Anesthetic Propofol Enhances Plasma γ-Tocopherol Levels in Patients Undergoing Cardiac Surgery

All chemicals were obtained from Sigma Chemical Co. (Milan, Italy) if not otherwise specified. Propofol was administered using a commercially available target-controlled infusion system (Diprifusor; AstraZeneca S.p.A, Milan, Italy).

Patients scheduled to undergo elective cardiac surgery with cardiopulmonary bypass (CPB) were enrolled in this prospective, randomized, controlled study after approval from the institutional review board (Centro Cardiologico Monzino, Milan, Italy) and written informed consent. Inclusion criteria were stable angina, left ventricular ejection fraction greater than 40%, and age 60–80 yr. Exclusion criteria were aortic valve stenosis, angina on arrival in the operating room, and acute myocardial infarction during the past 7 days. Preanesthetic medication included morphine (0.1 mg/kg) and atropine (0.07 mg/kg), administered intramuscularly 1 h before surgery. Intravenous cefazolin (30 mg/kg) was administered before instrumentation. Monitoring was performed by five-lead electrocardiography with continuous ST-segment analysis (leads II and V5), radial artery catheter, pulse oximetry, and triple-lumen catheter or pulmonary artery catheter (Swan-Ganz) inserted through the right internal jugular vein. Patients were assigned to one of two groups by means of a random computer-generated list. Anesthesiologists were unaware of the due treatment until the morning of surgery, after patient enrollment. In the propofol group, anesthesia was induced by propofol and remifentanil, simultaneously administered according to a target-controlled infusion protocol (2.5–4 μg/ml propofol, 10–12 ng/ml remifentanil). Anesthesia was maintained with propofol (1.5 μg/ml) and remifentanil (8–12 ng/ml), decreasing to 1.2 and 4 ng/ml, respectively, during CPB. In the control group, anesthesia was induced by thiopental (4–6 mg/kg) and maintained with inhalation of sevoflurane (1–2%) in an oxygen–air mixture and remifentanil (8–12 ng/ml, target-controlled infusion) for the entire procedure, with the exception of CPB, during which midazolam (50–100 μg · kg⁻¹· h⁻¹) was administered intravenously and remifentanil was reduced (4 ng/ml). During the procedure, the control group also received the propofol vehicle (100 ml Intralipid®; Fresenius Kabi Italia S.r.l., Verona, Italy), because it contains tocopherols. In both groups, tracheal intubation was facilitated by succinylcholine (0.1 mg/kg) and pancuronium (0.1 mg/kg). Lungs were ventilated at normocapnia in an air–oxygen mixture (fraction of inspired oxygen [Fio²]= 0.5). Patients received 300 U/kg heparin before cannulation, and activated clotting time was kept above 450 s during the extracorporeal procedure. CPB was performed with a roller or centrifugal pump and a hollow-fiber oxygenator in mild hypothermia (32°–34°C). The circuit was primed with 1.0 l Normosol® (Abbott Laboratories, North Chicago, IL), 5% glucose (500 ml), 18% sodium bicarbonate solution (100 ml), and 18% mannitol solution (100 ml), given just before the opening of the aortic cross clamp. Blood flow in the bypass was titrated to ensure a mean arterial blood pressure between 55 and 75 mmHg (at least at 2.4 l · min⁻¹· m⁻²). Additional boluses of norepinephrine and nitroglycerin were used to maintain the pressure within the range. Hematocrit was kept between 18% and 25%. The myocardium was protected by administration of cold antegrade and retrograde multidose blood cardioplegia. At the end of the procedure, all patients received an intravenous bolus of morphine (0.1 mg/kg) and were transferred to the intensive care unit without reversal of muscle relaxation. Here, both groups were ventilated in mandatory minute ventilation (Dräger Evita 4; Drägerwerk AG & Co., Lübeck, Germany) and extubated when they fulfilled the following criteria: adequate response to verbal stimuli, body temperature greater than 36°C, blood loss less than 100 ml/h, hemodynamic stability, respiratory rate between 10 and 15 breaths/min, partial pressure of carbon dioxide less than 45 mmHg, and arterial oxygen saturation greater than 96% with an Fio²of 0.5. Postoperative analgesia was obtained by remifentanil perfusion (0.1–0.25 μg · kg⁻¹· min⁻¹) until extubation, and with tramadol if required (visual analog score >3).

