Physical Activity and Mental Health of Patients with Pulmonary Hypertension during the COVID-19 Pandemic

The aim of the study was to analyze the effect of personal restrictions on physical activity, mental health, stress experience, resilience, and sleep quality in patients with pulmonary hypertension (PH) during the “lockdown” period of the COVID-19 pandemic. In total, 112 PH patients and 52 age-matched healthy control subjects completed a questionnaire on the topics of physical activity, mental health, resilience, and sleep quality. PH patients had significantly lower physical activity, mental health, and sleep quality compared to age-matched healthy controls. Physical activity positively correlated with mental health and sleep quality in the PH group. Mental wellbeing and life satisfaction could be predicted by total physical activity, sleep, stress level, and resilience. PH patients appeared as an especially vulnerable group, demanding interventions to promote an active lifestyle and protect mental health in these patients. This could be helpful in counseling on how to carry out physical activity while maintaining infection control.


Introduction
Pulmonary hypertension (PH) is a multifactorial chronic pulmonary disease which is defined by an elevated mean pulmonary arterial pressure, which untreated eventually can lead to right heart failure and death [1]. Depending on the clinical classification and risk stratification PH patients have quite a different prognoses, treatment options, and impairment in daily life [2]. Patients with PH experience symptoms such as shortness of breath, exertion, fatigue, chest pain that restrict physical activity, which in turn impairs quality of life and favors mental disorders such as depression [3][4][5][6].

Subjects Characteristics
In this cross-sectional study, subjects diagnosed with pulmonary hypertension were compared to a healthy control group concerning to their mental health represented by WHO-5, PHQ-4, L-1, physical activity, resilience, stress experience, and sleep quality during the COVID-19 associated lockdown. The control group was matched by age.
Subjects of the PH group had to have their residence inside of Germany and a medical diagnosis of PH. To be eligible for the control group, subjects must not have had a confirmed infection or symptoms of COVID-19, any comorbidities, or a residence outside of Germany. Only subjects declaring consent were included in this study.
PH patients of the University Hospital of Gießen and Marburg (UKGM) were directly contacted. The questionnaire was promoted during online consultation hours of the UKGM and in PH support groups (https://pulmonale-hypertonie-selbsthilfe.de/). If necessary, the questionnaire was sent to patients per mail due to lacking access to the internet and was entered into the online survey by hand.
In total, 251 PH patients followed the invitation to participate in the survey. There were 139 cases that were excluded, 111 of those for not completing the questionnaire, 25 for reporting to have no medical diagnosis of PH, and three because their residence was outside of Germany. This left us with questionnaires of 112 subjects being valid for the PH group. The mean age in the PH group was 54.4 ± 14.0 years with 77.7% being female. The average number of comorbidities was 2.6 ± 2.0. 52 healthy subjects were recruited into the control group in dependence of their eligibility criteria and age matching. The mean age of the control group was 52.3 ± 8.9 years with 67.3% being female. The sociodemographic characteristics of both groups are presented in Table 1.

