Reactions: persistent covid studies have created misunderstandings because of their limitations, says analysis

An analysis published in BMJ Evidence-Based Medicine reviews epidemiological research on persistent covid and concludes that many studies overestimate its prevalence due to overly loose definitions, absent or inadequate control groups and other methodological shortcomings.

26/09/2023 - 00:30 CEST
Expert reactions

Warren Tate - long covid EN

Warren Tate

Emeritus Professor, Department of Biochemistry, University of Otago

Science Media Centre New Zealand

There is no doubt many people have developed a serious post condition following their infection with SARS-C0V-2. The numbers are large because of the unprecedented SARS-Cov-2 infection numbers worldwide - now approaching 800 million. The BMJ paper highlights an important problem as we move forward to understand better the extent and long-term effects of the mixture of conditions classified as Long COVID.

The World Health Organization moved quickly to establish a clinical case definition for what was known to be a heterogeneous group, at the very least containing people with ongoing organ damage, and groups with a classic post viral fatigue condition reminiscent of Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS). Despite this, it was surprising that supposedly well researched papers have come out with a huge range of 10-50% of those with COVID, although more recently it has dropped in some publications below 10%, with variants like those of the Omicron group causing a lower incidence.

Hence, there has been a pressing need for better definition with precise controls and diagnosis of the Long COVID condition and in this sense the BMJ paper is a timely reminder. Particularly pressing is to have clear definition of the clinical phenotypes. A very recent paper defined four clinical phenotypes, a post viral fatigue syndrome like ME/CFS, a respiratory syndrome, a pain syndrome, and a neurosensory condition in a ratio of ~4:2;2;1. So, the need as outlined in this BMJ paper for better definitions and precise controls so targeted help for families, communities, and countries can be better defined, is timely.

The call not to divert precious funding towards Long COVID and away from other needs - while having vaildity - ignores the fact that for decades the collective group of post viral syndromes (ME/CFS) has been poorly resourced, and the affected largely ignored. This new interest in Long COVID has benefits for those people who have been living with such debilitating conditions for many years. But the BMJ paper principle is important in that these affected people may have even less support if the focus is totally on Long COVID. But it would be a tragedy if those affected with Long COVID become another subgroup of the ‘missing millions’.

The author has declared they have no conflicts of interest

Michael Baker - long covid EN

Michael Baker

Professor, Department of Public Health, University of Otago, Wellington

Science Media Centre New Zealand

I agree with the first message of this paper, which is that we need well designed studies to provide a valid measure of the long-term effects of acute COVID-19 infection (Long COVID).

Such studies need robust case definitions, adequate duration of follow-up, and suitable comparison groups. Now that the pandemic has been with us for more than three years, we should expect such high-quality studies. Ultimately, these studies should provide us with consistent evidence about key aspects of the epidemiology of Long COVID – including its prevalence, severity, and duration. The published literature is not yet giving us this clear consistent picture.

“The second message of this paper appears to be that there is a negligible risk of Long COVID, based on the selection of papers they have quoted.  That message is concerning and does not fit with mainstream scientific evidence, nor the experience of the large population of people living with Long COVID and the clinicians caring for them.  There is now overwhelming evidence that SARS-CoV-2 carries a significant risk of long-term effects.  This evidence doesn’t just come from epidemiological studies, but also from studies looking at the severe and lasting pathological changes that occur following SARS-CoV-2 infection.

“Even if the risk of Long COVID from a single infection is now relatively low with Omicron subvariants, and with vaccination fortunately pushing it down even further, it remains a serious problem that we need to manage. COVID-19 has infected most of world’s population and is continuing to cause multiple reinfections for many. Consequently, the population level impact of infection is large and growing. There is also the potential that some of the long-term effects of infection with this virus may take years to manifest themselves, as we have seen with other viral infections.