Blood was collected from the radial artery into pyrogen-free EDTA before induction of anesthesia (T⁰, baseline), 30 min after the beginning of CPB (T¹), after protamine administration and on-pump weaning (T²), at arrival in the intensive care unit (T³), and 24 h after anesthesia induction (T⁴). Plasma was obtained by centrifugation at 4°C and stored at −80°C until analysis. For studies with cultured WB, additional blood samples from a subgroup of patients (14 in propofol group and 15 in control group) were collected in heparin at T⁰and T⁴.

Aliquots of WB were cultured for 24 h at 37°C with or without 10 μg/ml bacterial lipopolysaccharide (Escherichia coli  0111:B4). Acetylsalicylic acid (30 μm) was added to samples to prevent COX-1 activity. After centrifugation (700g  for 15 min), prostaglandin E²(PGE²) and IL-10 levels were determined by commercially available enzyme immunoassays (Prostaglandin E²EIA Kit-Monoclonal; Cayman Chemical, Spi-bio, Montigny le Bretonneux, France, and Endogen Human IL-10 ELISA Kit; Pierce Biotechnology Inc., Rockford, IL, respectively). All samples were determined in duplicate.

Peripheral blood mononuclear leukocytes (PBMLs) were isolated by density gradient centrifugation over Ficoll-Paque (Amersham Biosciences Europe GmbH, Milan, Italy), as previously described,31from aliquots of blood collected in heparin at T⁰and T⁴from a randomly selected subgroup of 11 patients. After isolation, PBMLs were rinsed with phosphate-buffered saline supplemented with 0.5% bovine serum albumin and 2 mm EDTA and were counted. The pellet was stored at −20°C until analysis. For in vitro  studies, PBMLs obtained from venous blood of healthy donors were suspended in medium-199 supplemented with 10% human AB serum (5 × 10⁶/ml) and plated in tissue culture dishes for 90 min at 37°C.32Adherent cells were 85–90% monocytes, as determined by nonspecific esterase staining. Monocytes were incubated for 18 h with α- or γ-T, dissolved in ethanol. Lipopolysaccharide (100 ng/ml) was then added, and incubation was continued for further 24 h. For the evaluation of COX-2 activity in terms of PGE²synthesis, monocytes were washed and incubated for 30 min in Hank’s buffer (pH 7.4) containing 1 mg/ml bovine serum albumin and 10 μm arachidonic acid. Adherent monocytes were then harvested in lysis buffer, pH 6.8, and Western blot analysis was performed as described.32 

α-Tocopherol and γ-T concentrations were measured by high-performance liquid chromatography.33Plasma samples and PBMLs, resuspended in lysis buffer, were extracted with organic solvents, injected into a Discovery C18 reversed-phase column (250 × 4.6 mm, 5 μm; Supelco/Sigma-Aldrich, Bellefonte, PA), and eluted with methanol (100%), as mobile phase.

Plasma total malondialdehyde was detected by high-performance liquid chromatography with a method modified from Carbonneau et al . 34Briefly, plasma (200 μl) was subjected to alkaline hydrolysis (1 m NaOH, pH 13) at 60°C for 1 h. After acidification (2 m HClO⁴, pH 1) and centrifugation (13,000g  for 10 min), supernatants were treated with 2-thiobarbituric acid (0.2% wt/vol) for 1 h at 100°C. After cooling, the samples were centrifuged at 13,000g  for 10 min. Supernatants were injected into XBridge C18 reversed-phase column (150 × 4.6 mm, 5 μm; Waters Co., Milan, Italy) and eluted with 10 mm phosphate buffer (30% MeOH, pH 7), as mobile phase, at a flow rate of 1 ml/min. Malondialdehyde levels were measured by a fluorometer (Jasko FP1520; Tokyo, Japan: λ^Ex515 nm, λ^Em553 nm). The peak areas were integrated using commercial software (EZStart; ESA Biosciences, Chelmsford, MA). The sample concentrations were calculated from calibration curves using 1,1,3,3-tetraethoxypropane as standard. Calibration of the analytical procedure gave a linear signal over the malondialdehyde range of 0.25–4 μm (r = 0.9992), with a quantification and detection limits of 0.15 and 0.05 μm, respectively. The intraassay and interassay coefficient of variations were 2.4% and 9.2%.

Plasma individual antioxidant capacity (IAC), a parameter that provides a measure of the overall protection against oxidative damage, was measured by a commercially available spectrophotometric assay (OXY-Adsorbent Test Diacron®; Diacron International, Grosseto, Italy). Samples were tested for their capacity to neutralize massive oxidation by hypochlorous acid, and IAC values were determined by reading the absorbance at 505 nm.