Questionnaire and Outcomes
The survey contained several questions and validated questionnaires to evaluate study variables. In the beginning, sociodemographic data such as age, gender, relationship status, living situation, the number of residents per household, educational level, occupational status, worries about health, and satisfaction with sports behavior were gathered. The living situation covered the living area and the availability of a garden or a balcony. The number of residents per household was divided into the total number of residents and those younger than 18 years. The category comorbidity included sub-categories of comorbidities (apart from PH for the PH group). Every sub-category could be answered with "No", "Yes, medical diagnosis", or "Yes, self-assessment", although a self-assessed diagnosis was counted as a "No" in the analysis. The number of "Yes, medical diagnosis" was counted and summed up to create the variable "Comorbidities". Possible SARS-CoV-2 infection or symptoms were also evaluated.
Physical activity data were collected through items three to six of the BSA-F which was shown to be a valid tool for the measurement of physical activity and sports behavior [18]. The outcome scores were total physical activity, the activity of daily living, sports activity, and climbing stairs. The variable total physical activity is a sum of the activity of daily living and sports activity. The amount of different physical activity, exercise and sports modalities are calculated by multiplying the duration and frequency of the corresponding items and then summating them. The examined period for all scores was four weeks in the PH group. The scores total physical activity, activity of daily living and sports activity were converted into the unit minutes/week as designated and were thereby made comparable. The item climbing stairs was converted into unit floors/week. Mental health was assessed by using the WHO-5 Well-Being Index (WHO-5), the Patient Health Questionnaire-4 (PHQ-4), the L-1-scale, one question on perceived stress, and four items measuring state resilience [19]. The 5-item WHO-5 scale, which assesses subjective wellbeing but is also a valid tool for the screening of depression [20], was converted into a score ranging from 0 to 100 for comparison with other studies. Zero represents the worst mental health and a score ≤50 was used as sign for depression. Another valid and reliable score for the investigation of depression and anxiety in the general population is the PHQ-4 [3]. The PHQ-4 was added up to create a single score ranging between 0 and 12 with 0 indicating good mental health. The PHQ-4 consists of the Patient Health Questionnaire-2 (PHQ-2) measuring depression and the Generalized Anxiety Disorder 2 (GAD-2) quantifying anxiety. The overall life satisfaction was measured by the valid and reliable eleven-point L-1-scale [21] with 0 indicating no satisfaction and 10 representing a strong overall life satisfaction. Resilience was measured by four eight-point Likert-scales that were then averaged to a total score [19]. The subjective stress-level was assessed by a further eight-point Likert-scale rating the statement "I feel stressed out".
Sleep data were divided into general sleep quality and current sleep quality. Both scores were measured through a ten-point Likert-scale with 1 indicating bad sleep quality and 10 indicating excellent sleep quality.

Statistical Analysis
First, descriptive statistics were performed for both groups. Prior to this, we z-standardized all variables.
In the next step, we tested all observed variables regarding their distribution features. As none of the variables were normally distributed, the Mann-Whitney-U-Test was used to identify differences between the PH and the control group regarding mental health scores (WHO-5, PHQ-4, L-1), physical activity (total physical activity, the activity of daily living, sports activity, climbing stairs) resilience, stress experience, and quality of sleep.
We further used Spearman's rank correlation coefficients to determine the relationship between mental health (as indicated by WHO-5, L-1, and PHQ-4), physical activity (total physical activity, the activity of daily living, sports activity, climbing stairs), resilience, feelings of stress, and sleep quality for the PH and the control group separately.
Lastly, we analyzed the impact of resilience, total physical activity, sleep quality and feelings of stress as potential predictors of mental health using multiple regression analyses for the PH group as well as over both groups. We, therefore, further added four product terms to the model to test for an interaction between total physical, resilience, stress, sleep, and group membership. SPSS 22 was used for statistical analysis and partly for graphical illustration. For further illustration of the data, we used Python 3. p values < 0.05 will be considered significant. Bonferroni-correction was used to correct for multiple comparisons.

Mental Health Indicated by WHO-5, PHQ-4, and L-1
For the PH group, the level of mental health was represented through the WHO-5 scoring 11.9 ± 5.6, the PHQ-4 being 3.9 ± 3.0, and the L-1 being 5.6 ± 2.7 on average. In the present cohort, a prevalence of depression ranged from 36.4% and 53.6% derived from the WHO-5 score (score less than or equal to 50 as depressed), the PHQ-2 yellow flag (score greater than or equal to 3), and PHQ-2 red flag. Symptoms of anxiety were found in about one-third of the PH patients. The assessment of mental health revealed significantly lower life satisfaction (L1), well-being (WHO-5) as well as increased levels of depression and anxiety (PHQ-4) in the PH group regarding WHO-5, PHQ-4 and its subscores PHQ-2 and GAD-2 and L-1 (all p < 0.001) ( Table 2).

