Improved Physical Fitness Induced by Physiotherapist-led Exercise in Patients With Permanent Atrial Fibrillation Disappears and Health-related Quality of Life Is Impaired After Detraining – A 3 Months Follow-up

Background: Atrial brillation negatively impact physical tness and health-related quality of life in patients. We recently showed that physiotherapist-led exercise-based cardiac rehabilitation improves physical tness in patients with permanent atrial brillation, however little is known about the effect of detraining after nishing an exercise period. The purpose of the study was to examine the impact of 3 months of detraining on physical tness, physical activity level and health-related quality of life among patients with permanent atrial brillation, after ending a randomized comparison of physiotherapist-led exercise-based cardiac rehabilitation versus physical activity on prescription. Methods: Prospective 3-month follow-up study after a randomized multi-centre study. Of the 87 patients completing the intervention study, 80 (92%) participated in the detraining part (22 women; age 74 ± 5 years), 38 from the physiotherapist-led exercise-based cardiac rehabilitation group and 42 from the physical activity on prescription group. All patients were asked to refrain from organised exercise during the 3-months period of detraining. The primary outcome measure was maximal exercise capacity using an exercise tolerance test. Secondary outcomes measures were muscle function, physical activity level, and health-related quality of life using a muscle endurance tests, Short Form-36, and physical activity assessments (questionnaire and accelerometer), as in the intervention study. We used the Mann-Whitney U-test and X 2 test to analyse differences between the groups, and Cohen’s d to determine the effect size. A mixed effect model analysis was used to identify predictors of change in physical tness. Results: Compared to the physical activity on prescription, physiotherapist-led exercise-based cardiac rehabilitation showed a signicantly decreased exercise capacity (−9 ± 11 vs. −2 ± 12 W, P < .0001), reduction in shoulder exion repetitions (−4 ± 8 vs. 2 ± 7 repetitions, P = .001), and reduced health-related quality of life in the Short Form-36 dimension Role Emotional (−13 ± 39 vs. 6 ± 27 points, P = .006). Conclusion: In elderly patients with permanent atrial brillation detraining negatively impacted previously

Accordingly, we recently demonstrated that 12 weeks of physiotherapist-led exercise-based cardiac rehabilitation (PT-X) improved physical tness in elderly patients with permanent AF and several co-morbidities, compared to physical activity on prescription (PAP) (7). The physical tness improvements following PT-X were comparable to those achieved in patients with chronic heart failure, in whom such improvements were associated with decreased mortality and hospital admissions (8).
Although it appears that a period of PT-X improves physical tness in patients with permanent AF, little is knowns about the long-term effects and what happens upon detraining. Detraining is de ned as the cessation or reduction of training, or a decrease in physical tness caused by training cessation or reduction (9). In the English literature, we have found no report addressing the impact of detraining in patients with permanent AF.
In the present study, our primary aim was to investigate the impact of 3 months detraining in terms of physical tness, physical activity level, and HR-QoL among patients with permanent AF who had participated in a randomized study comparing PT-X and PAP (7). Our secondary aim was to identify predictors of outcome, measured as a change in physical tness.

Methods
We conducted a patient follow-up at 3 months after the cessation of an intervention comprising 12 weeks of either PT-X or PAP. The PT-X intervention involved central circulatory interval exercise and circuit training at two 60-min hospital-based sessions, along with two home-based exercise sessions per week.
PAP comprised 40-min sessions of active walking four times per week, and the patients recorded homebased exercises and PAP in a diary. The PT-X and PAP interventions have previously been described in detail (7).

Patiens
A total of 87 patients (24 women) who had completed the intervention part of the study (40 PT-X and 47 PAP)(7) were eligible to participate this detraining analysis. Participants completed a 3-month detraining period and follow-up assessment. The patients received both written and verbal information, and provided written informed consent to participate. The investigation conformed to the declaration of Helsinki, was approved by the Regional Ethics Committee of Gothenburg., and was retrospectively registred at ClinicalTrials.gov Identi er: NCT02493400.

Protocol
During the 3-month detraining period, the patients were asked to avoid participation in any systematic exercise program that could improve physical tness (aerobic capacity and muscular function). After the detraining period the patients in the PT-X group were motivated and supported to regain their exercise habits, and the PAP group were offered a period of PT-X.

