Effects of hatha yoga on self-reported health outcomes in a randomized controlled trial of patients with obstructive pulmonary disorders

Abstract

Yoga is gaining popularity as an alternative treatment but knowledge of its effects remains limited, particularly among patients with chronic conditions. This randomized controlled pilot-study investigated immediate and long-term effects of a hatha yogic exercises (YE) program as compared to a conventional training program (CTP) on health-related outcomes including anxiety, depression, stress, sleep quality, insomnia, and subjective health complaints (SHC). Patients with obstructive pulmonary disorders (N = 36) were randomized into YE (n = 19) or CTP (n = 17). The YE included a newly developed 12-week program, adapted to the patient group, with two weekly classes delivered by experienced certified instructors. CTP training involved an individually tailored CTP. Questionnaire data were collected at baseline, at 12 weeks, and at a 6-month follow-up. ANOVAs comparing YE and CTP showed statistically significant interaction effects for anxiety and stress. In both groups, effects on anxiety were weak. Decreased stress was found in the CTP only. Separate analyses of each intervention showed consistently and increasing sleep quality in the YE-group between baseline and the 6-month follow-up, and decreasing SHC between baseline and the 12-week follow-up. For CTP, statistically significant effects emerged for both stress and SHC between baseline and the two follow-ups. Overall, comparisons showed more consistent effects of CTP, but for stress only. The findings suggest that YE and CTP effect different outcomes. This points at the importance of deciding which outcome to target when choosing between treatment alternatives.

Keywords:

• anxiety
• depression
• sleep
• stress
• pulmonary disorders

Introduction

Many chronic diseases involve handling various disease-related health problems. For instance, many chronic conditions are coupled by problems relating to mental health, sleep, stress perceptions, and common subjective health complaints (SHC). Asthma and chronic obstructive pulmonary disease (COPD) are two conditions, which can be categorized as obstructive pulmonary disorders. Compromised breathing is a characteristic of obstructive pulmonary disorders and also constitutes a common stressor among patients (Maurer et al., 2008). Both COPD and obstructive asthma involve inflammation and airflow limitations in the lung tissue. But while COPD involves persistent airflow limitation and permanent structural changes of different causes (WHO, 2017), asthma involves inflammation and episodic smooth muscle contraction of genetic or environmental cause.

Worldwide 235 million people are estimated to suffer from asthma and estimates suggest that COPD will soon be among the top three leading causes of death (Hynninen, Breitve, Wiborg, Pallesen, & Nordhus, 2005). The worldwide prevalence of COPD and obstructive asthma has been estimated to 8–10% while the prevalence of asthma varies between 2% and 33% (Burney, Jarvis, & Perez-Padilla, 2015) with differences between countries. However, both asthma and COPD are considered to be largely underdiagnosed (Hynninen et al., 2005). When it comes to other health-related problems of obstructive pulmonary disorders, the prevalence of anxiety and depression in COPD-patients have been considered higher than those for other chronic conditions (Hynninen et al., 2005; Maurer et al., 2008; Usmani et al., 2017). Considering this along with estimated increases in obstructive pulmonary disorders, it is important to study systematically different rehabilitation options and their effects on health-related outcomes, including anxiety and depression.