Resilience, Stress Experience, and Sleep Quality
The level of resilience was rated on average 4.1 ± 1.6, subjective stress experience was rated on average 2.9 ± 2.1 in the PH group. The general sleep quality was rated as 5.7 ± 2.4 on average, while the current sleep quality was only 5.3 ± 2.5. The level of resilience as well as the subjective stress experience in the PH group did not differ significantly compared to the control group. The general and current sleep quality were both significantly lower in the PH group compared to the control group (both p < 0.001). The estimation of one's living standard showed a statistically significant increased score in the PH group regarding the item "worries about health" and a decreased score in the PH group regarding the item "satisfaction with sports behavior" (p < 0.001). Results are depicted in Figure 1 and Table 2. experience in the PH group did not differ significantly compared to the control group. The general and current sleep quality were both significantly lower in the PH group compared to the control group (both p < 0.001). The estimation of one's living standard showed a statistically significant increased score in the PH group regarding the item "worries about health" and a decreased score in the PH group regarding the item "satisfaction with sports behavior" (p < 0.001). Results are depicted in Figure 1 and Table 2.

Physical Activity
The average total physical activity of the PH group was 684.3 ± 954.3 (min/week), whereby 551.4 ± 816.9 (min/week) were accounted for by the activity of daily living and 129.4 ± 219.8 (min/week) for sports activities. The mean score of climbing stairs was 14.3 ± 23.0 (floors/week). All modalities of physical activity were significantly lower compared to the control group (all p < 0.001, Table 2).

Associations between Subjects' Characteristics, Physical activity, Mental Health (Reflected by L-1, WHO-5, PHQ-4), Resilience, and Stress Experience in PH Patients
The four activity indices were correlated with the mental health scores, resilience, stress experience, sleep quality, and sociodemographic data in the PH group (Table 3 for

Physical Activity
The average total physical activity of the PH group was 684.3 ± 954.3 (min/week), whereby 551.4 ± 816.9 (min/week) were accounted for by the activity of daily living and 129.4 ± 219.8 (min/week) for sports activities. The mean score of climbing stairs was 14.3 ± 23.0 (floors/week). All modalities of physical activity were significantly lower compared to the control group (all p < 0.001, Table 2). The four activity indices were correlated with the mental health scores, resilience, stress experience, sleep quality, and sociodemographic data in the PH group (Table 3 for = 0.002). These associations were not found in the control group. All correlations for the PH group can be found in Table 3. Table 3.

Associations between Subjects
Spearman's rank correlation coefficients for associations between activity scores, sociodemographic data, mental health, and sleep quality of PH patients. Significant associations are depicted in bold.

Regression Analysis: Prediction of the Mental Health by Resilience, Stress Experience, and Sleep Quality during the Lockdown
In a first step, we calculated regression models for the PH group only to investigate whether life satisfaction, mental well-and illbeing (represented by L1, WHO-5, and PHQ-4) could be predicted by total physical activity, sleep, stress, and resilience.
The explained by the model as a whole was R 2 = 0.682, F(4, 81) = 44.379, p < 0.001. These data revealed that especially resilience, sleep quality, and stress experience are relevant predictors of mental health outcomes and wellbeing. The patients' total physical activity only seem to impact variables related to life satisfaction (Figure 2A-C).
When considering the total sample and including the group variable in order to reveal differences between the PH group and healthy controls, linear regression analysis revealed that the mental health issues and mental wellbeing could be predicted by the diagnosis of PH, resilience, stress experience, total physical activity, and the current sleep quality. Total physical activity level had an albeit smaller impact on mental wellbeing.
Regarding life satisfaction, results showed that the L1-score was significantly lower for PH patients  When considering the total sample and including the group variable in order to reveal differences between the PH group and healthy controls, linear regression analysis revealed that the mental health issues and mental wellbeing could be predicted by the diagnosis of PH, resilience, stress experience, total physical activity, and the current sleep quality. Total physical activity level had an albeit smaller impact on mental wellbeing.
Results further showed that the WHO-5 score was significantly lower for the PH group  Results further showed that the WHO-5 score was significantly lower for the PH group compared to healthy controls (B = −0.511, beta = −0.246, p = 0.001). Like for the PH group only, the WHO-5 score was significantly moderated by the individual resilience (B = 0.522, beta = 0.539, p < 0.001), subjective stress level (B = −0.179, beta = −0.172, p = 0.001), and sleep quality (B = 0.269, beta = 0.278, p < 0.001) in the total sample analysis. Total physical activity had also a small significant impact on wellbeing (B = 0.096, beta = 0.097, p = 0.044) The total variance explained by the model as a whole was R 2 = 0.740, F(5, 127) = 73.393, p < 0.001. The inclusion of product terms reflecting interaction by group did not explain significant additional variance in the WHO-5 score, R 2 = 0.738, F(9, 127) = 40.809, p < 0.001, revealing no significant interactions between PH and control group.
Regarding mental health, results showed that the PHQ-4 score was significantly higher for PH patients compared to healthy controls (B = 0.638, beta = 0.295, p < 0.001). Moreover, increases in PHQ-4 scores correlated significantly with decreases in resilience (B = −0.505, beta = −0.504, p < 0.001), decreases in sleep quality (B = −0.196, beta = −0.196, p = 0.004) and increases in stress (B = 0.235, beta = 0.219, p < 0.001). The total variance explained by the model as a whole was R 2 = 0.640, F(5, 128) = 46.508, p < 0.001. The inclusion of the product terms did explain significant additional variance in the PHQ-4 score, R 2 = 0.685, F(9, 128) = 31.85, p < 0.001, revealing that significant interactions between group and resilience (B = −0.251, beta = −0.201, p = 0.034) as well as group and stress experience (B = 0.299, beta = 0.218, p = 0.014) reflecting a lower resilience as well as a higher stress experience during lockdown in patients with PH leading to a stronger rate of change of mental disorders in PH patients ( Figure 2C,D).

