Long COVID: A review of current knowledge and perspectives

Dr Manali Mukherjee, assistant professor in the Division of Respirology, McMaster University’s Department of Medicine, translational scientist, and immunologist, discusses long COVID in the context of other known postviral syndromes and emerging data.

This interview was recorded at the end of December 2022. The transcript was edited for clarity.

James Douketis, MD: Hello everyone. My name is Jim Douketis. I’d like to welcome you to another edition of the McMaster Textbook of Internal Medicine McMaster Perspective series.

I’m delighted to be joined today by Assistant Professor Manali Mukherjee, who is an assistant professor in the Department of Medicine at McMaster. She’s heavily involved in research relating to coronavirus disease 2019 (COVID-19) and in particular to long COVID. This is a bit of a gray zone in which the knowledge is rapidly evolving.

We’re going to lean on you, Manali, to share your expertise, but perhaps we can get started by stepping a little bit back and looking at COVID-19 and long COVID in the context of other postviral syndromes. Maybe you can just in your own words explain to us what are the key features of long COVID and how does this compare to the features of other postviral syndromes that may have an overlap in terms of their characteristics?

Manali Mukherjee, MSc, PhD: Thank you so much, Jim, and that’s a great question to start with. Before I even go into it, I’d like to remind the definition as it stands today. The World Health Organization (WHO) definition has been changing quite remarkably because it’s a very dynamic space that we’re looking into.

Currently the definition of long COVID stands as anybody with a confirmed or probable case of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection who is having some persistent symptoms at least 3 months after the initial COVID-19 and that persists for another 2 months at least and thereafter. These symptoms in the majority are fatigue, cough, dyspnea, and neurocognitive [disorders]; there’s a myriad, a constellation of symptoms, but these are the main ones that are being reported.

These symptoms may have been lingering on since their acute phase of COVID-19 or could be new onset, which means the patient or the individual could have been completely asymptomatic and then suddenly gets this onset of symptoms, where they do not feel 100% of their health. They may relapse back or they may actually completely get over it and then suddenly have a new other set of symptoms. Say, if they had fatigue before and they are fine, back to work and feeling good, and then maybe they start getting more inflammatory situations, like conjunctivitis, diarrhea, or maybe skin rashes. Every case is quite unique. You’re seeing a number of case reports and I have been speaking to a lot of persons with lived experiences, so I’m talking on that level.

Since I’ve been talking to a lot of people with lived experiences, one of the things that I ended up interacting with a lot is those who have had lived experiences with other postviral syndromes, for example, after a bad bout of influenza, cytomegalovirus (CMV), or Epstein-Barr virus (EBV). Their symptoms, like postexertional malaise or myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), and even fibromyalgia, these are very gray zones; as I understand, it’s very difficult for physicians to really put a diagnosis down and to help the patients. But these are there.

If I take a stand back and I look at it from the immunology perspective, any infection leads to a hyperinflammation and the body always tends to go back towards homeostasis. Evolutionarily speaking, our immune system can obviously fight an external virus or bacteria, but it also has protective barriers, which can bring the inflammation down after the threat has been dealt with. So it is not new, or it is not completely a new finding, that after an infection like COVID-19 some individuals are having these postviral-like symptoms, where it’s lingering on for longer than expected.

The problem is the numbers that we have seen. The number of people it’s happening to is much bigger because—let’s just face it—COVID-19 has infected way more people. Right now, coast to coast we are looking at 60% of the Canadian population, irrespective of vaccination, who have been infected in some way or the other. If we think it may be 10%—which is a conservative number [of how many people infected by SARS-CoV-2 may develop long COVID] that the WHO has given and it is kind of correct—so if there is 10% to 20% of this total, we are looking at numbers north of 300 right now who will have long COVID. [May 2023 update: According to a Statistics Canada report, in August 2022 ~1.4 million Canadians were reporting long COVID symptoms 3 months post recovery. From what is currently understood, most of these individuals will recover but a little over one-third will remain symptomatic beyond 12 months.]

