The research team searched six large medical databases and three preprint servers (platforms that show studies that have not yet been published in a peer-reviewed scientific journal, see infobox). The analysis included studies that prospectively investigated the occurrence of post-COVID-19 conditions.
In prospective studies, people are observed for a certain period of time into the future. In the case of Long COVID, for example, this means that the health status of people after a COVID-19 infection is systematically monitored for a certain period of time.
This creates a standardized data basis that can be used to determine how many of the infected people have recovered and how many are still suffering from symptoms at the end of the study period.
The opposite of a prospective study is a retrospective study, in which patient files are searched retrospectively for data that provide information on a specific question.
The advantage of a prospective study is that it is clear from the start which data is to be recorded and, accordingly, all relevant data is available in the same format at the end – this is not always the case with retrospective studies.
In the meta-analysis on the global incidence of Long COVID, the focus was on studies that covered a period of at least twelve weeks after the acute COVID-19 infection – in accordance with the WHO definition.
A total of 194 prospective studies were included with over 480,000 participants worldwide. The aim was to record the global prevalence of symptoms and functional limitations that persist at least three months after a confirmed COVID-19 infection.
More than half of the infected people still had at least one Long COVID symptom after one year.
The analysis showed that more than half of those affected by COVID-19 still reported at least one symptom months later. In the period from 12-26 weeks after infection, 56.5 % of those affected in studies that observed this period reported at least one symptom, after 27-39 weeks it was 50.9 % and after 40-52 weeks 32.6 %. Of note, not all studies looked at the infected persons over the entire time period.
These fluctuations show: Long COVID is not a static phenomenon. Depending on the time, symptom and study quality, the results sometimes differ significantly.
The analysis examined the frequency of over 20 symptoms. These were particularly frequent:
- Fatigue (26.9-30.7 %)
- Concentration problems (up to 29.9 % after 1 year)
- Sleep disorders (up to 29.4 %)
- Dyspnea (shortness of breath) (14.8-20.5 %)
- Depression (up to 40.4 %)
The quality of the results was often rated as low – mainly because the studies were structured very differently, did not use standardized rules for diagnosis and often examined people from hospitals.
The meta-analysis showed evidence that women, people with a severe course and those requiring intensive care during the acute infection were more likely to report long-term consequences. However, many of these subgroup comparisons were not statistically significant – the data situation therefore remains inconsistent.
Almost a quarter of those affected in the studies analyzed stated that they were still unable to work 1 year after infection.
The included studies often indicated functional limitations in those affected. The frequency of functional limitations was calculated from the studies in which the infected persons still had at least one symptom more than 1 year after their initial COVID-19 infection.
- 8-50.2% stated that they would not be able to return to their job after one year
- 3-32.4% have not returned to their profession after one year
- 2-17.0% reported difficulties in carrying out everyday activities
- 0-16.3% reported mobility restrictions
- 8-2.0% reported difficulties with self-care
How credible are these results?
This meta-analysis is the largest of its kind to date and it shows that Long COVID is not a marginal phenomenon. Nevertheless, it is difficult to clearly determine the frequency of Long COVID. The studies are too heterogeneous, the measurement methods too varied and too many aspects are still unclear.
The results of the meta-study provide comparatively high prevalence figures; many previous studies, especially population-based studies (including the general population), estimate a lower prevalence for Long COVID (5-20%). This difference can be explained by several methodological factors:
- The analysis includes almost exclusively studies with PCR-confirmed COVID-19 cases. These cases were mostly symptomatic or more severely affected and therefore more susceptible to long-term consequences.
- Many of the included studies recruited patients from hospital or rehabilitation contexts, which leads to an overrepresentation of severe courses. This also increases the reported long-term symptoms.
- There were large differences in study design, follow-up duration and symptom definitions. This heterogeneity leads to wide confidence intervals and makes comparisons difficult.
- The data often comes from groups that were already involved in the medical system - people with a mild or asymptomatic course are therefore underrepresented.
- Many symptoms were recorded by self-report, which increases the risk of over- or underestimation.
The authors therefore emphasize that their results are not representative population figures, but paint a specific clinical picture.
Conclusion
Long COVID is real, stressful and affects millions of people worldwide - even more than a year after infection. The meta-analysis provides valuable figures, but also an important insight: we need more accurate data, better studies and targeted support for those affected.