Tests
The testing procedures followed the same protocol as previously described in detail (7). The tests were performed after the detraining period by physiotherapists blinded to the randomisation. The primary outcome measure was maximum exercise capacity, assessed using a symptom-limited ergometer cycle test (Monark ergometer 839e; Monark, Varberg, Sweden) (10). The workload began at 25 W, and was increased by 25 W every 4.5 min until the patient's rated perceived exertion (RPE) was 17 (very heavy) on the Borg scale (11). At each workload, the patient's heart rate and blood pressure were measured.
The secondary outcome measures included HR-QoL, muscular endurance, and physical activity level. HR-QoL was measured using the Swedish version of the Short Form-36 Health Survey Questionnaire (SF-36) (12). The muscular endurance test involved unilateral isoinertial shoulder exion in a sitting position with a dumbbell (2 kg for women, and 3 kg for men); bilateral isometric shoulder abduction in a sitting position with a dumbbell (1 kg for both women and men); and an unilateral isoinertial heel-lift with a straight knee, on a 10° tilted wedge, with shoes on (13).
Physical activity level was measured using an accelerometer (Actigraph® GT3x+; Actigraph, Pensacola, Florida, USA). Patients were instructed to wear the accelerometer throughout the whole day for 7 days, except when taking a shower or bath. (14) Accelerometer data were calculated using the algorithm validated by Choi et al (15). The accelerometer measured the patient's MET-minutes per week (MET level × minutes of activity × events per week), and one metabolic equivalent of tasks (1 MET) was equal to an oxygen uptake (VO 2 ) of 3.5 mL × kg − 1 × min − 1 . Physical activity was also self-reported using the Swedish version of the Short Form International Physical Activity Questionnaire (IPAQ).(16) IPAQ categories were low (1), medium (2), and high (3) physical activity levels. For self-reported physical activity level during detraining, we also used the Saltin-Grimby's 6-grade physical activity level scale, which has been previously validated (17).

Statistics
Statistical analyses were performed using the Statistical Package for Social Science (IBM Corp. Released 2013. IBM SPSS Statistics for Windows, Version 22.0.; Armonk, NY, USA). Ratio and interval data are presented as mean (SD), ordinal data as median (range), and nominal data as absolute and relative numbers. The Mann-Whitney U-test and X 2 test were used to analyse differences between groups. Cohen's d was calculated to determine the effect size using the mean difference between groups and pooled standard deviation of the mean differences. A d value of ~ 0.2 indicated a small effect size, ~ 0.5 medium, and ~ 0.8 large.(18) Differences were considered statistically signi cant if P was < .05.
We performed mixed effect model analysis to identify predictors of change in physical tness. Independent variables were evaluated in a pairwise manner for possible collinearity, with Spearman's r ≥ 0.7. For variable pairs having one categorical variable with few categories, we additionally checked for overlapping boxplots. When both variables were categorical with few categories, we checked cross tables for > 80% observations in diagonal, and cells with < 5 observations. Three models were created, and each variable was checked in a stepwise manner and carried forward when P < .20. Model 1 included intervention, time, and intervention × time. Model 2 was model 1 with the addition of possible confounders: age, sex, and body mass index (BMI). Model 3 additionally included secondary explanatory variables for physical tness at baseline, mean value for heel-lift with the left and right leg, SF-36 PF (physical function), IPAQ category, and moderate-to-vigorous physical activity measured by accelerometer. The nal model 3 comprised the model 1 variable, signi cant model 2 variables (age and sex), and signi cant secondary explanatory variables (SF-36 PF and moderate-tovigorous physical activity measured by accelerometer). Variables in models were considered statistically signi cant if P < .05.

