Article Text

Download PDFPDF

Antibiotic pharmacokinetics in infected pleural effusions
  1. David T Arnold1,
  2. Liam Read2,
  3. Oliver Waddington2,
  4. Fergus W Hamilton3,4,
  5. Sonia Patole1,
  6. Jessica Hughes1,
  7. Alice Milne1,
  8. Alan Noel2,
  9. Mark Bayliss2,
  10. Nicholas A Maskell1,
  11. Alasdair MacGowan2
  1. 1Academic Respiratory Unit, University of Bristol, Bristol, UK
  2. 2Antimicrobial Reference Laboratory, North Bristol NHS Trust, Westbury on Trym, Bristol, UK
  3. 3Department of Infection Science, North Bristol NHS Trust, Bristol, Bristol, UK
  4. 4Population Health Sciences, University of Bristol, Bristol, UK
  1. Correspondence to Dr David T Arnold, Academic Respiratory Unit, University of Bristol, Bristol, UK; arnold.dta{at}


Pleural infection is usually treated with empirical broad-spectrum antibiotics, but limited data exist on their penetrance into the infected pleural space. We performed a pharmacokinetic study analysing the concentration of five intravenous antibiotics across 146 separate time points in 35 patients (amoxicillin, metronidazole, piperacillin-tazobactam, clindamycin and cotrimoxazole). All antibiotics tested, apart from co-trimoxazole, reach pleural fluid levels equivalent to levels within the blood and well above the relevant minimum inhibitory concentrations. The results demonstrate that concerns about the penetration of commonly used antibiotics, apart from co-trimoxazole, into the infected pleural space are unfounded.

  • Pleural Disease
  • Bacterial Infection
  • Pneumonia
  • Respiratory Infection

This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Pleural infection is a serious clinical condition with a mortality of up to 15%.1 Patients spend extended periods in the hospital for pleural fluid drainage and intravenous antibiotic administration.1 2 There has been concern around the ability of antibiotics to reach therapeutic levels at the site of infection.2


We recruited patients presenting to a single UK tertiary pleural centre with pleural infection. Patients were consented for regular pleural fluid sampling (via chest tube) with synchronous blood sampling timed with antibiotic administration (1, 2, 4 and 6 hours post dose for a 6-hourly (four times a day) regimen; 1, 4 and 8 hours for an 8-hourly (three times a day) regimen; 1, 4, 8 and 12-hourly (two times a day) for a 12-hourly regimen). The study received ethical approval from the East of Scotland Research Ethics Service (REC reference: 19/ES/0075).

Analytes were assayed using liquid chromatography mass spectrometry assays (LC-MS/MS). Deproteinised samples were injected into a Shimadzu LC system, and gradient separation was performed on a 50×2.1 mm ID 2.6 µm Kinetex XBC18 column using mobile phases containing water/acetonitrile/formic acid (0.1%,v/v). Analytes were selectively detected using a positive ion electrospray source on AB Sciex 4000 QTrap MS. Concentrations in test samples were calculated by Sciex Analyst software using calibrators prepared in the same biological matrix.

Total drug concentrations were plotted with the area under the curve (AUC) from time zero to the next dose determined by the trapezoidal rule. The penetration ratio (PR) for each patient was obtained by dividing the AUC for pleural fluid by the AUC for the blood. Pearson’s correlation coefficient was used to assess the correlation between antibiotic concentrations and pleural fluid pH. A paired t-test was used to compare concentrations between loculated and non-loculated effusions.


Thirty-five patients were recruited between October 2019 and March 2022. Six patients met the criteria for frank empyema with the others meeting at least one criterion for pleural infection.3 There was a male predominance (77%) with a median age of 76 (IQR 65–79). Mean values (and SD) from baseline pleural fluid analysis were pH of 7.07 (0.22), protein of 39 g/L (12 g/L) and lactate dehydrogenase of 2346 IU (3543 IU). Pleural fluid culture was positive in 7 patients (20%). Full pharmacokinetic analysis could be performed for the antibiotics shown in table 1 with summary curves shown in figure 1.

Table 1

Antibiotics assayed, dosing schedule and number of time points assessed

Figure 1

Mean (±SD) intravenous antibiotic concentration-time curves in the blood (dotted line) and pleural fluid (solid line): (A) amoxicillin, (B) metronidazole, (C) piperacillin (piperacillin-tazobactam), (D) tazobactam (piperacillin-tazobactam), (E) trimethoprim (co-trimoxazole), (F) sulphamethoxazole (co-trimoxazole) and (G) clindamycin. Minimum inhibitory concentrations (MICs) (EUCAST) of amoxicillin: Staphylococcus aureus 1 mg/L, Haemophilus influenza 2 mg/L, Streptococcus Milleri group 0.125 mg/L and Streptococcus pneumoniae 0.06 mg/L. MICs (EUCAST) of metronidazole: prevotella 1 mg/L, Gram-positive anaerobes (class) 0.25 mg/L and fusobacterium 0.125 mg/L. MICs (EUCAST) of piperacillin-tazobactam: S. aureus 2 mg/L, H. influenza 0.5 mg/L, Streptococcus Milleri group 1 mg/L, Escherichia coli 8 mg/L and S. pneumoniae 0.064 mg/L. MICs (EUCAST) of trimethoprim-sulfamethoxazole: S. aureus 0.25 mg/L, H. influenza 0.5 mg/L, S. pneumoniae 1 mg/L, E. coli 0.5 mg/L. MICs (EUCAST) of clindamycin: S. aureus 0.25 mg/L, Streptococcus Milleri group 0.25 mg/L and S. pneumoniae 0.25 mg/L. The MICs listed for common causative organisms are derived from the EUCAST website (, which describes the MIC of ‘wild-type’ and, therefore, fully sensitive bacteria.


