key: cord-0834612-e3aypqxu
authors: Parker, J. K.; Kelly, C. E.; Gane, S. B.
title: Molecular Mechanism of Parosmia
date: 2021-02-08
journal: nan
DOI: 10.1101/2021.02.05.21251085
sha: 7907ddf5128a01759c8dd6359c2f31f49378bb82
doc_id: 834612
cord_uid: e3aypqxu

The molecular stimuli that trigger parosmia have been identified. Parosmia is a debilitating disease in which familiar smells become distorted and unpleasant. Often a result of post infectious smell loss, incidences are increasing as the number of COVID-19 cases escalates worldwide. Little is understood of its pathophysiology, but the prevailing hypothesis for the underlying mechanism is a mis-wiring of olfactory sensory neurons. We identified 15 different molecular triggers in coffee using GC-Olfactometry as a relatively rapid screening tool for assessment of both quantitative olfactory loss and parosmia. This provides evidence for peripheral causation, but places constraints on the mis-wiring theory.

coffee. Likewise, dimethyl trisulfide (0.01 ug/L 13 ) is an exceptionally potent compound detected by 148 12/15 non-parsomic participants but only by 4 parosmic participants, and only once reported as a 149 trigger. 150

Furthermore, a handful of compounds were occasionally detected by but never reported as triggers. 151 (E)-β-Damascenone, a key odour-active compounds in coffee with a low odour threshold (0.01 152 ug/kg 13 ), was detected by 6 parosmic participants and always described as jammy and fruity. 153

Likewise, 4-ethylguaiacol was detected by 7 parosmic participants and always described as spicy, 154 sweet and smoky, but never parosmic. 155

Cluster analysis. Agglomerative hierarchical cluster analysis was carried out on the combined 156 intensity data from all parosmic participants of all potential triggers reported >3 times. It showed four 157 significant clusters of compounds (Fig. 2) . A structure activity pattern starts to emerge, suggesting, for 158 example, that some participants might perceive thiols more intensely and others may perceive 159 pyrazines more intensely. When the same analysis was carried out using the data from the non-160 parosmics, the same clusters did not emerge, but N was small, and we cannot draw any conclusions 161 from this. It is likely that in the general population, there are significant differences in the relative 162 perception of these compounds. 163

Correlation between ligand structure and odour receptor (OR)? Identifying a small number of 164 common molecular triggers for parosmia raised the obvious question of an olfactory receptor 165 similarity. To determine whether the clusters are associated with any of the known ligand odour 166 receptor pairs, we searched the ODORactor database 17 . We found no obvious segregation of triggers 167 by olfactory receptor (Fig. 3) . Most of the triggers activated (with > 50% probability) either OR1G1 168 or OR52D1. We then compared molecules never reported as triggers such as disubstituted pyrazines, 169 indole, skatole, cresol and found these to activate the same ORs, making it unlikely that these 170 olfactory receptors are the source of the parosmic signal. OR1G1 is known to be very broadly tuned 171 and bind odorants of different chemical classes 18 . However, only a fraction of the known olfactory 172 receptors have been deorphaned, and further identification of ligand-OR pairs is required. 173 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 Faecal odours. Parosmic participants often comment that the smell of faeces is never as unpleasant as 174 before, often smelling like other parosmic food, or even more pleasant and biscuity 6 , presenting the 175 interesting corollary that foods smell of faeces yet faeces smell of food. One normosmic and two 176 parosmic researchers carried out GC-O on a 50% faecal slurry in water. Whilst the normosmic scored 177 the intensity of the volatiles normally associated with faeces (indole, skatole and p-cresol 19 ) as close 178 to strongest imaginable, the parosmic researchers were unaware of these foul smells, rather reporting 179 many of the potent triggers already mentioned in coffee. This provides a neat explanation as to why 180 the changes in valence for faecal samples is reversed. In the absence of signals from the usual faecal 181 compounds, parosmic participants detect other potent volatiles in the sample, normally masked by the 182 faecal compounds for normosmics, which may elicit any number of odour percepts depending on the 183 sensitivity of the parosmic to the other compounds present. 184

Participant observations. 185 i.

