Intraoperative dexamethasone alters immune cell populations in patients undergoing elective laparoscopic gynaecological surgery

Abstract

Background. Anaesthetists use dexamethasone principally for its anti-emetic effect. The purpose of this study was to characterize the effects of a single intraoperative dose of dexamethasone on cellular and metabolic components of the immune system in patients undergoing laparoscopic surgical procedures.

Methods. In this prospective double-blind trial, female patients undergoing elective major laparoscopic surgery were randomized to receive saline (Control group, n=16) or dexamethasone 4 mg (Dexamethasone group, n=16) i.v. after the induction of anaesthesia. Inflammatory markers and immune cell counts were examined at 24 and 48 h and 6 weeks after surgery. The changes from baseline preoperative values were compared between groups using a Mann–Whitney U-test, and linear mixed models were used to validate the findings.

Results. No differences in concentrations of serum glucose and interleukin-6 were observed between groups after surgery. The increase in C-reactive protein concentration at 24 h after surgery was greater in the control group [median (interquartile range), 33 (25–65) vs 17 (7–26) mg dl⁻¹; P=0.018]. Extensive changes in the counts of white cells, including most lymphocyte subsets, were observed 24 h after surgery, and dexamethasone appeared to attenuate most of these changes. Changes at 48 h and 6 weeks did not differ between groups.

Conclusions. In female patients undergoing elective laparoscopic gynaecological surgery, dexamethasone administration appears to attenuate inflammation and to alter immune cell counts at 24 h, with no effects identified after this time. The importance of these changes for postoperative immune function is unknown.

Trial registration. Australia and New Zealand Clinical Trials Registry (ACTRN12608000340336).

Surgical trauma is characterized by a tightly integrated sequence of neurohumoral and immunological processes.¹ When this is marked, it can manifest as a clinical entity called the systemic inflammatory response syndrome.² This inflammation is a double-edged sword. It is essential to facilitate removal of dead tissue and to promote remodelling followed by healing. However, the processes involved can produce a degree of immunosuppression and, in severe instances, immunoparesis. Recent advances in our understanding of the immune responses to tissue injury,² coupled with an increased understanding of how genomic determinants may direct inflammatory responses,³ has led to the theory that the inflammatory milieu in the postoperative period can be a harmful and potentially modifiable condition. The T-cell exhaustion and lymphocyte anergy that are often observed are believed to create an environment that renders the patient vulnerable to both infection and the recurrence of malignancy.^4,5 The importance of preoperative inflammation to outcome in patients undergoing surgery for malignancy is well established.^6,7 However, its role in altering healing and infective outcomes is less clear, but does appear to be important.^6,8,9 The role played by postoperative inflammation-related immune suppression in terms of postoperative infection risk and malignancy recurrence, and the role that anaesthesia techniques may play, are currently subjects of intense speculation and investigation.^5,10–12

Glucocorticoids are compounds with protean effects upon the human immune system,¹³ being widely used in the perioperative period for multiple benefits, including an improvement in analgesia in surgical patients,¹⁴ and decreased swelling in dental and maxillofacial surgery.¹⁵ The principal anaesthesia indication for their use is in the prevention of postoperative nausea and vomiting,¹⁶ and the synthetic glucocorticoid dexamethasone is both effective and recommended by international consensus for this indication.¹⁷ Glucocorticoids, however, have both genomic and non-genomic actions, which lead to complex effects upon cellular components of the immune system.¹⁸ They are, for example, pro-apoptotic in the treatment of haematological malignancy,¹⁹ but prolong neutrophil survival through an anti-apoptotic effect.²⁰ Although concern has been expressed that such effects in the perioperative period might increase the risk of postoperative infections, and wound infections in particular,²¹ the effect of intraoperative anti-emetic doses of dexamethasone on postoperative cellular immunity has not been elucidated.

The purpose of this study was to examine the effect of a single intraoperative dose of dexamethasone, in patients undergoing laparoscopic surgery, on the cellular components of the immune system in terms of the changes in peripheral cell counts up to 6 week after surgery. Gender appears to influence the pattern of inflammatory responses observed, and in order to limit this confounding we performed this study in female patients only.²² Our hypothesis was that dexamethasone would alter immune cell populations in the postoperative period.

