Secondary Logo

Journal Logo

Featured Articles: Original Clinical Research Report

Prophylactic Cefazolin Dosing in Women With Body Mass Index >35 kg·m−2 Undergoing Cesarean Delivery: A Pharmacokinetic Study of Plasma and Interstitial Fluid

Eley, Victoria A. PhD*,†; Christensen, Rebecca PhD*,†; Ryan, Rochelle MBBS*,†; Jackson, Dwane MBBS*,†; Parker, Suzanne L. PhD; Smith, Matthew MBBS§; van Zundert, Andre A. PhD*,†; Wallis, Steven C. PhD; Lipman, Jeffrey MD†,‖; Roberts, Jason A. PhD‡,¶,#

Author Information
doi: 10.1213/ANE.0000000000004766


See Article, p 196


  • Question: Are cefazolin prophylaxis strategies in obese pregnant women adequate?
  • Finding: Dosing simulations suggest redosing of cefazolin at 2 hours.
  • Meaning: Current antibiotic prophylaxis recommendations may not be adequate for all obese pregnant women.

Maternal obesity is increasing,1 and cesarean delivery (CD) rates in Western countries are considered high.1,2 Women who are obese are more likely to be delivered by CD than nonobese women3 and have an increased risk of surgical site infection (SSI),4–6 which is thought to complicate between 2.8% and 10% of CDs.4,7,8 Prophylactic antibiotics reduce the incidence of wound infections and endometritis following CD by over 60%.9

Effective surgical antibiotic prophylaxis requires maintenance of antibiotic concentrations above the minimum inhibitory concentration (MIC) of potential pathogens. This must be achieved not just in the blood, but in the tissues susceptible to infection and for the duration of surgery.10 Physiological changes that occur in an obese pregnant patient, such as increased volume of total body water and increased renal clearance, may lead to subtherapeutic concentrations of antibiotic in the interstitial fluid (ISF) of adipose tissue, the susceptible tissue in SSI.11

Until April 2019, the Australian Therapeutic Guidelines recommended 2 g of cefazolin be administered within 60 minutes of skin incision for CD.12 The recently updated guidelines now suggest increasing the dose to 3 g in women of >120 kg weight. There is limited evidence to support this dose adjustment, which is also suggested by US-based organizations.10 A Statement from the Royal Australian and New Zealand College of Obstetricians and Gynaecologists (due for revision in 2019) currently recommends 1 g of cefazolin be administered before skin incision, increasing to 2 g in women weighing >100 kg.13 The British National Formulary (United Kingdom) states that women should be offered cefuroxime before skin incision for CD, with no weight adjustment recommended.14 It is apparent that clinical practice guidelines across the world remain in a state of flux,15 indicating a paucity of evidence available for application in this specific population.

To date, pharmacokinetic studies of cefazolin undertaken in obese pregnant women have measured drug concentrations in plasma and homogenized adipose tissue samples, with inconsistent methodology and results.16–21 Microdialysis for measuring drug concentrations in the ISF of tissues is now considered the gold standard approach for this purpose.22

We aimed to describe the plasma and ISF pharmacokinetics of intravenous cefazolin in women with a body mass index (BMI) >35 kg·m−2 undergoing elective CD at term, using ISF microdialysis catheters. We then sought to develop a population pharmacokinetic model and perform Monte Carlo dosing simulations to determine optimized dosing regimens for cefazolin prophylaxis in obese pregnant women.


This manuscript adheres to Strengthening the Reporting of Observational studies in Epidemiology (STROBE) guidelines. Ethical approval was obtained from the Human Research Ethics Committee of the Royal Brisbane and Women’s Hospital (HREC/15/QRBW/549) and written informed consent obtained from all participants. This open-labeled pharmacokinetic study was undertaken at the Royal Brisbane and Women’s Hospital, a tertiary referral hospital with 4500 annual deliveries. Data were collected between July 2016 and February 2017. Women were included if they had a BMI >35 kg·m−2 at delivery, age ≥18 years, were booked for an elective CD at a gestation of ≥37 weeks, and commencing under neuraxial anesthesia. Women were excluded if they had elevated serum creatinine (>70 µmol·L−1),23 allergy to cefazolin, or had received cefazolin in the preceding 72 hours.

