CICM Online CCR Journal logo CICM logo

Full Text View

Original Article

Cytokine and lipid metabolome effects of low-dose acetylsalicylic acid in critically ill patients with systemic inflammation: a pilot, feasibility, multicentre, randomised, placebo-controlled trial

Luca Cioccari*, Nora Luethi*, Thy Duong, Eileen Ryan, Salvatore L Cutuli, Patryck Lloyd-Donald, Glenn M Eastwood, Leah Peck, Helen Young, Suvi T Vaara, Craig J French, Neil Orford, Jyotsna Dwivedi, Yugeesh R Lankadeva, Michael Bailey, Gavin E Reid, Rinaldo Bellomo, (* equal first authors)

Crit Care Resusc 2020; 22 (3): 227-236


  • Author Details
    • Luca Cioccari* 1, 2, 3
    • Nora Luethi* 1, 2
    • Thy Duong 4
    • Eileen Ryan 5
    • Salvatore L Cutuli 1, 6
    • Patryck Lloyd-Donald 1
    • Glenn M Eastwood 1, 2
    • Leah Peck 1
    • Helen Young 1
    • Suvi T Vaara 1
    • Craig J French 2, 7
    • Neil Orford 2, 8
    • Jyotsna Dwivedi 9
    • Yugeesh R Lankadeva 10
    • Michael Bailey 2, 11
    • Gavin E Reid 4, 5
    • Rinaldo Bellomo 1, 2, 11
    • (* equal first authors) 12
    1. Department of Intensive Care, Austin Hospital, Melbourne, Vic, Australia.
    2. Australian and New Zealand Intensive Care Research Centre, School of Public Health and Preventive Medicine, Monash University, Melbourne, Vic, Australia.
    3. Department of Intensive Care Medicine, Inselspital, University Hospital, University of Bern, Bern, Switzerland.
    4. Department of Biochemistry and Molecular Biology, University of Melbourne, Melbourne, Vic, Australia.
    5. School of Chemistry, Bio21 Molecular Science and Biotechnology Institute, University of Melbourne, Melbourne, Vic, Australia.
    6. Dipartimento di Scienze dell’Emergenza, Anestesiologiche e della Rianimazione; UOC di Anestesia, Rianimazione, Terapia Intensiva e Tossicologia Clinica; Fondazione Policlinico Universitario Agostino Gemelli IRCCS; Istituto di Anestesia e Rianimazione; Università Cattolica del Sacro Cuore, Rome, Italy.
    7. Department of Intensive Care, Western Hospital, Melbourne, Vic, Australia.
    8. Department of Intensive Care, University Hospital Geelong, Vic, Australia.
    9. Department of Intensive Care, Bankstown Hospital, Sydney, NSW, Australia.
    10. Preclinical Critical Care Unit, Florey Institute of Neuroscience and Mental Health, University of Melbourne, Melbourne, Vic, Australia.
    11. Centre for Integrated Critical Care, University of Melbourne, Melbourne, Vic, Australia.
    12. (*equal first authors).
  • Competing Interests

    The trial was supported by the Intensive Care Trust Fund (Austin Hospital, Melbourne). The active and placebo drugs were sponsored by Bayer Pharmaceuticals, Germany. The funding bodies did not have input into the design, management or reporting of the trial.