Interleukin-10 levels were determined in plasma by a commercially available enzyme immunoassay, as described for WB experiments. All samples were determined in duplicate.

Sample size was calculated from a pilot study on the primary outcome (effect of propofol on γ-T levels in patients who underwent cardiac surgery). Twenty-one subjects per group allowed a 90% statistical power to detect a between-group difference of 0.15 μg/ml in peak levels of γ-T, with an α error of 0.05. Results are expressed as mean ± SD, unless otherwise stated. Variables whose distribution was markedly skewed (IL-10 levels, both in plasma and in WB, PGE²levels in WB, European System for Cardiac Operative Risk Evaluation [EuroSCORE], duration of intubation, and blood loss) were log¹⁰transformed before analyses. All other variables were nearly normally distributed.

Baseline data were compared between the two groups by Student t  test or Fisher exact test, when appropriate. To minimize the intersubject variability, individual variations of analytes in plasma, during and after surgery, were analyzed as delta versus  baseline. However, absolute values were also analyzed to check the consistency of the results. The time course of the analytes was compared between groups by a repeated-measures analysis of covariance with a group × time factorial design. The group effect, the time effect, and the group × time interaction were assessed. As post hoc  analyses, we tested the within-group difference of each time point versus  baseline, controlling for multiple testing by the Bonferroni method. The association between PGE²production and γ-T was assessed by the Pearson correlation on log-transformed data. The in vitro  effect of increasing doses of α- or γ-T on PGE²synthesis by monocytes was assessed by analysis of covariance with isoform (α- or γ-T) × concentration factorial design. The different effect of the two isoforms was assessed by testing the isoform × concentration interaction. Analyses were performed using SAS statistical package version 8.2 (SAS Institute, Cary, NC). All P  values reported are two-sided and are considered statistically significant at less than 0.05.

Sixty-five patients undergoing cardiac surgery were assessed for eligibility (fig. 1). Two patients refused to participate and 19 did not meet the inclusion criteria. After randomization, the baseline blood sample from one patient was unusable, and therefore, it was excluded from data analysis. Patients assigned to the propofol or control group were comparable for age, body mass index, risk factors, and type of surgical procedures (table 1). A nonsignificant imbalance was observed for sex distribution. The biochemical values in plasma did not significantly differ between the two groups (table 2). The two treated groups were also comparable for PGE²and IL-10 levels measured in WB cultured with or without lipopolysaccharide for 24 h (table 3). Changes in blood cellular profile assessed at baseline and 24 h after anesthesia induction were similar in the two groups (P = 0.22, 0.34, and 0.67 for leukocytes, erythrocytes, and platelets, respectively). No significant differences between the two groups were found in the perioperative course or in the outcomes, except for a borderline statistical difference in the time of intubation (table 4). No patients died or experienced major complications (myocardial infarction, acute respiratory distress syndrome, cerebral accidents, or cardiogenic or septic shock; table 4).

The α- and γ-T levels in plasma measured before (T⁰), during (T¹, T²), and after surgery (T³, T⁴) are shown in figures 2A and B. α-T levels decreased significantly in both groups 30 min after ischemia induction and on pump start (T¹), and remained low during and after surgery (T², T³) and the first postoperative day (T⁴). No significant difference between groups was observed (fig. 2A). In contrast, the two groups differed markedly in γ-T levels, which progressively and significantly increased in the propofol group, returning to basal values 24 h after surgery (fig. 2B). In the control group, γ-T levels remained roughly unchanged during and after surgery.

Because an association between sex and tocopherol levels have been reported,33the time course of tocopherols was reanalyzed, after adjusting for sex, to control for the excess of women in the propofol group. Similar results for the treatment effects were obtained (F = 0.72, P = 0.40 for α-T and F = 41.4, P ≤ 0.0001 for γ-T).

γ-Tocopherol levels measured at T⁴in PBMLs from a subgroup of patients were higher in the propofol group (1.42 ± 0.72 ng/10³PBMLs [mean ± SD], n = 4 in propofol group and 0.67 ± 0.37, n = 7 in control group; P = 0.045). An analogous significant difference between groups was not observed in α-T levels (7.15 ± 6.52 ng/10³PBMLs in propofol group and 5.00 ± 4.23 ng/10³PBMLs in control group; P = 0.51). The small sample size, however, limits the statistical power of this comparison.

