Wearable data reveals distinct characteristics of individuals with persistent symptoms after a SARS-CoV-2 infection, 2025, Ledebur

https://www.nature.com/articles/s41746-025-01456-x

Wearable data reveals distinct characteristics of individuals with persistent symptoms after a SARS-CoV-2 infection

• Katharina Ledebur,
• Marc Wiedermann,
• Christian Puta,
• Stefan Thurner,
• Peter Klimek &
• Dirk Brockmann

npj Digital Medicine volume 8, Article number: 167 (2025) Cite this article

Abstract

Understanding the factors associated with persistent symptoms after SARS-CoV-2 infection is critical to improving long-term health outcomes.

Using a wearable-derived behavioral and physiological dataset (n = 20,815), we identified individuals characterized by self-reported persistent fatigue and shortness of breath after SARS-CoV-2 infection.

Compared with symptom-free COVID-19 positive (n = 150) and negative controls (n = 150), these individuals (n = 50) had higher resting heart rates (mean difference 2.37/1.49 bpm) and lower daily step counts (mean 3030/2909 steps fewer), even at least three weeks prior to SARS-CoV-2 infection.

In addition, persistent fatigue and shortness of breath were associated with a significant reduction in mean quality of life (WHO-5, EQ-5D), even before infection.

Here we show that persistent symptoms after SARS-CoV-2 infection may be associated with pre-existing lower fitness levels or health conditions.

These findings additionally highlight the potential of wearable devices to track health dynamics and provide valuable insights into long-term outcomes of infectious diseases.

 

Not sure whether this tells us much about ME/CFS and similar given:
“What was surprising was that, for both groups, heart rate and step count eventually returned to pre-infection levels,” Ledebur notes. “Despite ongoing shortness of breath and/or fatigue, affected individuals took as many steps as they did before infection,” says Ledebur.

-----
https://www.eurekalert.org/news-releases/1077453

News Release 20-Mar-2025
Smartwatch data: Study finds early health differences in long COVID patients

People who later experienced persistent shortness of breath or fatigue after a SARS-CoV-2 infection were already taking significantly fewer steps per day and had a higher resting heart rate before contracting the virus, according to a CSH study

Peer-Reviewed Publication
Complexity Science Hub




image:

Mean resting heart rate (top) and mean daily steps relative to the average of individuals with Covid-19 infection but without persistent symptoms (bottom) before, during, and after a Covid-19 infection. The pink line represents patients who developed persistent symptoms, while the blue line represents those who did not.

view more


Credit: Complexity Science Hub

[Vienna, 19.03.2025]—Between April 2020 and December 2022, over 535,000 people in Germany downloaded and activated the Corona Data Donation App (CDA). Of these, more than 120,000 voluntarily shared daily data from their smartwatches and fitness trackers with researchers, providing insights into vital functions such as resting heart rate and step count.

“These high-resolution data served as the starting point for our study,” explains CSH researcher Katharina Ledebur. “We were able to compare vital signs in 15-minute intervals before, during, and after a SARS-CoV-2 infection.”

Higher Resting Heart Rate And 3,000 Fewer Steps

People with persistent symptoms showed significant differences in activity levels and resting heart rate compared to others. “In the three weeks before infection, patients who later reported persistent symptoms took an average of just 5,075 steps per day—about 3,030 fewer than SARS-CoV-2 patients who recovered without persistent symptoms,” says Ledebur.

Resting heart rates also showed differences before infection. Those who later developed persistent symptoms had a resting heart rate 2.37 beats per minute higher than individuals who recovered without lingering effects. Additionally, they exhibited a more pronounced and prolonged phase of slowed resting heart rate—known as bradycardia—lasting up to 18 days after infection.

“What was surprising was that, for both groups, heart rate and step count eventually returned to pre-infection levels,” Ledebur notes. “Despite ongoing shortness of breath and/or fatigue, affected individuals took as many steps as they did before infection,” says Ledebur.