Yogic exercises (YE) refer to a multimodal form of modern hatha yoga exercise involving body, breath, and mind (Garber et al., 2011; Schmalzl, Powers, & Henje Blom, 2015). Depending on style, the physical exertion can be intermittent and moderate (Larson-Meyer, 2016). YE involve body awareness and self-regulation techniques that use psychophysical and mindful exercises as a form of biofeedback (Schmalzl et al., 2015). Combining a focus on body, breath and mind makes YE a form of meditation in motion (Cramer, 2016; Desikachar, Bragdon, & Bossart, 2005; Govindaraj, Karmani, Varambally, & Gangadhar, 2016; Uebelacker et al., 2010). Assumed positive effects of YE include improved health and well-being and thus YE are often recommended for mental health problems. Aligning with assumptions, a review synthesizing systematic reviews of chronic disease management demonstrates that yoga seems most effective for reducing symptoms of anxiety (Cramer, Lauche, Klose, & Dobos, 2017; Hofmann, Andreoli, & Curtiss, 2016), have small to moderate effects on depression, (Cramer et al., 2017; Sivaramakrishnan et al., 2019) and reduce pain (Cramer et al., 2017). Moreover, research has suggested that YE as an add-on intervention module decreases depressive disorders (Cramer, Lauche, Langhorst, & Dobos, 2013), and stress (Zou et al., 2018). Other meta-analytic findings have suggested effects of yoga to be superior of those of psychoeducation on depression but not on acute, chronic and/or treatment-resistant mood and anxiety disorders (Vollbehr et al., 2018). As for obstructive pulmonary disorders, meta-analytic findings of relaxation training which forms part of YE, has shown small to moderate improvements in anxiety and depression among COPD-patients (Volpato, Banfi, Rogers, & Pagnini, 2015). Also, YE with meditative movements have shown moderately positive effects in terms of improved sleep quality among older healthy individuals with sleeping problems (Wu, Kwong, Lan, & Jiang, 2015).

Treatment of obstructive pulmonary disorders such as COPD and asthma can include both pharmacologic and nonpharmacologic components (Vestbo et al., 2013), with pulmonary rehabilitation being an important nonpharmacologic component (Spruit et al., 2013). Physiotherapeutic interventions that form part of pulmonary rehabilitation include individualized physical training programs with strength and endurance training. Such interventions have been found effective and are considered as an established and evaluated treatment option as they decrease disease progression (Goldberg, Hillberg, Reinecker, & Goldstein, 2004; Sheel, Foster, & Romer, 2011; Spruit et al., 2013; Watz et al., 2014), increase functional capacity, and reduce dyspnea as well as depression in patients with pulmonary disorders (American Thoracic Society/European Respiratory, 2002; Goldberg et al., 2004; Young, Dewse, Fergusson, & Kolbe, 1999). In research comparing different treatment and rehabilitation alternatives, established individualized physical training programs can be referred to as conventional training programs (CTPs). However, patients with obstructive breathing disorders often seek symptom relief through non-pharmacological interventions such as yoga. With recent national activity guidelines from the UK and US including recommendations of physical yoga as a muscle strengthening activity which recognize YE as a form of physical activity (Department of Health US and UK, 2019), patients may become increasingly likely to look for YE as an alternative. With such recommendations and with patients asking for and themselves seeking out other rehabilitation and treatment options than the conventional ones, it becomes important to study systematically such alternatives. This means that randomized controlled trials (RCTs) including alternative interventions, such as breathing control, pulmonary rehabilitation exercises, and YE, while also evaluating the immediate and long-term effects (Usmani et al., 2017), are needed for patients with obstructive pulmonary disorders. Such a systematic study also includes investigating the feasibility of a YE programs and the safety of participating among different patient groups.

Specifically, this randomized controlled pilot study aimed to investigate the immediate and long-term effects of a newly developed hatha yoga program (YE) as compared to a CTP on secondary outcomes including self-reports of anxiety and depression, stress, sleep-related measures, and SHC among patients with obstructive pulmonary disorders (asthma and COPD). Overall, YE was hypothesized to have effects comparable to those of CTP on the secondary outcomes. Another aim involved investigating the feasibility of the YE-program in terms of safety and injuries along with negative effects on the secondary outcomes. Considering existing findings, the YE-program was expected to have no negative effects and to be feasible, safe, and involve no injuries.

Materials and methods

Design

Participants included in this study were recruited for a RCT examining effects of modern hatha YE compared to a CTP on functional capacity and lung function in patients with obstructive pulmonary disease, including diagnosed COPD and asthma. Primary RCT outcomes on walk distance have been reported elsewhere (Papp et al., 2017). The present study focuses on secondary outcomes and feasibility and safety relating to these outcomes.

Participants

Participants included patients with diagnosed obstructive pulmonary diseases (COPD and asthma) who participated in the RCT and provided self-reports in questionnaires.