A case of parosmia not preceded by anosmia: One participant identified 45 aromas, of which 186 just 2 were triggers. The more intense one was 2-ethyl-3,5-dimethylpyrazine and the second a 187 mild and unidentified compound. They had a TDI score of 37, had tested positive for COVID-188 19 antibodies but reported no loss of sense of smell. This is an unusual case where parosmia 189 was not preceded by anosmia. 190 ii.

A case of parosmia improving with no concomitant increase in olfactory function: After 4 191 months one participant showed no improvement in threshold score, a decrease in the number 192 of GC-O aromas detected (22 to 16), and a 5-fold decrease in intensity scores (750 to 147) yet 193 reported an improvement in parosmia. Further work on the temporal aspect of parosmia is in 194 progress. 195 iii.

A case of excellent recovery from parosmia with significant improvement of olfactory 196 function: After several years, one participant reported a "new normal" olfactory function, 197 scoring 35 on the TDI test. This indicates regrowth of a broad range of healthy OSNs, yet two 198 of the 52 aromas detected were still identified as triggers of parosmia: 2-methyl-3-furanthiol 199 and one unidentified compound. 200 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 iv.

A case of a functional ansomia with parosmia: At the other extreme one participant was 201 within 3 months of onset and had a low TDI score (15) indicating functional anosmia. Only 202 five aromas were detected, all scored as barely detectable, and only one, 2-furanmethanethiol, 203 had parosmic character. Further investigations showed their parosmia was triggered more by 204 different compounds present in onion and garlic. 205 v.

A case of two distinct parosmic characters: one participant reported several distinct parosmic 206 smells. One, which was described as plastic, chemical and burning rubber, was associated 207 with sulfur compounds from Cluster A, and a second, described as sickly sweet, smoky and 208 woody, was associated with pyrazines. Further investigation showed that onion and garlic 209 gave a third parosmic character. These groupings are consistent with HCA clusters and a sub-210 theme emerging from the "AbScent Parosmia and Phantosmia Support" group on Facebook 6 . 211 vi.

A case of sulfur triggers only: one participant attended just 4 weeks after onset of parosmia 212 and was our most "fresh" parosmic. Their reactions at the GC-O were quite extreme when a 213 trigger was encountered, with three thiols scoring 90-100 on the intensity scale. With a TDI 214 score of 27, 20 GC-aromas were detected but only five of these were triggers, either thiols or 215 disulfides. Pyrazines were detected but not reported as triggers. This is counter to any 216 hypothesis which suggests that in the early days of parosmia, all (or many) compounds are 217 triggers. 218 vii.

A case of parosmia resulting from post-traumatic brain injury who reported the same 219 molecular triggers as the post infectious participants. This participant found meat a far worse 220 trigger than coffee, and this is reflected in detection of the meaty thiol rather than the coffee 221 thiol. 222 viii.

One participant assessed three different concentrations of coffee, spanning a factor of 10000. 224

At the regular concentration, 11 aromas were detected, of which 5 were triggers. When the 225 stimuli were diluted by 100, only 2 compounds were detected (barely detectable) but they still 226 had the same character as in the regular coffee extract-one was still parosmic and the other 227 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 was not. Dilution of the stimulus did not reverse the distortion, and this is backed up by many 228 who still report parosmia character for weak and barely detectable aromas. When the stimuli 229 were concentrated, those that were undistorted remained undistorted and there was no 230 evidence that concentration of the stimuli could create more distortions. This is backed up by 231 other participants reporting strong but undistorted aromas. On concentration, new aromas 232 both distorted (4) and non-distorted (5) were detected as their concentrations became 233 suprathreshold. Thus, for this parosmic, the concentration of the stimulus did not seem to 234 change its character or determine its valence, which is often the case for perception of highly 235 potent odorants, particularly sulfur compounds which on dilution can change from pungent to 236 a pleasant fruity character. 237

