The aim of the study was to examine the PA and sleep quality of patients with post-COVID before and after an inpatient post-COVID rehabilitation program. This study investigated the interested parameters by accelerometery as an objective and more valid measurement before and after rehabilitation in a longitudinal design. In summary, the patients didn’t experience a better sleep quality or being more physically active after rehabilitation. According to the groupwise comparison, some differences could be detected.
4.1 Physical activity
The current data shows no significant difference in the PA behaviour of post-COVID patients before and after rehabilitation. At both measurement points, the patients are almost 14 hours inactive during the day. The high inactivity time of the examined post-COVID patients can be explained by the high rates of fatigue and exercise intolerance at T1 and T2. At T1, 97% of patients are reporting symptoms of exercise intolerance and 91% symptoms of fatigue. This prevalence is not significantly decreasing after rehabilitation discharge. Further, sleep periods during the daytime are also classified as inactivity time and they are common in patients experiencing symptoms of fatigue. The sensor wearing position can also lead to higher inactivity time. If the patient is doing standing activities, where mainly the upper body and arms are moving, it is not possible for the sensor to detect this movement. Compared with other studies including post-COVID patients, the assessed inactivity time is higher. Plekhanova, et al. (29) and Benitez, et al. (28) reported inactivity times up to 12.6 hours per day. In contrast, in accelerometer studies with healthy study populations, the inactivity time is substantially lower than that of post-COVID patients with around 8.2 hours per day (81, 82). Even patients with COPD are spending less time inactive during the day (~8 h) than the included post-COVID patients (83). Previous research demonstrated that patients with chronic diseases have also a substantial lower amount of MVPA per day compared to healthy controls (84). Contrary to the high inactivity time within the current study sample most of the patients could achieve the WHO recommendations for MVPA for adults with chronic diseases, which includes any bodily movement produced by skeletal muscles requiring an energy expenditure ≥ 3 METs (21). Only six patients didn’t spend 150 minutes of MVPA during their week. This is in line with Benitez, et al. (28) but contrary to the results of Plekhanova, et al. (29) and van Bakel, et al. (30) which revealed, that the level of MVPA of post-COVID patients is not reaching the WHO recommendations. Considering the diverse and complex symptoms of post-COVID patients and the heterogeneity of the study results, nevertheless, it should be critically questioned whether the general WHO’s activity recommendations are suitable for such a specific sample, rather personalised concepts are needed. Furthermore, the WHO PA recommendations are mostly based on self-report data and a comparison with accelerometery data is difficult. The differences in the assessed PA may be explained by different accelerometer devices and methodological aspects. The choice of sensor wearing position, acceleration metrics and scoring algorithm influencing the PA estimates. Another factor could be the characteristic of the included study population. The current study has an interval between the acute SARS-CoV-2 infection and T1 of ~402 days. Compared to the other studies, which examined the patients after three to six months after acute infection, the patients are suffering from the post-COVID symptomatic almost three-fold longer.
The current data suggests that the rehabilitation program should more address behavioural changes according to PA and activity-related health competencies. There is evidence, that a higher amount of PA could reduce the post-COVID symptomatic and support the recovery process (29). More important, high inactivity times are associated with pro-inflammatory processes and have negative effects on functional capacity (33, 85). High levels of MVPA may reduce the negative effects of sustained inactivity but the current evidence shows that it does not eliminate it completely (86, 87). It is also necessary to change behaviour by interrupting long inactivity times with short periods of any PA intensity (e.g., walking during a phone call, standing) (88, 89). According to current health psychology theories and models (e.g., health action process approach (90), transtheoretical model of behavioural change (91), The Physical Activity-related Health Competence (92)), the improvement of physical resilience and functional capacity during rehabilitation is necessary but not sufficient to induce behavioural change (increase in habitual PA and decrease in inactivity time) within the scope of aftercare. The Physical Activity-related Health Competence is an integrative model describing personal determinants of PA (92). Apart from physical functionality, self-efficacy, knowledge, self-regulatory skills and exercise related attitudes are necessary to obtain health-related PA after rehabilitation in a long-term. To assess and strengthen the individual determinants of the Physical Activity-related Health Competence within clinical settings (e.g., during rehabilitation) may improve the rehabilitation outcomes in a long-term (93).