An intravenous cannula was inserted to facilitate blood sampling. To allow ISF sampling, a microdialysis catheter (CMA 60; Microdialysis AB, Stockholm, Sweden) was inserted in the upper abdominal subcutaneous tissue after the establishment of neuraxial anesthesia, as close to the site of incision as possible, but outside of the sterile field. Cefazolin was prepared as a single 2 g intravenous bolus dose in 20 mL of sterile 0.9% saline and was administered by a slow push after the establishment of neuraxial anesthesia, within 30 minutes of skin incision. At the time our study was conducted, this was the practice of our institution and consistent with the recommendation of the Australian Therapeutic Guidelines.24

Blood samples were taken predose and then at 10-, 30-, 60-, 90-, and 180-minute time points postdosing. Blood samples were immediately stored on ice. Within 6 hours of collection, blood samples were centrifuged at 3000 rpm for 10 minutes, and the resulting plasma was stored at −80°C until analysis. Microdialysate samples were collected at 15, 30, 45, 60 minutes postdose and then every 30 minutes until the completion of surgery. Microdialysate and perfusate samples were immediately kept on ice and then stored at −80°C. The recovery of cefazolin in the microdialysate solution was interpolated from the loss of internal standard (cephalothin) across the microdialysis membrane into ISF.25 This was performed in each patient by measuring concentrations of cefalothin in both the perfusate and microdialysate and calculating the efficiency of the microdialysis catheter as the percentage loss of cefalothin from the perfusate. A mean efficiency of 26% was applied to the cefazolin concentrations.

Total and unbound concentrations of cefazolin in plasma were measured from 1 to 500 mg·L−1 by a validated ultrahigh performance liquid chromatography-tandem mass spectrometry (UHPLC-MS/MS) method on a Nexera2 UHPLC system coupled to a 8030+ triple quadrupole mass spectrometer (Shimadzu, Kyoto, Japan).26 For unbound concentrations, the free fraction was isolated from plasma by ultrafiltration at 37°C using Centrifree devices (Merck, Darmstadt, Germany). Concentrations of cefazolin (0.1–20 mg·L−1) and cefalothin (1–100 mg·L−1) in microdialysis and perfusate were measured using the same technique. Clinical samples were assayed in batches alongside calibrators and quality controls and results were subject to batch acceptance criteria.27

Patient demographic information was collected, including maternal age, gestation at delivery, height, and weight (prepregnancy and at delivery). The prepregnancy BMI and BMI at delivery were calculated. Clinical details documented included surgical duration, volume of intravenous fluids administered, serum creatinine, albumin, liver function tests, hematocrit, pre- and postoperative hemoglobin concentrations, and estimated blood loss.

Statistical Analysis

There was no traditional sample size calculation for this observational pharmacokinetic study. Summary data are presented as median (interquartile range [IQR]) or mean (standard deviation [SD]). Statistical analyses were performed using Graphpad Prism, version 7 (Graphpad, San Diego, CA) and Microsoft Excel (Microsoft Office, Microsoft, Redmond, WA).

Pharmacokinetic Model Development

Pharmacokinetic modeling was performed using Pmetrics 1.5.0 with RStudio 0.99.902 and digital compiler Gfortran 5.2 (The Fortran Company, Chandler, AZ). For the population pharmacokinetic analysis, 3- and 4-compartment models were fitted using plasma (total and unbound) and microdialysis cefazolin concentration data, using nonparametric adaptive grid subroutines from the Pmetrics package for R (Laboratory of Applied Pharmacokinetics, Los Angeles CA). Primary pharmacokinetic parameters of clearance (CL), volume of distribution for the central (unbound, Vc), and ISF microdialysis compartment (Vm) were calculated (see Supplemental Digital Content, Figure 1,, for schematic of model). Elimination from the central compartment and intercompartmental distribution were modeled as first-order processes. Unbound cefazolin concentrations were related to total cefazolin concentrations assuming albumin is the sole binding site for cefazolin in plasma and taking into account protein binding,28 as follows:

In these equations, Cunbound is the unbound cefazolin concentration in mg·L−1, Ctotal is the total cefazolin concentration in mg·L−1, Bmax is the maximum binding concentration of cefazolin in mg·L−1, KD is equilibrium dissociation constant in mg·L−1 for cefazolin binding to albumin, ALB is the serum albumin concentration in g·L−1, N is the number of binding sites, with 0.6 binding sites for cefazolin per molecule of albumin, MWcef is the molecular weight of cefazolin (454.51 g·mol−1), MWalb is the molecular weight of albumin (66,500 g·mol−1), Koff is the first-order dissociation rate constant in h–1, and Kon is the second-order association rate constant in l·mg−1·h−1. Cunbound and Ctotal were measured independently and modeled with independent error functions such that any concentration- or time-related changes in protein binding could be described over the dosing interval. Intercompartmental distribution was described as rate of transfer from the unbound compartment to a tissue compartment, Kct; rate of transfer from a tissue compartment to the unbound compartment, Ktc; rate of transfer from the unbound compartment to a peripheral compartment, Kcp; rate of transfer from a peripheral compartment to the unbound compartment, Kpc; and rate of elimination from the unbound compartment, Ke.