Recruitment

Of 127 volunteers, 87 failed to meet inclusion criteria. The causes for exclusion covered (total n = 33) hypertension according to guidelines, upcoming surgery (n = 2), mental disorder (n = 5), chronic illness (n = 13), lack of time (n = 6), language problems (n = 5), time unsuitable (n = 7), and no specific personal reason (n = 16). This yielded 40 eligible patients who were randomized with 20 in each group. Drop-outs were as follows: YE = 3, CTP = 1.

Briefly, after providing written informed consent, individuals were screened for eligibility according to inclusion criteria (aged 35–85 yrs; according to electronic patient records diagnosed with obstructive pulmonary disease with mild to severe obstructions with GOLD 1–3, FEV₁/FVC <0.70 or diagnosed asthma with FEV₁ and FEV₁% of predicted respiratory function of 30% ≤FEV₁ ≤90%), and exclusion criteria (severe cardiovascular, neurological, orthopedic or rheumatologic disease or injury, surgery during the past 6 months, recent (previous 6 weeks) change of medication, medication affecting attention, unable to walk more than 200 m). Table 1 shows study sample demographics, diagnoses, and other characteristics important for the diagnoses.

Table 1 Descriptive statistics regarding demographics, diagnosis, and other characteristics at baseline (T1) for the hatha yoga (YE) intervention and the conventional training program (CTP).

	YE (n = 19)	CTP (n = 17)	p
Demographics
Age (years; M ± SD; range)	61 ± 10.1 (40–76)	67 ± 9.8 (43–84)	0.06
Gender (woman/man; n)	14/5	9/8	0.30
Diagnosis
COPD (n; %)	8 (42%)	3 (18%)	0.07
Asthma (n; %)	11 (58%)	14 (82%)	0.16
Other characteristics
Former smoker/never smoked/smoker	11/7/1	8/9/0	0.74/0.50/–
BMI (M ± SD; range)	25.5 ± 4.2 (23.5; 27.6)	28.1 ± 6.4 (24.8; 31.4)	0.16
FEV1, L (M ± SD; range)	1.86 ± 0.59 (1.58; 2.15)	1.75 ± 0.52 (1.48; 2.01)	0.55

Note: BMI: Body mass index, kg/m²; FEV₁: forced expiratory volume in 1 s, in liters. Three patients in YE had both COPD and asthma and one patient in the CTP (placed in COPD group).

Intervention

The YE-program (for details, see Papp et al., 2017) was developed by the first author, who is an experienced certified and registered yoga teacher and yoga therapist. Briefly, the intervention included two weekly classes, 60–70 min each, during 12 weeks. Classes were led by an experienced yoga teacher (not involved in the research project) who was trained by the first author. Participants were given a leaflet with the most important poses and a home training program on DVD along with recommendations to follow the program. The weekly classes included adapted hatha yoga poses (standing, seated, prone, supine, and back-bending) and breathing exercises emphasizing the synchronization of breathing and poses. Time spent on each yoga pose was gradually increased, with each pose being held between 5 and 40 s and each breathing exercise being performed for a longer duration with fewer pauses. Variations of poses (using walls, chairs, and the floor) were gradually introduced. Each class ended with relaxation and body scanning. The majority of the patients were novice to hatha yoga. All in the YE group were instructed to refrain from other YE and asked not to add any other physical activity during the intervention period.

Participants randomized to the active control group followed a CTP designed by a physiotherapist (second author). The CTP included low intensity cycling and strength training with two 60- to 70-min weekly sessions during the 12-week intervention period, and a recommended program for home training (DVD and paper instructions). Every participant had their individually adjusted load and repetitions. All in the CTP were instructed to refrain from yoga. Both CTP and YE sessions took place at the hospital. For both groups, the maximal dose included 24 sessions. After the final session, all were given physical activity on prescription and encouraged to continue their intervention programs.

Procedure

Participants provided self-reports in questionnaires at baseline (T1, a week before intervention start), immediate end-point (T2, within a week after ending the intervention), and at a follow-up (T3, 6 months after ending the intervention). The participants were blind to their results during the intervention and follow-ups. Afterwards some requested and received their individual data.