Methods

This prospective, double-blinded randomized controlled trial received approval from the Human Research Ethics Committee of King Edward Memorial Hospital for Women (1554/EW). The trial was registered with the Australia and New Zealand Clinical Trials Registry (ACTRN12608000340336). Women were recruited at pre-assessment clinics and from preoperative wards. Patients eligible for selection comprised adult females, ASA I or II, of age 18–60 yr, and undergoing elective major laparoscopic gynaecological surgery expected to require at least 90 min of operative time, requiring a hospital stay to include at least the first postoperative night. Patients were excluded if they were currently or recently taking immunosuppressive agents, had known or suspected malignancy, hypertension, diabetes mellitus, a history of peptic ulceration, chronic pain syndrome requiring regular opioid consumption, a predicted requirement for i.v. patient-controlled analgesia, or known hypersensitivity to dexamethasone or granisetron. Patients were randomized before surgery, in a 1:1 ratio using a computer-generated random number sequence, and allocated to one of two groups using sealed opaque envelopes. A standardized anaesthetic technique was used for all participants. This comprised fentanyl 1 μg kg⁻¹ i.v. and midazolam 2 mg i.v. before induction, with target-controlled propofol infusion commenced at an effect site concentration of 6–8 μg ml⁻¹, and titrated thereafter to maintain anaesthesia. Entropy monitoring (Datex Ohmeda, GE Healthcare, Giles, UK) was used to ensure comparable and appropriate depth of anaesthesia, with propofol infusions targeted to state entropy of 40–55. Intraoperative opioids were administered at the attending anaesthetist’s discretion.

Airway management comprised tracheal intubation after the administration of a non-depolarizing neuromuscular blocking agent and pressure-controlled ventilation with an air–oxygen mix. An inspired oxygen concentration of 60% was targeted unless there was a clinical imperative to target a different concentration. Tracheal extubation occurred at the end of surgery after antagonism of neuromuscular block using neostigmine and glycopyrrolate. All patients received parecoxib 40 mg i.v., and postoperative analgesia comprised titration of i.v. fentanyl in the recovery room and thereafter paracetamol 1 g every 6 h, ibuprofen 400 mg orally every 8 h, and oxycodone 5–15 mg orally for breakthrough pain.

Each patient received granisetron 1 mg and an assigned study drug (saline or dexamethasone 4 mg; Control and Dexamethasone groups, respectively) i.v. after the induction of anaesthesia. The study drug was prepared by an observer who was not connected to the study and diluted to a total volume of 4 ml with 0.9% sodium chloride in an unmarked syringe. The patient, the patient’s anaesthetist, and all investigators were blinded to the study drug identity.

Interventions

After an overnight fast from solids but unrestricted water consumption, a baseline blood sample was obtained (T₀) upon insertion of a 20-gauge cannula for the induction of anaesthesia. Further blood samples were obtained at 24 h (T₁), 48 h (T₂), and 6 weeks after surgery (T₃). At each time point, 4 ml of blood in EDTA and a 4 ml clotted serum sample were collected. The serum samples were immediately centrifuged at 2500g for 20 min. One 400 µl aliquot was used for immediate analysis of serum glucose concentration, whereas the other 400 µl aliquot was stored at −80°C for later analysis of C-reactive protein (CRP) and interleukin-6 (IL-6) concentrations. At each time point, a serum glucose measurement, full blood count, serum CRP, serum IL-6, and lymphocyte subset analysis was performed as described under ‘Laboratory methods’. Pain and nausea verbal analog scale (VAS) values were collected upon discharge from the postanaesthesia recovery unit (PACU), and the surgical wounds were inspected on postoperative days 1 and 2. All patients also received a formal wound assessment at 6 week follow-up, including documentation of all infections occurring in that interval, at the surgical outpatient department.

Laboratory methods

Serum glucose measurement

This assay was performed on fresh serum within 1 h of collection using the Glucose Architect™ system (Abbott Diagnostics, Lake Forest, IL, USA).