We selected the final model on the basis of minimizing the likelihood of the nested models (−2 × log-likelihood [−2 × LL]) and the Akaike Information Criterion (AIC); the latter includes a penalization according to the number of parameters in the model. We used a reduction of AIC for statistical comparison of nested models. We also factored bias (mean-weighted predicted-observed error) and imprecision (bias-adjusted, mean-weighted squared predicted-observed error) into the selection of the final model. Shrinkage was calculated as the percentage of total variance in each model probability distribution due to the variance in the posterior parameter probability distribution.29

Covariate Screening

Covariate model building was performed using sequential assessment of biologically plausible clinical characteristics. Continuous data were used for all covariates tested, with the exception of sex, which was male/female. The association of the covariates versus parameters was assessed by regression with linear or nonlinear associations able to be used in the covariate model testing depending on which was more statistically significant. Inclusion was based on a statistically significant improvement in the AIC and −2 × LL value. The covariates evaluated against pharmacokinetic parameters were maternal age, fetal gestational age at delivery, height, prepregnancy weight and BMI, preoperative weight and BMI, sex, difference in weight and BMI from prepregnancy to delivery, prepregnancy, preoperative and change in hemoglobin, hematocrit, serum platelet count, γ-glutamyl transferase, aspartate transaminase, alanine transaminase, length of surgery, estimated blood loss during surgery, and serum concentrations of lactate dehydrogenase, albumin, bilirubin, urea, and creatinine.

Probability of Target Attainment

Monte Carlo dosing simulations (n = 1000) were performed using the final covariate model in Pmetrics to determine the probability of target attainment (PTA) using the pharmacokinetic-pharmacodynamic target of unbound drug concentrations remaining above the MIC for 100% of the dosing interval time (100% fT>MIC) for clinically relevant MICs (0.125–32 mg·L−1) during various loading and maintenance doses during CD surgery of up to 3 hours for an obese patient with a weight at delivery of 119 kg, based on the mean weight of our sample.

Fractional Target Attainment

MIC data for Staphylococcus aureus from the European Committee on Antimicrobial Susceptibility Testing (EUCAST) database30 were used to determine the fractional target attainment (FTA). The FTA identifies the likely success of treatment using cefazolin by comparing the pharmacodynamic exposure (PTA) against a MIC distribution up to the epidemiological cutoff (from 0.125 to 2 mg·L−1). In this analysis, the FTA was calculated using various doses to achieve a target that was pre-established before analysis of 95% unbound concentrations >2 mg·L−1 in the ISF throughout 3 hours of surgery. Five levels of patient weights (at delivery) were tested (90, 110, 119, 130, and 150 kg) reflecting the distribution of values in our study population.


Twelve women were included. Participant characteristics are reported in Table 1. They had a mean (SD) age of 32.8 (4.8) years and a median (IQR) BMI at delivery of 41.5 (39.7–46.6) kg·m−2. Five women had a body weight at delivery that was >120 kg. The mean (SD) duration of surgery was 75.8 (21.0) minutes, range 54–125 minutes. Eleven women received 2 g of cefazolin according to hospital recommendation and 1 received 3 g according to treating anesthetist preference. This woman also received a general anesthetic after failure of neuraxial anesthesia. Results from this participant were consistent with those of the rest of the group and therefore were included in the analysis.

Table 1. - Characteristics of 12 Women Delivering by Elective Cesarean Delivery at Term, With BMI >35 kg·m−2
Characteristics Mean SD Range
Prepregnancy weight (kg) 106.9 20.8 82.0–139.0
Prepregnancy BMI (kg·m−2) median (IQR) 36.8 33.3–42.2 31.0–51.1
Weight at delivery (kg) 119.1 18.8 93.7–150.0
BMI at delivery (kg·m−2) median (IQR) 41.5 39.7–46.6 35.1–55.0
Difference in weight (kg)a 12.2 7.4 −2.6 to 21.4
Difference in BMI (kg·m−2)a 4.2 2.8 −1.0 to 8.3
Height (cm) 166.3 4.6 161.0–175.0
Gestational at delivery (median, IQR) 38.7 37.9–39.0 37.0–39.4
Preoperative hemoglobin (g·L−1) 123.0 12.0 103.0–142.0
Postoperative hemoglobin (g·L−1), n = 9 112.0 15.0 91.0–144.0
Difference in hemoglobin (g·L−1)b, n = 9 12.0 9 −20.0 to 26.0
Hematocrit (%) 40.0 3.1 32.0–43.0
White cell count (×109·L−1) 10.6 2.8 6.8–15.7
Serum platelet count (×109·L−1) 244 65 128–335
Sodium (mmol·L−1) 136 1.7 134–139
Chloride (mmol·L−1) 84.2 42.2 4.2–108
Urea (mmol·L−1) 23.8 64.7 2.5–208
Serum creatinine (µmol·L−1) 53.5 11.2 36–70
Serum albumin (g·L−1) 27.3 4.9 21–38
Total bilirubin (µmol·L−1) 8.0 1.2 6–10
γ-glutamyl transferase (U·L−1) 15.7 7.5 9–30
Alanine transaminase (U·L−1) 16.5 9.6 7–37
Aspartate transaminase (U·L−1) 19.2 9.1 11–41
Lactic acid dehydrogenase (U·L−1) 223 39.1 179–297
Estimated blood loss during surgery (L) 0.63 0.33 0.3–1.3
Blood loss replacement (crystalloid) (L) 1.28 0.57 0.6–2.5
Data are presented as mean and SD except where noted as median, IQR.
Abbreviations: BMI, body mass index; IQR, interquartile range; SD, standard deviation.
aDifference is between the prepregnancy value and the value at delivery.
bDifference is between preoperative and postoperative values.