Measures

Anxiety and depression

Symptoms of anxiety and depression was measured using the Hospital Anxiety and Depression Scale (HADS; Lisspers, Nygren, & Soderman, 1997; Zigmond & Snaith, 1983), a well-established measure with adequate psychometric properties (Bjelland, Dahl, Haug, & Neckelmann, 2002). HADS includes two subscales, anxiety and depression, with seven items each. Ratings were made along a 4-point scale ranging from (0) to (3). Sum scores were computed with higher scores indicating more symptoms of anxiety and depression. Scores above 11 indicate clinical anxiety or depression, scores between 8 and 10 are borderline, while scores below 7 have no clinical relevance (Snaith, 2003).

Perceived stress

Overall stress perceptions were measured using the 10-item Perceived Stress Scale (PSS; Cohen, Kamarck, & Mermelstein, 1983; Nordin & Nordin, 2013). PSS was developed to measure whether individuals experience their lives as stressful, unpredictable and uncontrollable. Items were rated along a 5-point scale ranging from Never (0) to Very often (4). A sum score was computed with high scores indicating more perceived stress.

Subjective health complaints

A check-list including 15 common mental and physical complaints (Tibblin, Bengtsson, Furunes, & Lapidus, 1990) was used to measure SHC (Eriksen & Ihlebaek, 2002; Tibblin et al., 1990). For each complaint, participants were asked to indicate how often during the past two weeks they had had a complaint. Response alternatives ranged from Never (0) to Almost every day (4). Sum scores were computed with higher scores indicating more SHC.

Sleep and insomnia

Pittsburgh Sleep Quality Index (PSQI; Buysse, Reynolds, Monk, Berman, & Kupfer, 1989) with its seven domains was used to measure sleep quality over the last 2 weeks. Sum scores including all domains were computed with higher scores indicating poorer sleep quality; scoring 5 or more indicates poor sleep while 8 is the accepted clinical criterion.

Insomnia

The insomnia severity index (ISI; Bastien, Vallieres, & Morin, 2001) including seven items evaluating sleeping problems was used to measure insomnia. Scores between 8 and 14 indicate some sleeping problems, while scores 15–21 indicate moderate sleep disorder, and scores above 22 reflect severe insomnia.

In addition to the questionnaire asking about mental health, participants reported their age, gender, and smoking status (current smoker, former smoker, and never smoked). Objective assessments of length and weight were used to calculate body mass index (BMI). Spirometry was used to measures lung function (FEV₁ in liters). Finally, YE participants and the yoga teacher were asked to report any injuries.

Statistical analyses

Power calculations regarding sample size were performed for the primary study outcomes (e.g., 6-min walk distance, dyspnea-related distress-index) and required 24 individuals in each group for 80% power (two-sided test). Thus, no power calculations were performed for the secondary analyses reported here.

Fishers, Chi², and t-tests for independent groups were performed to examine group differences at baseline. Repeated measures ANOVAs including both 2 (CTP, YE) × 2 (T1, T2), and 2 (CTP, YE) × 3 (T1, T2, T3) were performed for each self-report measure. With significant time effects, planned post hoc tests were performed (T1 vs. T2; T1 vs. T3) and Student’s t-test or Wilcoxon Rank-Sum Test (for skewed variables) within and between groups. We also performed intention-to-treat-analyses and missing data were computed with the last observation carried forward method.

Analyses were performed for both the full sample and for those participating in at least 12 of 24 sessions. In total, 29 (YE: 13; CTP: 16) participants attended minimum 12 (of 24) training sessions, and were considered to have been sufficiently exposed to either intervention or its training effects.

Statistical significance was defined as p < 0.05. Due to the number of analyses, Bonferroni correction was performed to disentangle robust effects. The analyses were conducted using SPSS 23 (IBM SPSS Statistics for Windows, Version 23.0. Armonk, NY and STATA 14 (College Station, TX, USA).

Results

Table 1 presents baseline characteristics for both groups and shows no statistically significant group differences in these variables. Regarding health-related outcomes, there were no statistically significant group differences at baseline (Table 1).