  • Abstract
    OBJECTIVE: The systemic inflammatory response syndrome (SIRS) is a dysregulated response that contributes to critical illness. Adjunctive acetylsalicylic acid (ASA) treatment may offer beneficial effects by increasing the synthesis of specialised proresolving mediators (a subset of polyunsaturated fatty acid-derived lipid mediators).
    DESIGN: Pilot, feasibility, multicentre, double-blind, randomised, placebo-controlled trial.
    SETTING: Four interdisciplinary intensive care units (ICUs) in Australia.
    PARTICIPANTS: Critically ill patients with SIRS.
    INTERVENTIONS: ASA 100 mg 12-hourly or placebo, administered within 24 hours of ICU admission and continued until ICU day 7, discharge or death, whichever came first.
    MAIN OUTCOME MEASURES: Interleukin-6 (IL-6) serum concentration at 48 hours after randomisation and, in a prespecified subgroup of patients, serum lipid mediator concentrations measured by mass spectrometry.
    RESULTS: The trial was discontinued in December 2017 due to slow recruitment and after the inclusion of 48 patients. Compared with placebo, ASA did not decrease IL-6 serum concentration at 48 hours. In the 32 patients with analysis of lipid mediators, low-dose ASA increased the concentration of 15-hydroxyeicosatetraenoic acid, a proresolving precursor of lipoxin A4, and reduced the concentration of the proinflammatory cytochrome P-dependent mediators 17-HETE (hydroxyeicosatetraenoic acid), 18-HETE and 20-HETE. In the eicosapentaenoic acid pathway, ASA significantly increased the concentration of the anti-inflammatory mediators 17,18-DiHETE (dihydroxyeicosatetraenoic acid) and 14,15-DiHETE.
    CONCLUSIONS: In ICU patients with SIRS, low-dose ASA did not significantly alter serum IL-6 concentrations, but it did affect plasma concentrations of certain lipid mediators. The ability to measure lipid mediators in clinical samples and to monitor the effect of ASA on their levels unlocks a potential area of biological investigation in critical care.
    TRIAL REGISTRATION: Australian New Zealand Clinical Trials Registry (ACTRN 12614001165673).
  • References
    1. Levy MM, Fink MP, Marshall JC, et al. 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Crit Care Med 2003; 31: 1250-6
    2. Churpek MM, Zadravecz FJ, Winslow C, et al. Incidence and prognostic value of the systemic inflammatory response syndrome and organ dysfunctions in ward patients. Am J Respir Crit Care Med 2015; 192: 958-64
    3. Bosmann M, Ward PA. The inflammatory response in sepsis. Trends Immunol 2013; 34: 129-36
    4. Trauer J, Muhi S, McBryde ES, et al. Quantifying the effects of prior acetyl-salicylic acid on sepsis-related deaths: an individual patient data meta-analysis using propensity matching. Crit Care Med 2017; 45: 1871-9
    5. Sossdorf M, Otto GP, Boettel J, et al. Benefit of low-dose aspirin and non-steroidal anti-inflammatory drugs in septic patients. Crit Care 2013; 17: 402
    6. Eisen DP, Reid D, McBryde ES. Acetyl salicylic acid usage and mortality in critically ill patients with the systemic inflammatory response syndrome and sepsis. Crit Care Med 2012; 40: 1761-7
    7. osche W, Boettel J, Kabisch B, et al. Do aspirin and other antiplatelet drugs reduce the mortality in critically ill patients? Thrombosis 2012; 2012: 720254
    8. Schror K, Rauch BH. Aspirin and lipid mediators in the cardiovascular system. Prostaglandins Other Lipid Mediat 2015; 121: 17-23
    9. Serhan CN. Treating inflammation and infection in the 21st century: new hints from decoding resolution mediators and mechanisms. FASEB J 2017; 31: 1273-88
    10. Serhan CN, Chiang N. Endogenous pro-resolving and anti-inflammatory lipid mediators: a new pharmacologic genus. Br J Pharmacol 2008; 153 Suppl 1: S200-15
    11. Serhan CN. Pro-resolving lipid mediators are leads for resolution physiology. Nature 2014; 510: 92-101
    12. Ramon S, Baker SF, Sahler JM, et al. The specialized proresolving mediator 17-HDHA enhances the antibody-mediated immune response against influenza virus: a new class of adjuvant? J Immunol 2014; 193: 6031-40
    13. Cilloniz C, Pantin-Jackwood MJ, Ni C, et al. Lethal dissemination of H5N1 influenza virus is associated with dysregulation of inflammation and lipoxin signaling in a mouse model of infection. J Virol 2010; 84: 7613-24
    14. Morita M, Kuba K, Ichikawa A, et al. The lipid mediator protectin D1 inhibits influenza virus replication and improves severe influenza. Cell 2013; 153: 112-25
    15. Astarita G, Kendall AC, Dennis EA, Nicolaou A. Targeted lipidomic strategies for oxygenated metabolites of polyunsaturated fatty acids. Biochim Biophys Acta 2015; 1851: 456-68
    16. Murphy RC. Specialized pro-resolving mediators: do they circulate in plasma? J Lipid Res 2015; 56: 1641-2
    17. Masoodi M, Eiden M, Koulman A, et al. Comprehensive lipidomics analysis of bioactive lipids in complex regulatory networks. Anal Chem 2010; 82: 8176-85
    18. Masoodi M, Lee E, Eiden M, et al. A role for oleoylethanolamide in chronic lymphocytic leukemia. Leukemia 2014; 28: 1381-7
    19. Masoodi M, Pearl DS, Eiden M, et al. Altered colonic mucosal polyunsaturated fatty acid (PUFA) derived lipid mediators in ulcerative colitis: new insight into relationship with disease activity and pathophysiology. PLoS One 2013; 8: e76532
    20. Ryan E, Reid GE. Chemical derivatization and ultrahigh resolution and accurate mass spectrometry strategies for “shotgun” lipidome analysis. Acc Chem Res 2016; 49: 1596-1604
    21. Yamada M, Kita Y, Kohira T, et al. A comprehensive quantification method for eicosanoids and related compounds by using liquid chromatography/mass spectrometry with high speed continuous ionization polarity switching. J Chromatogr B Analyt Technol Biomed Life Sci 2015; 995-996: 74-84
    22. Gabbs M, Leng S, Devassy JG, et al. Advances in our understanding of oxylipins derived from dietary PUFAs. Adv Nutr 2015; 6: 513-40
