Presented in part at the American Society of Anesthesiology Annual Scientific Meeting, October 2010, San Diego.
The influence of propofol or desflurane on postoperative cognitive dysfunction in patients undergoing coronary artery bypass surgery*
Version of Record online: 18 APR 2011
© 2011 The Authors. Anaesthesia © 2011 The Association of Anaesthetists of Great Britain and Ireland
Volume 66, Issue 6, pages 455–464, June 2011
How to Cite
Royse, C. F., Andrews, D. T., Newman, S. N., Stygall, J., Williams, Z., Pang, J. and Royse, A. G. (2011), The influence of propofol or desflurane on postoperative cognitive dysfunction in patients undergoing coronary artery bypass surgery. Anaesthesia, 66: 455–464. doi: 10.1111/j.1365-2044.2011.06704.x
- Issue online: 16 MAY 2011
- Version of Record online: 18 APR 2011
- Accepted: 22 February 2011
We investigated the influence of either propofol or desflurane on the incidence of postoperative cognitive dysfunction in a randomised trial of 180 patients undergoing coronary artery bypass surgery. The primary outcome was incidence of postoperative cognitive dysfunction at 3 months, defined as ≥ 1 SD deterioration in two or more of 12 neurocognitive tests. Secondary outcomes included early postoperative cognitive dysfunction (between days three and seven), delirium on day one, morbidity and length of hospital stay. Early postoperative cognitive dysfunction was significantly higher with propofol compared with desflurane (56/84 (67.5%) vs 41/83 (49.4%), respectively, p = 0.018), but this effect was not seen at 3 months (10/87 (11.2%) vs 9/90 (10.0%), respectively. There was no difference in delirium (7/89 (7.9%) vs 12/91 (13.2%), respectively, length of hospital stay (median (IQR [range]) 7 (6-9 [4-15]) vs 6 (5-7 [5-16) days, respectively or other morbidities. Desflurane was associated with reduced early cognitive dysfunction.
Postoperative cognitive dysfunction (POCD) is a common morbidity associated with coronary artery bypass surgery, ranging in frequency from 20% to 60% for between 6 weeks to 3 months after surgery [1–3]. It can be an enduring problem for patients, as the presence of POCD at 3 months after surgery has a high chance of persisting into the long term [3, 4]. Much research has been devoted to investigating aspects of cardiopulmonary bypass that have been considered to be the most likely cause of POCD, with conflicting results. The focus on cardiopulmonary bypass, and consequent aortic manipulation required for aortic cannulation and cross-clamp application, has been challenged by the finding that POCD occurs in non-cardiac surgery patients , as well as in coronary artery surgery performed without cardiopulmonary bypass (off-pump surgery) .
Few studies have investigated the possibility that anaesthetic agents may be neurotoxic and contribute to POCD. In human studies it is not ethically possible to separate the effect of the anaesthetic drug from the effects of surgery and inflammation. However, in animal studies there is evidence of harmful effects of anaesthetic agents on memory and learning [7, 8], and it is possible that despite the introduction of modern and short-acting anaesthetics, these drugs, capable of inducing profound unconsciousness, could also cause delayed recovery of brain function or even long-term deficit.
Conversely, by intracellular mechanisms, these anaesthetic agents may be protective in the setting of ischaemia-reperfusion injury. In the past decade, considerable research has focused on myocardial protection elicited by anaesthetic agents during ischaemia and reperfusion injury [9–12]. In both animal and human studies, volatile anaesthetic agents have been shown to reduce the extent of myocardial injury, and improve functional recovery following ischaemia-reperfusion injury. Cardiopulmonary bypass with cardioplegic arrest represents a human model of ischaemia-reperfusion, as the current cardioplegic techniques used to protect the heart are not absolute, and ischaemia-reperfusion occurs to some extent.