Results
Of the 87 patients who nished the main study, 80 (92%) completed the detraining period (22 women; 38 PT-X and 42 PAP). The remaining 7 patients (median age, 77 years; age range, 67-85 years; 2 PT-X and 5 PAP) withdrew their participation from the follow-up analysis. Of these 7 patients, 6 did not participate in follow-up due to medical causes or unwillingness, while one patient died within 3 months after the intervention due to causes unrelated to the intervention. All 80 patients who agreed to participate in the follow-up completed a detraining period according to the instructions and were included in the analysis. Figure 1 presents a ow-chart of the inclusion process. Table 1 summarizes the patients' demographic data and clinical characteristics, and Table 2 the patient's pharmacological therapies. The two intervention groups remained well balanced for the detraining period. The only signi cant difference was a slightly higher left ventricular ejection fraction in the PAP group, but the level was normal in both groups (58% vs 54%. P = .0014). The groups did not signi cantly differ in terms of pharmacological therapy. Around 90% received some form of stroke prophylaxis (mainly oral anticoagulation), > 80% required heart rate regulation (mainly beta-blockers), and the majority used reninangiotensin-inhibiting therapy.
We compared the data from the tests performed after the intervention period with those obtained after the 3-month detraining period.

Physical tness
Detraining led to a signi cantly greater reversal of physical tness in the PT-X group compared to the PAP group in terms of maximum workload (− 9 vs 2 Watts; P = .000014), and muscle endurance in shoulder exion in the left arm (− 4 vs 2 repetitions. P = .00078) ( Fig. 2A, B and Table 3). The mixed effect model con rmed that physical tness decreased after the detraining period in the PT-X group but not in the PAP group ( Table 5). The same result was obtained after adjustments for confounders (age and sex), and for the secondary explanatory variables, including the SF-36 physical function (PF) and moderate-to-vigorous physical activity measured by accelerometer (Fig. 3, Tables 5 and 6).

Physical activity
After detraining, the PAP group reported signi cantly higher energy expenditure compared to the PT-X group: 1863 kcal vs − 59 kcal (P = .036); 3464 MET/min/week vs − 96 MET/min/week (P = .040). However, these differences were not corroborated by accelerometer data. The two groups did not signi cantly differ in self-reported physical activity during the detraining period as measured by the Saltin-Grimby scale (Table 3).
Health-related quality of life Detraining led to a signi cantly greater reduction of HR-QoL in the PT-X group compared to the PAP group when measured using the SF-36 dimension Role Emotional (RE) (− 13 vs 6 points; P = .0058) ( Table 4).

Discussion
The present results showed that 3-month detraining almost completely reversed the improvement in physical tness (2 mL × kg − 1 × min − 1 measured as maximum work load) that had been achieved through 12 weeks of PT-X. Furthermore, the PT-X group exhibited impaired HR-QoL, with decreased scores on all SF-36 dimensions, a moderate decrease in Role Emotional (− 13 points), as well as clinically and socially signi cant decreases (− 5 to − 6 points) in Role Physical, Vitality, and Social Functioning. These ndings indicate that both physical tness and HR-QoL deteriorate if PT-X-induced improvements are not preserved. In the PAP group, detraining did not affect physical tness, but had contrasting effects on Role Emotional (improved by 6 points) and Role Physical (deteriorated by − 5 points).
Several studies among patients with cardiovascular diseases other than AF show that patients have di culty continuing exercise and maintaining a su cient physical activity level after nishing exercisebased cardiac rehabilitation (19,20). We previously found that elderly patients, with permanent AF and several co-morbidities, could achieve signi cant improvements in physical tness. However, our present results demonstrated that these improvements disappeared after 3 months of detraining, with the additional cost of impaired HR-QoL. Although a 3-month period of PT-X improves physical tness in older patients with permanent AF (7). and it has been shown that older adults require a longer time to retrain physical tness lost after detraining (21), it is rare for patients with AF to have access PT-X (22,23). How to best maintain improvement in physical tness is an active eld of research and development (6).

