Physical activity, acute severity and long-term consequences of COVID-19: an 18-month follow-up survey based on a Swedish national cohort

Introduction

The COVID-19 pandemic has highlighted the importance of physical activity (PA) in maintaining physical and mental well-being while enduring lockdowns and periods of social isolation.1 2 COVID-19 has been found to cause different acute symptoms3 of varying severity; the vast majority of patients have a mild infection, but some have a more severe infection requiring acute hospitalisation.4 For some, the symptoms from the acute COVID-19 infection remain, with a multitude of long-term consequences.5 6 Limitations in usual activities have been reported as the third most prevalent symptom (14.9%) up to 1 year after infection in populations with both hospitalised and non-hospitalised individuals.7 In a small Swedish study of individuals with a post-COVID-19 condition, 62% reported having difficulties being physically active 2 years postdischarge from hospital, though there was a significant improvement from the 4-month follow-up.8

Being regularly physically active is a well-known strategy recommended to promote health and prevent disease.9 Regular PA can also strengthen the immune system, which has been suggested to be beneficial when exposed to viral communicable diseases,10 including COVID-19.11 When infected with COVID-19, physically inactive individuals are at higher risk of more severe illness, hospitalisation and death than individuals who are regularly physically active.11 12

Several studies13–18 have reported changes in PA during the pandemic. Although multiple studies in Swedish populations13–15 have reported at least a slight increase in PA, other studies from the UK,16 Italy17 and USA18 have reported a decrease. This may be partly explained by different approaches to restrictions during the pandemic.14 The changes in PA also seem to depend on age,13 16 sex,13 17 occupation13 and the different waves of the pandemic. A Swedish study13 reported that more individuals changed their lifestyle habits during the first wave than during the second wave, with 26.4% reporting a negative change in daily activity during the first wave, compared with 20.3% during the second. Furthermore, studies have shown that decreased PA is associated with reduced life satisfaction14 and a negative impact on mental health,13 17 18 and that social inequality in PA increased during the pandemic.15

In the working-age population, falling ill with COVID-19 often means having to take sick leave from work.19 For some, this results in long periods of sick leave and, sometimes, periods of recurrent sick leave.20 21 In Sweden, compensation for sick leave is comprehensive, tax funded and administered by the Swedish Social Insurance Agency (SSIA). Our previous study found that nearly 12 000 individuals in Sweden were on sick leave due to COVID-19 within the framework of sickness benefits during the first pandemic wave.20 These individuals comprise the cohort of the current study.

Based on the national cohort of individuals registered for sick leave due to COVID-19 during the first pandemic wave, we investigated how PA and changes in PA levels in regard to acute disease severity relate to perceived difficulties in performing daily life activities 18 months after a COVID-19 infection.

Results

Physical activity

PA level was lower in hospitalised responders than in non-hospitalised responder both prior to COVID-19 (p=0.035) and at the 18-month follow-up (p=0.008; table 2).

View this table:

Table 2

Current and previous physical activity levels in the responders

In the whole group of responders, the PA level decreased from prior COVID-19 infection to the 18-month follow-up (p<0.001; figure 1). Summing up the changes in PA, the level of PA decreased in 1726 (33.6%) responders, was sustained in 3105 (60.5%) responders and increased in 300 (5.8%) responders. Of the individuals with the highest level of PA (SGPALS 4) prior to COVID-19, approximately one-third sustained their level of PA until the 18-month follow-up. In the subgroup analysis based on acute disease severity, the PA level decreased in the non-hospitalised group (p<0.001), with 1244 responders (33.5%) decreasing their level of PA, 2282 (61.5%) sustaining their level of PA, and 184 (5.1%) increasing their level of PA. A significant decrease in PA was also found in the initially hospitalised responders (p<0.001), with 482 (33.9%) decreasing their level of PA, 823 (58.0%) sustaining their level of PA and 115 (8.1%) increasing their level of PA. There was no difference in change in PA over time between hospitalised and non-hospitalised responders (p=0.193).

Figure 1

Changes in physical activity (PA) levels according to the Saltin-Grimby PA Level Scale (SGPALS) from prior to COVID-19 infection (left side) to the 18-month follow-up (right side) n=5131 individuals. SGPALS: 1, physically inactive; 2, light PA for at least 4 hours/week; 3, regular moderate PA and training for at least 2–3 hours/week; and 4, regular vigorous physical training for competitive sports several times per week.

Long-term consequences of COVID-19 in daily life and relationship to PA

At the 18-month follow-up, 2434 responders (46%) reported still having difficulties performing daily life activities due to COVID-19. Logistic regression was performed to explain these long-term difficulties. The model showed that a decrease in PA increased the odds (OR 5.58, 95% CI 4.90 to 6.34) of having difficulties performing daily life activities. Furthermore, having been initially hospitalised due to COVID-19 and taking sick leave during the year prior to COVID-19 also increased the odds, whereas age and sex did not contribute significantly to long-term difficulties in daily life activities (figure 2). The model could significantly (p<0.001) distinguish between responders who did or did not report perceived difficulties in performing daily life activities at the 18-month follow-up.