Intervention effects

Results of repeated measures ANOVAs comparing YE and CTP showed statistically significant time effects on anxiety, sleep quality, and SHC with levels of self-reported health across these measures improving over time (Table 2). These overall time effects remained in the analyses of participants completing minimum 12 sessions (Table 3). Significant interactions (time × group) were found for anxiety and stress (Table 2). Separate analyses (t-tests) of the interventions showed significant immediate effects (T1–T2) on anxiety in both groups. Albeit the results for anxiety did not remain significant when adjusting for multiple comparisons, the results seemed slightly more consistent in the CTP. However, the analysis of participants who completed minimum 12 sessions showed a significant group × time interaction for stress only and not for anxiety (Table 3).

Table 2 Results (means, standard deviations, and significance levels) for self-rated health outcomes of analyses comparing hatha yoga (YE; n = 19) with a conventional training program (CTP; n = 17) and for the different points in time, including T1 (baseline), T2 (immediately after 12-week intervention), and T3 (6-month follow-up after ending the intervention) (N = 36).

Outcome measure	Anova time p**	Anova group ×  time p**	YE T1 M ± SD	YE T2 M ± SD	YE T3 M ± SD	YE T1–T2 p	YE T1–T3 p	CTP T1 M ± SD	CTP T2 M ± SD	CTP T3 M ± SD	CTP T1–T2 p	CTP T1–T3 p	YE/CT T1–T2 p	YE/CT T2–T3 p	YE/CTP T1–T3 p
Anxiety	0.0013	0.022	5.68 ± 4.08	4.84 ± 3.76	6.11 ± 4.62	0.031	0.60	6.18 ± 5.91	3.94 ± 4.31	4.35 ± 4.12	0.012	0.045	0.29	0.25	0.06
Depression	0.26	0.17	4.11 ± 3.35	3.84 ± 3.52	4.47 ± 3.45	0.15	0.80	4.53 ± 3.26	3.53 ± 3.34	3.41 ± 3.18	0.31	0.16	0.96	0.10	0.21
PSQI	0.0016	0.15	7.16 ± 4.16	6.89 ± 3.74	5.11 ± 2.83	0.42	0.007	7.71 ± 5.39	7.12 ± 5.25	6.88 ± 5.16	0.47	0.30	0.92	0.01	0.04
ISI	0.18	0.14	9.47 ± 5.37	9.21 ± 5.74	9.18 ± 5.57	0.56	0.49	11.18 ± 7.66	9.71 ± 7.56	9.24 ± 7.00	0.30	0.07	0.59	0.37	0.08
Stress	0.08	0.001	14.95 ± 8.06	15.37 ± 7.62	16.47 ± 7.67	0.69	0.15	15.06 ± 9.71	11.59 ± 7.65	11.59 ± 6.77	0.025	0.005	0.04	0.39	0.002
SHC	0.0001	0.096	16.00 ± 9.36	13.32 ± 8.23	14.47 ± 8.76	0.008	0.13	18.35 ± 10.86	13.59 ± 9.22	13.53 ± 8.69	0.002	0.007	0.34	0.32	0.07

Note. p = p-Value; p** = results of ANOVAs; p-values in bold denote values remaining significant after Bonferroni correction (adjusting for mass significance after multiple testing). PSQI: Pittsburgh sleep quality index; high scores reflect poor sleep with scores above 5 indicating poor sleep while 8 is the accepted clinical criterion; ISI: Insomnia severity index; high scores reflect poor sleep with scores between 8 and 14 indicating some sleeping problems, 15–21 indicating moderate sleep disorder, scores above 22 reflect severe insomnia; Stress: perceived stress scale; high scores reflect more stress; SHC: subjective health complaints; high scores reflect more complaints.

Table 3 Results (mean, standard deviations, and significance levels) for self-rated health outcomes of analyses comparing hatha yoga (YE; n = 13) with a conventional training program (CTP; n = 16), and for the different points in time, including T1 (baseline), T2 (immediately after 12-week intervention), and T3 (6-month follow-up after ending the intervention) including patients participating in minimum 12 out of 24 training sessions of YE or CTP (N = 29).

