General anaesthesia produces a reversible loss of short-term memory, which soon resolves after cessation of anaesthetic drug action. Anaesthetic drugs have been blamed for post-operative olfactory dysfunction, but it has never been proven that they actually cause hyposmia or anosmia.1 Although altered perception of smells has been described during inhalational induction in children, it is unknown whether this represents a deficit of the olfactory function or memory.2 At least to our knowledge, no published data have addressed the olfactory effects of anaesthetic agents.
Olfactory memory is an important cerebral function in mammals, serving functions such as food seeking, protection, reproduction and infant–mother bonding.3 Further, explicit memory enables humans to attribute associative meaning to odours and to effectively respond to previously experienced events.4 Olfactory memory involves the piriform cortex, associated with identification of odours, the amygdala and the entorhinal cortex, which pairs odours with the proper memories. The associated neurotransmitting systems [adrenergic and gamma-aminobutyric acid (GABA)ergic] are known targets of anaesthetic drugs, implying a possible interaction of the latter with olfactory memory.5,6 Based on the extended hormonal effects of anaesthetic drugs, another possible mechanism is that olfactory memory impairment is mediated by distinct effects on hormones known to regulate associative memory, namely, melatonin and oxytocin.7,8 To test the hypothesis that anaesthesia impairs olfaction, we studied the effects of general anaesthesia with two different anaesthetic agents (propofol vs. sevoflurane) and epidural anaesthesia for minor procedures on post-operative olfactory function and memory and plasma melatonin and oxytocin levels.
Ethical approval for this study (protocol number 3758) was provided by the Educational Bureau of the Naval Hospital of Athens, Athens, Greece (President E. Douzinas) on 20 July 2004.
We enrolled 60 patients of both sexes, aged 21–69 years, with the American Society of Anesthesiologists I and II status, who were scheduled for minor surgery of anticipated duration 40–120 min, after written informed consent was obtained. Exclusion criteria for enrolment were history of smoking, alcoholism or use of other addictive substances, history of encephalopathy, mental retardation, psychiatric disease (including depression, suicidal behaviour, schizophrenia and dementia), use of psychiatric drugs, history of neurosurgical or ear-nose-throat surgery, malignant hyperthermia, allergic rhinitis, nasal polyps, recent airway infection or past airway infection with residual hyposmia and history of hyposmia or other olfactory deficit. Surgical procedures included hernia repair, varicose vein surgery in the lower limbs, minor gynaecological procedures, arthroscopy of the knee joint, minor orthopaedic surgery in the lower limbs and minor urological procedures.
The patients were randomly (envelope randomisation) allocated to receive epidural anaesthesia (group E, n = 20), general anaesthesia with propofol (group P, n = 20) or general anaesthesia with sevoflurane (group S, n = 20). Twelve patients were excluded due to conversion of regional to general anaesthesia (one), conversion to major surgery (one) and inappropriate (<40 or >120 min) duration of surgery (10). Finally, data from 16 patients in each group were completed and analysed. Twelve hours pre-operatively (time-point 0), all patients were subjected to the baseline olfactory tests and blood sampling and were afterwards pre-medicated with hydroxysine 50 mg by mouth. All surgical procedures were performed between 08: 00 a.m. and 12: 00 (noon). Upon arrival in the theatre, peripheral venous access was established and patients received dimethindene 4 mg, metoclopramide 10 mg and ranitidine 50 mg via intravenous infusion and standard monitoring was applied. In groups P and S, train-of-four twitches and bispectral index were also monitored. In group E, lumbar epidural anaesthesia was performed by injecting ropivacaine 7.5 mg ml−1 (Naropeine; Astra Zeneca, London, UK) and fentanyl 25 μg through the epidural catheter in an appropriate dose to achieve surgical anaesthesia. In groups P and S, general anaesthesia was induced with propofol 2 mg kg−1, rocuronium bromide 0.6 mg kg−1 and remifentanil 1 μg kg−1 infused in 1 min and followed by endotracheal intubation and mechanical ventilation (tidal volume 7–9 ml kg−1, FiO2 0.40, respiratory rate adjusted to maintain end-tidal CO2 35–45 mmHg). Anaesthesia was maintained in group P with continuous intravenous infusion of propofol 1% (5–7 mg kg−1 h−1; Diprivan; Astra Zeneca) and in group S with sevoflurane (end-tidal concentration 1.8–2.2%; Sevorane; Abbott, Queenborough, UK), in order to maintain bispectral index in all cases in the range of 40–60. Analgesia was provided with continuous intravenous infusion of remifentanil started at 0.15–0.25 μg kg−1 min−1 and adjusted to maintain mean blood pressure between 80 and 120% of pre-operative values. Additional boluses of rocuronium bromide 0.1 mg kg−1 were given when train-of-four was more than 1. Approximately, 30 min before anticipated end of operation, patients in groups S and P received parecoxib 40 mg intravenously. Awakening, reversal of paralysis and extubation followed standard procedures and patients were admitted to the post-anaesthesia recovery room. End of anaesthesia was considered the time at which spontaneous breathing had recovered, the trachea was extubated, the patient could say his name and Aldrete score was at least 9. For operations conducted under epidural anaesthesia, the end of anaesthesia was noted as the end of operation (last suture).