Summary. In summary, we have identified for the first time specific molecules which trigger 238 parosmia. These experiments demonstrate that there is a common set of molecular triggers responsible 239 for distortions and sense of disgust in coffee, and they also trigger parosmia in other chemically 240 related foods. However, not all molecules in this set are triggers in all parosmic participants. These 241 molecules tend to be potent, have very low detection thresholds and in isolation are neither distorted 242 nor unpleasant for non-parosmic participants. However, odour activity is not the defining factor since 243 (E)-β-damascenone has an exceptionally low odour threshold, is one of the more potent compounds in 244 coffee 20 , and was always perceived as jammy and fruity by those parosmic participants who detected 245 it. Most of the trigger molecules found in coffee belong to one of four distinct groups: thiols, 246 pyrazines, disulfides, methoxypyrazines but there are no known odour receptors which are specific for 247 the described trigger molecules. In addition, individual case studies suggest that parosmia symptoms 248 are independent of olfactory function and parosmia may occur in patients with objectively normal 249 olfactory function. Parosmic odour quality is not necessarily related to odour concentration. Recovery 250 from parosmia can be associated with either improvement or stasis of olfactory function. As we stated 251 in our hypothesis, in those with poor olfactory function, parosmia may be enhanced by a lack of 252 contribution from other more desirable and less potent aroma compounds, but we found that the 253 molecular triggers alone are the key drivers of parosmia and individually responsible for the 254 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021.

perception of disgust. This explains why those with normal olfactory function, who also perceive the 255 more desirable less potent aroma compounds, still experience parosmia. 256

Little is known about the pathophysiology of parosmia. Like the aetiologies of smell loss, both central 258 and peripheral mechanisms have been proposed 21 and can broadly be thought of as the central theory, 259 the ephaptic theory and the mis-wiring theory. Although there is now some doubt about the role of the 260 bulb in olfactory identification 22 , these theories take the standard model of glomerular activation 261 pattern within the olfactory bulb as the motif of recognition within the CNS. 262

The central theory is based on the changes occurring in the integrative centres in the brain. A decrease 263 in olfactory bulb volume 23, 24 and a significant loss of grey matter volume 25 has been demonstrated in 264 parosmic participants. Further evidence of central mechanisms has been published recently showing 265 different fMRI activation patterns in parosmic participants compared to hyposmics 26 . Increased 266 activation in the thalamus and the putamen was observed in the parosmic participants, the latter being 267 of relevance since it is connected to the olfactory cortical networks and has been associated with the 268 perception of disgust. Also, stronger activation was observed in the ventral striatum which is 269 associated with odour valence. Whilst there is good evidence in humans for the central theory of 270 parosmia, a purely central causation seems unlikely based on our evidence that parosmia is triggered 271 by a group of highly specific molecules at the periphery. 272

The "mis-wiring" theory posits aberrant targeting of OSN to the glomerulus during regeneration from 273

insult. This has been observed in mice with impaired olfactory function induced by ciliopathies 27-28 , 274 physical lesioning 29-30 , and induced chemical degeneration 31 but not yet in humans. However, it has 275 been adopted as the likely mechanism for the perception of distorted olfactory percepts in parosmia. It 276 is further suggested that the change in hedonic valence is due to broad activation of the olfactory bulb 277 sending a disordered and unmoderated array of signals to the central neural processing system which 278 invokes a strong sense of disgust. Our data neither support nor refute the mis-wiring hypothesis, but 279 certainly place constraints on it. We have demonstrated the requirement to account for the non-280 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101/2021.02.05.21251085 doi: medRxiv preprint stochastic nature of the OSNs involved in any proposed mis-wiring theory. Whilst mis-wiring is 281 attributed to a loss of axonal pathfinding mechanisms 27 , in light of our results, this theory needs to 282 explain why some OSNs are relaying the "correct" undistorted signal, whereas others are not, even 283 early in the patient's recovery. 284

The ephaptic theory summarised by Hawkes 32 suggests that demyelination of the OSNs allow the 285 activation of other, non-stimulated OSNs adjacent to the activated OSN by current flow in the 286 extracellular fluid: "a form of short circuiting". This too would result in a broader activation of the 287 olfactory bulb and must be able to account for the non-random nature of the OSNs involved. 288

Of course, these theories are not mutually exclusive. Alterations in innervation of the glomeruli, or 289 ephaptic activation of afferent nerves, and therefore the whole bulb, would result in the alteration of 290 downstream central processing. The plasticity of this part of the pathway is much more limited. The 291 "wiring" of the brain in recognising and acting on certain patterns of activation within the bulb 292 remains after the peripheral insult. Therefore, a broader activation of the glomeruli would activate 293 more odour object patterns within the cortex. 294