Measurement of IL-6

Ineterleukin-6 concentrations were measured at T₀, T₁, and T₂, using enzyme-linked immunosorbent assay (ELISA) with a commercial ELISA assay (Human IL-6 Quantikine HS ELISA; R&D Systems, Inc., Minneapolis, MN, USA). Values are expressed as picograms per millilitre. The coefficient of variation of the assay for the entire sample batch was 10.8%.

Full blood count

A full blood count analysis was performed on the EDTA blood sample using a three-colour fluorescence multiangle polarized scatter separation instrument (CELL-DYN Sapphire™ system; Abbott Diagnostics).

Lymphocyte subset analysis

Assessment of T, B, and natural killer cells and lymphocyte count

EDTA samples were processed using a single-platform one-tube lysis-no-wash method for lymphocyte subset assessment. Fifty microlitres of each well-homogenized blood sample was mixed with 10 μl of Multitest 6-colour TBNK reagent (CD3 FITC/CD16⁺ 56 PE/CD45 PerCPCy5.5/CD4 PECy7/CD19 APC/CD8 APCCy7; Becton Dickinson catalogue no. 644611) in a Becton Dickinson (BD) TruCount Tube (catalogue no. 340334). Each tube contained a lyophilized pellet that released a known number of fluorescent beads during sample preparation. After thorough vortexing, the samples were incubated for 15 min at room temperature in the dark. FACSLyse Red Cell Lysis Buffer (BD 349202) 0.5 ml was added to each tube and incubated for 15 min at room temperature in the dark before processing for flow cytometry analysis, with extra vortexing immediately before acquisition on a BD FACSCanto II flow cytometer (BD Biosciences, San Jose, CA, USA). BD FACSCanto clinical software was used to calculate the absolute lymphocyte count and the percentages and absolute counts of total T cells (CD3⁺), CD4⁺ T cells (CD3⁺ CD4⁺), CD8⁺ T cells (CD3⁺ CD8⁺), B cells (CD19⁺), and natural killer (NK) cells (CD3⁻ CD16/56⁺).

Assessment of memory B cells

EDTA samples were processed using a one-tube lysis-wash method for memory B-cell assessment. Two hundred and fifty microlitres of each well-homogenized blood sample was mixed with 3.0 ml lysis buffer (BD PharmLyse Red Cell Lysis Buffer; BD 555899; prepared as per manufacturer instructions) and mixed for 10 min on a Unico TTR200 Test Tube Rocker followed by three cell washes in phosphate-buffered saline (PBS) to remove lysed red cells. The washed leucocytes were resuspended in 250 μl of buffer. Fifty microlitres of cell suspension was mixed with IgM FITC (BD 555782; 10 μl)/CD19 PECy7 (BD 341093; 5 μl)/CD27 APC (BD 337169; 2.5 μl)/CD45 APCH7 (BD 641399; 2.5 μl ), mixed by vortexing and incubated for 15 min at room temperature in the dark, followed by the addition of 0.5 ml buffer before processing for flow cytometry analysis, with extra vortexing immediately before acquisition on a BD FACSCanto II flow cytometer using BD FACSDiva 6.0 software to determine the following: (i) IgM memory B cells, defined as CD19⁺ CD27⁺ IgM⁺, reported as a percentage of CD19⁺ B cells; (ii) switched memory B cells, defined as CD19⁺ CD27⁺ IgM⁻, reported as a percentage of CD19⁺ B cells; and (iii) naive B-cells, defined as IgD⁺/CD27⁻.