Post-operative analgesia in group E included continuous epidural infusion of ropivacaine 2 mg ml−1 and fentanyl 3 μg ml−1 at a rate 4–6 ml h−1 with boluses of 3 ml and lock-out of 30 min. Post-operative analgesia in groups P and S included parecoxib 40 mg twice daily intravenously and dextropropoxyphene 75 mg intramuscularly as needed. Olfactory tests and blood sampling were repeated 0.5 h (time-point 1) and 3 h (time-point 2) following end of anaesthesia.
Olfactory function test
For detection of the acuity threshold, the following procedure was applied. The odorant n-butyl alcohol was prepared in a series of 10 dilutions beginning with 4% v/v in deionised water and progressing in successive thirds. We used n-butyl alcohol for olfactory threshold testing because it is a potent stimulus for the olfactory nerve at concentrations that have no impact on the trigeminal nerve.9
The procedure of olfactory test consists of 10 steps. In each step, the odorant and a blank are presented to the participant. The test progresses from weaker to stronger concentrations of odorant. An odorant bottle was presented to the participant accompanied by an identical bottle that contained distilled water only. The participant sniffs each one approximately for 9 s and then chooses which one smelled stronger. If the participant was incorrect at one concentration, the next higher concentration was presented. When a correct choice was made, the same concentration of odorant was presented to the participant until four consecutive correct responses were given. A bottle of distilled water was presented alongside each of the four consecutive presentations of the same concentration of odorant. Threshold was defined as the butanol concentration correctly chosen over water in four consecutive trails. The corresponding number of the concentration was taken as the threshold value.10
Olfactory memory test
The University of Pennsylvania Smell Identification Test (UPSIT) was administered to participants to assess olfactory identification ability (The Smell Identification Test; Sensonics Inc., Haddon Heights, New Jersey, USA). The test is in multiple-choice format, with four written response alternatives for each odour. The odours are released when the labels are scratched. The examiner scratched each target patch and instructed participants to smell the patch and then select the name of the released odour from among four alternatives.11 The UPSIT is considered as a linear, unbiased unidimensional Rasch measure of human smell recognition abilities.12
Melatonin levels were measured in plasma by ELISA (IBL, Hamburg, Germany). Analytical sensitivity was at 1.6 pg ml−1 and specificity at 1.2–2.5%. Intra-assay precision was 8.8–151.7 (3–11.4% of control values) and inter-assay precision was 5.6–134.3 (6.4–19.3% of control values). Value variation was 4.0–77.5 pg ml−1.
Oxytocin was determined by ELISA (Assay Design, Ann Arbor, Michigan, USA; sensitivity: 11.7 pg ml−1, range: 15.6–1000 pg ml−1), according to manufacturer's instructions. The ELISA reader used for the above measurements was ELX-800 by Biotek (Winooski, Vermont, USA).
Demographic data were compared with one-way analysis of variance (ANOVA) between groups. Due to absence of normal distribution, olfactory memory test and hormones were analysed with non-parametric tests. Mistaken recognition of odours in the olfactory memory test was compared between the three groups with the Kruskal–Wallis test and the individual differences were identified by using the Mann–Whitney U-post-hoc test. Changes in hormones between groups were compared by applying the Kruskal–Wallis test and the Mann–Whitney U-post-hoc test after calculating the ratios of values at time-points 1 and 2 to baseline measurements, so that inter-patient variability in baseline hormone levels could be compensated. Changes in hormones within each group were analysed using the non-parametric test of Friedman's two-way ANOVA. Correlation of melatonin values to the UPSIT scores was analysed by calculating Spearman's coefficient of correlation. Statistical significance was set at a P value of less than 0.05. Statistical analysis was performed using the SPSS v.15.0 software.