    23. O’Neal HR Jr, Koyama T, Koehler EA, et al. Prehospital statin and aspirin use and the prevalence of severe sepsis and acute lung injury/acute respiratory distress syndrome. Crit Care Med 2011; 39: 1343-50
    24. Eisen DP. Manifold beneficial effects of acetyl salicylic acid and nonsteroidal anti-inflammatory drugs on sepsis. Intensive Care Med 2012; 38: 1249-57
    25. Du F, Jiang P, He S, et al. Antiplatelet therapy for critically ill patients: a pairwise and bayesian network meta-analysis. Shock 2018; 49: 616-24
    26. Tsai M-J, Ou S-M, Shih C-J, et al. Association of prior antiplatelet agents with mortality in sepsis patients: a nationwide population-based cohort study. Intensive Care Med 2015; 41: 806-13
    27. Boyle AJ, Di Gangi S, Hamid UI, et al. Aspirin therapy in patients with acute respiratory distress syndrome (ARDS) is associated with reduced intensive care unit mortality: a prospective analysis. Crit Care 2015; 19: 109
    28. Wang Y, Zhong M, Wang Z, et al. The preventive effect of antiplatelet therapy in acute respiratory distress syndrome: a meta-analysis. Crit Care 2018; 22: 60
    29. Kor DJ, Carter RE, Park PK, et al. Effect of aspirin on development of ARDS in at-risk patients presenting to the emergency department: the LIPS-A randomized clinical trial. JAMA 2016; 315: 2406-14
    30. Zaslona Z, Palsson-McDermott EM, Menon D, et al. The induction of Pro-IL-1beta by lipopolysaccharide requires endogenous prostaglandin E2 production. J Immunol 2017; 198: 3558-64
    31. Strassmann G, Patil-Koota V, Finkelman F, et al. Evidence for the involvement of interleukin 10 in the differential deactivation of murine peritoneal macrophages by prostaglandin E2. J Exp Med 1994; 180: 2365-70
    32. Netea MG, Puren AJ, Dinarello CA. A short course of oral aspirin increases IL-18-induced interferon-gamma production in whole blood cultures. Eur Cytokine Netw 2000; 11: 379-82
    33. Kiers D, van der Heijden WA, van Ede L, et al. A randomised trial on the effect of anti-platelet therapy on the systemic inflammatory response in human endotoxaemia. Thromb Haemost 2017; 117: 1798-1807
    34. Hotchkiss RS, Monneret G, Payen D. Sepsis-induced immunosuppression: from cellular dysfunctions to immunotherapy. Nat Rev Immunol 2013; 13: 862-74
    35. van der Poll T, van de Veerdonk FL, Scicluna BP, Netea MG. The immunopathology of sepsis and potential therapeutic targets. Nat Rev Immunol 2017; 17: 407-20
    36. Serhan CN, Arita M, Hong S, Gotlinger K. Resolvins, docosatrienes, and neuroprotectins, novel omega-3-derived mediators, and their endogenous aspirin-triggered epimers. Lipids 2004; 39: 1125-32
    37. Dalli J, Colas RA, Quintana C, et al. Human sepsis eicosanoid and proresolving lipid mediator temporal profiles: correlations with survival and clinical outcomes. Crit Care Med 2017; 45: 58-68
    38. Lewis JP, Yerges-Armstrong LM, Ellero-Simatos S, et al. Integration of pharmacometabolomic and pharmacogenomic approaches reveals novel insights into antiplatelet therapy. Clin Pharmacol Ther 2013; 94: 570-3
    39. Chiang N, Bermudez EA, Ridker PM, et al. Aspirin triggers antiinflammatory 15-epi-lipoxin A4 and inhibits thromboxane in a randomized human trial. Proc Natl Acad Sci U S A 2004; 101: 15178-83
    40. Ellero-Simatos S, Beitelshees AL, Lewis JP, et al. Oxylipid profile of low-dose aspirin exposure: a pharmacometabolomics study. J Am Heart Assoc 2015; 4: e002203
    41. Yerges-Armstrong LM, Ellero-Simatos S, Georgiades A, et al. Purine pathway implicated in mechanism of resistance to aspirin therapy: pharmacometabolomics-informed pharmacogenomics. Clin Pharmacol Ther 2013; 94: 525-32
    42. Morin C, Sirois M, Echave V, et al. 17,18-epoxyeicosatetraenoic acid targets PPARgamma and p38 mitogen-activated protein kinase to mediate its anti-inflammatory effects in the lung: role of soluble epoxide hydrolase. Am J Respir Cell Mol Biol 2010; 43: 564-75
    43. VanRollins M. Epoxygenase metabolites of docosahexaenoic and eicosapentaenoic acids inhibit platelet aggregation at concentrations below those affecting thromboxane synthesis. J Pharmacol Exp Ther 1995; 274: 798-804
    44. Im DS. A promising anti-inflammatory and anti-thrombotic drug for sepsis treatment. Arch Pharm Res 2011; 34: 339-42
    45. Moore CG, Carter RE, Nietert PJ, Stewart PW. Recommendations for planning pilot studies in clinical and translational research. Clin Transl Sci 2011; 4: 332-7
    46. Leijte GP, Kiers D, van der Heijden W, et al. Treatment with acetylsalicylic acid reverses endotoxin tolerance in humans in vivo: a randomized placebo-controlled study. Crit Care Med 2019; 47: 508-16
    47. Winning J, Neumann J, Kohl M, et al. Antiplatelet drugs and outcome in mixed admissions to an intensive care unit. Crit Care Med 2010; 38: 32-7
    48. Harr JN, Moore EE, Johnson J, et al. Antiplatelet therapy is associated with decreased transfusion-associated risk of lung dysfunction, multiple organ failure, and mortality in trauma patients. Crit Care Med 2013; 41: 399-404
    49. Wehbeh A, Tamim HM, Abu Daya H, et al. Aspirin has a protective effect against adverse outcomes in patients with nonvariceal upper gastrointestinal bleeding. Dig Dis Sci 2015; 60: 2077-87
The systemic inflammatory response syndrome (SIRS) is defined as two or more abnormalities in respiration, temperature, heart rate or white blood cell count. 1 SIRS may occur in various conditions related, or not, to infection. 2 While a certain amount of inflammatory response may be beneficial, a dysregulated host response can result in life-threatening organ dysfunction and even death. 3 Acetylsalicylic acid (ASA) might provide beneficial effects by attenuating systemic inflammation 4 and protecting against the injurious effects of uncontrolled SIRS and sepsis, 5, 6 and it has been associated with decreased mortality in critically ill patients. 7