In most human studies of pharmacological preconditioning, volatile anaesthetics have been compared with the intravenous agent propofol [11, 13]. Propofol appears less efficacious than volatile anaesthetics, such as isoflurane, desflurane or sevoflurane, in reducing markers of myocardial damage, hospital recovery and longer term outcome [13–15]. The exact mechanisms of cardioprotection are still not fully elucidated, but there is evidence that the volatile anaesthetic drugs increase the activity of the mitochondrial K+ATPase channel and reduce opening of the mitochondrial permeability transition pore during reperfusion [16, 17]. Propofol is a known scavenger of reactive oxygen species and may reduce the burst of reactive oxygen species that occurs with reperfusion, which can damage mitochondrial integrity and cell viability .
As there is potential for cerebral ischaemia if cardiopulmonary bypass flow is reduced or inadequate, or focal hypoperfusion distal to cerebral artery stenoses, and focal ischaemia secondary to atheromatous microemboli, there is potential for anaesthetics to be protective during cardiac surgery if similar mechanisms are involved for brain as for myocardial protection. Consequently, a balance must exist between potential direct neurotoxicity and protection during ischaemia-reperfusion injury for anaesthetic drugs. The measurable outcome is the summation of these opposing effects and cannot be differentiated in human trials.
The aim of this study was to identify whether different classes of anaesthetic could affect the development of postoperative cognitive dysfunction. The null hypothesis was that the type of anaesthetic would not influence the incidence of postoperative cognitive dysfunction in patients undergoing coronary artery bypass graft surgery requiring cardiopulmonary bypass.
The study was approved by the Melbourne Health Human Research Ethics committee and written informed consent obtained from all participants. A parallel randomised trial with 1:1 allocation ratio was conducted to compare the incidence of POCD after coronary artery bypass surgery using cardiopulmonary bypass, where propofol or desflurane was administered as the primary anaesthetic.
Patients were eligible for enrolment if they were more than 18 years of age, undergoing primary elective coronary artery bypass surgery with cardiopulmonary bypass, and not undergoing additional procedures such as valve surgery, and were able to speak sufficient English to complete the neurocognitive tests. Exclusion criteria included dialysis dependent renal failure, liver transaminases more than 1.5 times normal, pre-existing diagnosis of schizophrenia, dementia, recent stroke, known disorder affecting cognition, severe anxiety states, recent alcohol abuse or a history of chronic opioid or other psychotropic drug use. Not all surgeons participated in the study and patients were withdrawn before randomisation if their surgery was changed to a non-participating surgeon.
All patients were from a single institution. Pre-operative demographic and neurocognitive testing was performed in a quiet setting located either in the pre-operative clinic or on the ward. Postoperative data were collected in hospital before discharge (days 3–7), and the 3-month neurocognitive testing was performed on an outpatient basis. The battery of 12 neuropsychological tests assessed cognitive function under four broad-based categories of memory, attention and concentration, perceptomotor ability and executive function.
- 1 and 2Trailmaking A and B (Trails A and B); endpoint: time to completion.
- 3Controlled Word Association test; endpoint: number correct.
- 4Stroop test; endpoint: number correct.
- 5Letter cancellation; endpoint: time to completion.
- 6 and 7Grooved Pegboard Dominant and Non-Dominant Hands; endpoint: time to completion.
- 8 and 9Rey Auditory Learning test; endpoint: learning sum of trials 1–5, and delayed trial 7 minus trial 5.
- 10 and 11Digit Span Forward and Digit Span Backward; endpoint: number correct.
- 12Symbol Digit Modalities test (written only); endpoint: number correct.
Where available, parallel forms were used to limit the effects of learning between assessments.
Patients were allocated to receive either propofol or desflurane as the primary anaesthetic agent for the duration of surgery. For initiation of anaesthesia, the desflurane group received inhaled sevoflurane. Following intubation of the patient’s trachea, the volatile anaesthetic was changed to desflurane which was titrated as necessary to maintain adequate anaesthesia depth for the entire operative period including cardiopulmonary bypass. A bispectral index (BIS) of between 40 and 60 was aimed at. Anaesthesia in the propofol group was induced with propofol (Fresol 1%; Pharmatel Fresenius Kabi Pty Ltd, Hornsby, NSW, Australia) using target concentration infusion from 1.5 to 3 μg.ml−1 utilising the Marsh pharmacokinetic profile, and adjusted as necessary to maintain an adequate depth of anaesthesia. No volatile anaesthetic agent was permitted at any stage in the propofol group. Depth of anaesthesia monitoring using BIS was used in all patients.