Demographic data of patients included in the analysis are presented in Table 1. No differences were found in age, sex distribution, BMI or duration of anaesthesia between different groups of anaesthesia. During all anaesthetics, mean arterial pressure, heart rate and pulse oxymetry were stable and no anaesthetic complication was described.
Olfactory function and memory
All patients identified correctly the n-butyl alcohol compared to blank at all time-points (acuity threshold 4% n-butyl-alcohol v/v). No patient mentioned having dysosmia or phantosmia. When comparing the preservation of olfactory memory, post-operative wrong recognition of odours, which were initially recognised correctly, was significantly more frequent in group S compared to groups E (time-point 1, P = 0.019; time-point 2, P = 0.026) and P (time-point 1, P = 0.047; time-point 2, P = 0.035; Fig. 1). No difference was found between groups in the UPSIT scores, but a statistically significant decrease in the UPSIT scores was noted in group S (P = 0.009; Table 2). No single odorant could be identified for which the answers differed between groups.
Melatonin and oxytocin changes
Melatonin levels did not differ significantly between groups. However, the ratios of melatonin change significantly decreased in group S (P = 0.047), but did not change in the other groups. Reduction of melatonin levels in group S was significant compared to that in groups E (time-point 2, P = 0.019) and P (time-point 2, P = 0.021; Fig. 2). In addition, melatonin levels significantly correlated to the UPSIT scores (r = + 0.309, P < 0.001).
Oxytocin levels did not alter statistically significantly post-operatively in any group of anaesthesia and also did not differ between groups (Fig. 3).
Post-operative misinterpretation of odours was more common with sevoflurane anaesthesia compared to propofol or epidural anaesthesia. As far as we know, this is the first study to demonstrate post-operative olfactory memory changes and to report that they are not associated with higher threshold for odour sensitivity. Interestingly, sevoflurane anaesthesia reduced plasma melatonin levels, which could possibly represent an underlying humoral mechanism.
Based on the data published by Henkin,1 1.8% of patients referred to a national centre of USA for smell and taste disturbances had no other obvious a etiological factor than preceding anaesthesia. Of course, this was only an observation and the diagnosis was made by exclusion; no prospective study exists for the incidence of post-operative olfactory dysfunction. If the hypothesis is true, it is a rather rare phenomenon – 59 patients referred to a national centre of USA in 19 years – as Dr Henkin1 admits in his correspondence to Dr Adelman.13 Therefore, the population needed to investigate post-operative olfactory dysfunction is probably substantially larger than in the present study. Post-operative olfactory memory deficits have not been systematically studied before either. From the present study, it is obvious that post-operative olfactory memory impairment is more likely after sevoflurane anaesthesia, but its clinical significance needs further investigation.
The interest in the role of melatonin in memory has been recently renewed, following reports that this hormone facilitates short-term memory and short-term olfactory social memory.14,15 Interestingly, animal studies show that melatonin strongly correlates with the associative strength of memory, especially during stress.7 This relationship could rely on the notable concentration of melatonin binding sites in entorhinal cortex and the CA1 field in hippocampus, sites important in pairing odours with the appropriate memory. Other studies also confirm the role of melatonin in consolidation of memories.16 The role of anaesthesia, and especially of sevoflurane, in reducing melatonin levels has been shown in other studies as well.17,18 Unfortunately, data on memory tests are not available to examine the relevance of melatonin to post-operative memory performance. In a previous study, we showed that sevoflurane did not alter melatonin levels following minor gynaecological procedures.19 These conflicting results could be due to the shorter duration of anaesthesia, the limited population studied or a specific effect of the procedure. However, Hanania and Kitain20 reported two cases in which melatonin prevented and reversed post-operative delirium, which strengthens the role of this hormone in post-operative preservation of cognitive function. The above findings could represent a theoretical basis for the mediating role of melatonin in the herein described impaired associative olfactory memory following sevoflurane anaesthesia.