All total and unbound plasma concentrations of patients were above the lower limit of quantification of 1 mg·L−1 with all microdialysate concentrations above 0.1 mg·L−1. The plasma and microdialysis assay methods were validated for precision and accuracy using the Food and Drug Administration (FDA) criteria for bioanalysis.27 The clinical samples were assayed for cefazolin in 4 separate batches for total and unbound concentrations of plasma and for cefazolin and cefalothin in 5 separate batches for microdialysis and perfusate concentrations. Precision and accuracy data were calculated from quality control samples assayed within these batches with the results as follows. Precision was 5.3%, 4.7%, and 4.1% and accuracy 0.0%, 2.5%, and −2.7% at total plasma cefazolin concentrations of 20, 100, and 400 mg·L−1, respectively. Unbound plasma cefazolin concentration precision was 5.4%, 3.6%, and 6.3%, and fraction of cefazolin unbound was 15.0%, 19.1%, and 54.6% at total concentrations of 20, 100, and 400 mg·L−1. For the microdialysis assay, precision was 5.2%, 1.8%, and 0.8% and accuracy was 9.1%, −0.4%, and −0.5% at cefazolin concentrations of 0.3, 3, and 10 mg·L−1, respectively. A description of the bioanalytical validation has been published.26

Table 2. - Pharmacokinetic Model Data Based on Plasma (Total and Unbound) and Interstitial Fluid Cefazolin Concentrations Obtained From 12 Women With Body Mass Index >35 kg·m−2 of Obese Patients Undergoing Elective Cesarean Delivery at Term
Pharmacokinetic Parameter Units Mean SD Range 95% Confidence Interval CV% Median Shrink%
Minimum Maximum Lower Upper
CL L·h−1 25.0 7.0 15.1 38.4 19.0 29.7 27.8 26.1 0.153
Vc L 10.5 2.1 6.53 12.9 8.8 12.6 19.9 10.8 0.045
Vm L 25.0 12.1 7.15 49.8 14.7 33.2 48.4 23.6 0.055
Kcp h−1 3.1 1.8 0.663 6.6 1.9 3.9 58.8 2.56 0.031
Kpc h−1 3.6 2.9 0.145 9.0 1.7 5.5 79.5 2.40 0.107
Kct h−1 12.1 10.7 1.51 34.8 3.2 20.3 88.5 6.19 0.130
Ktc h−1 10.1 7.7 0.548 17.9 1.1 17.3 76.2 14.8 0.171
K on l·mg–1· h−1 49.2 18.9 20.9 74.9 33.7 68.7 38.5 45.5 0.002
K off h−1 1413.0 275.0 842.0 1910.0 1224.0 1612.0 19.4 1430 0.006
Abbreviations: CL, clearance of unbound cefazolin; CV, coefficient of variation; SD, standard deviation; Shrink, model shrinkage; Vc, central volume of distribution of cefazolin; Vm, volume of distribution of cefazolin in the tissue compartment; Koff, rate of dissociation binding from albumin (first-order); Kon, rate of association binding to albumin (second-order); Kcp, rate of transfer from the unbound compartment to a peripheral compartment; Kct, rate of transfer from the unbound compartment to a tissue compartment; Kpc, rate of transfer from a peripheral compartment to the unbound compartment; Ktc, rate of transfer from a tissue compartment to the unbound compartment.

Median concentration-time pharmacokinetic profile for cefazolin for total plasma, unbound plasma, and interstitial fluid from 12 women with body mass index >35 kg·m−2 undergoing elective cesarean delivery at term. Error bars are standard deviation. Total plasma concentration (circle), unbound plasma concentration (diamond), interstitial fluid concentration (triangle).

The minimum total plasma concentration of cefazolin was 31.7 mg·L−1 and the maximum was 164.1 mg·L−1, observed in 2 different participants. The minimum unbound plasma concentration was 7.7 mg·L−1 and the maximum was 73.1 mg·L−1, observed in 2 different participants. For each timepoint up to 150 minutes, the median across subjects of unbound ISF concentrations remained above 2 mg·L−1. The minimum unbound ISF concentration was 0.7 mg·L−1, observed in 1 participant. In 2 participants, the ISF concentration of cefazolin was not maintained above 2 mg·L−1 for the duration of sampling. The mean (± standard error [SE]) penetration of drug (calculated as area under the concentration-time curve for the unbound fraction of drug [fAUC]tissue/fAUCplasma) into the ISF was 0.884 ± 1.11. Cefazolin concentration-time profiles for unbound plasma, total plasma, and microdialysate are shown in the Figure. The pharmacokinetic data are summarized in Table 2.