The following techniques were common for both groups. Anaesthesia co-induction/sedation was with fentanyl (2–5 μg.kg−1) and midazolam (0.025–0.05 mg.kg−1) followed by a combined intravenous infusion of fentanyl (1.5 μg.kg−1.h−1) and midazolam (0.025–0.05 mg.kg−1.h−1) with FIo2 = 1. In addition to routine anaesthetic monitoring, all patients received intra-arterial pressure monitoring, pulmonary artery pressure measurement via a thermodilution catheter, nasopharyngeal temperature monitoring and transoesophageal echocardiography, consistent with standard care at the Royal Melbourne Hospital. For control of blood pressure, metaraminol, glyceryl trinitrate and beta-blockers were allowed to control systemic arterial blood pressure within a range of 100–140 mmHg. Dobutamine was administered for low cardiac output (cardiac index less than 2.0 l.min−1.m−2), and noradrenaline infusion was used to treat persistent low systemic vascular resistance. The cardiopulmonary bypass circuit was primed with 2 l crystalloid solution and maintained at a temperature of 35 °C. Cardiopulmonary bypass was performed by cannulation of the ascending aorta and venous cannulation with a single two-staged right atrial-caval cannula. Temperature was allowed to drift progressively to 34 °C until rewarming was initiated before commencement of the last distal anastomosis, with heat exchanger temperature not exceeding 37 °C. The haemoglobin concentration was maintained above 7 g.dl−1. The ascending aorta was cross-clamped and cardiac arrest induced by administration of tepid blood cardioplegia (20–25 °C) with both antegrade and retrograde delivery. Further doses of maintenance cardioplegia were given following completion of each graft anastomosis. For postoperative sedation, a low-dose propofol infusion (50–150 mg.h−1) was commenced for both patient groups at the end of surgery until suitable for tracheal extubation in the intensive care unit when the patients were awake and co-operative, warm, haemodynamically stable, and with Paco2 < 6.7 kPa and Pao2 > 13.3 kPa with a FIo2 of 0.4.
The primary outcome was the incidence of POCD at 3 months after surgery, defined as a deterioration of ≥ 1SD from baseline, in at least two of the 12 neurocognitive tests [19, 20].
A number of secondary outcomes were collected. The incidence of postoperative cognitive dysfunction before discharge, was as defined above. This assessment was performed on the day before hospital discharge wherever possible and the timings ranged from day three up to day seven post-surgery. Delirium was assessed using the Confusion Assessment Method  performed on the first postoperative day. Details of the duration of mechanical ventilation and intensive care unit and hospital length of stay (from the operative day) were collected. Postoperative morbidity was collected including death, myocardial infarction (at least two of the following criteria: cardiac enzyme level elevation with either troponin I > 20 μg.l−1 or CK-MB > 30 μg.l−1, serial ECG with new Q waves in two or more leads, or new wall motion abnormality on echocardiography [22, 23]), ventricular or supraventricular arrhythmia, inotrope use to support low cardiac output, mechanical cardiac support; a bleeding rate of > 200 ml.h−1; re-operation; ventilation > 24 h, use of continuous positive airway pressure support after extubation, pneumothorax or pleural infusion requiring intervention and drainage, radiological pulmonary oedema, respiratory failure requiring re-intubation of the patient’s trachea; superficial wound or skin requiring treatment, deep sternal, chest, urinary or abdominal infection requiring treatment; stroke, transient ischaemic attack, coma, agitation or delirium; serum creatinine increase > 1.5 times baseline value, haemofiltration; and homologous blood product transfusion. Finally, depression was assessed using the cardiac depression index score .
The sample size was determined using a study performed on a similar surgical cohort in our own institution , where the incidence of cognitive dysfunction was 38% using techniques similar to the propofol group (propofol, midazolam and fentanyl infusions and aortocoronary anastomoses) vs 3.8% in the comparative group (which employed epi-aortic ultrasound, pedicle Y-graft technique and epidural analgesia). As this trial was non-randomised and involved differences in surgical and anaesthetic techniques, we adopted a more conservative estimate of effect size related to the anaesthetic as a single intervention. We estimated that a 50% reduction in observed POCD was potentially possible. A minimum sample size of n = 90 in each group was required based on 80% power to reject the null hypothesis using an alpha value of 0.05. A blinded interim analysis was performed on the primary endpoint after 90 patients with a stopping rule of significance at a value of p < 0.001, and the study continued to completion.