One remarkable feature of this paper severely damages its credibility. It includes a section titled The most well-designed studies provide reassuring estimates.  In it the authors include just two studies to support that sweeping statement: one was in children, and in the other they reported and commented on the analysis of findings for people younger than 50 years. In doing this highly selective ‘mini meta-analysis’ they are contradicting the core message of their paper. This very biased treatment of the subject suggests that these authors have an underlying view of Long COVID, rather than the evidence-informed perspective they are promoting.

The author has declared they have no conflicts of interest

Amanda Kvalsvig - long covid EN

Amanda Kvalsvig

Associate Professor, epidemiologist in the Department of Public Health, University of Otago Wellington

Science Media Centre New Zealand

The authors of this study usefully point out that because the umbrella term Long COVID is a broad definition, it's challenging to design population studies that give a precise measure of Long COVID prevalence in a population. The discussion about pitfalls in study design is good. But this paper is highly selective in terms of the studies that the authors choose to highlight.
The lead author Høeg is linked with think tank groups that have been strongly advocating against public health measures throughout the Covid-19 pandemic. These links may explain the more optimistic view of this author group about Long COVID risks.
The authors have omitted to mention that alongside population surveys there is an extensive and active field of research to understand the impacts of this virus on the body. The science about longer-term effects of Covid-19 infection is now very much better understood than it was even a year ago. The research shows that COVID-19 is a multisystem disease that can cause microclots, changes to the immune system, viral persistence in tissues, and other effects even in mild cases; this major paper in Nature Reviews Immunology provides a detailed and recent overview. These effects are the basis for well-described longer-term effects of COVID-19 such as impaired brain function, extreme fatigue, and stroke. These tissue-level effects can also be silent (causing no symptoms) but they are known risk factors for heart disease and other conditions, raising concerns for population health in the future, with potentially large numbers of people experiencing poor health in years to come.
Even the research that is mentioned in this article is not as reassuring as the article seems to suggest: 

  • Antonelli et al. reported lower risk of Long COVID for Omicron compared with Delta infection, depending on age and time since vaccination. But they noted that Omicron variants have caused far higher case numbers and the conclusion of their article is that 'future numbers with Long COVID will inevitably rise.'
  • Studies that use serology to identify past infection have an important limitation that some people (including children) don’t produce long-lasting antibodies despite having had the infection, leading to unreliable results in Long COVID research.
  • Studies using ONS data are not designed to detect and measure the longer-term effects that are described in other research.

All pandemics and major epidemics cast a long shadow of chronic disease that can continue to emerge decades after the initial emergency phase. In 2023 we are still experiencing high levels of COVID-19 infection and reinfection in Aotearoa New Zealand. Reducing COVID-19 infection rates is still the only reliable way to reduce the longer-term impacts on population health.

The author has not responded to our request to declare conflicts of interest

Mona Jeffreys - long covid EN

Mona Jeffreys

Associate Professor (Research) in Epidemiology, Victoria University of Wellington

Science Media Centre New Zealand

The paper is a careful assessment of the pitfalls of epidemiological studies. It is worth remembering that some of these were conducted during a time of significant upheaval.

The authors make good points regarding the lack of consistency in case-definitions, and other potential sources of bias. It is worth noting that all definitions, including the one used in Aotearoa New Zealand, are broadly consistent, other than the American one (CDC). 

These are useful reminders about how to carry out careful epidemiological investigations. 

However, the authors do not appear to have been informed by the Myalgic Encephalomyelitis/ Chronic Fatigue Syndrome literature; they appear to prefer to focus on a set of pre-defined, persisting, single post-COVID symptoms, rather than considering Long COVID as a term for a syndrome or collection of symptoms which can fluctuate in presence/ absence as well as severity.

"There is no evidence for the “unintended consequences” listed in the key messages box, these are not mentioned in the paper, and appear to be the authors’ impressions, rather than evidence-based statements. 

The author has not responded to our request to declare conflicts of interest
How methodological pitfalls have created widespread misunderstanding about long COVID
  • Peer reviewed
  • People
BMJ Evidence-Based Medicine
Publication date

Høeg TB, Ladhani S, Prasad V.

Study types:
  • Peer reviewed
  • People
The 5Ws +1
Publish it