March 7, 2022

expert reaction to study looking at changes in brain regions related to smell following COVID-19

Research, published in Nature, has looked at changes in brain regions related to smell following COVID-19.

Dr Laura Benjamin, Wellcome Clinical Career Development Fellow, Principal Clinical Research Fellow/Honorary Consultant in Stroke Neurology, UCL; and British Neuroscience Association member, said:

“This elegant work exploits an invaluable longitudinal clinical brain imaging cohort (UK biobank) and helps us unravel some crucial questions.  Notably, Douaud et al. provide the first insights into mild (non-hospitalised) Covid-19 brain involvement and uniquely take account of pre-existing brain changes.  The authors demonstrated COVID-19 specific changes in the brain regions that control smell and memory networks.  Although these changes compared with the control group are modest (~2% difference), a higher frequency of cognitive impairment occurs in the COVID-19 group during an average interval of 5 months, suggesting some physical effects from this marginal change.  Although not definitive, the anatomical involvement around the nasal tract and brain spinal fluid interface points to a viral entry route; other studies support these routes.

“We do not know the impact of vaccination and whether these changes are reversible or reached a plateau.  Long-term follow-up of this and other brain imaging cohorts to evaluate the trajectory of brain involvement in COVID-19 may give further insight.”

Dr Alastair Noyce, Reader in Neurology and Neuroepidemiology, Lead at Preventive Neurology Unit, and Honorary Consultant Neurologist, Queen Mary University of London, said:

“This is a well conducted study using data from UK Biobank in participants who have been scanned twice on average 3 years apart.  Half of them had COVID during the interval and these participants were matched with participants who did not test positive for COVID.  The analysis is robust, particularly the efforts to ensure that the two groups were similar at baseline, except for the diagnosis of COVID, which in turn addresses confounding.  Some of the most striking results relate to involvement of the parts of the brain concerned with olfaction (sense of smell), which is a recognised symptom of COVID.  Earlier studies suggested that COVID’s effects on smell were outside the brain, but these results indicate possible changes in the olfactory centres in the brain and changes in connected areas.  I hasten to add that this, as the authors acknowledge, does not mean direct invasion of the brain by COVID, and there are several possible explanations.  However, it is interesting because a putative mechanism for neurodegenerative diseases is entry via the olfactory structures and then spread to other brain structures.  Whilst this article on its own should not generate overt concern, it reasserts the need for dedicated research and surveillance on the potential risk of long-term sequelae of COVID infection.”

Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“This is an important piece of research, and generally it deals appropriately with the statistical aspects, though they are inevitably rather complicated.  However, there are some issues in interpreting the findings, that should be borne in mind.

“This seems to be the first research study that compared measurements on people brains both before and after Covid-19 infection.  It’s pretty obviously more informative to look at the state of people’s brains both before and after infection, than to look only afterwards.  If people who have had Covid-19 had detectable differences in their brains after infection, compared to people who had not had Covid-19, then that could possibly be because those differences were present even before they had Covid-19.  In that case it wouldn’t be possible to get much idea of how they were associated with the infection.  The participants in this new study had scans made of their brains both before and after infection (if they had been infected), as part of a wider study (UK Biobank).  So the researchers could look at changes in their brains, and compare them with changes in the brains over similar periods of time in ‘control’ participants who had not had Covid-19.  Several of the brain areas that the researchers examined did show changes for the worse, that were more marked (on average) in people who had had Covid-19 than in the control group.

“As well as their findings from brain scans, the researchers also looked at changes in performance on some cognitive tests (relating to aspects of thinking, reasoning, memory and so on) that the participants took on two occasions, before and after a Covid-19 infection in those who had Covid-19, and at similar times in the control group.  They did find some differences, on average, in some of the changes in cognitive test scores, between the people who had been infected and those who had not.  I’d say that the statistical evidence for these cognitive differences is perhaps a little weaker than for the differences found in the scans, though evidence of an association is certainly there.

“That’s all concerning, but it doesn’t entirely get us off the hook of not knowing what really caused those differences in the results of scans and cognitive tests.  That’s because the study is observational. Inevitably, there were other differences between the people who got Covid-19 and those who didn’t.  The researchers matched the controls to the people who had been infected in terms of several potentially important factors (age, sex, ethnicity, and the time interval between the two brain scans), but that doesn’t mean that the two groups were perfectly matched on every relevant factor.  So the differences in changes in their brains could have been caused, in whole or in part, by differences in other factors.  The researchers, rightly, do acknowledge this in their report.