Apart from its well known anti-inflammatory properties owing to, at least in part, its irreversible inhibition of cyclooxygenase (COX)-1, ASA also modulates COX-2 activity. 8 Aspirin-acetylated COX-2 acts on arachidonic acid (AA), eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) and promotes resolution and counter-regulation of excessive inflammation by increasing the synthesis of polyunsaturated fatty acid (PUFA)-derived specialised proresolving mediators (SPMs). SPMs are bioactive lipid mediators — produced by cells of the innate immune system via enzymatic conversion of essential fatty acids including EPA, DHA and AA 9 — and are categorised into four families: resolvins, lipoxins, maresins and protectins. 10 SPMs stimulate phagocytosis of cellular debris by macrophages, counter the release of proinflammatory cytokines 11 and have been shown to promote antiviral B lymphocytic activity. 12 Results from previous studies in animals and humans support the hypothesis that SPMs can stop the cytokine storm, enhance resolution of inflammation, and improve outcomes in cases of severe infection. Animal models of influenza A suggest a protective role for SPMs, 13 and treatment with SPM improved survival and reduced severity of infection in mice. 14 SPMs are synthesised locally (at the site of infection) and they are rapidly metabolised. 15 Therefore, whether SPMs are detectable in low volumes of plasma has been a matter of debate. 16 However, early detection and quantification of SPMs is now feasible using refined mass spectrometry methods 17, 18, 19 and can thus provide real-time information on changes in inflammatory status.