Outcome measure	Anova time p**	Anova group ×  time p**	YE T1 M ± SD	YE T2 M ± SD	YE T3 M ± SD	YE T1–T2 p	YE T1–T3 p	CTP T1 M ± SD	CTP T2 M ± SD	CTP T3 M ± SD	CTP T1–T2 p	CTP T1–T3 p	YE/CTP T1–T2 p	YE/CTP T2–T3 p	YE/CTP T1–T3 p
Anxiety	0.003	0.10	4.08 ± 3.40	3.15 ± 2.76	4.08 ± 2.90	0.08	1.00	6.00 ± 6.06	3.63 ± 4.24	4.13 ± 4.15	0.01	0.06	0.27	0.49	0.15
Depression	0.32	0.17	3.08 ± 3.09	2.92 ± 3.01	3.62 ± 3.15	0.26	0.69	4.38 ± 3.30	3.19 ± 3.12	3.19 ± 3.15	0.21	0.15	0.84	0.15	0.22
PSQI	0.003	0.45	6.08 ± 1.89	5.23 ± 2.20	4.15 ± 1.52	0.14	0.003	7.50 ± 5.50	7.06 ± 5.42	6.56 ± 5.15	0.68	0.21	0.59	0.14	0.14
ISI	0.20	0.13	7.46 ± 4.70	6.92 ± 5.22	8.08 ± 5.59	0.37	0.38	10.94 ± 7.84	9.25 ± 7.56	9.00 ± 7.16	0.20	0.10	0.67	0.26	0.09
Stress	0.26	0.002	10.77 ± 4.32	11.46 ± 4.65	13.23 ± 6.27	0.55	0.04	14.50 ± 9.74	11.25 ± 7.77	11.25 ± 6.85	0.05	0.009	0.08	0.21	0.002
SHC	0.0001	0.18	12.15 ± 6.59	9.69 ± 4.77	10.38 ± 4.59	0.04	0.16	17.63 ± 10.78	13.00 ± 9.19	12.75 ± 8.34	0.003	0.01	0.38	0.21	0.09

Note: p = p-Value; p** = results of ANOVAs; p-values in bold denote values remaining significant after Bonferroni correction (adjusting for mass significance after multiple testing). PSQI: Pittsburgh sleep quality index; high scores reflect poor sleep with scores above 5 indicating poor sleep while 8 is the accepted clinical criterion; ISI: Insomnia severity index; high scores reflect poor sleep with scores between 8 and 14 indicating some sleeping problems, 15–21 indicating moderate sleep disorder, scores above 22 reflect severe insomnia; Stress: perceived stress scale; high scores reflect more stress; SHC: subjective health complaints; high scores reflect more complaints.

When it comes to stress, subsequent analyses showed that both immediate (T1–T2) and long-term effects (T1–T3) were statistically significant in the CTP-group. However, only long-term effects remained when adjusting for multiple testing. Results for participants completing minimum 12 sessions also reproduced this significant interaction with long-term effects (T1–T3) being statistically significant in CTP-group only (Table 3). Yet, none of these findings withstood adjustment for multiple testing, probably due to low statistical power in these additional analyses.

Separate analyses of the two interventions also showed some clear differences, with significant long-term effects on sleep quality in the YE-group, which remained when adjusting for multiple testing and also emerged when analyzing YE-participants completing minimum 12 sessions. No such effect emerged for the CTP-group. For SHC, statistically significant immediate effects (T1–T2), which remained when adjusting for multiple testing, were found for both interventions (Table 2). However, in the CTP-intervention, long-term effects (T1–T3) were found as well. Including those participating in minimum 12 sessions, yielded similar findings for the CTP-group (Table 3), but with only the immediate effect sustaining adjustment for multiple testing.

For the study variables investigated here, no significant differences emerged between participants completing 12–24 sessions and the others (Tables 2 and 3). However, more individuals participating in minimum 12 sessions belonged to the CTP, with YE having lower adherence. Regarding exercise maintenance, after ending the intervention until the 6-month follow-up, some patients reported that they continued their programs (CTP: n = 4; HY: n = 7), while some reported discontinuing their programs (CTP: n = 4; HY: n = 5). Most reported doing some other training (CTP: n = 8; HY: n = 10). However, not all participants reported their activity.