The groupwise comparisons revealed significant differences in inactivity time between male and female patients. At the beginning of rehabilitation, male patients had a significant higher amount of inactivity time during their day than female patients. This result is in line with Plekhanova, et al. (29) reporting a difference in inactivity time of 0.7 h between males and females. The current data shows a higher increase in physical activity after rehabilitation in male patients compared to female patients. As this observation is not statistically significant further research with a longer follow-up-interval is needed to confirm this result. Regarding the age of the participants, the current results are consistent with the results of van Bakel, et al. (30). Younger post-COVID patients (<55 years) tend to spend less time inactive and achieved a significantly higher amount of vigorous PA before rehabilitation compared to older patients (≥55 years). Another study with post-COVID patients didn’t reveal a significant difference between age groups even if there are tendencies that younger patients have a higher amount of MVPA and lower inactivity time (30). One reason for this difference could be the group allocation. In the present study, younger patients were defined as under 55 years whereas in van Bakel, et al. (30) younger patients were under 67 years. According to the literature, PA decreases with age as the number of physical and structural barriers increases (94). Nevertheless, in particular older patients with post-COVID should be encouraged to practice PA appropriate to their abilities in order to obtain health benefits. The analyses of group differences regarding pre-existing diseases revealed just some significant differences. In general, the differences suggest that post-COVID patients without pre-existing cardiovascular, respiratory, or metabolic disease are slightly more physically active than post-COVID patients with one of the mentioned comorbidities. This is in line with previous research in COPD patients. Sievi, et al. (95) and Mantoani, et al. (96) could show, that COPD-patients with comorbidities have a lower level of PA than patients without any comorbidity. The current data suggests that the severity of acute infection does not influence the level of PA in the post-acute state, since there weren’t any significant differences between these two groups before and after inpatient rehabilitation.
4.2 Sleep quality
Like PA no changes in sleep quality of post-COVID patients could be detected before and after rehabilitation. Overall, the patients show poor sleep quality. Several reported parameters are below the recommendations for adults of the National Sleep Foundation (79). A good sleep quality is characterized by a sleep duration of six to 10 hours, a sleep latency of less than 30 minutes, a sleep efficiency over 85%, and a WASO-time less than 51 minutes. At T1 as well as at T2 the patients have a median sleep duration of 5.5 hours, the sleep efficiency is Mdn=67 % and the median of the WASO time is ~2 hours. Only the sleep latency is within the recommended range with a median of 24 minutes. This is in line with the subjective assessed sleep quality. At T1 79 % of patients indicate to suffer from sleep disturbances. This prevalence is not significantly decreasing after rehabilitation discharge. The unchanged sleep quality after rehabilitation may be explained by the measurement procedure. The patients wore the sensor and filled out the questionnaires two weeks after rehabilitation discharge. Possibly, the patients first had to get used to their home environment and everyday life again after being discharged from rehabilitation. Some persons were still on sick leave a few days after rehabilitation. Thus, the time of measurement could have fallen on the first working days, which may also result in a change in the sleep-wake rhythm. Another explanation for the ongoing poor sleep quality of post-COVID patients is the persistently high burden of disease after rehabilitation. The prevalences of e.g., exercise intolerance, fatigue, joint and muscle pain, and mental disorders are high with 60-87% of patients still suffering from these symptoms. Previous research reported a poor sleep quality of post-COVID patients, too (50, 97, 98). Jarosch, et al. (50) examined sleep quality of post-COVID patients with polysomnography and compared the results with healthy controls. The sleep quality of post-COVID patients was significantly impaired. Furthermore, the post-COVID patients in the study by Mekhael, et al. (99) experienced a significantly shorter sleep duration compared to healthy controls. In general, regarding the impairment of the immune system functionality due to sleep deprivation and poor sleep quality (52, 54, 100), the post-COVID rehabilitation program should focus more on a better treatment of sleep disturbances. Potential approaches to address sleep disturbances during rehabilitation is sleep-related psychoeducation, cognitive behaviour therapy and inducing sleep structuring techniques (77). Furthermore, for post-COVID patients with ME/CFS, the PACING technique may also lead to an improvement in sleep quality (5, 15, 101).