“Also, they made statistical adjustments with the aim of allowing for some factors (so-called ‘potential confounders’), that differed between the infected people and the controls, and were potentially also related to brain and cognitive changes.  But, as they again acknowledge, adjustments of that kind can’t be perfect in an observational study, so that (in their words) “one cannot make claims of disease causality with absolute certainty”.  Another possibility could be that the cause and effect partly works in a different direction – that is, there could be differences in some factors that make some individuals more likely to be infected with Covid-19, and also that independently make them more likely to show greater deterioration in their brains.  This may not be very likely, but the study can’t rule it out entirely.

“What I’m saying here is definitely not that the brain and cognitive changes cannot be caused by the infection.  That remains an important and concerning possibility, and the researchers give several arguments as to why they believe it may well be true, even though they can’t be certain.  I’m just pointing out that it’s not the only possibility, and that’s part of the reason why the researchers are rightly calling for more research to investigate why and how the changes happen.  More research is also needed, as they say, to investigate whether the changes continue in the longer term.

“The press release mentions that the researchers also compared brain changes in people who had had pneumonia compared to people who had not, to see whether the differences they found between Covid-19 patients and controls might be associated with respiratory diseases in general, rather than specifically with Covid-19.  They did find some differences in brain changes between people who had and had not had pneumonia, but the differences were not the same as in the Covid-19 comparison.  So, even though we still can’t be certain that the differences in those infected with Covid-19 were caused by Covid-19, they do appear to be fairly specific to Covid-19.”

Dr Rebecca Dewey, Senior Research Fellow in Neuroimaging, University of Nottingham, said:

“I do feel the press release accurately reflects the article findings.  I believe this is very good quality research and that the conclusions are backed up by the data found in the study.  The study reports a large sample size and uses measures that are objective.  The article shows that great effort has been taken to explore any confounding factors that might alter the results, and is very frank and honest about the sizes of effects.  Statistically, they have been highly stringent in their methods.  Evidence in this area is just emerging and little large-scale evidence such as this has yet been produced.  However, these findings are consistent with both the understanding that inflammation affecting the whole body has been associated with these changes, and also consistent with less substantial bodies of work that have been published so far.

“As I always say for Biobank studies, the methodology was designed and approved by a consortium of some of the best scientists in the relevant fields.  There is a huge advantage from the fact that these people were recruited and scanned before COVID was even a thing.  The same people were then scanned at a later date, and so changes reported use each person as their own control subject, making the findings really strong.  This comes with only one caveat; that the original pool of participants in Biobank (the 40,000 people scanned before the pandemic) is a self-selecting sample, with potential differences in characteristics from those of the general population.  I can’t comment on whether or not these characteristics are likely to associated in any way with the likelihood of catching or becoming particularly ill from COVID.  However, the authors searched comprehensively for bias in participant characteristics between those who had and hadn’t experienced COVID and could not find any.

“One of the really interesting findings in this article is the presence of the cognitive differences.  Not only were there robust and significant differences between those who’d had COVID and those who hadn’t, but these differences were also found to be related to the imaging findings in the people who had had COVID.  This is some diligent analysis and good reporting.

“Implications-wise, it’s really hard to tell.  If the findings were based on imaging data alone, I would say that we have much less reason to worry about this as the brain is so plastic that it is likely to compensate in the absence of any conditions preventing this.  However, I find the significant association between the imaging findings and the cognitive tests highly compelling.  The authors took great effort to show that this effect was not present prior to COVID.

“These sorts of changes are seen after many forms of disease onslaught, and even that of healthy ageing.  The key difference shown here is that they appear to be happening faster than with ageing alone.”

Dr Max Taquet, NIHR Oxford Health BRC Senior Research Fellow, University of Oxford, said:

“This is the first large-scale study to investigate the actual changes in the brain that can occur after a COVID-19 infection.  It is well established that COVID-19 infection is associated with subsequent risks of neurological and psychiatric problems in some people including brain fog, loss of taste and smell, depression, and psychosis.  But why this occurs remains largely unknown.  This study starts to shed light on this important question by showing that brain regions connected to the “smell centre” of the brain can shrink after COVID-19 in some people.  These brain changes were not observed in every patients and they were mostly subtle.

