We prospectively evaluated changes in SB and MVPA before, during, and 3 years after the COVID-19 pandemic in CVD patients. We found no significant changes in MVPA across timepoints (2018: 1.9 [0.9, 2.8] h/day versus 2020: 1.9 [1.0, 2.9] h/day versus 2023: 1.7 [0.6, 2.8] h/day), suggesting that CVD patients maintained their MVPA patterns. However, SB substantially increased during the pandemic, and despite a small, but significant decrease 3 years thereafter, SB levels did not return to pre-pandemic levels in 52.4% of CVD patients (2018: 7.8 [5.8, 9.8] h/day versus 2020: 8.7 [6.6, 10.8] h/day versus 2023: 8.5 [6.4, 10.5] h/day). The inability to return to pre-pandemic SB levels was associated with lower baseline SB levels, larger increase during lockdown, presence of residual complaints after COVID-19 infection, and the presence of cardiac arrhythmias at baseline. These findings provide important information on the long-term effects of the COVID-19 pandemic on habitual physical (in)activity levels of CVD patients and may be used to optimise secondary prevention strategies in CVD patients.

To our knowledge, this is the first study to prospectively investigate the effects of the COVID-19 pandemic on SB years after initial measures were taken. Our approach allowed us to compare intra-individual levels of SB before, during, and after the COVID-19 lockdown. Previous studies reported an acute increase in SB during the lockdown [10, 12, 24], as individuals were instructed to work from home, PA opportunities were restricted, and social interactions reduced [12]. Nevertheless, research on whether SB and PA levels return to pre-pandemic levels years after the pandemic, and potential explanations for this behaviour change, is lacking. Our post-pandemic findings are alarming. Although SB levels were slightly decreased post- versus peri-pandemic, time spent sedentary remained 10% higher compared to pre-pandemic levels. Previous studies reported an annual, age-related increase of SB ranges of 0 to 1% [25, 26], suggesting that the post-pandemic SB levels are likely attributable to the COVID-19 lockdown and not to aging alone. We observed that 52.4% of our population showed an increase of > 30 min in SB at post- versus pre-pandemic assessment. At group level, MVPA levels did not change over time, despite increases in odd jobs during lockdown restrictions [10, 11]. However, individual MVPA patterns may have varied. The increases in SB may induce deleterious health outcomes, such as increased cardiometabolic risk diseases and all-cause mortality [27]. Van Bakel et al. showed that patients’ lifestyles were impeded by lack of social contact, fear of a SARS-CoV‑2 infection and limited PA possibilities [11]. These determinants may have contributed to the development of changes in sedentary and physical activity habits.

Several factors were associated with the inability to return to pre-pandemic SB levels. The binary logistic regression model had an explained variance of 87%, indicating good fit. Those with higher baseline SB levels were more likely to return to or remain at their pre-pandemic SB levels; since these levels were already high (e.g. > 10 h/day 23% of the study population at baseline), further increases would be unlikely. These patients would remain at the same SB levels, and therefore were defined as ‘returned to pre-pandemic SB levels’. Moreover, larger increases in SB during the lockdown were associated with a higher risk for the inability to return to pre-pandemic SB levels. Large increases in SB during the lockdown may be due to anxiety for infection and recommendations to work from home, limiting active transportation and PA during working hours, however, this was not validated in the questionnaire. Independent of these reasons, upon lifting the lockdown measures, CVD patients may have adapted their lifestyle to these SB patterns, and substantially changing them seems difficult [11]. To our knowledge, the finding that the presence of a cardiac arrhythmia diagnosis at baseline was an associated factor with not returning to pre-pandemic SB levels was not earlier reported in literature. Further research on this topic is warranted. Finally, we also found that the presence of residual complaints after COVID-19 was associated with a lack of return to pre-pandemic SB levels. Patients with a (self-reported) diagnosis of long COVID experience limited exercise capacity and, in many cases post-exertional malaise. These symptoms promote a sedentary lifestyle due to discomfort during PA. At the very least, these observations highlight the difficulty to target a sedentary lifestyle following substantial changes in SB as observed during the lockdown.

Physical inactivity increases the risk for cardiometabolic diseases, obesity and all-cause mortality, especially in this CVD population [14, 28]. CVD patients are characterised by a more sedentary lifestyle, while even PA at light or moderate intensity can provide significant health benefits and prevents further decline in cardiovascular health. This highlights that our observation of the excessive post-pandemic SB levels is clinically relevant, and that a focus should be on reengaging this population in PA and decrease SB. The sensitivity analyses showed no differences in patterns of changes over time when separated for week and weekend days. However, sedentary times were significantly higher at each individual timepoint, suggesting that CVD patients are overall more physically active during weekend days. More insight on physical activity patterns and individual habits could increase knowledge on the motivation of patients on changing their SB or PA patterns. Therefore, more research is needed to optimize lifestyle programmes with evidence-based guidelines. Targeted lifestyle programmes could offer a solution to improve daily life activity patterns [29, 30].

Strengths of this study are the large study population and longitudinal nature of the study design. 66% of participants included in the previous study responded to the invitation. Participants did not differ from those lost to follow-up in MVPA and SB levels and other patient characteristics during pre-pandemic assessment. Limitations to this study include the use of questionnaires, which rely on self-reported MVPA and SB levels, making questionnaires prone recall bias and PA overestimation and SB underestimation. However, an important advantage is that this study involves within-subject analyses, whilst the validated SQUASH and SBQ show good reproducibility [16, 31]. Therefore, valid representation of change over time is presented. Moreover, relatively many SB datapoints are missing due to incomplete questionnaires. A potential explanation for this is that the SBQ was obtained after the SQUASH. The nature and context of PA levels are dependent on individual habits and patterns, such as longitudinal (cardiovascular) health status, hospitalisation(s), and psychological and social factors concerning the COVID-19 pandemic. Whilst we were unable to correct for these factors, we were able to add retirement status as a factor to the binary logistic regression model. This additional analysis showed that retirement status was not associated with not returning to pre-pandemic SB levels. Finally, the observational design precludes causal inference. We therefore cannot draw any causal conclusions from the results.