The allocation sequence was generated by computer random number generation, and the allocation was placed in sequentially numbered opaque sealed envelopes by a non-investigator. Enrolment and data collection were performed by trained research staff who were not involved in the care of the patients. Assignment to the intervention group was performed by the treating anaesthetist once the patient arrived in the operating room. The treating clinicians were not blinded to the assignment group, but all other staff involved in both the collection and collation of data, and administration of neurocognitive testing, were blinded to group allocation. Participants were not informed of assignment, but may have detected the smell of gaseous induction of anaesthesia in the desflurane group and therefore it cannot be assured that they were blinded. Assignment was only decoded after completion of the 3-month follow-up of the last patient.
The incidence of patients having POCD was assessed with the chi-squared test and significance set at a value of p < 0.05. For secondary outcomes, categorical variables were assessed using the chi-squared test and continuous variables were assessed using t-tests. Significance was set at a value of p < 0.01 to reduce the risk of type-1 error due to multiple comparisons.
Independent predictors for primary and secondary outcomes were identified using multiple logistic regression analysis. The outcome variables were: POCD at hospital discharge; POCD at 3 months postoperatively; Confusion Assessment Method defined delirium at 24 h; the presence or absence of any peri-operative morbidity; and postoperative hospital length of stay greater than the modal 5 days. For all patients the presence of the following variables were noted: age; EuroSCORE; sex; body mass index > 30 kg.m−2; ejection fraction; the presence of diabetes mellitus; daily intake of sedatives or anti-depressants; pre-operative haemoglobin, platelet and creatinine values; operative time greater than the median time of 320 min; and group allocation. Each of these independent variables was included in a forced entry univariate logistic regression on the different outcome variables. All significant predictor variables (p < 0.1) for each outcome variable were then included in a backward stepwise (likelihood ratio) multivariate logistic regression analysis to identify the independent risk factors for each of the different outcome variables. Significance was set at a value of p < 0.05. Analysis was performed using PASW Statistics V18.0 (SPSS Inc, Chicago, IL, USA).
The trial was conducted between September 2007 and January 2010. The CONSORT participant flow diagram is shown in Fig. 1. One hundred and eighty-two patients were randomly assigned, and 180 received correct allocation. The two patients who received neither propofol nor desflurane as their primary anaesthetic were not studied. A further four patients were not studied for analysis as they refused to participate in the final outcome assessment. Eighty-seven patients receiving propofol and 90 patients receiving desflurane were analysed for the primary endpoint. No patients in either group died during the hospital or follow-up periods. Baseline and intra-operative variables are shown in Table 1. Following induction of anaesthesia and before cardiopulmonary bypass, the mean (SD) pulmonary artery pressure was lower in the desflurane group (16 (4.1) compared with propofol, 19 (5.3) mmHg, p = 0.003), but other variables were not different.
