Is Long COVID Really 'Long Everything'?

F. Perry Wilson, MD, MSCE


June 21, 2023

This transcript has been edited for clarity.

Welcome to Impact Factor, your weekly dose of commentary on a new medical study. I'm Dr F. Perry Wilson of the Yale School of Medicine.

These days, when a patient presents with symptoms of a viral respiratory infection, we can usually determine which virus is to blame within hours, thanks to lab testing that has become fairly routine. But it doesn't stop us, nor should it, from guessing beforehand. We've all learned that viruses have tells, after all. Flu announces its presence with deep muscle aches. RSV, with wheezing. But COVID, to me at least, always felt a bit apart from these other pathogens. The sometimes permanent loss of the sense of smell is such a specific and bizarre finding. And then, of course, there is long COVID, a syndrome that has been devilishly difficult to define clearly but seems to crystallize our modern, post–vaccine-era concerns about the virus.

But is this postillness syndrome unique to COVID infection? Or have we simply failed to understand that in reality, there is long everything?

The study we are looking at this week acknowledges a host of post-COVID conditions that occur with some frequency after a COVID hospitalization. But the heart of any epidemiologic study is the control group. Should we compare individuals hospitalized with COVID to healthy people from the general population?

This paper, by Kieran Quinn and colleagues, argues strongly against that.

They give us three potential control groups against which to compare people hospitalized with COVID (N = 26,499):

  • Individuals hospitalized with influenza in the pre-COVID era: 17,516

  • Individuals hospitalized with sepsis in the pre-COVID era: 282,473

  • Individuals hospitalized with sepsis during COVID but who were not COVID infected: 52,878

The study leverages the remarkable universal electronic health record system of Ontario, Canada, to compile what amounts to complete data capture on nearly 400,000 adults who were hospitalized with one of the conditions of interest and, importantly, survived that hospitalization. Can't get long COVID if you don't survive COVID, after all.

Patients hospitalized with COVID are quite different from those hospitalized with flu or sepsis. They are substantially younger, with a mean age of 61 compared with 74 or 75. They are less likely to be female; it has become quite clear that men tend to do worse with COVID than do women. It's worth noting that only 15.3% of the COVID group had received a dose of vaccine, compared with 35% of the sepsis group hospitalized during the pandemic. COVID patients were markedly less likely to carry a cancer diagnosis or hypertension, and they had a lower frailty score. In other words, the COVID patients, prior to getting COVID, were healthier.

Nevertheless, patients with COVID were more likely to end up in the ICU and require mechanical ventilation. So I want to be clear on this as we move forward: COVID is worse than flu, even worse than sepsis — at least for hospital survivors.

We need to take severity of illness into account when we think about long COVID. The syndrome may represent the unique pathophysiologic consequence of COVID itself or it might represent the sequela of any severe illness; it's just that COVID happens to be a particularly severe illness.

The researchers used something called "propensity score overlap weighting" to account for the baseline differences among patients as well as for the stuff that happened during the hospitalization — essentially accounting for the particular severity of COVID. This is a rather novel method to account for traditional confounders, which I really like and my lab has used in a few earlier papers. You get very balanced groups after the process. Sorry for the inside-baseball.

Okay, the big question. If we account for the fact that people who get hospitalized with COVID tend to be healthier than people hospitalized for flu and sepsis, and also the fact that people hospitalized with COVID tend to have a worse course than people hospitalized for flu or sepsis, who would have worse outcomes after discharge?

Rather than creating some arbitrary long-COVID definitions, the researchers looked at a variety of specific disease outcomes, which is much more informative.

And here's the rub: There was almost no difference between those who survived a hospitalization with COVID and those who survived the other conditions. They had similar long-term rates of heart attack, dementia, depression, and stroke.

The risk for venous thromboembolism was higher among COVID patients compared with those hospitalized with influenza, a finding that may be real; COVID does seem to be a relatively prothrombotic infection. But overall, the message of this study is yes, bad things happen after COVID, but bad things happen after any severe illness.

It's of course critical to mention that the outcomes the authors looked at were data that can easily be extracted from the electronic health record — hard outcomes like stroke — but we don't have data on more subtle presentations of long COVID (think fatigue, brain fog).

And, of course, the study doesn't say anything about patients who had COVID but never required hospitalization. There may well be a long-COVID syndrome in that population, but finding an appropriate control group for outpatients with COVID would be incredibly difficult.

I do have one niggling concern about this study, though. The overall 1-year mortality rate was dramatically higher in the control groups than the COVID groups.


Around 6% of those who survived their COVID hospitalization died within the following 12 months. Around 12% of those who survived a flu hospitalization, and 25% who survived a sepsis hospitalization, died in the next 12 months. That makes sense, of course; the COVID group was much younger and healthier overall.

But the authors treated death as a "competing risk" in their analysis, which means that people who die don't contribute to the rate of other conditions. You can't have a stroke if you die of something else before you have a stroke. It's like saying that ingesting cyanide reduces the risk of getting struck by lightning. Sure it does, but not really in the way you want.

This framework would bias the results against COVID by deflating the rate of events in the control groups. Censoring for death, rather than treating it as a competing risk, would reveal this difference, but the authors don't report that analysis.

So, does long COVID exist? Yes, absolutely. Is long COVID different from long flu or long sepsis? That is less clear. The real question is whether there is a syndrome unique to COVID that we can identify. That has proven somewhat difficult. Future research should certainly investigate clotting phenomena, but it should also consider whether the higher death rate of those hospitalized for other conditions makes COVID look worse than it should.

F. Perry Wilson, MD, MSCE, is an associate professor of medicine and director of Yale's Clinical and Translational Research Accelerator. His science communication work can be found in the Huffington Post, on NPR, and here on Medscape. He tweets @fperrywilson and his new bookHow Medicine Works and When It Doesn'tis available now.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.