The different effect of the two anesthesia regimens on γ-T levels in plasma was not mirrored by changes in the oxidative status (figs. 2C and D). Measurement of malondialdehyde, as an index of lipid peroxidation, showed no appreciable difference between the two groups of treatment. Malondialdehyde levels markedly increased during surgery with a progressive decline after ischemia induction and on-pump beginning (T¹) in both groups (fig. 2C). As a reflection of the increase in malondialdehyde, IAC levels significantly declined during surgery, again with no difference between the two groups (fig. 2B). After surgery, IAC levels remained lower than at baseline, although a trend toward baseline values was detected (fig. 2D).

Changes of plasma IL-10 levels in the propofol and control groups before, during, and after surgery are shown in figure 3. IL-10 increased during surgery, with no significant difference between groups. A complete return to baseline values was observed at T⁴.

The time courses of α- and γ-T, malondialdehyde, IAC, and IL-10 were also analyzed using actual instead of delta values, and superimposable results were obtained: Specifically, the effect of propofol on γ-T levels was associated with a P  value of 0.005.

Prostaglandin E²production, as an index of COX-2 activity, was measured in WB collected at T⁰and T⁴and then cultured for 24 h either in the absence or in the presence of 10 μg/ml lipopolysaccharide. The reduction of lipopolysaccharide-induced PGE²production, relative to preoperative levels, was significantly more marked in the propofol group than in the control group (fig. 4A). In contrast, no significant difference between groups was detected when IL-10 levels were measured in the same experimental system (fig. 4B).

Of interest, significant negative correlations between PGE²production induced by lipopolysaccharide in WB at T⁴and γ-T levels were found, both in plasma at the end of surgery (T³; r =−0.44, P = 0.04) and in PBMLs (r =−0.65, P = 0.04).

To assess the effect of α- and γ-T on COX-2 activity, we performed in vitro  experiments (n = 10) on adherent monocytes exposed to lipopolysaccharide (100 ng/ml) in the presence or absence of α- or γ-T. Lipopolysaccharide markedly increased PGE²synthesis from 0.30 ± 0.34 to 4.8 ± 1.2 ng/ml. The increase was dependent on COX-2 induction (figs. 5A and B). γ-T, preincubated with monocytes for 18 h before lipopolysaccharide addition, concentration-dependently reduced PGE²levels (fig. 5A), without affecting COX-2 expression (fig. 5B). The effect of the same concentrations of α-T on COX-2 activity was less apparent (fig. 5A): The slopes of COX-2 activity inhibition versus  tocopherol concentration were significantly different for the α- and γ-T isoforms (P = 0.016 for the interaction term).

This study shows, for the first time, that propofol selectively enhances the levels of γ-T in the plasma of patients undergoing cardiac surgery and that this effect may account for the reduction of PGE²synthesis observed in their WB cultured with lipopolysaccharide. This finding is reinforced by the in vitro  observation that γ-T inhibits lipopolysaccharide-induced PGE²synthesis in human adherent monocytes.

The increase of γ-T after propofol infusion observed in this study is in accordance with data obtained in animal models of ischemia–reperfusion and septic shock.35,36The comparison with control patients, who received the vehicle only, ruled out the possibility that this increase was dependent on the tocopherol content of the propofol vehicle.37 

γ-Tocopherol was still raised in mononuclear leukocytes 1 day after the induction of anesthesia, which suggests that these cells may represent a suitable compartment for γ-T accumulation, as already observed for other tissues.12,38This feature may account for the lower PGE²levels observed in cultured WB from propofol patients 1 day after the intervention. The antiinflammatory effect of γ-T observed ex vivo  is supported by the results obtained in vitro , in adherent monocytes. In these cells, in agreement to what has been shown in murine macrophages and Caco2 cells,11,39γ-T counteracts the synthesis of PGE²mediated by COX-2 induction. γ-T acts posttranscriptionally on COX-2 activity, leaving COX-2 protein levels unaltered, as already reported for other cell types.39 

Plasma levels of malondialdehyde, a marker of peroxidative stress, sharply increased 30 min after the beginning of bypass surgery, reflecting the switch of reactive oxygen species and isoprostane production during ischemia–reperfusion.24,26Along with the increased pro-oxidant status, antioxidant defenses declined, as shown by the time course levels of IAC, an index of the overall protection against oxidative damage. The elevation of systemic malondialdehyde and the decrease of IAC levels during bypass surgery were similar under the two anesthesia regimens, despite the selective and marked increase of γ-T occurring in the propofol group. Therefore, our findings do not support a greater antioxidant effect of propofol, at the doses used in this study, with respect to the control treatment. We cannot, however, exclude the possibility that propofol infused at higher doses for a longer time period would exert an antioxidant effect, as reported by others.21,40 