|Propofol n = 89||Desflurane n = 91|
|Age; year||63.8 (10.6)||61.8 (10.4)|
|Male||80 (90%)||73 (80%)|
|Weight; kg||86.5 (17.4)||83.4 (15.4)|
|Height; cm||171 (11)||171 (8)|
|BMI; kg.m−2||29.6 (6.2)||28.4 (4.4)|
|Ejection fraction ≥40%||85 (96%)||83 (91%)|
|Diabetes requiring any treatment||64 (72%)||73 (80%)|
|NYHA Class 1-2||49 (55%)||59 (65%)|
|Class 3-4||40 (45%)||32 (35%)|
|Preoperative haemoglobin; g.dl−1||14.3 (1.4)||14.2 (1.7)|
|Preoperative creatinine; mmol.l−1||96 (22)||90 (20)|
|Preoperative sinus rhythm||72 (81%)||85 (93%)|
|Oral hypoglycemic medication||18 (21%)||12 (13%)|
|Antidepressant medication||5 (6%)||12 (13%)|
|Pre CPB haemoglobin; g.l−1||10.8 (1.3)||10.8 (1.3)|
|Pre CPB heart rate; b.min−1||65 (14)||64 (12)|
|Pre CPB mean arterial pressure; mmHg||80 (11)||79 (12)|
|Pre CPB mean pulmonary artery pressure; mmHg||16 (4)||19 (5)|
|Pre CPB cardiac index; l.min−1.m−2||2.4 (0.6)||2.5 (0.6)|
|Post CPB heart rate; b.min−1||81 (9)||80 (12)|
|Post CPB sinus rhythm||83 (93.2%)||86 (94.5%)|
|Post CPB mean pulmonary artery pressure; mmHg||20 (8)||20 (4)|
|Post CPB cardiac index; l.min−1.m−2||2.7 (0.6)||2.9 (0.5)|
|Highest glucose; mmol.l−1||8.8 (2.0)||8.6 (1.5)|
|CPB Time, min||96 (76–118 [47–233])||92 (77–110 [27–193])|
|Aortic cross-clamp time; min||74.5 (60–95 [24–201])||72 (60–89 [18–152])|
|Temperature during CPB; max||36.3 (36–36.5 [35.6–37.5])||36.3 (36.1–37.0 [35.4–37.5])|
|Temperature during CPB; min||34.0 (33.7–34.3 [32.7–35.7])||34.0 (33.6–34.3 [32–35.1])|
|Number of distal grafts||3 (2–4 [1–5])||3 (2–3 [1–6])|
|Other cardiac surgery||3 (3%)||2 (2%)|
|Insulin infusion||60 (67%||65 (71%)|
|Aprotinin use||1 (1%)||1 (1%)|
|Tranexamic acid use||9 (10%)||11 (12%)|
|Total midazolam dose; mg||13.3 (5.1)||13.4 (4.2)|
|Total fentanyl dose; μg||1315 (405)||1188 (359)|
|Operative time; min||335 (66)||332 (70)|
|EuroSCORE||3 (1–5 [0–11])||3 (2–4 [0–14])|
Intention to treat analysis was performed. The incidence of neurocognitive dysfunction pre-discharge was significantly higher for propofol (56/84 (67.5%) compared with desflurane (41/83 (49.4%), p = 0.018), but there was no difference on this outcome at 3 months (10/87 (11.2%) compared with 9/90 (10.0%), respectively, p = 0.748).
Mean (SD) cardiac depression index scores were significantly reduced at 3 months from baseline scores (p < 0.001) for both groups, but were not different between groups (propofol baseline 102 (22) compared with desflurane 102 (18), p = 0.992; and at 3 months propofol 88 (21) compared with desflurane 89 (22), p = 0.751). The incidence of delirium on day 1 was not significantly different between groups (propofol 7.9% compared with desflurane 13.2%, p = 0.250).
Morbidity and length of stay are shown in Table 2. There were no differences in postoperative ventilation time, intensive care unit and hospital length of stay. There were no significant differences in the incidence of morbidity between groups other than the incidence of superficial wound or skin infection, which was higher in the propofol group (18.0% compared with 5.5%, p = 0.009).