Figure 2

Variables in the model explaining the odds of perceiving difficulties in performing daily life activities. The model explained between 15% (Cox & Snell R²) and 20% (Nagelkerke R²) of the variance in the outcome.

Discussion

The results of this study show that individuals with less severe acute disease had a higher level of PA, both prior to COVID-19 and at the 18-month follow-up, than individuals with more severe acute disease. In both the total study population and both groups based on acute disease severity, the PA level was decreased at the 18-month follow-up compared with the level prior to COVID-19. Furthermore, a decrease in PA over that time was associated with perceived difficulties performing daily life activities 18 months after COVID-19.

The PA level of responders was rather high before COVID-19. The study population also seems to be quite physically active compared with other cohorts in which PA has been assessed by the same measure.24–26 Given that regular PA has been shown to be beneficial in protecting individuals against severe illness and hospitalisation for COVID-19,11 12 we could anticipate that this population would not be severely affected by COVID-19 infection. Nevertheless, one-fourth of the current study population was initially hospitalised. Being more physically active during the pandemic may have entailed being in places or contexts with a higher risk of transmission of COVID-19, such as gyms or team sports, whereas less active individuals may have been more isolated in their own homes. Furthermore, exercising while already infected with COVID-19 may have worsened the course of the disease.27 In the polio epidemics, being very physically active28 or pregnant,29 both of which affect the immune system, increased the risk of paralysis among poliomyelitis cases. However, in the current study population’s subgroups based on acute disease severity, hospitalised responders had significantly lower PA levels than non-hospitalised responders, supporting previous findings of the benefits of PA in regard to COVID-19 severity.11 12

Summing up the changes in PA, there was a decrease in the level of PA in both the total study population and the groups based on acute disease severity. In the general Swedish population, PA increased during the pandemic,13–15 in contrast to the current findings in a cohort of individuals who had COVID-19. The decrease in PA seen in the current study can be speculated to relate to the disease itself and deconditioning.30 However, a Swedish study14 reported that, in a general population, 36% reported increased PA and 30% decreased PA during the pandemic; 71% and 89%, respectively, attributed this change to COVID-19-related restrictions.14 The relatively mild approach to restrictions in Sweden may have had diverse consequences for PA in different groups of individuals. Recommendations, such as working from home, may have given them more free time and, therefore, increased PA for some, whereas cancelled sporting events and competitions may have had the opposite effect, especially among individuals with a high level of PA. In the current study, only one-third of individuals with the highest level of PA prior to COVID-19 sustained their PA level until the 18-month follow-up.

Furthermore, the strictest recommendations, which involved isolation and were applied to people >70 years of age and those infected, as a means to protect the most vulnerable and constrain transmission, most likely had an influence on the levels of PA. A study from the US showed an association between self-isolation and negative mental health that was not seen with social distancing.18 Previous studies13 17 18 also reported an association between negative mental health effects and a decrease in PA during the pandemic. Taken together, these findings suggest that pandemic-related restrictions and mental health could also play a role in the decrease in PA in the current cohort.

In this study, in-hospital care due to COVID-19 was used as a proxy for acute disease severity. As no one in Sweden was denied in-hospital care due to lack of capacity and healthcare in Sweden is tax funded, this proxy for acute disease severity seems reasonable and relies on the assessment of educated healthcare professionals. Furthermore, need for hospital care is included in the WHO Clinical Progression Scale,31 which supports the interpretation of in-hospital care as a proxy for COVID-19 disease severity.

Using the SGPALS,23 the current study showed a decrease in PA among individuals who were on sick leave due to COVID-19 during the first wave of the pandemic in Sweden. Notably, only one-third of individuals with the highest level of PA prior to COVID-19 sustained their PA level until the 18-month follow-up. Because making up for deconditioning takes time,32 this decline in PA may have long-term implications, and returning to pre-COVID-19 activities and levels may be difficult even for the most highly trained individuals, such as elite athletes.33

PA is important for maintaining health and can protect against both non-communicable chronic diseases34 and communicable diseases,35 such as COVID-19. Therefore, the results of this study, with a decreased PA level regardless of acute disease severity, are worrisome, especially considering the risk of future pandemics.

A strength of the present study was that a validated instrument, the SGPALS, was used for questions regarding PA. Furthermore, the different steps of the scale were clearly defined in the questionnaire, making it easier for individuals to answer according to their true activity levels. This lends weight to the results and facilitates comparisons with other studies and populations. However, the retrospective assessment of PA level, asking responders to report their PA level 18 months previous, involves risk or recall bias, which is a limitation.

Another limitation of the study was the response rate of nearly 50%, and responders were found to be older and having had more severe acute disease than non-responders. Thus, differences between responders and non-responders need to be considered when interpreting the findings.

In conclusion, PA levels were reduced 18 months after COVID-19 infection. A decrease in PA over that time was associated with perceived difficulties performing daily life activities 18 months after COVID-19. As PA is important in maintaining health and deconditioning takes time to reverse, this decline may have long-term implications for PA and health.