Statistical analysis

A power analysis, based on pilot data of differences in total neutrophil counts, recommended that a total of 16 patients per group should be recruited to provide a power of 0.9 to detect a difference between group mean neutrophil counts of 0.7×10⁹ litre⁻¹ with a pooled sd of 0.5 at a significance level of 0.05; this equates to a 23% difference between means. There were a number of significant differences in some parameters between groups at baseline (before surgery and the study intervention (either drug or control); Supplementary material, Table S1), and we therefore decided that changes from baselines values would be a more robust measure to evaluate. Changes in cell counts and concentrations of serum glucose, CRP, and IL-6 were calculated between baseline values at T₀ and those at T₁, T₂, and T₃. These changes were compared between the Control and Dexamethasone groups to identify the time points when changes from baseline differed between the two groups. We believe this approach to be preferable to the use of a linear mixed model to examine whether the two groups were different in an overall fashion during the whole 6 week study period because this would not permit identification of the time at which significant changes occur. Nonetheless, we used a linear mixed model (as below) to validate the findings of the univariate analysis. Categorical variables were analysed using the χ² test. Testing with the Kolmogorov–Smirnoff test indicated that the continuous variables were not normally distributed, and non-parametric testing was used for all continuous variables. Differences between the two groups (Control and Dexamethasone) in terms of changes in each of the measured continuous parameters from the baseline value used the Mann–Whitney U-test for independent samples. A Bonferroni correction was applied to account for comparisons of changes at multiple (n=3) time points in the univariate analyses. A corrected P-value of <0.05 was taken to indicate a significant finding. Finally, a sensitivity analysis using a linear mixed model with scale identity correlation matrices was used to confirm some of the significant differences between the Dexamethasone and Control group. All analyses were performed using SPSS^® for Windows Version 23 (IBM Analytics, IBM Corporation, Armonk, NY, USA).

Results

Between February 2010 and March 2013, a total of 32 patients were recruited to the study, and the modified CONSORT diagram (Fig. 1) represents the patient involvement. One patient from the control arm was withdrawn from the analysis because of failure to comply with the study protocol, leaving data on 31 patients to be analysed. Two patients in each group did not complete cell count testing at 48 h, and five patients in the Control group and four in the Dexamethasone group did not complete cell count testing at 6 weeks after surgery. The two groups were well matched in terms of patient and procedural characteristics (Table 1). Nausea scores on admission to the PACU were comparable, and no patients in either group experienced vomiting or dry-retching in the PACU. No differences were observed between groups in terms of pain scores, either at rest or on coughing, in the PACU median VAS at rest 1, interquartile range (IQR) 0–5 in the Control group vs median (IQR) 3 (0–4) in the Dexamethasone group, P=0.74; median (IQR) VAS on coughing 3 (0–5) in the Control group vs 4 (0–6) in the Dexamethasone group, P=0.95. No difference was observed between the groups in terms of postoperative pyrexia or wound infections up to 6 weeks after surgery. Four patients in the Control group developed wound, urinary tract, or vaginal infections, compared with five patients in the Dexamethasone group (P=0.72). More patients in the Dexamethasone group received antagonism of neuromuscular block than those in the Control group [16 (100%) vs 11 (73%), P=0.043].

The changes in concentration of serum glucose and IL-6 from baseline (T₀) were comparable between both groups at T₁, T₂, and T₃ (Table 2). The increase in CRP concentration observed at 24 h (T₁) in the Control group was significantly reduced in the Dexamethasone group [median (IQR) 33 (25–65) vs 17 (7–26) mg litre⁻¹, P=0.003]. No difference was observed, however, at T₂ and T_3. The counts of all cells examined were significantly different in the Dexamethasone group at 24 h (T₁) apart from platelet, basophil, and NK cell counts (Table 2). The increases in total white cell and neutrophil counts observed at 24 h (T₁) in the Control group were greater in the Dexamethasone group, [median (IQR) 1 (0.2–2.7) vs 3.8 (2.2–5.6)×10⁹ litre⁻¹, P=0.003; and 1.5 (0.6–3.2) vs 3.8 (2.2–5.6)×10⁹ litre⁻¹, P=0.021 respectively; Table 2]. The observed decrease in eosinophil count at 24 h (T₁) was much less marked in the Control group compared with the Dexamethasone group [median (IQR) −0.03 (−0.08 to − 0.01) vs −0.13 (−0.22 to − 0.08)×10⁹ litre⁻¹, P=0.012; Table 2]. The decrease in lymphocyte count at 24 h (T₁) was much greater in the Control group than in the Dexamethasone group [median (IQR) −6.3 (−0.8 to − 0.01) vs −0.01 (−0.3 to 0.3)×10⁶ litre⁻¹, P=0.009], and this pattern was reflected across all lymphocyte subsets (Table 2). The T- and B-cell counts were more markedly reduced in the Control group [median (IQR) −471 (−678 to −214) vs −17 (−184 to 169)×10⁶ litre⁻¹, P=0.009; −45 (−108 to 7) vs 57 (33–125)×10⁶ litre⁻¹, P=0.003]. All of these significant changes at 24 h (T₁) had resolved by T₂.

