before transplantation	2 (10%)
Serum creatinine concentration, mg/dL	0.72 (0.56–1.03)
Calculated creatinine clearance (Cockcroft–Gault equation), mL/min	99.57 (88.6–107.3)b
Serum albumin concentration, g/dL	2.90 (2.65–3.15)
Total operation time, h	9.6 (8.5–10.2)
Total bleeding, mL	11 158 (6032–14 125)
Total intraoperative infusion and transfusion, mL	16 900 (11 248–21 675)
Crystalloid, mL	3175 (2238–3900)
Colloid, mL	2375 (1500–3938)
Red blood cell, mL	4320 (2340–5640)
Fresh frozen plasma, mL	6660 (3600–8400)
Platelet component, mL	800 (400–1000)

^aChronic hepatitis B and D, and autoimmune hepatitis.

^bRecipients who underwent renal replacement therapy were excluded.

From the 20 participants, 224 blood samples were collected. Figures 1 and 2 show the plasma concentrations of antibiotics and the cumulative amount of bleeding during surgery for each participant. Preoperative creatinine clearance and intraoperative clearance of ampicillin and sulbactam are shown in Supplementary Table (available as Supplementary data at JAC-AMR Online). Participants whose bleeding exceeded 20 000 mL or whose mean bleeding speed exceeded 2400 mL/h (participants 005, 009, 013, and 018) showed lower plasma concentrations. Most participants achieved a PTA > 0.8; however, participants with bleeding > 20 000 mL tended to have a lower PTA. The calculated median clearances (CL) of ampicillin and sulbactam were 4.82 L/h (IQR, 3.60–6.09) and 4.63 L/h (IQR, 3.20–6.35), respectively. Pearson correlation coefficients between preoperative creatinine clearance and intraoperative CL were 0.74 for ampicillin and 0.71 for sulbactam when the two recipients who received renal replacement therapy were excluded. Additionally, the calculated median Vd was 12.57 L (IQR, 9.88–19.74) for ampicillin and 12.16 L (IQR, 10.16–17.47) for sulbactam. The median of the body-weight-normalized Vd was 0.20 L/kg (IQR, 0.16–0.28) for ampicillin and 0.20 L/kg (IQR, 0.17–0.28) for sulbactam, respectively. Assuming protein binding rates of 28% for ampicillin and 38% for sulbactam,⁸ 14 (70%) and 11 (55%) participants maintained 100% PTA for plasma concentration of ampicillin and sulbactam during surgery, respectively. According to Figures 1 and 2, among the four participants whose intraoperative blood loss exceeded 20 000 mL, two (50%) for ampicillin and three (75%) for sulbactam exhibited a PTA below 0.8. Participants with intraoperative bleeding greater than 20 000 mL were significantly more likely to achieve a PTA of less than 0.8 (P = 0.03).

Figure 1.

Individual plasma concentrations of ampicillin during surgery. Blue triangles indicate the start of infusion of ampicillin–sulbactam every 3 h over 30 min. Orange rectangles indicate the MIC of Enterobacterales divided by the unbound fraction of ampicillin (32 mg/L/0.72). *Patients with maintenance dialysis preoperatively. ABPC, ampicillin; ft > MIC, free time above the MIC.

Figure 2.

Individual plasma concentrations of sulbactam during surgery. Blue triangles indicate the start of infusion of ampicillin–sulbactam every 3 h over 30 minutes. Orange rectangles indicate the MIC of Enterobacterales divided by the unbound fraction of sulbactam (16 mg/L/0.62). *Patients with maintenance dialysis preoperatively. SBT, sulbactam; ft > MIC, free time above the MIC.

Discussion

In this study, we found that the plasma concentrations of ampicillin and sulbactam among 20 participants who received LDLT remained above the MIC of common SSI pathogens despite a median of 11.2 L of bleeding during transplantation when 2 g of ampicillin and 1 g of sulbactam were administered every 3 h without additional redosing after 1500 mL of bleeding. The findings suggest that no additional redosing of ampicillin and sulbactam was needed after excessive bleeding until bleeding reached 20 000 mL, at which point increasing the clearance of both ampicillin and sulbactam increased.