These hypotheses have to explain four characteristics: that parosmia arises almost uniquely in settings 295 of widespread synchronous neuronal destruction either post infection or post traumatic brain injury, is 296 triggered by one of a number of common odorants, is of novel odour character, and that this character 297 is almost always unpleasant. Whereas the mis-wiring theory is consistent with the first, and the central 298 theory may explain the novel odour character and the change in valence, it remains for us to 299 determine why only a few potent molecules elicit such a strong parosmic response. 300

These trigger molecules share the trait of having an extremely low threshold in human olfaction, so 301 they are detectable in very low concentrations. Such low olfactory thresholds may be attributed to a 302 higher binding affinity for their respective olfactory receptors (ORs), but there are other possible 303 explanations: higher rate of expression of specific ORs on the OSN cell surface, stronger activation of 304 OSN depolarisation by the OR due to reduced habituation at the receptor or cellular levels, 305 overrepresentation of that OSN within the mucosal OSN population, zonal expression (not 306 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101/2021.02.05.21251085 doi: medRxiv preprint demonstrated in humans) resulting in improved access to the nasal airstream, glomerular level or 307 higher configurations resulting in greater salience of the glomerular activation, or the odorant could 308 activate more than one olfactory receptor and the combined activation is summed as a lower 309 detectable threshold within the bulb. 310

The common molecular structures, low odour thresholds and physiochemical grouping of the 311 molecular triggers of parosmia strongly suggest that this is an olfactory receptor-level phenomenon, 312 although we are unable to identify any specific olfactory receptors responsible from publicly available 313 databases. The fact that some parosmic participants only report distortion with some of the groups 314 suggests that there may be up to four separate olfactory receptors involved, or more when we consider 315 other food and household items that elicit distortions. So why then these groups of OSNs? Several 316 mechanisms could account for the role of specific receptors and their neurons. 317

1. These OSNs are regenerating because they are selectively damaged by the insult and others are 318 preserved. This may account for the presence of parosmia in objectively normosmic volunteers, but it 319 is unlikely that the insult from brain injury and viral infection would lead to the same preferential 320 damage to particular OSNs. 321 2. The specific OR is predominant within the regenerating OSN population for one of two reasons. In 322 one scenario, instead of a purely stochastic OR selection process in the olfactory mucosa, these 323 "parosmic ORs" are preferentially selected for expression in OSNs either normally or just in the post-324 insult olfactory mucosa, thus increasing the number of these neurons in the olfactory mucosa as a 325 whole. Alternatively, there is evidence in mice that activated OSNs have a longer lifespan 33 . Because 326 of their ease of activation, these OSNs tend to survive longer and therefore make up a greater 327 proportion of the overall OSN population. 328 3. The specific OR are not over-represented but merely more easily activated, so although many 329

OSNs regenerate and aberrantly innervate the glomeruli, since only a few afferent neurons pass to 330 each individual glomerulus, these few molecules are powerful enough to activate many glomeruli 331 simultaneously at physiological odorant concentrations. 332 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101/2021.02.05.21251085 doi: medRxiv preprint 4. Since axon guidance is at least partially OR-dependant 34 these specific OR-expressing OSNs could 333 be more likely to demonstrate aberrant targeting of glomeruli in an as-yet unknown way. 334 5. It is possible that parosmia does not arise from the activation of the glomeruli per se, but the 335 disruption of the network of interneurons, mitral, glial and tufted cells which is thought to act as a 336 habituation and modulatory network in the bulb. Disruption of input innervation and sporadic re-337 innervation could cause feedback loops to interfere with the previous web of inhibition and promotion 338 of signal at this level and this disordered activation is experienced as unpleasant. 339

Parosmia is a tetrapartite symptom: it is a triggered, short lived, altered smell sensation which almost 340 universally elicits the basic emotion of disgust. Our finding that this is reliably triggered by a common 341 group of low threshold odorants advances our understanding of this debilitating condition and 342 constrains the pathophysiological hypothesis space. 343

That there are reliable molecular triggers of parosmia point to an olfactory receptor level pathology 344 which agrees with the fact that the sensation is triggered by smells and follows the usual pattern of 345 habituation and attenuation expected in an otherwise intact olfactory system. The selective 346 regeneration of only a few OSNs also explains how the odour percept is altered (if the central dogma 347 of glomerular activation patterns within the olfactory bulb is accepted). If only some of these 348 glomeruli are activated out of the previous broader activation pattern recognised as coffee, this will be 349 perceived as a novel smell. 350