|Propofol n = 89||Desflurane n = 91||p value|
|CAM Delirium day 1||7 (7.9%)||12 (13.2%)||0.245|
|ICU ventilation time; min||443 (354–710 [180–2482])||410 (325–550 [133–9860])||0.206|
|ICU LOS; h||24 (22–47 [15–264]||23 (21–47 [15–192])||0.482|
|Hospital LOS; day||7 (6–9 [4–15])||6 (5–7 [5–16])||0.161|
|In hospital death||0||0||1.000|
|Postoperative myocardial infarction||0||0||1.000|
|Ventricular arrhythmia||2 (2.2%)||8 (8.8%)||0.100|
|Atrial fibrillation||37 (41.6%)||30 (33%)||0.232|
|Inotrope use||8 (9.0%)||11 (12.1%)||0.499|
|Mechanical circulatory support||0||2 (2.2%)||0.497|
|Bleeding > 200 ml.h−1||8 (9.0%)||9 (9.9%)||0.836|
|Re-operation for bleeding||1 (1.1%)||3 (3.3%)||0.621|
|Re-operation sternal debridement||0||3 (3.3%)||0.246|
|Required CPAP post extubation||24 (27.0%)||22 (24.2%)||0.668|
|Ventilation > 24 h||1 (1.1%)||3 (3.3%)||0.621|
|Pleural effusion requiring drainage||3 (3.3%)||7 (7.7%)||0.330|
|Acute pulmonary oedema||2 (2.2%)||3 (3.3%)||1|
|Respiratory failure requiring re-intubation||3 (3.3%)||2 (2.2%)||0.680|
|Superficial wound or skin infection||16 (18.0%)||5 (5.5%)||0.009|
|Chest infection||3 (3.4%)||4 (4.4%)||1|
|Abdominal infection||2 (2.2%)||1 (1.1%)||0.619|
|Delirium during hospital stay||16 (18%)||18 (19.8%)||0.757|
|Creatinine Increase > 50% from baseline value||7 (7.9%)||5 (5.5%)||0.524|
|Haemofiltration||1 (1.1%)||2 (2.2%)||1|
|Homologous blood product transfusion||24 (27.0%)||36 (39.6%)||0.073|
Independent predictors of primary and secondary outcomes are shown in Table 3. Propofol was an independent predictor for cognitive dysfunction before hospital discharge (p = 0.011, OR 2.31) but not at 3 months.
|Outcome variable||Independent predictors||p value||Odds ratio (95% CI)|
|Cognitive deficit at discharge||Group (propofol)||0.011||2.31 (1.21–4.43)|
|Maximal intra-operative temperature||0.012||2.72 (1.24–5.95) (per 1 °C)|
|Cognitive deficit at 3 months||Age||0.009||1.08 (1.02–1.15) (per year)|
|Ventilation time||0.035||1.001 (1.00–1.002) (per min)|
|CAM defined delirium||Body mass index > 30 kg.m−2||0.043||2.80 (1.30–7.56) (per kg.m−2)|
|Peri-operative morbidity||Group (propofol)||0.022||2.60 (1.15–5.92)|
|EuroSCORE||0.023||1.25 (1.03–1.52) (per 1 unit in EuroSCORE)|
|Ventilation time||0.013||1.003 (1.001–1.005) (per min)|
|Hospital LOS > 5 days||Age||0.002||1.06 (1.02–1.1) (per year)|
|Female sex||0.032||5.34 (1.16–24.7)|
|Ventilation time||0.008||1.002 (1.001–1.004) (per min)|
This study has shown that the choice of primary anaesthetic agent (desflurane or propofol) did not significantly affect the incidence of POCD at 3 months following coronary artery bypass surgery. There was, however, a significantly higher incidence in POCD before hospital discharge with propofol. The only difference in postoperative morbidity outcomes was a higher incidence of superficial wound infection with propofol.
The reporting of POCD at between the third and seventh postoperative days is controversial, as there is the potential for concurrent drugs, for example analgesics, to affect cognitive performance. It is well-known that the incidence of POCD in this time period is very high [1, 19, 20]. These early assessments are perhaps best viewed as a period of differential rates of cognitive recovery, but where some individuals continue to have persistent cognitive dysfunction. The dysfunction at 3 months is more likely to represent a permanent state, though this is not absolute as continued healing may occur. A lower incidence of POCD before hospital discharge may reflect more rapid or better quality of cognitive recovery from the state of deep coma induced by anaesthesia.
Cognitive function is considered to reflect the integrity of the brain. Measurement of POCD is based on patients completing a series of neurocognitive tests, and examining the change in scores from a pre-operative baseline value to the postoperative measurements. As there are many different types of cognitive tests, a consensus was reached to try and standardise measurement, to improve compatibility between studies. This study utilised a 12 test battery of neurocognitive tests with POCD defined as ≥ 1 SD deterioration in performance in two or more tests from pre-operative baseline values. Our sample size estimates were based on a previous study using a similar test battery, and interventions that were consistent with the propofol arm of the study . The incidence of POCD in those studies was much higher than in this study. One reason for this difference is likely to be the time of the assessment that in the other studies was typically 6 weeks rather than the 3 months used in this study. In both studies, the in-hospital incidence of POCD was high, with significant recovery over time. The differences between groups in this current study at 3 months were small, and on these findings a much larger study would be required to detect any differences.