Again, the long COVID, the data that is coming out—we will go to that later. But what about how it looks? Sometimes it also looks like postexertional malaise, so it is similar to those who had it from any other virus. There’s a lot of work that is going on to just see the similarities, so that you don’t end up reinventing the wheel. Many, many researchers are working with those who work on ME/CFS and they are trying to get their expertise, knowledge, and understanding of what they know from those postviral diseases.

There are certain cases of the Middle East respiratory syndrome (MERS) and SARS-CoV-1 as well, but you didn’t reach these kinds of numbers. And there is obviously a little bit of public hysteria also to add to it, because people just didn’t know what was going on. I think the medical community as a whole was all taken together to take care of keeping our intensive care units (ICUs) little less... to keep the people alive. And then the ones who were alive... you were not looking at them, right? I mean, think about it. You were just trying to keep patients off the ICU and trying to keep the patients alive. So I guess there was a little bit off balance there and anxiety crept in.

James Douketis: Certainly, we’re slowly focusing more on some of the long-term consequences now because, as you said, we’re past that acute phase, where we were dealing with the more critically ill or those who were hospitalized. But the numbers that you referred to earlier are really concerning. What I’d like to now ask you about is the paper that we were reviewing [see Publications of the Week: Long COVID] that was just published in the Journal of the American Medical Association (JAMA). They looked at a large number of populations and the numbers that they came up with for the different syndromes that you nicely outlined related to long COVID were hovering around 3% to 7%. And the numbers were worse in women than in men.

When you looked at these data at first blush, what was your view? Were these numbers higher or lower than expected? Are we missing cases? What do you think of the overall scope of the problem based on the prevalence? At the end of the day, if a patient recently had COVID-19, what would you advise that the doctor tells them? “You’ve got a 5% or 10% chance of developing these symptoms”? Can you maybe comment a little bit on the paper and numbers?

Manali Mukherjee: If I remember, the paper is a systematic review and meta-analysis, so it is taking into picture all the different papers that have been published and is getting an idea out of that.

If you look at it conservatively, yes, it’s kind of on the right brink. But if we dig a little bit deeper, the questions always arise about the root papers. Whenever I read a meta-analysis, I always go back to the root papers. And one slight problem that I’ve always had is that if you do a cross-sectional study and the sample size is not great, then the answers become a problem. So that’s why a meta-analysis really helps.

In this meta-analysis in all the root papers the definition of long COVID based on the time that it needs to be seen is the crux of the matter: if you are looking at a 3-month window versus a 6-month window versus a 12-month window—we know people even at the 24-month window who were still suffering. But the truth, at least from my dataset or what I have been understanding and what I’ve been seeing as well, is that 75% to 80% of individuals who have had COVID-19 and were convalescent do get better over 12 months. So there is a significant attenuation of the patient-reported outcomes and along with that the molecular mediators of inflammation also subside. But it’s in a good 75% to 80% of individuals; let’s just say 75%, because that’s what my dataset was saying. But it’s also true that I was biased in my dataset, because most of them were hospitalized, so they would be seeing the doctors more. That bias is also there. But in general, the numbers seem to be right, especially the numbers that look into the phenotypes.

The dataset that came out of the National Health Service (NHS) United Kingdom survey, which was I think late summer this year [2022]… they looked at very similar phenotypes: a cardiorespiratory phenotype, an inflammatory phenotype, and a neurocognitive phenotype, and there is a lot of overlap within that. The percentages of the phenotypes that this paper has published are round about very similar to what is being reported even with the different cross-sectional cohorts and with all their biases.

So round about the numbers are kind of there: 50% to 60% of the patients invariably report fatigue, there is no doubt to it. I have been looking at fatigue not just with patient-reported outcomes, because it’s such a vague term to understand; what fatigue means for me might vary from what fatigue for another person is. So we use a fatigue assessment score, which is one of the scales that’s used very commonly in rheumatology. It has a confirmed minimal clinically important difference (MCID) and a threshold of clinical relevance. Based on that also, yes: 50% to 60%, so it’s very similar to what the study has said. So I can agree with that.