As for feasibility, all participants managed to perform the YE and CTP programs. Regarding safety, neither instructors nor test leaders observed any injuries and none of the participants reported any injury.

Discussion

Using an RCT-design with an active control group to investigate immediate and long-term health-related effects of a 12-week YE-program adapted to patients with obstructive pulmonary disorders, we found no clear and consistent long-term effects on secondary mental health outcomes in terms of anxiety or depression. This was contrary to our overall hypothesis. Aligning with previous research comparing yoga with no treatment, progressive relaxation or medication suggesting that yoga seems the superior alternative (da Silva, Ravindran, & Ravindran, 2009), we found an immediate effect suggesting reduced anxiety. But, with the changes being small, this finding has to be considered more of a trend. Additional effects on sleep, stress and SHC varied between interventions with YE having positive effects on sleep quality and SHC, while CTP had positive effects on stress and SHC.

Regarding the separate analyses of the interventions, there were positive long-term effects of YE on sleep quality. This long-term effect remained in additional analyses, also when adjusting for multiple testing, including the smaller sample of participants attending minimum 12 sessions and receiving a reasonable YE-dose. Overall, this effect on sleep follows previous research on individuals with sleeping problems (Rubio-Arias, Marin-Cascales, Ramos-Campo, Hernandez, & Perez-Lopez, 2017; Wu et al., 2015, 2018). However, with long-term effects only it is unclear whether the effect follows from the YE-intervention. Yet, the descriptive statistics (Tables 2 and 3) suggest a gradient with sleep quality increasing (as reflected in decreasing scores) over time from the immediate to the long-term follow-up. With YE-participants being encouraged to continue their programs beyond the 12-week intervention, and with about 36% doing so while others did some other exercise, makes it unclear whether this effect relates to YE.

As for SHC, separate analyses of the interventions showed immediate effects in both groups with the effect remaining in the CTP-group (Table 2). These effects remained when analyzing the smaller group getting the minimum dose (Table 3), but only in the CTP-group. The differences in SHC findings may relate to the randomization, which unfortunately resulted in the YE-group including more COPD-patients. Due to their disease severity, COPD-patients typically suffer from more health complaints than asthma patients, which means that COPD-related complaints are more difficult to improve and any improvements are more difficult to maintain over time. However, before the intervention, the CTP-group with more asthma-patients had higher SHC levels (Table 2). This may relate to the CTP-group including older individuals (higher mean age and a wider age range; Table 1). Obviously, any age-related deteriorations reflected in more SHC may also have influenced the findings.

Regarding perceived stress, the YE-program may stimulate parasympathetic activity but in the present study, no effect was found on perceived stress. Perhaps additional measurements including biomarkers of stress are needed to identify such an effect (Corey et al., 2014). Yet, other studies have shown that YE with physical exercises (asanas) reduce stress (Bussing, Michalsen, Khalsa, & Sherman, 2012; Hylander, Johansson, Daukantaitė, & Ruggeri, 2017). Perhaps a reason for not finding any effects of YE on stress may relate to the YE-group including more COPD-patients, with the stressors of their chronic disease being more difficult to target than stressors of asthma among the patients in this study but also of stressors of non-patient groups included in other studies. However, perceived stress levels are comparable between groups while SHC levels are higher in the CTP-group. This suggests that overall stress and complaints may be less of an issue in this particular YE-group with more COPD-patients. Obviously, other factors such as patients perhaps considering the YE stressful and unfamiliar, may also explain the present findings.

For depression and insomnia, neither YE nor CTP had any significant effects. For depression this may seem surprising considering that previous research suggests that yoga is superior to no treatment and to progressive relaxation for depression (da Silva et al., 2009). Here however, the non-significant effects on depression probably relate to the symptom level being of no clinical significance, meaning that there were floor effects. The same holds for insomnia, with both groups having some insomnia issues but not of any severity.