What this does not explain is the presumably hypothalamic disgust response to this particular altered 351 odour. The miswiring hypothesis posits that broader, unregulated patterns of glomerular activation are 352 de novo perceived as unpleasant and disgusting, but this has not been demonstrated. Certainly, in the 353 normal nose, novel smell percepts are not usually automatically disgusting, so the mere novelty of the 354 percept is unlikely to be enough to explain this. 355

In this paper we identify the first common molecular triggers of human parosmia, characterised by 356 their physiochemical properties and sharing a low odour threshold for humans. We demonstrate that 357 parosmia is an olfactory dysfunction only partially correlated with olfactory loss, and provide 358 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101/2021.02.05.21251085 doi: medRxiv preprint evidence to support its arising in the periphery of the olfactory system. This information is vital to the 359 understanding of the pathophysiology of this increasingly widespread condition and will be important 360 in guiding further research and future therapies. 361

The authors would like to acknowledge all those who participated in this study, Aidan Kirkwood for 363 assistance with the participants, Peter Jackson for his assistance in sourcing and handling the faecal 364 sample, and Professor Barry Smith for useful discussion and review of the manuscript. 365

Authorship contribution 366 JP contributed to conception, acquisition, analysis and interpretation of data, manuscript draft and 367 review; CK contributed to conception, participant management, data acquisition and review; SG 368 contributed to conception, interpretation of data, manuscript draft and review (SG). 369 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint Committee. All parosmic participants were recruited via Facebook support groups or local ENT 375 consultants, and non-parosmic participants from within the Department of Food and Nutritional 376

Sciences or through private Facebook pages. The initial study was carried out with pre-COVID-19 377 parosmic participants (N=15) and non-parosmic participants (N=15) between October 2019 and 378

March 2020. This was supplemented with post-COVID-19 parosmic participants (N=15) between 379

July and September 2020. All volunteers completed a screening questionnaire (Supplementary Table  380 S2) before attending a study day in the Olfaction Laboratory at the University of Reading. Selection 381 was based on the participants listing coffee as a key trigger, and answering "often" at least once to 382

two key questions which discriminate most efficiently between parosmic participants and those with 383 quantitative olfactory disorders 35 : 384 1) Are odours that are pleasant to others, unpleasant to you? Never/rarely/often/always 385 2) Is the taste of food different to what you expect? Never/rarely/often/always 386 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10. 1101 /2021 Procedure at the odour-port. Subjects were familiarised with the instrument, instructed to breathe 425 normally during the run, and advised that they could stop at any time, particularly if they felt dizzy or 426 light-headed. As the aromas eluted from the column, 3 bits of information were requested from the 427 subjects: an odour description, an odour intensity and an indication of whether the odour elicited a 428 parosmic response. Since the description and identification of aromas in the absence of any other cues 429 is difficult, all participants were presented with a flavour wheel (Supplementary Figure S1 ) before 430 they started which they could use as a reference during the GC-O run. It had been developed by 2 431 experts who sniffed samples of the same coffee (both at regular strength and concentrated) by GC-O. 432

The words were categorised into food and non-food, and colour coded for quick reference. This was 433 of more use to non-parosmic participants, as parosmic participants found it hard to describe many of 434 the aromas, even with the help of the flavour wheel. Many resorted to using the terms "new coffee", 435 "that parosmia smell", "trigger number 1" or "trigger number 2". As each aroma eluted, parosmic 436 participants were prompted to highlight anything that had a parosmic character or trigger. Intensity 437 was scored on a 158 mm horizontal general labelled magnitude scale (gLMS) with anchors at "barely 438 detectable", "weak", "medium", "strong", "very strong" and "strongest imaginable" corresponding to 439 intensity scores of 1.4, 6, 17, 35, 51 and 100 respectively (et al. 2004 ). This was chosen over the more 440 common visual analogue scale to allow for instances where parosmic participants, in particular, 441 wanted to extend upwards the range of scores. It is a logarithmic scale which better relates the 442 psychophysics of perception to the concentration of the stimulus (Stevens's Law). Time of elution 443 was recorded manually by the researcher. All subjects carried out the GC-O of coffee twice, once 444 before lunch and once after a 45 min lunch break. During the second run, the focus was on refining 445 the descriptors as well as obtaining a duplicate intensity rating. During the first run, the subjects 446 recorded the descriptors in their own words, prompted only by the flavour wheel, whereas during the 447 second run, there was more discussion between the researcher and the subject, to verify the odour 448 character and identity of the compound eluting. 449