Pharmacokinetic Model

A 3-compartment model was described empirically based on the data from plasma (total and unbound) and microdialysis. Inclusion of albumin was also empiric to describe the dissociation/association rates for protein binding. The inclusion of a fourth compartment was accepted on the basis of a decrease in AIC of 58.6 and LL of 63.1. The inclusion of preoperative weight as a covariate on tissue volume was accepted on the basis of a decrease in AIC of 1.0 and LL of 1.0, as well as improvements to the slope and bias on the observed versus predicted plots. A final 4-compartment model, incorporating an additive error, adequately described the total and unbound plasma as well as microdialysate cefazolin concentrations (Supplemental Digital Content, Figure 1, Plasma albumin concentrations were included empirically to describe the protein binding relationship. Delivery weight (not prepregnancy) was supported for inclusion as a covariate in the model on ISF volume of distribution (Vm; r2 = 0.26) as it improved the likelihood of the model (as −2 × LL and AIC), as well as the bias and imprecision of the observed-predicted plot (Supplemental Digital Content, Figure 2, The diagnostic plots to confirm the goodness-of-fit of the model were acceptable (Supplemental Digital Content, Figure 2,

Dosing Simulations

Table 3. - Fractional Target Attainment for 6 Cefazolin Dosing Regimens for 5 Different Patient Weights (at Delivery)
Cefazolin Dose and Frequency
Weight (kg) 2 ga 2 g, 2 g at 2 hb 3 gc 3 g, 1 g at 3 hd 3 g, 2 g at 2 he 3 g, 3 g at 2 hf
90 81.4 99.4 86.2 96.6 99.5 99.8
110 78.9 98.9 83.8 95.1 99.2 99.6
119 78.0 98.6 82.8 94.3 99.0 99.5
130 76.6 98.2 81.7 93.4 98.8 99.4
150 74.4 97.7 80.0 91.9 98.2 99.1
Calculated based on the minimum inhibitory concentration wild-type European Committee on Antimicrobial Susceptibility Testing distribution of Staphylococcus aureus31 for a target trough unbound concentration of 2 mg·L−1 throughout 3 h of surgery.
a2 g, 0–30 min preincision.
b2 g, 0–30 min preincision followed by readministration of 2 g at 2 h.
c3 g, 0–30 min preincision.
d3 g, 0–30 min preincision followed by readministration of 1 g at 3 h.
e3 g, 0–30 min preincision followed by readministration of 2 g at 2 h.
f3 g 0–30 min preincision followed by readministration of 3 g at 2 h.

Monte Carlo simulations (n = 1000) were performed using a 3-hour dosing period. From these, estimates of PTA and FTA were prepared. Supplemental Digital Content, Figure 3,, shows the PTA of unbound cefazolin concentrations in plasma (A) and ISF (B) for 5 different dosing regimens in an obese woman with a body weight of 119 kg undergoing elective CD. The FTA of 6 different cefazolin dosing regimens for a MIC distribution for S. aureus is presented in Table 3. These were repeated for 5 different patient weights: 90, 110, 119, 130, and 150 kg. Three regimens achieved >95% FTA in all simulated patient weights, all required redosing at 2 hours after the initial dose.


Our results suggest that redosing of cefazolin after 2 hours may be required to be certain of achieving target concentrations of >2 mg·L−1 in the ISF of subcutaneous tissue in women weighing between 90 and 150 kg where the CD surgical duration is 3 hours. These simulation-based preliminary results require validation in further studies.

At the time our study was designed, the Australian Therapeutic Guidelines recommended administration of a 2 g cefazolin bolus within 60 minutes of skin incision for CD, with redosing every 2 half-lives (or approximately 4 hours).12 Our simulations demonstrate that this would result in an FTA of 81.4% for any patient in excess of 90 kg body weight. A single 3 g dose would result in an FTA of 86.2% if there is no redosing. Our study demonstrates that administration of an increased bolus dose and more frequent redosing of cefazolin will result in improved FTA of >95% of isolates for S. aureus in patients undergoing CD across a range of body weights, if the surgery lasts up to 3 hours after the initial dose. Thus, even the recently updated Therapeutic Guidelines appear inadequate for obese pregnant women, recommending an initial bolus dose increase to 3 g for patients weighing >120 kg, with no reduced redosing interval.24 Although significant variation in the tissue penetration of cefazolin has been reported,32,33 the mean penetration of drug into the ISF (0.88) in our population was similar to that reported in other nonobese, nonpregnant populations.33

The maximum operating time in our 12 patients was 125 minutes. Our dosing simulations were undertaken assuming a 3-hour dosing period from the initial bolus dose. Unpublished data from our institution demonstrated 12.7% of obese women undergoing CD (n = 63), with a median (IQR) BMI of 44.0 (41.0–48.0) kg·m−2 had a surgical duration >120 minutes.31 The longest surgical duration in a complicated operation was 388 minutes. An Australian study published in 2017 also showed that operative time for CD increases as maternal BMI increases.34 In our study, the first dose occurred between 0 and 30 minutes before incision. To maintain appropriate antibiotic concentrations in the ISF between the first dose and wound closure, our results suggest redosing with 2 g of cefazolin at 2 hours after the first dose, if wound closure is not complete. If the initial dose was 2 g, redosing with 2 g will provide an FTA of >95% across all weights up to 150 kg. Redosing of 3 g will provide an FTA of >99% across those weights if the initial bolus dose was 3 g.