The idea that in humans, anaesthetics are potentially neurotoxic yet paradoxically neuroprotective during an organ injury such as ischaemia-reperfusion, cannot be answered by our data, as both groups received an equivalent dose of anaesthetic drug. Nor can we exclude the potential for propofol administered in the intensive care unit to reduce a potential neuroprotective signal from desflurane, as all patients received propofol as their intensive care unit sedation drug (this was a consensus decision by the treating clinicians). Further research could be aimed at combining regional and general anaesthesia to reduce the concentration of anaesthetic drug required to produce general anaesthesia, and different intensive care unit sedation regimens to answer this question further. Previous studies have not found a substantial difference in the incidence of POCD with different anaesthetic agents, or in comparison of regional with general anaesthesia (in non-cardiac surgery) . However, neuroprotective studies often focus on giving a large dose of anaesthetic before cardiopulmonary bypass to induce an isoelectric EEG ; the idea that more is better may be flawed if there is neurotoxic potential from the anaesthetic drug.
The patient groups were well-matched and had very similar outcomes. There was no mortality and morbidity rates were equal to or better than other published studies [28, 29] and less than the EuroSCORE-predicted mortality. The only significant morbidity difference was the incidence of superficial skin or wound infection, which was higher with propofol. The brand of propofol used in this study did not contain an antimicrobial agent. During the course of a cardiac surgery operation, many changes of propofol syringes are required, with potential interruption of the aseptic barrier on each occasion. We cannot determine from our data whether the increased incidence of infection is due to a property of the drug, or the method of delivery.
In examining the role of anaesthetic agents in POCD it is not possible to separate the effects of anaesthesia from surgery and the resultant inflammatory process invoked during healing. The role of the anaesthetic agent and differences between anaesthetic agents may be small in any contribution to POCD. It is possible that POCD at 3 months is unrelated to the choice of anaesthetic. This is supported by the multiple logistic regression analysis where age and ventilation time were the only independent predictors. In addition, the incidence of POCD at 3 months increases proportionally with age, as does the incidence of severe ascending aortic and arch atheroma . Cognitive dysfunction was, however, present in a large proportion of patients in the first week following surgery, which may have implications for quality of recovery outcomes such as home-readiness, return to work or other activities of daily living and emotive recovery. These are important patient outcomes, which could influence their rate and quality of recovery, as well the social support required during that recovery period. The future focus of research on anaesthetic agents in cardiac surgery should perhaps be aimed at cognitive recovery rather than long-term dysfunction.
The study has a number of limitations. It did not test either desflurane or propofol in isolation from other sedatives and analgesics . Neither anaesthetic has analgesic properties; propofol in particular requires a much higher concentration if used without other sedatives. The balanced anaesthesia regimen used in this study is consistent with clinical practice at the Royal Melbourne Hospital. We cannot exclude, however, that the fentanyl and midazolam, intensive care unit sedation with propofol, or postoperative analgesia did not contribute to the development of POCD. As with most studies of this type we were not able to blind the treating anaesthetist, as they were required to administer the anaesthetic. We did, however, blind the researchers involved in collecting and analysing the data, which should reduce observer bias. The clinical outcome of the patients was good, with no mortality and low morbidity rates, and the incidence of POCD reported in this study is at the low end of published studies.
The authors thank the many anaesthetists, surgeons, nursing and administrative staff who assisted in the conduct of the study. Specifically, we thank Matthew Sheppard, Research Assistant, University of Melbourne, Australia; and Linda Copland, R.N. research nurse, University of Melbourne, Australia, for assistance with data collection, and Baxter Healthcare for funding the study.
This was an investigator-initiated study, funded by Baxter Healthcare. The research groups of Colin Royse and Stanton Newman also have received funding for unrelated projects from Baxter Healthcare. No other competing interests are declared.