Gas chromatography-mass spectrometry (GC-MS). An extract from a coffee prepared with one 450 sachet in 3 mL of boiling water was extracted as above and analysed by GC-MS to aid identification 451 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101/2021.02.05.21251085 doi: medRxiv preprint of aroma compounds detected by GC-O and confirm their presence in the coffee extract. A7890A Gas 452

Chromatograph coupled to a 5975C series GC/MSD from Agilent was used, equipped with either of 453 the columns described above. The oven was held at 40 °C for 2 min, increased from 40 °C to 250 °C 454 at a rate of 4 °C/min and then kept constant at 250 °C for 5 min. Helium was the carrier gas at a flow 455 rate of 1.2 mL/min. Mass spectra were recorded in electron impact mode at an ionisation voltage of 70 456 eV and source temperature of 220 °C. A scan range of m/z 20-300 with a scan time of 0.69 s was 457 employed and the data were controlled and stored by the ChemStation software (Agilent, Santa Clara, 458 CA). 459

with the retention times of C6-C25 n-alkane series analysed on the same day using the same conditions 461 as for sample analyses (Supplementary Table S3 Confirmation of molecules as trigger or non-trigger. Three parosmic participants returned to assess 471 coffee on a polar column to confirm the identity of trigger compounds. Once identified, selected trigger 472 compounds were also presented to 2 parosmic participants in dilute form to verify their parosmic 473 character, using the sample preparation protocol described for the European test of olfactory 474 capabilities 36 . Aroma chemicals were diluted in mineral oil or propylene glycol and applied to small 475 discs (5 mm diameter) of absorbent paper in vials which were presented to the participants. They were 476 asked to sniff the vial and indicate whether each compound released "that parosmia smell" which they 477 had described previously. 478 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 Additional samples. All additional samples were prepared as for coffee with the following 479 modifications. Cocoa: 3g of cocoa powder was dissolved in 10 g boiling water, stirred and a 3 g aliquot 480 was used for extraction. Meat: a lean breast fillet, thickness 1 cm was grilled for 3 min on either side 481 using a Cuisinart grill (Stamford, CT) set on high. Finely chopped meat (3 g) was used for extraction. 482

A 50:50 slurry of peanut butter (3 g) was used for extraction. Finely diced red pepper (3 g) was extracted 483 at 40 °C prior to desorption. The faecal sample was thawed, mixed with an equal weight of water, and 484 3 g transferred to an SPME vial Chromatography conditions for all samples remained the same as for 485 coffee. 486

Statistics. Analysis of variance was carried out on age and TDI scores to determine whether there 487 were significant differences between pre-COVID-19 parosmic participants, post-COVID-19 parosmic 488 participants and non-parosmic participants. Post hoc pairwise comparisons were performed using 489

Fishers LSD at p=0.05. Agglomerative hierarchical clustering (AHC) using Ward's Euclidean 490 distance was carried out on intensity data for the 17 compounds most frequently identified in coffee as 491 molecular triggers of parosmia. All statistical analyses were carried out using XLSTAT statistical and 492 data analysis solution (Addinsoft 2020). 493 494 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 Pre-screening Questionnaire Part 2 for those reporting smell disorders Does your problem with your sense of smell or taste relate to your (select one answer) Sense of Smell Sense of Taste Both smell and taste Can you tell the difference between salt and sugar? Yes/No When did you problem start? (select one answer) Less than 3 months ago 3-12 months ago 1-2 years ago More than 2 years ago How did the problem start? (select one answer) Slowly Suddenly I don't know How has your problem changed since it started? (select one answer) There's been an improvement No change It's got worse Not sure What do you think might have caused it? (select one or more answers) Accident Nasal polyps Cold or infection Surgery I was born without a sense of smell Other I don't know Do you have any of these signs/symptoms? (select one or more answers) Stuffy nose . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101/2021.02.05.21251085 doi: medRxiv preprint 

Flavour wheel provided to all participants during GC-O session

Predominance of an altered sense of smell or taste among long-497 lasting symptoms in patients with mildly symptomatic COVID-19