The primary objective of this randomised controlled trial was to test the feasibility and safety of ASA treatment as well as its biological effects on conventional inflammatory markers in critically ill patients with SIRS. The secondary objective was to analyse SPMs in low volumes of human plasma in a clinical setting using modern mass spectrometry methods. 20 We hypothesised that short term treatment with low-dose ASA would lead to a significant reduction in serum levels of interleukin (IL)-6 at 48 hours after randomisation, and detectable changes in plasma levels of SPMs.


Study design, screening and randomisation

We conducted a multicentre, double-blind, randomised, placebo-controlled trial across four interdisciplinary intensive care units (ICUs) in Australia. All ICU admissions with two or more positive SIRS criteria were assessed for eligibility. Eligible patients were adults (18 years or older) who were expected to stay in the ICU for at least 48 hours. We excluded patients transferred to ICU after cardiac surgery or neurosurgery, those with active bleeding, coagulopathy or other contraindications to ASA, and patients already taking ASA for primary or secondary prevention. Details of the inclusion and exclusion criteria are provided in Table S1 (Online Appendix). Eligible patients were randomly assigned in a 1:1 ratio to receive either ASA 100 mg or placebo administered twice daily via oral route or nasogastric tube. Randomisation was performed by the site investigators using sealed opaque envelopes. Blinding regarding the trial regimen was ensured by supply of the study drugs as visually identical tablets. All site investigators, patients and treating clinicians were unaware of the trial-group assignment and sequence. Study drugs were continued until ICU discharge, up to a maximum of 7 days, or until the development of any exclusion criteria. All other aspects of patient care were at the discretion of the treating clinicians.

Outcome measures

The primary outcome was serum IL-6 level at 48 hours after randomisation. Secondary outcomes included serum levels of other inflammatory biomarkers (IL-10, E-selectin and C-reactive protein) at 48 hours, length of ICU and hospital stay, duration of mechanical ventilation, and vital status at ICU and hospital discharge. Exploratory outcomes included plasma PUFA-derived lipid mediator levels, which were measured once daily over the first 72 hours in one centre (Austin Hospital). Safety outcomes included the requirement for blood transfusions and the frequency of renal replacement therapy during the intervention period.

Blood sampling and processing

For batch analysis of IL-6, IL-10 and E-selectin, blood samples were obtained at the time of randomisation and after 48 hours. An additional plasma sample was taken daily for the first 72 hours in the subgroup of patients included in the lipid mediator analysis. Blood was collected from the central venous or arterial catheter using a BD Vacutainer system (Z Serum Sep Clot Activator and 3E 3K-EDTA, respectively; BD Diagnostics, Franklin Lakes, NJ, USA). For biomarker analysis, blood samples were centrifuged at 5000 rpm for 5 minutes immediately after collection, and then the supernatant serum and/or plasma was aspirated into 2 mL Eppendorf Tubes and stored directly at –80°C.

Analysis of IL-6, IL-10 and E-selectin levels

Quantification of serum IL-6, IL-10 and E-selectin levels was performed by enzyme-linked immunosorbent assays (ELISAs) using commercially available kits as per the manufacturer’s instructions (Thermo Fisher Scientific, Vienna, Austria). All tests were performed in duplicate. Reactions were developed using chromogen (TMB substrate; Invitrogen, Carlsbad, CA, USA) at 100 µL/well, and then stopped by the addition of 100 µL/well stop solution (2 M H2SO4) and immediately read (λ = 450 nm and 550 nm) on a spectrophotometer (Synergy H1 Hybrid Microplate Reader, BioTek Instruments, Winooski, Vt, USA). The intra-assay variation for all serum cytokine and E-selectin ELISAs was < 10%.