Other effects of sevoflurane could also account for the impaired odour discrimination. Because the neurotransmitters of the olfactory memory system are in common with those affected by general anaesthetics (e.g. γ-aminobutyric acid), a direct effect of sevoflurane on olfactory memory structures cannot be excluded. For instance, subanaesthetic doses of sevoflurane were shown to cause memory loss by a specific action on amygdala, perhaps through the GABAA receptors.21 Current knowledge on the molecular actions of individual anaesthetics in distinct cerebral regions does not suffice to explain the differential actions of anaesthetic drugs in different cerebral neuronal systems.22 Nevertheless, the fact that both propofol and sevoflurane produce their anaesthetic effect through some common GABAergic mechanisms in the amygdale23 strengthens the possibility that a different mediator should account for the herein described differences in post-operative olfactory memory. Such a potential mediator could be melatonin, as is suggested by our results. We tested whether the three different anaesthetic techniques differentially affected oxytocin, because oxytocin has proved to facilitate discrimination of odours.8,24 However, a regulating role of oxytocin in the post-operative olfactory memory is clearly not supported by the findings of this study.
The olfactory memory impairment described here persisted for 3 h after the end of operation, despite that time of exposure to anaesthetics cannot be considered long. We cannot know when olfactory memory would return to baseline, but if melatonin was the single determinant of olfactory memory impairment, then according to the results of Ram et al.,17 this would be expected to be reversed within 24 h. Surely, further research is warranted to this point, as well as to the effect of longer lasting procedures on post-operative olfactory memory. Nevertheless, our results clearly show that impaired post-operative olfactory memory is not related to anaesthetic mishaps, but rather to a specific effect of sevoflurane. This effect should be more analytically studied in populations prone to olfactory memory disorganisation, such as the elderly.25 Olfactory function and memory are also impaired in schizophrenia, dementia and depression;26,27 currently, it is not known how surgery affects olfaction in those patients. Therefore, impaired post-operative olfactory memory should not be attributed solely to the anaesthetic in elderly patients or before psychiatric disorders are excluded. A limitation of our study is the lack of post-operative cognitive capacity measurement, which could exclude the possibility that the olfactory memory deficit is due to a generalised post-operative cognitive dysfunction. However, a recent meta-analysis of the recovery profile of anaesthetics, which revealed an advantage of sevoflurane compared to propofol in the early recovery and no differences regarding the readiness to home discharge, supports that it is hardly likely that the reduced olfactory memory post-operatively is due to a cognitive dysfunction.28 Nevertheless, future studies examining post-operative olfactory memory should also include measurement of the cognitive function. Although the significant correlation between melatonin and the UPSIT supports a relationship between these variables, this could reflect parallel post-operative changes, and not necessarily a causative relationship. Considering the lack of published data to support the role of melatonin in olfactory memory and the fact the changes of olfactory memory shown here are of limited clinical extent, these data should be considered as preliminary. Perhaps, an interventional study with melatonin could help clarify the role of melatonin in the olfactory memory in the immediate post-operative period.
In conclusion, these results strengthen the assumption that a relationship exists between post-operative olfactory memory impairment and sevoflurane, despite preserved olfactory acuity. In this sense, whether melatonin supplementation could reverse the post-operative olfactory memory deficit remains to be shown. The observed association between impaired odour discrimination ability and sevoflurane suggests that post-operative discrimination of odours deserves greater attention from health professionals.
The present work was financially supported by the Special Account for Scientific Research of the University of Athens and the Hellenic Navy General Staff. No sponsorship was used.
The study was supported by the 2nd Department of Anaesthesiology, School of Medicine, University of Athens, ‘Attikon’ Hospital, Athens, Greece.
1 Henkin RI. Altered taste and smell after anesthesia: cause and effect? Anesthesiology 1995; 83:648–649.
2 Fukumoto M, Arima H, Ito S, et al
. Distorted perception of smell by volatile agents facilitated inhalational induction of anesthesia. Paediatr Anaesth 2005; 15:98–101.
3 Firestein S. How the olfactory system makes sense of scents. Nature 2001; 413:211–218.
4 Wilson DA, Stevenson RJ. In: Wilson DA, Stevenson RJ, editors. Learning to smell: olfactory perception from neurobiology to behavior
. Baltimore: The Johns Hopkins University Press; 2006. pp. 222–223.
5 Kaba H, Nakanishi S. Synaptic mechanisms of olfactory recognition memory
. Rev Neurosci 1995; 6:125–141.