Previous investigations have shown that detraining causes central and peripheral alterations in athletes
and recently trained healthy individuals (24)(25)(26). These alterations are multifactorial, and include a reduced VO 2max due to reduced blood volume and higher heart rate response, which reduce stroke volume and affect cardiac output (26). The peripheral alterations include reduced muscular capillarisation and oxidative enzyme activities, and a decreased arterial-venous oxygen difference and reduced oxygen delivery to the cells, which affect mitochondrial ATP production (24). Our present ndings also revealed signi cant loss of muscular endurance, in line with previously reported results of detraining in patients with and without cardio vascular disease (27)(28)(29). Our results were comparable to those of Volaklis et al (28) and Ratel et al (29), which revealed that detraining led to reversal of improvements of VO 2peak and muscular strength among patients with cardio vascular disease and in older adults (24,25,(27)(28)(29). Our study showed that detraining had similar effects in patients with permanent AF.
Compared to after the intervention, patients in the PT-X group showed reduced HR-QoL after detraining, according to the scores on all dimensions of SF-36. In some dimensions, the post-detraining scores were lower than the scores before PT-X. In three of these dimensions, the SF-36 scores exhibited a reduction of at least 5 points, which is considered a clinically and socially relevant difference. Notably, the PT-X group showed a 13-point deterioration in the Role Emotional dimension, which is considered a moderate change (30), and which was signi cantly larger compared to the change seen in the PAP group.
Risom et al (6) found no evidence that PT-X improved HR-QoL. This was also our nding in our main study, in which the patients self-rated scores were similar to in the normative Swedish population of the same age range(31) Teixeira-Salmela et al (27) reported that HR-QoL increased with exercise training in the elderly population, and this increase persisted during detraining. They proposed that patients felt better about their physical abilities and, therefore, adopted a more active lifestyle. That situation differs from the situation in our present study, in which patients were asked to avoid structured exercise. Clearly, adherence to these instructions led to a reduction of HR-QoL in the PT-X group. Our results are congruent with the ndings of an observational study, in which elderly individuals (> 65 years of age) participated in a detraining period after a period of structured exercise (32). The participants received instructions similar to those given in the present study, and the results revealed signi cant deterioration of all dimensions of the HR-QoL as measured by SF-36, as in our study (32). These ndings clearly con rm that after cessation of PT-X, it is crucial to support patients to maintain their physical tness and HR-QoL.
Our results also showed that participating in PT-X was the most important factor for improved physical tness, and that the detraining-induced decline in physical tness in the PT-X group can partly explain the deterioration of physical scores on SF-36. However, the PT-X group participants also exhibited declines of scores related to mental well-being and social life (33). This may indicate the importance of PT-X participation for elderly patients with permanent AF. Although the intervention part of our study con rms a bene t of PT-X among elderly patients with permanent AF and co-morbidities, the overall evidence supporting PT-X is low (6). This extended study contributes additional knowledge, and corroborates that patients require support to preserve the improvements achieved through a period of PT-X.

Methodological aspects and limitations
In this study, we investigated the impact of detraining among patients with permanent AF after a period of PT-X or PAP. Allowing patients to participate in detraining may be ethically questionable due to the high evidence of bene ts from improved physical tness. However, it is well known that it is di cult to maintain good exercise habits, and the need for support in lifestyle changes is an important issue. The present ndings provide solid evidence supporting the need to continue exercising after a period of successful PT-X, and can thus serve to motivate patients to pursue this goal.

Conclusions
Although PT-X improves physical tness also in elderly patients with AF and concomitant diseases, a subsequent period of detraining has negative effects not only on physical tness but also on HR-QoL. The importance of continued exercise after PT-X is emphasized and should be part of the strategy.  Data are presented as mean (SD) or n (%). Abbreviations: PT-X, physiotherapist-led exercise-based cardiac rehabilitation; PAP, physical activity on p y p p y y prescription; LV-EF, left ventricular ejection fraction; TIA, transient ischemic attack; COPD, chronic obstructive pulmonary disease; CABG, coronary aortic bypass grafting; SD, standard deviation. Data are presented as n (%).
Abbreviations: C's d, Cohen's d; PT-X, physiotherapist-led exercise-based cardiac rehabilitation; PAP, physical activity on prescription; bpm, beats per minute; SBP, systolic blood pressure; DBP, diastolic blood pressure; HR, heart rate; Reps, repetitions; IPAQ, International Physical Activity Questionnaire; Kcal, kilocalorie; MET, metabolic equivalent; MPA, moderate physical activity; MVPA, moderate-tovigorous physical activity.  .007 In all models, time is used as a categorical variable to allow for non-linear relationships.  TABLE 6 Estimates of exercise capacity in Watts for the two intervention groups (PT-X and PAP) over the three time-points, using model-based mean (LS mean) and 95% con dence interval (CI)