The AUC8hr for the dosing interval was marginally higher in the pleural fluid compared with the blood (62.8 vs 48.5, PR 1.1) and above the minimum inhibitory concentration (MIC) of causative organisms for the entirety of the dosing schedule.


The AUC8hr for metronidazole was similar between the pleural fluid and blood (128.5 vs 151.5, PR 0.84).


Piperacillin levels peaked rapidly in the plasma but fell sharply at the 4 hour timepoint.4 The AUC8hr was similar between the pleural fluid and blood (465 vs 648, PR 0.72). The AUC8hr for tazobactam was similar between the pleural fluid and blood (55.7 vs 59.5, respectively) with a ratio to piperacillin of 11.6%.


The trimethoprim and sulfamethoxazole constituents of co-trimoxazole were measured separately. The AUC12hr was much lower for trimethoprim in the pleural fluid compared with the blood (13.5 vs 47.2), with a PR of 0.29. The ratio for sulfamethoxazole was more equivalent (AUC12hr 465.0 in the blood vs 425.8 in the pleural fluid, PR 0.92).


The AUC6hr for clindamycin was similar between the pleural fluid and blood (41.3 vs 51.3, PR 0.81).

pH and loculation

Across the compounds tested, there was no correlation between the overall pooled mean antibiotic concentrations and the pH of the pleural fluid on sampling. For example, the mean concentration of amoxicillin was 7.9 mg/L (SD 10.2) with a correlation coefficient of −0.145, (p=0.49) (see online supplemental table S1). There was also no mean difference between the blood and pleural fluid antibiotic concentrations depending on the presence and severity of loculations within the effusion (for amoxicillin, metronidazole, piperacillin-tazobactam). For example, the mean difference in amoxicillin concentration was 0.59 mg/L (SD 17.5) in non-loculated effusions versus −1.1 mg/L (SD 11.4) in loculated effusions (p=0.77) (see online supplemental table S2).

Supplemental material


This is the largest study of antibiotic pharmacokinetics performed in infected pleural effusions. We have demonstrated that commonly used antibiotics such as amoxicillin, metronidazole, piperacillin-tazobactam and clindamycin reach levels within the pleural fluid equivalent to that in the blood and above the MIC for bacteria known to cause pleural infection. The trimethoprim element of co-trimoxazole did not reach the pleural fluid adequately, raising concerns about its use in established or developing pleural infection.

Despite antibiotic recommendations in guidelines and across study protocols, studies on the pleural penetration of many of these antibiotics have not been performed.5 The most widely cited evidence on antibiotic concentrations in infected pleural spaces is obtained from rabbit models of turpentine-induced pleural infection.6 In 1987, Shohet demonstrated that gentamicin had diminished efficacy in pleural infection, leading to reduced aminoglycoside use for pleural infection.7 Given the stark differences in conclusions compared with human pharmacokinetic studies, results from animal models should be interpreted with caution.2

Studies of antibiotic pharmacokinetics in human pleural infection are limited in sample size, methodology and antibiotic regimens. A review by Lau et al reported that across the relevant five studies, most antibiotics had been assessed following a single dose in one to three patients with pleural infection. Thys et al measured aminoglycoside levels after a single intravenous dose in patients with uninfected versus infected purulent pleural infection (n=19 and n=11, respectively). In the infected effusions, levels were either undetectable or significantly lower than plasma, demonstrating the importance of performing pharmacokinetic studies using infected pleural fluid.

This study has demonstrated that the pleural concentrations of amoxicillin, metronidazole, piperacillin-tazobactam and clindamycin are equivalent to that of the blood levels.6 Pleural fluid trimethoprim levels were much lower than in the blood across the dosing schedule. There is no previous literature on the pharmacokinetics of trimethoprim within the infected pleural fluid. Extrapolating from other encapsulated infections, there is evidence of poor penetration of co-trimoxazole into the gall bladder during acute cholecystitis.8 When compared with the MIC for gram positives that commonly cause pleural infection (eg, Streptococcus pneumoniae and the Streptococcus virdans group) levels for trimethoprim were inadequate.9

We were not able to perform sub-group analysis for individual antibiotics for factors such as the patient’s age, fluid pH or degree of pleural loculation. Second, like many pharmacokinetic studies, comparative MICs have been extrapolated from studies of wild-type bacteria as opposed to those with resistance patterns.


The commonly used antibiotics such as amoxicillin, metronidazole, piperacillin-tazobactam and clindamycin reached levels equivalent to the blood within infected pleural fluid. Low penetration of trimethoprim into the pleural space raised concerns about the use of cotrimoxazole for patients with pleural infection or parapneumonic effusions.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants and was approved by East of Scotland Research Ethics Service (19/ES/0075). Participants gave informed consent to participate in the study before participating.


We would like to thank the patients who participated in this work.



  • NAM and AM are joint senior authors.

  • X @gushamilton

  • Contributors DTA, NAM, FWH and AM had the idea for the study and developed the protocol. LR, OW, AN and MB were responsible for sample processing and analysis. SP, JH and AM were responsible for recruiting participants and database logging. DTA performed the statistical analysis with AM. All authors were involved in approving the final manuscript.

  • Funding DTA is funded by a National Institute for Health Research (NIHR) Doctoral Research Fellowship (DRF-2018-11-ST2-065) for this research project. This publication presents independent research funded by the National Institute for Health Research (NIHR). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.