Discussion
The present data showed that during the COVID-19 pandemic PH patients had significantly lower physical activity, mental health, and sleep quality compared to age-matched healthy subjects. Being physically active positively correlated with mental health and sleep quality in the PH group. The inclusion of product terms reflecting interaction by group did not explain significant additional variance in the WHO-5 score, revealing no significant interactions between PH and control group. Using multiple regression, data revealed for the PH group that mental health issues could be predicted by total physical activity, sleep, stress level, and resilience indicating that especially resilience, sleep quality and stress experience are relevant predictors of mental health outcomes and wellbeing. The patients' total physical activity levels, however, only seem to impact life satisfaction. When comparing both groups, we found that in PH patients as well as healthy controls lower resilience, higher stress experience, lower sleep quality as well as reduced physical activity leading to a diminishment of life satisfaction and mental wellbeing. Furthermore, the present data demonstrate that lower resilience and higher stress experience were even leading to a stronger increase of mental illbeing in PH patients compared to healthy controls.
Previous studies documented that PH patients are significantly less active compared to healthy subjects in non-pandemic living conditions [22,23]. During the lockdown period of the pandemic, PH patients seemed to further reduce their activity levels to about 50% compared to accelerometry data from Gonzales-Saiz et al. [22]. Here, one has to keep in mind that accelerometer data are difficult to compare with questionnaire data. However, due to social desirability, most people are even more positive about themselves in surveys than they are [24]. Interestingly, these results are in contrast to recent findings in healthy subjects. Here, it has been demonstrated for a large group of participants that lockdown restrictions did not lead to a decrease in sports activity levels in previously low active subjects [16]. From a therapeutic point of view, this seems to be problematic for PH patients, because a daily activity is an important prognostic factor for the symptomatology and progression of the disease [25].
Concerning mental health issues, PH patients show a significantly decreased mental wellbeing compared to healthy controls. The present data further implicate a prevalence of depression ranging from 36.4% and 53.6% during the pandemic. In contrast, during non-pandemic times, depression prevalence of 7.5 to 55% was reported, with an average prevalence of depression of 36% [5,26,27]. Hence, the prevalence of depressive symptoms of PH patients during the pandemic was found in the upper half of the occurrence under normal circumstances. These findings suggest that social distancing and self-isolation only slightly favor the development of depressive symptoms. However, the more significant factor seems to be the functional limitations of PH patients [26]. Mental impairments that promote depression, are represented by anxious symptoms which were found in about one-third of the PH patients. Compared to the given control group, this prevalence is significantly higher. However, previous studies found anxiety disorders in PH patients in a range of 13 to 45.5% in non-pandemic situations [27][28][29]. Accordingly, the level of anxiety seems to be not significantly higher during the COVID-19 pandemic.
A potential driver of mental disorders seems be the worse sleep quality of PH patients. The self-reported general sleep quality indicates an about 25% lower level compared to subjects of the control group. Similar results were found for the current sleep quality during the COVID-19 pandemic, where the PH group scored 27% lower levels compared to the control group. PH patients seem to suffer also during non-pandemic conditions from a reduced sleep quality compared to healthy subjects [26]. Quality of sleep and the mental health scores WHO-5 and L-1 positively correlated with total physical activity, total activity and climbing stairs. These associations once again underline the close connection between different lifestyle factors. However, it has to be stated that the participation in sports