Two Persistent Symptoms

For the study, researchers divided smartwatch and fitness tracker data into four phases: before infection, during the acute infection (0 to 4 weeks after a positive test), the subacute phase (5 to 12 weeks after a positive test), and the post-acute phase (beyond 12 weeks).
Participants also completed monthly and weekly surveys about Covid-19 tests, their subjective well-being, and ongoing symptoms.

“Using these symptom reports, we found that among 11 symptoms analyzed, only shortness of breath and fatigue persisted beyond the acute phase of SARS-CoV-2 infection. Other symptoms, such as cough, headache, or fever, showed no long-term effects,” Ledebur explains.
Overall, 2.6% of all SARS-CoV-2-positive individuals reported prolonged shortness of breath, 10.4% experienced long-lasting fatigue, and 1.8% had both symptoms.

Raising Awareness And Protecting At-Risk Individuals

The findings of this study suggest that persistent symptoms may be linked to lower fitness levels or preexisting health conditions.
“However, this does not mean that a higher resting heart rate, lower step count, or preexisting conditions alone are responsible for these symptoms,” Ledebur emphasizes. “Rather, they highlight the need for special attention and protective measures for affected individuals,” says Ledebur, who is also a PhD candidate at the Medical University of Vienna.

Exceptional Data Quality

Thanks to the high-resolution smartwatch data from the Corona Data Donation App (CDA), researchers were able to compare not only different patients but also individual changes over time. “If you examine people after they have been infected, you cannot assess individual changes because you lack baseline information,” explains Peter Klimek from CSH and the Medical University of Vienna.

“Despite the advantages of such data, there are limitations: Men are overrepresented, while both adolescents and individuals over 65—who have a higher risk—are underrepresented,” Ledebur notes. Additionally, people who wear smartwatches or fitness trackers may generally be more health-conscious.

“Nevertheless, the study highlights the enormous potential of smartwatches to provide valuable, objective insights into the physiological and behavioral effects of an infection and to help identify individuals at higher risk,” the researchers conclude.

The CDA was launched by the Robert Koch Institute under the leadership of Dirk Brockmann, CSH external faculty member and professor at TU Dresden.

About the study

The study “Wearable Data Reveals Distinct Characteristics of Individuals with Persistent Symptoms after a SARS-CoV-2 Infection” by K. Ledebur, M. Wiedermann, C. Puta, S. Thurner, P. Klimek, and D. Brockmann was published in npj Digital Medicine (doi: 10.1038/s41746-025-01456-x).

About CSH

The Complexity Science Hub (CSH) is Europe’s research center for the study of complex systems. We derive meaning from data from a range of disciplines—economics, medicine, ecology, and the social sciences—as a basis for actionable solutions for a better world. Established in 2015, we have grown to over 70 researchers, driven by the increasing demand to gain a genuine understanding of the networks that underlie society, from healthcare to supply chains. Through our complexity science approaches linking physics, mathematics, and computational modeling with data and network science, we develop the capacity to address today’s and tomorrow’s challenges.

Journal
npj Digital Medicine

DOI
10.1038/s41746-025-01456-x

Method of Research
Data/statistical analysis

Subject of Research
People

Article Title
Wearable data reveals distinct characteristics of individuals with persistent symptoms after a SARS-CoV-2 infection

Article Publication Date
19-Mar-2025

COI Statement
The authors declare no competing interests.

Disclaimer: AAAS and EurekAlert! are not re

 

Persistent symptoms following Sars-Cov-2 infection were defined as reporting "both shortness of breath and fatigue at least five times after their positive SARS-CoV-2 test."

Based on these criteria you would expect that they exclude patients with fatigue and shortness of breath at the study start but I only found this:

So perhaps pre-existing fatigue and shortness of breath were counted as a persistent symptom following infection, even if it was not related to the infection? The supplementary material for example shows that 24% of the persistent symptoms group had asthma compared to 9% in the control groups.

 

Seems to be the case for the proportion that report fatigue and headache, which did not increase much in the COVID-19[+]PS group as it was already high. For shortness of breath there is a strong increase.