The main strength of our study lies in the use of microdialysis to characterize cefazolin concentrations in the ISF of tissues, and to the best of our knowledge, this is the first study in obese pregnant women undergoing CD to do so. We obtained prepregnancy weight and BMI, as well as that at delivery and were able to evaluate these variables in the models. Our study does have limitations. There was only 1 woman in our sample with a BMI at delivery >50 kg·m−2. In our institution, approximately 0.6% of pregnant women have a BMI >50 kg·m−2, and it is likely that our results cannot be extrapolated to these more extreme levels of BMI. We selected a target of 100% ƒT>MIC based on recommendations for surgical prophylaxis.35 Altering this target would of course change the obtained FTA. Additionally, we only examined the pharmacokinetic relationship for S. aureus, the most common pathogen in SSI.36 Other relevant pathogens may have a higher minimum inhibitory concentration at which 90% of isolates are inhibited (MIC90), and therefore FTA will be reduced. Our study did not evaluate the effect of these dosing regimens on the incidence of SSI and an adequately powered randomized controlled trial is required to do so. While our sample size of 12 appears small, we analyzed 76 ISF samples and 58 plasma samples, comparable to other studies using this approach. Our sample size is larger than many other microdialysis studies for this drug33,37 and provides valuable data for optimizing dosing regimens in the current setting. By using the most accurate measure of tissue concentrations (microdialysis) and applying dosing simulations for a much larger population, our approach has the benefit of being an ethical and efficient use of resources, peer-accepted practice, and justifies robust conclusions.

Previous study in this area has applied varying methodology to yield inconsistent results. The pharmacokinetics of cefazolin doses of 2–4 g have been evaluated in obese pregnant women undergoing CD.16–21 These studies, however, have utilized tissue samples, such as homogenized subcutaneous, myometrial, or adipose tissue.16–21 The US FDA states that microdialysis sampling of the subcutaneous ISF, as utilized in our study, is the best way to describe tissue concentrations of antibiotics and should be used to guide therapy.38 Consistent with our study, Brill et al32 demonstrated reduced tissue penetration in nonpregnant obese patients, using microdialysis sampling of the ISF. Based on this, they recommended weight-based dose adjustments in obese patients. The reduced penetration may be due to an increased volume of subcutaneous adipose tissue in patients with obesity or reduced blood flow in the adipose tissue.39

There are limitations of applying BMI in pregnancy, both in clinical practice and the research setting.40 Arguably, the BMI status of the woman at delivery, rather than the prepregnancy or “booking-in” BMI, is more representative in terms of drug handling at delivery. In our study, a 4-compartment model best described plasma ISF drug concentrations. The participant’s actual body weight at delivery was supported for inclusion in the model as being descriptive of the volume of distribution of the ISF compartment. Prepregnancy weight, BMI, and the change in weight and BMI from prepregnancy values did not improve the model. This indicates that in obese pregnant women, it is reasonable to adjust cefazolin dose based on actual body weight at delivery. Current guidelines vary significantly and do not specify if weight-based dose adjustments should be made on prepregnancy or delivery weight, which may introduce further variability when implemented clinically.12–14

Our results suggest that even recently updated recommendations for SSI prophylaxis for elective CD may be inadequate for obese pregnant women. A pragmatic approach to practice change would be to continue administration of the recommended 2 g bolus, with a second 2 g bolus administered 2 hours following the first dose, if surgery continues. Alternatively, administering a 3 g bolus initially, followed by a 3 g bolus at 2 hours would achieve >99% FTA in women between 90 and 150 kg. While the latter would achieve higher FTA, it would require greater practice change which could be considered less desirable. The clinical efficacy of such a dosing strategy requires a controlled evaluation, using the clinical end point of SSI.


Name: Victoria A. Eley, PhD.

Contribution: This author helped with protocol design, ethics application, writing of the manuscript.

Name: Rebecca Christensen, PhD.

Contribution: This author helped with protocol design, ethics application, data collection, writing of the manuscript.

Name: Rochelle Ryan, MBBS.

Contribution: This author helped with protocol design, ethics application, data collection, writing of the manuscript.

Name: Dwane Jackson, MBBS.

Contribution: This author helped with protocol design, ethics application, data collection, writing of the manuscript.

Name: Suzanne L. Parker, PhD.

Contribution: This author helped with data analysis, interpretation of results, and writing of the manuscript.

Name: Matthew Smith, MBBS.

Contribution: This author helped collect the data and write the manuscript.

Name: Andre A. van Zundert, PhD.

Contribution: This author helped design the protocol and write the manuscript.