The data shows sex related group differences in sleep parameters. Even if both groups were showing a poor sleep quality, female patients did spend significantly more time in bed during the night and slept longer at both measurement points. This difference could be explained by sex differences regarding sleep. In general, females are more likely to need more sleep during the night caused by a different hormonal state (78, 102). However, the difference could be explained by higher prevalence of mental impairments and fatigue in women in the current sample. The analysis of Müller, et al. (37) with the same cohort revealed sex-based differences in the fatigue symptomatic, which may also lead to longer sleep duration and time in bed. Nevertheless, neither male nor female patients show sleep parameters indicating a good sleep quality. Regarding age, younger patients (<55 years) have a significant longer sleep duration with around 5.91 hours per night at T1 and 6.03 hours per night at T2. Patients ≥55 years have a sleep duration of 5 hours per night at T1 and 5.23 hours per night at T2. Further, younger patients have a significantly better sleep efficiency (~70 %) than older patients (~64 %) before and after rehabilitation. According to the literature, the sleep duration is decreasing with increasing age and the deep sleep stage gets shorter (103). Older patients wake up more often during the night and the current data seems to confirm this observation as the difference in WASO time is significantly shorter in younger patients than in older. It is important to recognise that the general findings from the literature may be based on broad populations studies. However, our study focuses on a specific population and a limited time period. For more profound insights, further longitudinal studies may be necessary to understand the specific mechanisms and factors influencing sleep quality in post-COVID patients.
4.3 Strength and Limitations
This study provides a device-based assessment of PA and sleep quality of post-COVID patients before and after rehabilitation. The current data contributes to more knowledge about the habitual PA and sleep behaviour of post-COVID patients. The paired sample size is also above the computed sample size of 115 patients (incl. 25% Drop-out Rate) as mentioned in the study protocol (64). Nevertheless, under consideration of an observational study design without control group, the results should be interpreted with caution. Future research should include a control group to strengthen the validity of the current results and to provide a comparison with a healthy study population, one with other chronic diseases or patients without rehabilitation intervention. Furthermore, the sample sizes of the groupwise comparisons are not well balanced. When comparing the current results with other studies it should be considered that within our study sample most of the participants are working in the healthcare-sector and are suffering from post-COVID for a long time (~400 days). Hayden, et al. (104) stated, that the effectiveness of rehabilitation is higher soon after acute COVID-19 and healthcare-workers are susceptible for sleep disturbances due to rotating working hours and night shifts (53). Therefore, in future studies, pre-existing sleep disturbances should be assessed. Moreover, it is not possible to analyse causal associations between post-COVID and PA and sleep quality based on the available data. The available data provides a very general overview of the PA behaviour of post-COVID patients and does not depict specific forms, patterns, and intensities of PA. This emphasises the need for more precise data collection to enable a more differentiated analysis of PA behaviour of post-COVID patients. Another limitation is the wearing position of the sensor. A waist-worn sensor is not able to distinguish between different body positions (e.g., sitting vs. standing) and activities with the arms or hands can’t be recorded. In addition, the time interval between the two measurement points is quite short, as effects on PA and sleep quality may only become apparent in the longer term, and thus, another measurement point e.g., 12 months after rehabilitation is needed. Additionally, this allows longitudinal analyses over three measurement points and to estimate e.g., within-person impacts of the rehabilitation phase and post-rehabilitation phase on PA and sleep quality. Furthermore, the rehabilitation program was not explicitly designed to improve patients’ sleep quality. However, since it follows a holistic approach and sleep disorders are one of the most common post-COVID symptoms, a separate examination of patients’ sleep is legitimate.