Extraction and analysis of SPMs

After thawing on ice, 0.4 mL of plasma spiked with internal standards was quenched with 1.2 mL of ice-cold methanol (containing 0.01% butylated hydroxytoluene) and left to equilibrate for 15 minutes to allow protein precipitation. The samples were then centrifuged for 5 minutes at 10 000 rpm, and the supernatant was diluted into 4.8 mL of Milli-Q water. Lipid mediators were then enriched by using solid-phase extraction (96-well plates, Trajan Scientific and Medical, Melbourne, Australia). The wells were conditioned with 0.5 mL of methanol and then 0.5 mL of Milli-Q water before samples were loaded. Next, the wells were washed twice with 0.5 mL of 5% methanol and eluted with 0.4 mL of methanol. The eluants were dried to completion under a gentle stream of nitrogen. Samples were reconstituted in 40 µL of methanol (containing 0.01% butylated hydroxytoluene), and then 10 µL of reconstituted sample was injected for reversed-phase high-performance liquid chromatography (HPLC) electrospray ionisation tandem mass spectrometry analysis (Shimadzu LC/MS 8060 triple quadrupole mass spectrometer, coupled with a Nexera HPLC system; Shimadzu, Kyoto, Japan). Mass spectrometry scans were performed using positive and negative ionisation mode polarity switching, based on a method previously developed by Yamada and colleagues for the Method Package for Lipid Mediators for Shimadzu, 21 and further extended as part of this current study. Data processing, including peak detection and integration, was performed using the LabSolutions LCMS version 5.56 software (Shimadzu). Each lipid was manually inspected at the expected retention time plus or minus 30 seconds. Peak abundances were exported and normalised to the internal standard compound CUDA (12-[[(cyclohexylamino)carbonyl]amino]-dodecanoic acid) before proceeding with statistical analysis. Lipids not detected, lipids that had saturated the detector and had a peak abundance of four orders of magnitude larger than the lipid mediators of interest, lipids derived from non-enzymatic oxidation, and lipids that have no reported biological function were excluded from the analysis. The remaining 81 lipids (Online Appendix, figure S2) were categorised into four groups according to their known or presumed biological functions: proinflammatory, anti-inflammatory, both proinflammatory and resolving, and proresolving. 22

Statistical analysis

The trial aimed to include 120 patients to reach a > 90% power to demonstrate a significant difference in IL-6 levels. All analyses were performed on an intention-to-treat basis. In the case of missing data, we reported the number of available observations and did not use imputation. Proportions were compared using the χ2 or Fisher exact test. Continuous variables are presented as medians with interquartile ranges (IQRs) and compared using the Mann–Whitney test. After logarithmic transformation, plasma biomarkers were analysed using the two-way RM-ANOVA (repeated measures analysis of variance) to test for the fixed effects of time, treatment allocation and treatment by time interaction. Analyses were performed using STATA/SE 14.0 (StataCorp, College Station, Tex, USA) and SAS 9.4 (SAS Institute, Cary, NC, USA). Unless otherwise stated, a two-sided P value of 0.05 was used to indicate statistical significance. No adjustments were made for multiple comparisons.

Ethics approval and patient consent

The study protocol was approved by the ethics committees of Austin Health HREC/14/Austin/219) and each of the participating institutions. Written informed consent was obtained from all participants or their legal representatives.


Patient flow and recruitment

Between March 2015 and December 2017, we screened 985 patients for eligibility and randomly assigned 50 patients to one of the study groups (a recruitment rate of 1.5 participants per month). The trial was discontinued prematurely, on 31 December 2017, due to slow patient recruitment and insufficient funding. The most frequent reasons for exclusion were that the patient was already taking ASA (259/985, 26.3%) and that the patient had a high risk of bleeding or a coagulopathy (244/985, 24.8%). The patient flow through the study is summarised in Figure 1. Of the 50 patients assigned to a study group, one patient in each group revoked consent, leaving a total of 48 patients (23 in the ASA group and 25 in the placebo group) for the final intention-to-treat analysis. With 48 patients included, the study had > 80% power to demonstrate a 30% relative difference in the primary outcome.

Patient characteristics

Overall, 40 patients (83.3%) fulfilled the Sepsis-2 criteria and, of those, 18 (45.0%) presented with septic shock. There were no significant differences in median (IQR) Acute Physiology and Chronic Health Evaluation (APACHE) III scores (53 [47–62] v 52 [42–64], respectively), numbers of patients receiving vasopressor therapy or other baseline characteristics between the ASA and the placebo group. Baseline demographics and clinical characteristics of patients, according to treatment allocation, are shown in Table 1.

All 48 patients received the assigned trial regimen. The median (IQR) time from randomisation to commencement of the study drug was 59 (11–115) minutes in the ASA group and 30 (15–65) minutes in the placebo group (P = 0.75). Median duration of study treatment was 4 (2–6) days in the ASA group and 6 (3–7) days in the placebo group (P = 0.11) and patients received a median of 7 (4–12) and 10 (5–14) ASA and placebo doses (P = 0.11), respectively (Online Appendix, table S2).

Primary and secondary outcomes

Seven patients were discharged from the ICU in less than 48 hours. Therefore, 41 patients (85.4%; 19 in the ASA group and 22 in the placebo group) had one complete set of blood samples taken at baseline and one at 48 hours after assignment to a study group. In these patients, serum concentrations of IL-6, IL-10 and E-selectin at 48 hours were not significantly different between study groups (Table 2). Furthermore, we found no differences in C-reactive protein levels at 48 hours, ICU length of stay, duration of mechanical ventilation, ICU mortality and hospital mortality between the groups.