Name: Steven C. Wallis, PhD.

Contribution: This author helped with data analysis and interpretation and writing of the manuscript.

Name: Jeffrey Lipman, MD.

Contribution: This author helped design the protocol and write the manuscript.

Name: Jason A. Roberts, PhD.

Contribution: This author helped with protocol design, data analysis, interpretation of results, and writing of the manuscript.

This manuscript was handled by: Ken B. Johnson, MD.



    1. Australian Institute of Health and Welfare. Australia’s mothers and babies 2016—in brief. Perinatal statistics series no. 34. Cat no. PER 97. 2018.Canberra: Australian Institute of Health and Welfare.
    2. World Health Organization. WHO Statement on Caesarean Section Rates. 2015.Geneva, Switzerland: World Health Organization.
    3. Sullivan EA, Dickinson JE, Vaughan GA, et al.; Australasian Maternity Outcomes Surveillance System. Maternal super-obesity and perinatal outcomes in Australia: a national population-based cohort study. BMC Pregnancy Childbirth. 2015;15:322.
    4. Moulton LJ, Munoz JL, Lachiewicz M, Liu X, Goje O. Surgical site infection after cesarean delivery: incidence and risk factors at a US academic institution. J Matern Fetal Neonatal Med. 2018;31:1873–1880.
    5. Krieger Y, Walfisch A, Sheiner E. Surgical site infection following cesarean deliveries: trends and risk factors. J Matern Fetal Neonatal Med. 2017;30:8–12.
    6. Alanis MC, Villers MS, Law TL, Steadman EM, Robinson CJ. Complications of cesarean delivery in the massively obese parturient. Am J Obstet Gynecol. 2010;203:271.e1–271.e7.
    7. Brown J, Thompson M, Sinnya S, et al. Pre-incision antibiotic prophylaxis reduces the incidence of post-caesarean surgical site infection. J Hosp Infect. 2013;83:68–70.
    8. Wilson J, Wloch C, Saei A, et al. Inter-hospital comparison of rates of surgical site infection following caesarean section delivery: evaluation of a multicentre surveillance study. J Hosp Infect. 2013;84:44–51.
    9. Smaill FM, Grivell RM. Antibiotic prophylaxis versus no prophylaxis for preventing infection after cesarean section. Cochrane Database Syst Rev. 2014:CD007482.
    10. Bratzler DW, Dellinger EP, Olsen KM, et al.; American Society of Health-System Pharmacists (ASHP); Infectious Diseases Society of America (IDSA); Surgical Infection Society (SIS); Society for Healthcare Epidemiology of America (SHEA). Clinical practice guidelines for antimicrobial prophylaxis in surgery. Surg Infect (Larchmt). 2013;14:73–156.
    11. Fischer MI, Dias C, Stein A, Meinhardt NG, Heineck I. Antibiotic prophylaxis in obese patients submitted to bariatric surgery. A systematic review. Acta Cir Bras. 2014;29:209–217.
    12. eTG complete. Antibiotics: Surgical Prophylaxis. 2019. Melbourne, Australia: Therapeutic Guidelines Ltd; Available at: Accessed January 6, 2019.
    13. Prophylactic Antibiotics in Obstetrics and Gynaecology. Royal Australian and New Zealand College of Obstetricians and Gynaecologists 2016. Melbourne, Australia. Available at: Accessed July 6, 2019.
    14. Medicines Complete [Internet] United Kingdom: British National Formulary. Obstetric and Gynaecological Surgery Antibacterial Prophylaxis.. 2019. Available at: Accessed May 8, 2019.
    15. Douville S, Callaway L, Amoako A, Roberts J, Eley VA. Reducing post-caesarean delivery surgical site infections: a narrative review. Int J Obstet Anesth. 2019 [Epub ahead of print].
    16. Pevzner L, Swank M, Krepel C, Wing DA, Chan K, Edmiston CE Jr.. Effects of maternal obesity on tissue concentrations of prophylactic cefazolin during cesarean delivery. Obstet Gynecol. 2011;117:877–882.
    17. Stitely M, Sweet M, Slain D, et al. Plasma and tissue cefazolin concentrations in obese patients undergoing cesarean delivery and receiving differing pre-operative doses of drug. Surg Infect (Larchmt). 2013;14:455–459.
    18. Swank ML, Wing DA, Nicolau DP, McNulty JA. Increased 3-gram cefazolin dosing for cesarean delivery prophylaxis in obese women. Am J Obstet Gynecol. 2015;213:415.e1–415.e8.
    19. Young OM, Shaik IH, Twedt R, et al. Pharmacokinetics of cefazolin prophylaxis in obese gravidae at time of cesarean delivery. Am J Obstet Gynecol. 2015;213:541.e1–541.e7.
    20. Maggio L, Nicolau DP, DaCosta M, Rouse DJ, Hughes BL. Cefazolin prophylaxis in obese women undergoing cesarean delivery: a randomized controlled trial. Obstet Gynecol. 2015;125:1205–1210.