Sensitivity analysis

Linear mixed models confirmed the significant differences in CRP concentrations and lymphocyte counts between patients who received dexamethasone and placebo (Supplementary material, Tables S2 and S3).

Discussion

This is the first in-depth examination of the effects of a commonly used i.v. anti-emetic dose of dexamethasone on inflammation and cellular components of the immune system in the postoperative period. The findings indicate that dexamethasone 4 mg administered during surgery decreased inflammation, as detected using CRP concentrations, and attenuated the changes in counts of all cells, apart from basophils and NK cells, that were seen in the Control group. The inflammation and cellular effects were seen only at 24 h after surgery, having completely resolved by 48 h. Postoperative serum glucose concentrations were unaffected by dexamethasone administration. These findings indicate that a seemingly small dose of dexamethasone has detectable effects upon the cellular components of the postoperative inflammatory response. It is uncertain whether these transient changes have relevance for inflammation-related outcomes in surgical patients.

Neutrophils and monocytes are key cellular components of the innate immune system, and they interact together very effectively to mediate cytotoxic pathogen killing,²³ with neutrophils in particular playing a crucial role.²⁴ Changes in cell populations and signalling pathways in the perioperative period are believed to contribute to postoperative immunosuppression, particularly changes observed in monocyte pathways.²⁵ The immune responses of NK cells and T lymphocytes (particularly cytotoxic CD8 T cells) are essential to control infection and to identify and eradicate premalignant change within host cells.^26,27 Glucocorticoids have a profound influence on cells of the immune system, particularly T cells, where their principal action is to induce a lymphopenia through an increase of apoptosis.²⁸ An increase in white cell count and a decrease in lymphocyte counts 24 h after surgery has been described previously . Surgery is known to induce leucocytosis and lymphopenia at 24 h after surgery,^29,30 which usually resolves within 3–7 days.^31,32 The extent of lymphopenia is proportionate to the extent of surgery³³ and is far less marked with laparoscopic than with open procedures.³⁴

Persistent leucocytosis and failure to normalize lymphopenia after trauma are associated with adverse outcome.³⁵ Our results demonstrate clear and substantial effects of dexamethasone on changes in neutrophil and lymphocyte counts 24 h after surgery. The neutrophilia at 24 h after surgery in the control group was augmented by dexamethasone, which is consistent with the findings of others.^36,37 Likewise, the exaggerated reduction in eosinophil numbers has also been demonstrated previously.³⁸ All lymphocyte subsets were decreased in the Control group at 24 h, with some recovering at 48 h. Dexamethasone prevented this decrease for all subsets, with the exception of NK cells, which were decreased to a similar extent in both groups. We would have expected to observe a greater lymphopenia in the Dexamethasone group, where the effect of surgery and glucocorticoid use would be expected to be synergistic. There is no obvious mechanistic explanation for this finding, but we have previously demonstrated in volunteers that dexamethasone-induced lymphopenia has resolved at 24 h post dose (unpublished observations, Tomás B. Corcoran, Lisa M. Hill, Evelyn M. Fletcher, Pui S. Loh, Martyn French, Kwok M. Ho). Natural killer cells have previously been shown to increase in the first 24 h after surgery, so we cannot account for the discrepancy³⁰ with our data.