 

Why did they only study 350 out of the 20,000+ participants?

 

Spoiler: Cohort creation details, long

Cohort creation
To study reported symptoms relative to a reported positive SARS-CoV-2 test (COVID-19[+]PS) and compared them with a negative control (COVID-19[−]) cohort we looked at 35,355 individuals who filled out the corresponding surveys. We excluded 3,660 individuals who reported a positive SARS-CoV-2 test in a previously distributed one-time questionnaire to avoid previous infections with SARS-CoV-2 in the population. We divided the population into two groups: those with at least one positive SARS-CoV-2 test and those with all negative tests. Both cohorts are limited to the period between October 21, 2021, and November 9, 2022. For individuals with positive tests, we used the date of their first positive test if multiple tests were positive. To avoid including individuals with re-infections, we excluded those reporting a test more than 28 days after the chosen test date, resulting in 7839 individuals. Among the 18,282 individuals reporting only negative tests, we randomly selected a test date. We also excluded users who reported shortness of breath more than four times before their test result, as a proxy for pre-existing respiratory conditions. To account for seasonal effects in symptom reporting, we adjusted the negative cohort to match the relative frequency of negative reports per calendar week and the temporal distance to the test week observed in the positive cohort. This yields a positive cohort (COVID-19[+]NS) of 7691 individuals and a negative control (COVID-19[−]) of 12,987 individuals (full cohort, Fig. 1f).

For the analysis of WHO-5 and modified EQ-5D, we selected individuals from COVID-19[+] who reported both shortness of breath and fatigue at least five times after their positive SARS-CoV-2 test. This subset, comprising 137 individuals, forms the persistent symptoms cohort (COVID-19[+]PS). We compared their survey responses with those of the remaining 7,669 individuals in the Positive Cohort (referred to as the positive control cohort (COVID-19[+]NS)) and the 12,987 individuals from the negative control (COVID-19[−]) cohort (differentiated cohort Fig. 1g).

To study differences and changes in physiological parameters (RHR) and behavior (step count) between the COVID-19[+]NS, COVID-19[−], and COVID-19[+]PS cohorts, we analyzed fine-grained wearable data collected by wearable devices. We standardized the data by aggregating high-frequency measurements into 15-minute intervals. Four phases relative to the reported SARS-CoV-2 test are defined: pre-phase (start of data collection until the week of the positive test), acute phase (zero to four weeks after the test), sub-acute phase (five to twelve weeks), and post-phase (twelve weeks or more after the positive test), aligning with recent definitions of COVID-19 disease stages24. After excluding individuals with less than 100 15-minute interval heart rate measurements across all phases, the persistent symptoms match (M−COVID-19[+]PS) cohort comprises 50 individuals. We created (1:3) age- and gender-matched positive match (M−COVID-19[+]NS) and negative match (M−COVID-19[−]) cohorts, each with 150 individuals selected from the COVID-19[+]NS and COVID-19[−] cohorts, respectively (match cohort, Fig. 1h). We chose a 1:3 matching, common in medical research, to increase statistical power and to improve the generalizability of our findings by ensuring that the control group more accurately represents the population from which the cases and controls were drawn.

 

Thank you!

So they only looked at first time infections?

Am I correct in saying that only 137 out of 7839 individuals with persistent symptoms after an infection reported both fatigue and shortness of breath?

This would certainly indicate that the category of ‘persistent symptoms’ included people that had those symptoms before the infection - because fatigue and shortness of breath are some of the most common persistent symptoms. As others have mentioned above.

Surely, we would expect a far larger share to have both symptoms?

So 87 out of 137 of those did not adequatly participate in the wearable part of the study?

I’m not sure this data can be used for anything if my interpretation is correct. There are just too many sources of bias, especially when you consider the skewed demographics.

 

At least, in that cohort, the finding that “Despite ongoing shortness of breath and/or fatigue, affected individuals took as many steps as they did before infection” is yet another nail to bang into the coffin of the BPS "fear of activity"/deconditioning model.