Anti-inflammatory and proresolving lipid mediators

For comprehensive lipid mediator analysis, we analysed 262 blood samples from 32 patients (15 in the ASA group and 17 in the placebo group) using mass spectrometry. We detected 81 different specific lipid mediators derived from three major bioactive essential fatty acid pathways (AA, EPA and DHA) (Online Appendix, figures S1 and S2).

In the AA pathway, low-dose ASA increased the concentration of the proresolving mediator 15-HETE (hydroxyeicosatetraenoic acid; P = 0.03 for treatment by time interaction), and reduced the concentration of the proinflammatory, cytochrome P-dependent mediators 17-HETE (P = 0.046), 18-HETE (P = 0.01) and 20-HETE (P = 0.01) (Figure 2).

In the EPA pathway, ASA transiently increased the concentration of the anti-inflammatory mediators 14,15-DiHETE (dihydroxyeicosatetraenoic acid; P = 0.04 for treatment by time interaction) and 17,18-DiHETE (P = 0.03) (Figure 2). We found no significant differences in the abundance of metabolites of DHA.

Safety outcomes

None of the participants experienced clinically relevant bleeding and there was no difference in proportion or number of participants receiving blood transfusions in the ICU (Table 2).


Key findings

In this multicentre randomised placebo-controlled trial of critically ill patients with SIRS, low-dose ASA did not reduce levels of classic inflammatory biomarkers, including IL-6, compared with placebo. Because the study was stopped prematurely, it was underpowered and therefore did not provide sufficient evidence to support or refute the primary hypothesis. However, we detected a wide array of complex lipid mediators of inflammation using mass spectrometry. ASA had a measurable effect on lipid mediators derived from the AA and EPA acid pathways, increasing the concentrations of individual proresolving and anti-inflammatory mediators, and decreasing the levels of specific proinflammatory and vasoregulatory mediators.

Relationship to previous studies

Observational data suggest that prehospital use of low-dose ASA is associated with reduced mortality in sepsis 6, 7, 23, 24, 25, 26  and patients with acute respiratory distress syndrome. 27, 28 However, in a recent multicentre randomised placebo-controlled trial including 390 patients, ASA (81 mg per day) failed to reduce rates of acute respiratory distress syndrome or levels of inflammatory cytokines. 29 Consistent with these findings, we found no difference in interleukin levels and other inflammatory markers in patients with SIRS.

Until recently, the presumed beneficial effects of ASA have been attributed to its anti-inflammatory actions, as it is a negative regulator of prostaglandin-induced cytokine production. 30, 31 However, ASA has also been shown to increase leukocytic cytokine production 32, 33 and to alleviate sepsis-induced immunosuppression. 34, 35 Yet, the biology of the resolution of inflammation appears more complex. In particular, the traditional concept of resolution of inflammation as a passive process was challenged by the discovery of proresolving lipids and their endogenous aspirin-triggered epimers. 36

Early detection and quantification of specific lipid mediators of inflammation using mass spectrometry has the potential to provide real-time information on changes in inflammatory status. However, although the technology for lipid mediator analysis has improved in terms of measurement specificity and sensitivity, 15, 21  the complexity of the biological plasma matrix still is a significant hurdle for the extraction and measurement of PUFA-derived lipid mediators, particularly SPMs, in low volumes of plasma. In this regard, using an optimised method for the extraction and analysis of lipid mediators, we were able to detect significant changes in key mediators of inflammation and resolution.

To our knowledge, only one study to date has measured SPMs in critically ill patients using liquid chromatography mass spectrometry techniques. 37 It was a single-centre, observational study, and included 22 patients with sepsis. The authors detected 30 bioactive lipid-derived mediators but did not assess the effects of ASA. Non-survivors had significantly higher levels of inflammation-initiating and proresolving mediators, including resolvin E1, resolvin D5 and protectin D compared with survivors. 37 Owing to the low mortality rate in our study, we could not replicate these findings.

In our study, low-dose ASA significantly increased the concentration of the proresolving mediator 15-HETE, increased the concentration of the anti-inflammatory mediators 14,15-DiHETE and 17,18-DiHETE, and decreased the level of the proinflammatory mediators 20-HETE, 17-HETE and 18-HETE. Such non-COX-mediated effects of ASA are increasingly recognised 38 and have been described in healthy volunteers. 39, 40, 41 Of note, 14,15-DiHETE and 17,18-DiHETE are the diol metabolites of the anti-inflammatory 14,15-EpETE (epoxyeicosatetraenoic acid) and 17,18-EpETE. 42, 43 Although we were unable to detect bioactive EpETEs in our study due to their instability, their diol counterparts are relatively stable and were transiently increased in ASA-treated patients. The presence of these metabolites indicates that the anti-inflammatory limb of the EPA pathway was activated in the ASA group, and to a greater extent than in the control group.