The causes of these changes in cell populations and their relevance for perioperative immune competence are speculative. We may simply have observed the known effect of glucocorticoid administration on cell populations, with these actions being nonetheless detectable despite a background of surgery-related inflammation. It is also possible that we may have identified an effect of dexamethasone on some component of the perioperative response that controls cell population changes. Certainly, the time course is consistent with the known pharmacology of dexamethasone.³⁶ Although changes in cell count do not necessarily reflect a change in immune function,³⁹ glucocorticoids have been shown to augment the anti-inflammatory function of monocytes,⁴⁰ aiding the resolution of inflammation.⁴¹ The increase in neutrophil counts that we observed at 24 h might reflect demargination of neutrophils from the endothelium and interference by dexamethasone with L-selectin expression.³⁶ These mechanisms might be associated with an increased risk of infection in the postoperative period, because neutrophils that cannot marginate to sites of inflammation are considered ineffectual for microbe clearance. The effect of dexamethasone on CRP concentrations has been described previously,⁴² but the effect on IL-6 was not as marked as others have shown,³⁶ perhaps because of the relatively limited inflammatory response in our patients and the short time frame after surgery in which peak concentrations are observed.⁴³ Hence, the sampling time points may have missed a difference that occurred between groups if the concentration peaked before the 24 h sampling time point.

The impact of perioperative inflammation on surgery-related outcomes is emerging as a significant and potentially modifiable factor.⁶ The use of neutrophil-to-lymphocyte ratios and the Glasgow Prognostic Score, both of which quantify measures of preoperative inflammation, can predict outcomes of surgery for resection of malignant tumours,⁷ including malignancy recurrence and infection.⁴⁴ The role that postoperative inflammation plays has been less well studied because of the large variations, influenced by surgical procedure, genetic, and co-morbidity-related factors. However, the pattern and magnitude of post-trauma and post-surgery inflammatory responses appear to be strongly predictive of adverse outcomes, with more intense, prolonged, and widespread inflammation having the most deleterious effect.^3,45–47 Postoperative infection is an outcome of particular concern that is influenced by perioperative inflammation, but its importance has been appreciated only recently. It affects 11.9% of patients undergoing surgical procedures, with an in-hospital mortality of 15.5%.^48,49 It prolongs hospital stay, has a long-lasting impact on patient outcome,⁵⁰ and is 10 times more common than myocardial ischaemia in high-risk cohorts.⁵¹ Any intervention that modifies the occurrence of postoperative infection is therefore likely to have a significant impact for a large number of patients. The less intense inflammatory response associated with laparoscopic surgical approaches is considered to be an important factor in improved outcomes in these patients compared with open surgical procedures.⁵² The role that anaesthesia techniques and specific agents can have on the postoperative inflammatory response, with the potential to confer outcome benefit, is becoming a research focus.⁵³ The surgical stress response is partly driven by the endocrine system, and dexamethasone may alter not only the pattern of endocrine responses, but also postoperative analgesia and opioid requirements.^54,55 Although a small dose of intraoperative glucocorticoid may not be considered sufficient to alter the inflammatory response, our study shows that we can identify an effect upon specific key cellular components of perioperative inflammation. Whether such effects are important remains uncertain.

The relevance of the influence of specific anaesthesia techniques and interventions, including dexamethasone administration, on important outcomes is the focus of much attention and current research.^10,46,56 The authors of multiple meta-analyses have asserted the safety of perioperative glucocorticoids, even in very large doses,^57–59 but these give an unwarranted impression of the safety of dexamethasone. Meta-analyses are unreliable, with a low positive predictive value for outcomes subsequently examined in large randomized controlled trials,⁶⁰ and at best they should generate hypotheses intended to be tested in large, properly blinded, appropriately powered multicentre randomized controlled trials.⁶¹ The majority of trials in these meta-analyses are small, single-centre trials, which did not include surgical site infection as a primary or even a secondary end point and which did not carry out postoperative surveillance. A large randomized controlled trial of high-dose dexamethasone in cardiac surgery (1 mg kg⁻¹) identified a reduction in infective complications in the postoperative period.⁶² A smaller trial in non-cardiac surgery patients, using a smaller dose (∼16 mg), did not identify any effect on infection.⁴² Although the latter was a well-performed study, it was underpowered, and at least 3000 patients (rather than the 381 enrolled) would have been required to produce a 25% reduction clinically important effect.⁶³ A further randomized trial (n=555) by the same group failed to demonstrate an association between dexamethasone use and infection.⁶⁴ Given the frequency of administration of anti-emetic dexamethasone to surgical patients,⁶⁵ the influence of such a dose in terms of postoperative immune integrity is extremely pertinent to these discussions and remains a concern for many.^21,66 Our data indicate that this dose has perceptible effects on the inflammatory response and cellular counts after surgery. A previous study has examined the effect of preoperative administration of dexamethasone into the epidural space on perioperative inflammation and lymphocyte counts.⁶⁷ That study is not directly comparable to ours, but similar findings in relationship to lymphocyte count and CRP concentrations were observed.