activities is generally low in PH patients, even if there is sufficient evidence available that regular exercise training has an overall positive effect on the physiological and psychological components of PH [25]. The positive relationship between sports activity and mental wellbeing and life satisfaction holds for the PH and the control group as revealed by the missing interaction in the multiple regression analysis which is not surprising given the tremendous evidence for physical activity and sports to promote wellbeing and buffer stress [30]. However, correlations for the control group failed to reach significance. Here, we speculate that this effect is driven by the rather small size of this sub-group.
The present data revealed that especially subjective stress experience as well as individual resilience seem to be strong predictors of mental illbeing of PH patients as well as healthy controls. We assume that subjects of the control group also experience more stress during the pandemic and that their resilience suffers. Therefore, they are approaching the lower level of PH patients. As one major antecedent of mental health issues, like depression, is life stress, both daily hassles [31] and major negative life events [32], it is not surprising that the subjective stress level is strongly associated with mental wellbeing in both groups. Our data revealed that lower resilience and higher stress experience even lead to a stronger increase of mental illbeing in PH patients than in healthy controls reflecting that the present pandemic hit PH patients even harder as their adaptive capacity and resources to react on the situation are lower and mental health might be impaired even more.
The observed impact of resilience on mental health seems plausible and has already been confirmed in healthy subjects [33]. Theory suggests that resilient individuals bounce back from negative experiences quicker and more effectively [34]. As it is demonstrated by a broad body of literature, physical activity is positively associated with a person's resilience [35]. Recent research showed that especially individuals with high trait anxiety, which may be a risk factor for developing clinically significant mental health problems, may preferentially show psychological, as well as physiological, benefit from physical activity [36]. In the present study, however, this could only partly be confirmed as we found only a small relationship between the activity level and life satisfaction and mental wellbeing.
A limitation of our study is that we used the hemodynamic PH definition of the current European Society of Cardiology (ESC)/European Respiratory Society (ERS) guidelines. The impact of the newly proposed hemodynamic PH criteria on mental health, physical activity, and resilience merit further investigation [2].

Clinical Implications
Given current research [25] and the present data, we suggest that patients should be encouraged to remain physically active even in pandemic times, although they should certainly do so in consideration of infection control. Vulnerable populations should also receive therapeutic support to improve their sleep quality and stress management as well as psychological resilience factors. All these interventions should also have a positive effect on mental wellbeing accompanied by less anxiety and depressive symptoms [37,38]. Of note, scheduling of appointments of PH patients in specialized outpatient PH expert centers was significantly reduced during the lockdown period [39].

Conclusions
Current data showed that PH patients showed significantly lower physical activity, mental health, and sleep quality compared to the healthy subjects during the lockdown period of the COVID-19 pandemic. While levels of depression seem to be only slightly affected during the COVID-19 pandemic, significantly lower resilience and higher stress experience lead to an albeit stronger diminishment of wellbeing in PH patients. Hence, it seems desirable to pay special attention to PH patients. Especially in this situation, patients should receive increased therapeutic support to improve lifestyle factors such as sleep quality, stress management and physical activity levels. This could be helpful, for example, in counseling on how to carry out physical activity while observing infection control. Funding: We got no funding for the study.