    21. Grupper M, Kuti JL, Swank ML, Maggio L, Hughes BL, Nicolau DP. Population pharmacokinetics of cefazolin in serum and adipose tissue from overweight and obese women undergoing cesarean delivery. J Clin Pharmacol. 2017;57:712–719.
    22. Marchand S, Chauzy A, Dahyot-Fizelier C, Couet W. Microdialysis as a way to measure antibiotics concentration in tissues. Pharmacol Res. 2016;111:201–207.
    23. Lowe SA, Bowyer L, Lust K, et al.; Society of Obstetric Medicine of Australia and New Zealand. The SOMANZ guidelines for the management of hypertensive disorders of pregnancy 2014. Aust N Z J Obstet Gynaecol. 2015;55:11–16.
    24. eTG complete [Internet]. 2018. Melbourne, Australia: Therapeutic Guidelines Ltd; Antibiotics: Surgical prophylaxis. Available at: Accessed October 1, 2018.
    25. Wang Y, Wong SL, Sawchuk RJ. Microdialysis calibration using retrodialysis and zero-net flux: application to a study of the distribution of zidovudine to rabbit cerebrospinal fluid and thalamus. Pharm Res. 1993;10:1411–1419.
    26. Parker SL, Guerra Valero YC, Roberts DM, Lipman J, Roberts JA, Wallis SC. Determination of cefalothin and cefazolin in human plasma, urine and peritoneal dialysate by UHPLC-MS/MS: application to a pilot pharmacokinetic study in humans. Biomed Chromatogr. 2016;30:872–879.
    27. US FDA [Internet] US Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research (CDER) 2001. Guidance for Industry: Bioanalytical Method Validation. Available at: Accessed January 3, 2019.
    28. Byrne CJ, Parton T, McWhinney B, et al. Population pharmacokinetics of total and unbound teicoplanin concentrations and dosing simulations in patients with haematological malignancy. J Antimicrob Chemother. 2018;73:995–1003.
    29. Savic RM, Karlsson MO. Importance of shrinkage in empirical Bayes estimates for diagnostics: problems and solutions. AAPS J. 2009;11:558–569.
    30. [Internet] European Committee on Antimicrobial Susceptibility Testing. MIC and zone diameter distributions and ECOFFs. Available at: Accessed January 3, 2019.
    31. Eley VA, van Zundert A, Callaway L. What is the failure rate in extending labour analgesia in patients with a body mass index ≥ 40 kg/m(2)compared with patients with a body mass index < 30 kg/m(2)? a retrospective pilot study. BMC Anesthesiol. 2015;15:115.
    32. Brill MJ, Houwink AP, Schmidt S, et al. Reduced subcutaneous tissue distribution of cefazolin in morbidly obese versus non-obese patients determined using clinical microdialysis. J Antimicrob Chemother. 2014;69:715–723.
    33. Douglas A, Udy AA, Wallis SC, et al. Plasma and tissue pharmacokinetics of cefazolin in patients undergoing elective and semielective abdominal aortic aneurysm open repair surgery. Antimicrob Agents Chemother. 2011;55:5238–5242.
    34. Dennis AT, Lamb KE, Story D, et al.; MUM SIZE Investigators. Associations between maternal size and health outcomes for women undergoing caesarean section: a multicentre prospective observational study (The MUM SIZE Study). BMJ Open. 2017;7:e015630.
    35. Zelenitsky SA, Lawson C, Calic D, et al. Integrated pharmacokinetic-pharmacodynamic modelling to evaluate antimicrobial prophylaxis in abdominal surgery. J Antimicrob Chemother. 2016;71:2902–2908.
    36. Worth LJ, Bull AL, Spelman T, Brett J, Richards MJ. Diminishing surgical site infections in Australia: time trends in infection rates, pathogens and antimicrobial resistance using a comprehensive Victorian surveillance program, 2002-2013. Infect Control Hosp Epidemiol. 2015;36:409–416.
    37. Andreas M, Zeitlinger M, Wisser W, et al. Cefazolin and linezolid penetration into sternal cancellous bone during coronary artery bypass grafting. Eur J Cardiothorac Surg. 2015;48:758–764.
    38. Chaurasia CS, Müller M, Bashaw ED, et al. AAPS-FDA Workshop White Paper: microdialysis principles, application, and regulatory perspectives. J Clin Pharmacol. 2007;47:589–603.
    39. Rossi M, Nannipieri M, Anselmino M, Guarino D, Franzoni F, Pesce M. Subcutaneous adipose tissue blood flow and vasomotion in morbidly obese patients: long term effect of gastric bypass surgery. Clin Hemorheol Microcirc. 2012;51:159–167.
    40. National Health and Medical Research Council. Melbourne, Australia. Clinical Practice Guidelines for the Management of Overweight and Obesity in Adults, Adolescents and Children in Australia. Available at: Accessed January 3, 2019.

    Supplemental Digital Content

    Copyright © 2020 International Anesthesia Research Society