We must acknowledge several weaknesses in this study. Firstly, as alluded to above, we sought simply to measure alterations in different subsets of immune cell counts, and we cannot therefore extrapolate our results as alterations in immune function. Secondly, this is a small study in patients undergoing laparoscopic surgery, with a limited inflammatory response. Hence, whether the use of such a small dose of dexamethasone in patients undergoing a large inflammatory response, such as that associated with open abdominal surgery, would produce similar alterations of cell populations cannot be commented upon. Thirdly, there were significant baseline differences between the groups in terms of cell counts. We therefore sought to examine changes from the baseline value rather than the absolute values at each time point. The use of a linear mixed model, which can control for such changes, confirmed the significance of the findings for both CRP concentrations and total lymphocyte counts. The number of comparisons performed, although stipulated a priori, might have inflated the type I error. However, the Bonferroni-corrected P-values in most instances indicated highly significant results, and the linear mixed model confirmed the differences between dexamethasone and placebo in our sensitivity analyses. Also, we chose to compare changes from baseline in each individual rather than absolute values for groups, at each time point. We believe this approach was preferable because it accounted for the inter-individual variability in these parameters and permitted identification of the exact time point when any differences were detected. Again, the linear mixed models confirmed the significant findings. Fifthly, the losses to follow-up at 48 h and 6 weeks might have introduced bias into the results for these periods. It may have accounted for the failure to identify differences at these times. Finally, it is possible that the differences between the groups in terms of smoking status and the antagonism of neuromuscular block might have influenced the results observed, particularly given the effects of smoking on T-cell responses.⁶⁸ These would need to be taken into consideration in any larger study.

In conclusion, this small study indicates that a commonly used anti-emetic dose of dexamethasone can decrease inflammation and produce detectable alterations in immune cell counts in surgical patients at 24 h after undergoing laparoscopic surgery with a small inflammatory stimulus. These differences had receded 48 h after surgery. Dexamethasone appeared to attenuate the surgery-related changes in counts of all leucocytes, with the exception of basophils and NK cells. The largest effects were observed in lymphocyte populations. It is uncertain whether these changes would be detected in the setting of the strong inflammatory responses after major surgery. Whether these changes would translate to alterations in immunocompetence or susceptibility to infection is unknown. We can, however, say that the administration of a low dose of dexamethasone to patients during elective laparoscopic gynaecological surgery decreases the inflammatory response and produces detectable but transient alterations in most immune cell populations.

Authors’ contributions

Study design: T.C., M.P., M.F.

Design of the substudy: K.M.H.

Composition of the study protocol: T.C., M.F.

Patient recruitment: N.A.M., D.L.

Establishment of laboratory assays: M.F.

Data analysis: T.C., M.P., K.M.H.

Wrote the manuscript: T.C., M.P., N.A.M., D.L., M.F., K.M.H.

Supplementary material

Supplementary material is available at British Journal of Anaesthesia online.

Acknowledgements

The authors would like to acknowledge the assistance of Mr Rom Krueger, Senior Scientist at the Flow Cytometry Laboratory, Royal Perth Hospital, and the research coordinators at King Edward Memorial Hospital for women, Subiaco, Western Australia.

Declaration of interest

T.C. is a member of the executive committee of the ANZCA Clinical Trials Network (Australia and New Zealand College of Anaesthetists (ANZCA) CTN). M.P. is an editor of Anaesthesia and Intensive Care and is on the Editorial Board for International Journal of Obstetric Anesthesia, Obstetric Anesthesia Digest, Anesthesiology, and Pain Medicine. The other authors declare no conflicts of interest.

Funding

Royal Perth Hospital (departmental funds); Health Department of Western Australia (Raine Foundation Clinical Practitioner Fellowship to T.C.); University of Western Australia (T.C.).

References