“These findings might help explain why some people experience brain symptoms long after the acute infection.  The causes of these brain changes, whether they can be prevented or even reverted, as well as whether similar changes are observed in hospitalised patients, in children and younger adults, and in minority ethnic groups, remain to be determined.  It is possible that these brain changes are not caused by COVID-19 but represent the natural progression of a disease that itself increased the risk of COVID-19.  That being said, attributing brain changes to COVID-19 in any definite way would be incredibly difficult if not impossible because we cannot just randomise people to catching or not catching COVID-19.  To inform us on this question, we therefore rely on excellent observational studies of which this study is an example.”

Prof Alan Carson, Professor of Neuropsychiatry, Center for Clinical Brain Sciences, University of Edinburgh, said:

“My main concern with the study is that it is based on a hypothesis that SARS-CoV-2 enters the brain via the olfactory nerve.  This was a widely held concern at the time of publication of the initial pre print of the results in March 2020 and the very specific, if somewhat unusual, ‘region of interest’ analysis used could be justified on these grounds.  However, there is now widespread agreement that this is not the case and in fact direct infection of the CNS is rare; the brain may be affected by other mechanisms such as immune, inflammatory, vascular or psychological/behavioural change but not direct infection.  This essentially render the purpose of the study void.

“The loss of sense of smell seems to be down to damage in support cells in the nose not the brain.

“I am very concerned by the alarming use of language in the report with terms such as ‘neurodegenerative’.  The size and magnitude of brain changes found is very modest and such changes can be caused by a simple change in mental experience, e.g. bigger changes were reported in the now famous study of cab drivers as they took the knowledge test.  What this study almost certainly shows is the impact, in terms of neural changes, of being disconnected from one’s sense of smell.  It serves to highlight that the brain connects to the body in a bidirectional relationship that is both structurally and functionally dynamic.  We may in that regard use the metaphor of the brain as a muscle.  But I don’t think it helps us understand the mechanisms underpinning cognitive change after covid infection.

“As a side note the cognitive examination results in biobank samples are somewhat unreliable.”

Prof David Nutt, Edmond J Safra Chair and Head of the Centre for Neuropsychopharmacology, Imperial College London, said:

“Many experts in psychiatry and neurology predicted from very early on in the epidemic that the covid19 virus would result in significantly neuro-psychiatric complications in some people.  This paper uses brain imaging to confirm our predictions.  It’s good to see the UK biobank initiative being used to deal with an immediate health question.  The large scale of the BioBank participants database brings the necessary power to reveal significant biological phenomena, as in the current paper.  What we now need is a concerted effort to deal with these brain disorders in the same way as we had massive engagement with ways to deal with the respiratory impact of the  virus for example with the rapid development of ventilators.  The UK government should have started a similar initiative for the brain two years ago: maybe it will now act?”

‘SARS-CoV-2 is associated with changes in brain structure in UK Biobank’ by name of first author et al. was published in Nature at 16:00 UK time on Monday 7 March 2022.

DOI: 10.1038/s41586-022-04569-5

Declared interests

Dr Laura Benjamin: “Research funding from Wellcome Trust and GlaxoSmithKline to study viral infection and brain disease.”

Dr Alastair Noyce: “I don’t have any conflicts of interest.  I have worked with UK Biobank data and run the PREDICT-PD study (predictpd.com) which aims to identify people at future risk of Parkinson’s.”

Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee.  My quote above is in my capacity as an independent professional statistician.”

Dr Rebecca Dewey: “I can’t think of any conflicts of interest.”

Dr Max Taquet: “I have been and I am collaborating with some of the authors of this paper.”

Prof Alan Carson: “My COIs are that I am a paid associate editor of Journal of Neurology, Neurosurgery and Psychiatry. Unpaid president elect of the Functional Neurological Disorders Society and I give independent testimony in Court on a range of topic.  I am also involved in a range of Studies on the CNS effects of COVID including COVID CNS and being chief investigator of Cognitive phenotypic in Long Covid study.  I do not think any of these areas impact on my interpretation of this study – indeed, if anything, my scepticism about these results is rather disadvantageous for my own research portfolio.”

None others received.