 

Good point! But they could spin it to argue that it’s possible to keep up your activity level despite of the symptoms, and that the symptoms will eventually be unlearned if they keep at it.

 

This study may be useful in showing that persistent symptoms after covid may be significantly dependent on pre-existing health impairment.

 

But you'd still have to account for the people who don't fit this pattern, who were in good health and busily active before Covid but then develop disabling Long Covid.

 

Can we say that based on 50 patients?

 

Oh yes, but the point is that trying to find subtle differences between LC patients and people who have had COivd without LC using -omics or other measures may be confounded by just picking up the fact that the LC people are on average more likely to show differences because they had been ill for some other reason before.

 

If they found differences that were already present before Covid infection then yes. 50 is a reasonable number to get a statistic out of.

 

Oh, no, they'll keep saying it without skipping a beat, with a slight modifier that changes nothing. Even if someone showed them multiple studies showing the same they'll have all sorts of pocket sand they can throw at it. They don't care, and neither does any random bystander being informed of this, MD or not. This is entirely a feelings and vibes ideology.

There are many pieces of evidence of this in the past, including explicit acknowledgement by some of the main quacks that the levels of activity are relatively normal. They. Don't. Care. They built bullshit models that their own colleagues, who have all the skills necessary to see through the bullshit, think are great, because the institutions bought them in full and don't have the integrity to go back on it, however much harm they cause.

They're at war with reality, having built a castle out of pure fantasy. And they loooove that castle. It's a great castle, with moats and gilded rococo everywhere. They'll be dragged kicking and screaming out of it, and not a second sooner.

 

Possible in my case, I suppose, but I certainly wasn't aware of anything. (The only thing I can think of, looking back, is that I had seen the GP a few months pre-Covid infection about some pain I was having around the area of one kidney, but tests showed nothing to worry about, they said it was probably muscle strain that would improve with time and exercise, so I carried on my usual exercise and it did go away.)

[edited to add] Coming back to this because I realised something overnight: the Covid infection which turned into Long Covid was almost certainly my second one. As far as I know, after my first Covid infection in early 2020, which felt like just a few days' heavy cold, I bounced back to normal health very quickly. But perhaps something went on lurking and then flared up disastrously with the second one.

 

I think it's useful to point out that this is not a Long-Covid study. Which is probably also why the authors don't use these words. It's simply a study of tracking people before and after a Covid infection.

Maybe it would be interesting to know how many of these people would actually be considered to have Long-Covid (that will vary depending on the definition)?

I think the way Rob Wüsts study was designed would mitigate some of those concerns. Working hours is certainly not a perfect measure, especially not if there is a pandemic which might cause some disruption in itself, but if you have a cohort that is fully employed before infection and can't work afterwards at all anymore it makes it more likely that you aren't just looking at people who had been ill the whole time.

You will find plenty of people who were perfectly healthy before their Covid infection. Of course that doesn't mean that the studies are always tracking those people. And especially those looking at hospitalised patients will likely have been tracking all sorts of different confounding factors.

 

Why? This data is pretty consistent with REACT estimates where 2.83% of people reported persistent covid symptoms during the omicron period, 2/3 reported mild fatigue, and 1/3 had shortness of breath. In the ONS survey, 4% of adults reported post-covid symptoms after the first omicron infection, 70% with fatigue, and 50% with shortness of breath.

The conclusion is not based on 50 or even 350 participants, matched cohorts were exclusively used only for the wearable data portion of the study. Data on the preexisting conditions, the relative frequency of symptoms, and the overall wellbeing were all based on the information from thousands of participants. There are no contradictions between the different sets of data.

There are studies on working hours after covid infection. For example, this study (https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0306122) showed that the negative impact of Long COVID on working hours is primarily from working zero hours (sickness leave) rather than working fewer hours. There are also worker inactivity estimates after covid infections based on the ONS survey dataset. https://academic.oup.com/eurpub/article/34/3/489/7616634

 

I did the maths on this one. 137/7839 = 1.75 %. So you’re right. Thank you for finding number to compare with.