Of interest, γ-T is characterized by selective antioxidant properties that are not shared by α-T and that may result in distinct biologic effects.7Although the biologic or clinical significance of our findings is far to be clear (indeed our study was not powered to assess differences in clinical outcomes), indications exist that a walnut diet, resulting in increased γ-T levels, improves endothelium-dependent vasodilation in hypercholesterolemic patients41and in healthy subjects.42In addition, a 6-week supplementation of γ-T (800 mg/day), resulting in doubled plasma levels of γ-T (similarly to what observed in propofol treated patients), reduced biomarkers of inflammation, i.e ., plasma C-reactive protein and urine nitrotyrosine, in subjects with metabolic syndrome.43 

Our data show that IL-10 levels increase in plasma during surgery, as observed in patients undergoing elective major surgery and coronary artery bypass graft surgery.44,45The increase of IL-10 may be due to a counterregulatory response to the proinflammatory status induced during and after the surgical intervention. Increased levels of proinflammatory cytokines such as tumor necrosis factor α, IL-6, and IL-8 play a key role in the inflammatory cascade that follows cardiac surgery and that can lead to adverse perioperative events,46whereas antiinflammatory cytokines such as IL-10 significantly limit these complications.47It has been reported that propofol reduces the release of proinflammatory cytokines both in vitro  48and in patients undergoing abdominal and coronary artery bypass graft surgery.20,49With respect to control anesthesia, propofol did not significantly suppress the increase of plasma IL-10 during surgery or affect its production in lipopolysaccharide-cultured WB, an experimental condition in which the synthesis of PGE²is, conversely, markedly reduced. This finding suggests that propofol, by increasing γ-T levels in cells, selectively targets the proinflammatory PGE², sparing the capacity of WB to synthesize IL-10. The positive effect of propofol on the balance between antiinflammatory and proinflammatory cytokines has been highlighted also in surgical settings that do not require CPB and that, therefore, are not associated with ischemia–reperfusion injury.50–52 

The increased γ-T levels observed in the propofol group could be attributed to the interaction of propofol with CYP3A4, for propofol is mostly eliminated by CYP3A4-mediated catabolism,53and tocopherols are metabolized by side-chain ω-oxidation and consecutive β-oxidation through a CYP3A-dependent mechanism.54,55The selective effect of propofol on γ-T levels may be explained by the different turnover and metabolic rates of the two tocopherol isoforms, possibly related to their chemical structure.55,56 

It should be mentioned that potential limitations may have affected the results of our study. Basic hemodynamic parameters could not be measured continuously in all patients because, according to the clinical practice of our institution, only selected patients had a Swan-Ganz catheter in place. Therefore, the influence of these parameters on γ-T or other analytes cannot be excluded. In addition, differences in anesthetic regimens other than propofol may have affected γ-T levels. To our knowledge, however, a specific effect of any of the drugs administered in the control treatment on γ-T has never been documented. Although a protective effect of sevoflurane against reperfusion injury has been reported,57this could be hardly put in relation with a decrease, or even with a lesser increase, of plasma γ-T.

Moreover, the sample size of the study did not allow enough statistical power to detect small differences in several of the comparisons reported. Specifically, the lack of significance in the comparison performed on α-T levels in PBMLs and on IL-10 production in WB should be considered with due caution.

The use of delta values in the statistical analysis also deserves some clarifications. Our approach was adopted a priori  to control the intersubject variability. In addition, when the analysis was performed using actual values, similar results were obtained: As expected from the reduced statistical power, P  values were generally higher (data not shown), but the significance of the main findings was unchanged.

In conclusion, our data show that γ-T levels are significantly increased in patients treated with propofol, compared with patients treated with the control anesthetic regimen. This effect was detected both in plasma and in mononuclear leukocytes and is associated with reduced COX-2 activity. These findings suggest for propofol a novel antiinflammatory effect, which may be relevant in controlling the inflammatory response related to tissue injury after reperfusion. Whether this antiinflammatory effect translates into a clinically significant outcome remains to be determined on the basis of properly powered clinical trials.

The authors thank Loredana Boccotti (Technician, Institute of Cardiology, University of Milan, Milan, Italy) for her expert technical assistance.