The respiratory phenotype, again, depends on how you are defining it. All the root papers had different definitions that I went into but in general [included] cough, dyspnea—mostly exertional dyspnea, so exercise induced. They all have a modified Medical Research Council (mMRC) scale of, say, 1 or 2 or more, but never 4, which means they are short of breath when they go up the stairs or walk up the hill. So, in the respiratory phenotype the percentage that they said is pretty much similar.

Then the inflammatory phenotype, which is basically rashes, swollen joints, and being febrile—some of them keep a low-grade fever—plus the fatigue. And then there’s the neurocognitive phenotype. The percentages that they’ve given of the phenotypes are very similar and if you notice, there is a lot of overlap. In the study that I am also doing the percentages are very similar, but the overlap—an individual having 3 or more of these phenotypes—is ~20%. So 20% of my cohort could fall into the fatigue-neurocognitive phenotype, or the fatigue-inflammatory phenotype, or the [fatigue]-respiratory phenotype. I think those people are the ones who still have a very high inflammation in their serum that I can detect. So I can believe exactly what’s coming up here in the study.

James Douketis: Do you think that these numbers are going to be different—or they are different based on your knowledge—depending on a patient, an individual’s vaccination status, or depending on the variant? Are we seeing lower incidences of these syndromes in, let’s say, vaccinated patients or people who have had the more recent Omicron or BA.5 variant? What do you think about that?

Manali Mukherjee: That is actually the question right now that is being really heavily looked upon and the COVID-19 Immunity Task Force is really trying to get to that. The early evidence that you’re getting is that it seems like with vaccination the numbers of the individuals who are reporting—I don’t call them patients because they do not have a diagnosis, but they are individuals who are reporting symptoms—are relatively lower.

One study came out showing that the chance of one getting a post-COVID complication if they had been vaccinated at least once or twice is ~60% less. But that was also a very limited study, because it was cross-sectional, the sample size was of concern. So I will not absolutely say it because initially, if we look at the data, they had said that those who were hospitalized or those who required oxygen supplementation or ICU admission were more often predicted to have post-COVID conditions, whereas later on the data came in saying—and my own experience with our ongoing study is—that home-recovered people can just as get very severe symptoms to the point that we have a few who have been diagnosed already, including one person who actually has lupus. Three have postural orthostatic tachycardia syndrome (POTS), one has vitiligo. They were previously healthy people.

James Douketis: That’s really interesting but worrisome at the same time.

We’ve learned a lot from you in a short period of time. Do you have any parting pearls? If you look into your crystal ball, what can you see in the future in regard to long COVID?

Manali Mukherjee: Before I even say that, I wanted to particularly pick up on one point, which I kind of missed out myself. It’s that this paper that you’re discussing had a very key point, and I wanted to end on that key point too. They said, “An estimated 15% continued to experience symptoms at 12 months.” I think this 12-month brink is extremely important because, like we were chatting, we hope people can come back to homeostasis after an infection, right? For some reason with COVID-19 it varies from individual to individual, but what we’re seeing at 12 months and more, if you are still having persistent symptoms at 12 months and thereafter, that’s when I think you just need to be seen or at least get some medical attention. That percentage that they have given, 15%, is more or less correct. So if there is a patient subset who are getting very worried at the third or fourth month post COVID they are very symptomatic, unwell, and they are 50% of their health, they most invariably will get better by 12 months. But there’s 15% of them who will not and who might require medical attention.

I think the 12-month brink is an extremely important point, and research going ahead should look at this 12-month cutoff. We should be looking at that, not the immediate 3 to 4 months. It’s the 12-month cutoff that is really important. That’s just what I would take as the crucial takeaway message from this paper.

James Douketis: Dr Mukherjee, I want to thank you very much for these insights. Clearly, we’re learning a great deal and we look forward to hearing more about your research, which is at the cutting edge. Good luck and please let us know if there are any further developments. We’ll have you back for an interview. But once again, thank you very much.

This is Dr Manali Mukherjee, Department of Medicine, McMaster University. Thank you very much for your time and for this interview.

Manali Mukherjee: Thank you so much. My pleasure.