6 Veyrac A, Nguyen V, Marien M, et al
. Noradrenergic control of odor recognition in a nonassociative olfactory learning task in the mouse. Learn Mem 2007; 14:847–854.
7 Rimmele U, Spillmann M, Bärtschi C, et al
acquisition under stress independent of stress hormone release. Psychopharmacology (Berl) 2009; 202:663–672.
8 Kavaliers M, Colwell DD, Choleris E, et al
. Impaired discrimination of and aversion to parasitized male odors by female oxytocin
knockout mice. Genes Brain Behav 2003; 2:220–230.
9 Murphy C, Gilmore MM, Seery CS, et al
. Olfactory thresholds are associated with degree of dementia in Alzheimer's disease. Neurobiol Aging 1990; 11:465–469.
10 Lehrner J, Gluck J, Laska M. Odor identification, consistency of label use, olfactory threshold and their relationships to odor memory
over human lifespan. Chem Senses 1999; 24:337–346.
11 Doty RL, Shaman P, Dann M. Development of the University of Pennsylvania Smell Identification Test: a standardized microencapsulated test of olfactory function. Physiol Behav 1984; 32:489–502.
12 Lange R, Donathan CL, Hughes LF. Assessing olfactory abilities with the University of Pennsylvania Smell Identification Test: a Rasch scaling approach. J Alzheimers Dis 2002; 4:77–91.
13 Adelman BT. Altered taste and smell after anesthesia: cause and effect? Anesthesiology 1995; 83:647–649.
14 Jean-Louis G, von Gizycki H, Zizi F. Melatonin
effects on sleep, mood, and cognition in elderly with mild cognitive impairment. J Pineal Res 1998; 25:177–183.
15 Argyriou A, Prast H, Philippu A. Melatonin
facilitates short-term memory
. Eur J Pharmacol 1998; 349:159–162.
16 Gorfine T, Yeshurun Y, Zisapel N. Nap and melatonin
-induced changes in hippocampal activation and their role in verbal memory
consolidation. J Pineal Res 2007; 43:336–342.
17 Ram E, Vishne TH, Weinstein T, et al
. General anesthesia for surgery influences melatonin
and cortisol levels. World J Surg 2005; 29:826–829.
18 Arai YC, Ueda W, Okatani Y, et al
. Isoflurane increases, but sevoflurane
decreases blood concentrations of melatonin
in women. J Anesth 2004; 18:228–231.
19 Fassoulaki A, Kostopanagiotou G, Meletiou P, et al
. No change in serum melatonin
, or plasma beta-endorphin levels after sevoflurane
anesthesia. J Clin Anesth 2007; 19:120–124.
20 Hanania M, Kitain E. Melatonin
for treatment and prevention of postoperative delirium. Anesth Analg 2002; 94:338–339.
21 Alkire MT, Nathan SV. Does the amygdala mediate anesthetic-induced amnesia? Basolateral amygdala lesions block sevoflurane
-induced amnesia. Anesthesiology 2005; 102:754–760.
22 Antognini JF, Carstens EE. Anesthesia, amnesia, and the amygdala: reducing the fear of intraoperative awareness. Anesthesiology 2005; 102:711–712.
23 Ren Y, Zhang FJ, Xue QS, et al
. Bilateral inhibition of gamma-aminobutyric acid type A receptor function within the basolateral amygdala blocked propofol-induced amnesia and activity-regulated cytoskeletal protein expression inhibition in the hippocampus. Anesthesiology 2008; 109:775–781.
24 Engelmann M, Ebner K, Wotjak CT, Landgraf R. Endogenous oxytocin
is involved in short-term olfactory memory
in female rats. Behav Brain Res 1998; 90:89–94.
25 Boesveldt S, Verbaan D, Knol DL, et al
. Odour identification and discrimination in Dutch adults over 45 years. Rhinology 2008; 46:131–136.
26 Nordin S, Murphy C. Impaired sensory and cognitive olfactory function in questionable Alzheimer's disease. Neuropsychology 1996; 10:113–119.
27 Atanasova B, Graux J, El Hage W, et al
. Olfaction: a potential cognitive marker of psychiatric disorders. Neurosci Biobehav Rev 2008; 32:1315–1325.
28 Gupta A, Stierer T, Zuckerman R, et al
. Comparison of recovery profile after ambulatory anesthesia with propofol, isoflurane, sevoflurane
and desflurane: a systematic review. Anesth Analg 2004; 98:632–641.