We did not detect any significant differences in levels of 14-HDHA (hydroxydocosahexaenoic acid; maresin precursor), 17-HDHA (protectin precursor), resolvin D1, resolvin E2 or resolvin E3. Besides, we did not detect measurable levels of ASA-triggered (acetylated-COX-dependent) lipid mediators such as 15(R)-HETE and 5(S),6(S)-lipoxin A4. These discrepancies to other studies may be due to several reasons. First, owing to rigorous exclusion criteria, our patients were less ill than those in other trials, as reflected by lower APACHE III scores. Second, we used a different dose of ASA (200 mg per day, as compared with 81 mg, 39, 40, 41 325 mg and 650 mg per day 39 in other studies), which may have been sufficient to induce some, but not all, of the previously reported effects. Third, we collected blood samples once per day, which may also explain the difference in our results, as PUFA metabolites are transiently formed, have limited stability and are present at very low concentrations. 15

Implications of the findings

The slow recruitment rate of 1.5 patients per month was primarily due to the exclusion of a significant proportion of patients with pre-existing ASA treatment or with evidence of coagulopathy. Of note, we observed no episode of considerable bleeding. Thus, our observations indicate that future trials of ASA in critically ill patients could, therefore, consider less stringent exclusion criteria. Following other reports, 44 our results indicate that the effect of ASA in patients with an acute inflammatory response may not be quantifiable by measuring levels of classic markers of inflammation and cytokines. Crucially, our findings indicate that it is feasible to detect and quantify numerous lipid mediators in a limited volume of human plasma, and that this can provide unparalleled insights into the complex system of PUFA metabolomes and the effect of ASA (or other interventions) on specific lipid mediators. Finally, they indicate that such biological effects may vary, depending on patient characteristics, ASA dose and timing, and the technology used to analyse lipid mediators, particularly SPMs.

Strengths and limitations

To our knowledge, this is the first randomised placebo-controlled trial providing a comprehensive assessment of the effects of ASA in critically ill patients with severe inflammation. All the enrolled patients received the allocated intervention of ASA or placebo, which were administered less than 1 hour after randomisation. We measured classic inflammatory markers such as C-reactive protein levels, white cell counts and inflammatory cytokine levels, and compared them to specific lipid mediators of inflammation, using advanced, cutting-edge technology. The multicentre, double-blind design of the trial increased the robustness of our findings.

Our study had, however, some limitations. The most important limitation was the small sample size due to slow recruitment and early termination, which decreased our power to detect a difference in the primary outcome. However, we were able to demonstrate apparent changes in key lipid mediators of inflammation in a small subset of patients. Our screening failure rate is in line with that of other larger randomised controlled trials of ASA. 29 We did not adjust for multiple comparisons, as we (and others 45 ) believe a higher type I error rate is acceptable for preliminary hypothesis testing in the setting of a pilot study. Another limitation is the exclusion of numerous patients due to rigorous exclusion criteria. However, several studies have shown that the effects of ASA differ in patients who have previously been exposed to the drug 46 and that the number of patients taking antiplatelet drugs at the time of ICU admission is substantial. 47 Therefore, excluding such patients seemed necessary and logical from a methodological point of view. The use of ASA in patients at risk of bleeding remains difficult to justify, although in some studies use of antiplatelet drugs was not associated with unfavourable bleeding but with improved outcomes, even in neurosurgical patients, 47 high risk trauma patients 48 and patients with severe non-variceal gastrointestinal bleeding. 49 Therefore, future studies may choose a more inclusive approach to enrolment.


In patients admitted to the ICU with SIRS, of which more than 80% were septic, low-dose ASA did not significantly alter IL-6 concentrations at 48 hours compared with placebo. However, ASA increased the concentration of specific proresolving and anti-inflammatory lipid mediators and decreased the concentration of proinflammatory mediators. Our findings provide additional data on the complex lipid mediator system in systemic inflammation, and support further investigations outside the more commonly measured cytokine system.
Acknowledgements: We thank the trial site research staff for their support with randomisation and data collection: Samantha Bates and Rebecca Morgan, Footscray Hospital, Western Health; and Tania Elderkin, University Hospital Geelong († Deceased March 2019).