Lasko et al. conducted Monte Carlo simulations using data from ten orthotopic liver transplant recipients who received ampicillin/sulbactam as surgical prophylaxis.¹¹ The study participants received 1–2 g/0.5–1 g of ampicillin/sulbactam every 2–3 h at the discretion of anaesthesiologists, mostly administered as an intravenous push over1 min. They found that blood product resuscitation had no impact on exposure to ampicillin, despite the administration of a median of 10.7 L of blood products. Although Lasko et al. did not report the amount of bleeding during the surgery, the findings of the present study and the report by Lasko et al. raise questions about the necessity of an additional dose of antibiotic agents following excessive bleeding, at least concerning ampicillin.

The current recommendation for redosing after bleeding exceeding 1500 mL was based on a PK study of cefazolin, where the tissue concentration of cefazolin was correlated with bleeding exceeding 1500 mL in elective surgical procedures involving spinal instrumentation.⁴ Compared with cefazolin, both ampicillin and sulbactam have a larger Vd. In our study, the median of the body-weight-normalized Vd was 0.20 L/kg for ampicillin and 0.20 L/kg for sulbactam. These are in line with previously reported values.⁹ In addition, this study did not involve extremely obese or lean patients. In contrast, the Vd of cefazolin was 0.14 L/kg during various surgeries.¹² Other previous studies indicated that the central Vd of cefazolin, not normalized with body-weight, was approximately 5 L^13–15. Taken together, our results may imply that ampicillin and sulbactam are even more likely to be eliminated via bleeding, as the smaller Vd value indicates that the drug has a higher proportion in the plasma.

In our study, 3 g of ampicillin/sulbactam every 3 h without an additional dose, regardless of the amount of bleeding, maintained the plasma concentration above the MIC without the accumulation of antibiotics if bleeding was < 20 000 mL. Even though ASHP guidelines recommend redosing ampicillin/sulbactam every 2 h, the present study indicated that a 3-h interval resulted in adequate plasma concentrations during LDLT when the amount of bleeding did not exceed 20 000 mL. This observation suggests a 3-h interval may be appropriate for maintaining optimal plasma concentrations of ampicillin and sulbactam in most patients. More data are required to define the appropriate cut-off value for shorter intervals or redosing.

This study has several limitations. First, given the nature of liver transplant surgery, we did not have a control group without excessive bleeding to thoroughly assess the impact of blood loss. Second, we used NCA to estimate PK parameters, making it difficult to simulate the optimal intraoperative dosage for patients receiving other types of surgeries. Future clinical studies that recruit patients with different scenarios will be needed. Third, we did not measure the local concentration of antibiotics in ascites or bile, nor consider clinical outcomes, such as the incidence of SSI, success rate of transplantation, and mortality.

In summary, even with excessive bleeding, administering 2 g of ampicillin and 1 g of sulbactam every 3 h maintained a sufficient PTA, suggesting that an additional dose of ampicillin/sulbactam after excessive bleeding might be unnecessary.

Funding

This work was supported in part by Japan Society for the Promotion of Science (grant Number JP 23K16252 (KO)).

Transparency declarations

K.O. received lecture and medical advice honoraria from Thermo Fisher Scientific, Becton, Dickinson and Company, Eiken Chemical, CO., and Kyorin Pharmaceutical, Co. H.H. received speaker honoraria from MSD and Shionogi. U.K. is in collaboration with Nihon Kohden Corporation and Nipro Corporation. Neither of these collaborations has any relationship with the present study. All other authors declare no conflict of interest.

Author contributions

W.Y., K.O., and T.Y. were involved in conceiving the study, medical data and blood sample collection, analysis, discussion of results, and writing the manuscript. T.Y. and R.Y. contributed to the measurement of plasma concentration of antibiotics and the design of pharmacokinetic analysis. T.K. and Y.H. were involved in blood sample collection and anaesthesia. N.A. and J.K. contributed to obtaining informed consent and performance of surgery. N.A., T.K., Y.H., S.H., H.H., J.D., T.T., R.Y., J.K., S.O., T.T., K.H., K.U., and T.T. were involved in manuscript preparation.

Informed consent statement

Written informed consent has been obtained from the patient to publish this paper.

Supplementary data

Supplementary Table is available as Supplementary data at JAC-AMR Online.