With this study reporting on secondary health-related outcomes, the statistical power calculations were based on the primary outcomes. To get clear significant effects, larger samples are obviously needed. This is particularly the case for the minimum-dose analyses, which should be considered preliminary but still informative regarding adherence effects. Regarding the YE dose, newer consensus guidelines recommend a higher dose for targeting anxiety and depression (de Manincor, Bensoussan, Smith, Fahey, & Bourchier, 2015). Unfortunately, this was unknown when designing and developing the YE-program. However, with patients suffering from pulmonary conditions, including COPD and asthma, it is also important to evaluate whether patients are able to adequately carry out the program. This means that the YE-dose has to be balanced to characteristics of patients and their diseases. Here, all participants managed to carry out their programs, which probably related to the individual adaptions and possibilities to perform exercises in different ways. Still, the patients had fairly good mental health and reported surprisingly few mental health problems (Table 2). Thus, giving little room for any change over shorter time period. This also suggests that we, due to careful screening, and despite the severity of the diagnosis, recruited a healthier patient-group while patients with more problems probably chose to refrain from participation. Yet, there were more drop-outs in the YE-group and the YE-participants also attended fewer sessions compared to CTP participants. Starting with a smaller sample with most having asthma, any drop-out decreased groups sizes, which might have affected the results. This means that additional studies on larger patient groups are needed before drawing any conclusions. However, any future intervention study has to take into consideration that the recruitment and screening of patients with a specific diagnosis typically takes time, which is a challenge when trying to achieve larger sample sizes.

As for the long-term effects found in the present study, it is unclear whether any of these followed from YE, CTP or any unknown factors (e.g., other exercise programs). However, all study participants were given their individualized programs and were encouraged to continue exercising according to their programs. But only a minority in each group did so. Yet, with longitudinal attrition, exact figures remain unclear. Regarding the health-related outcomes studied, we included self-report measures of variables typically investigated in yoga research. Importantly, previously validated self-report measures were used, with their reliabilities (where applicable) being satisfactory. Yet, due to the smaller sample sizes, the combination of two diagnoses three patients in YE-group and one in the CTP-group had both asthma and COPD (they were included in the COPD-group). Also the fact that the age distributions varied between groups (wider range and older mean age in the CTP-group), additional studies are needed. In particular, future research has to investigate why participants drop-out from YE and understand factors relating to maintenance of YE. For instance, expectations may play a role. Unfortunately, information on patient expectations were not collected which means that it is impossible to know whether this had any effect on the present results. Despite the YE program being adapted to the patient group, the program may have been too strenuous. However, with additional adjustments during class, this seems unlikely. Instead, such adjustments are likely to have increased program feasibility and helped avoiding injuries, which along with exclusion criteria for comorbidity, allowed for a safe intervention.

Taken together, the present study has to be considered a feasibility study, which shows that it is possible to recruit and engage patients with obstructive pulmonary disorders, without them suffering from any consistently negative effects relating to mental health outcomes, and with YE having a better effect on sleep and CTP having a better effect on stress. From a clinical perspective, the findings suggest that patients should be asked about their needs and priorities regarding different health-related issues to then choose accordingly between different treatment and rehabilitation alternatives.

Conclusions

This study showed significant differences between YE and CTP, with follow-up analyses suggesting that YE had positive effects on sleep quality and SHC while positive effects of CTP emerged for stress. This positive effect on stress, favoring the CTP, also emerged in the initial and overall analyses. Aside the long-term effects on sleep quality and immediate effects on SHC, no other consistently significant effects were found of YE on the investigated secondary health-related outcomes, which also included symptoms of anxiety, depression, insomnia, and perceived stress. This may relate to the overall patient group reporting no clinically significant mental health problems, and few sleeping problems. Yet, due to the rather small sample sizes and combination of two diagnoses, there is a need of additional research on YE in this patient group.

Abbreviations
YE	=	hatha yogic exercises
FEV1	=	forced expiratory volume in one second

Acknowledgments

Thanks to all patients who volunteered participation. Additional thanks to our statistics helper and to the yoga instructors for their valuable time, support, and dedication. All data are stored at Karolinska Institutet, Division of Family medicine and Primary Care.

Disclosure statement

No potential conflict of interest was reported by the authors. First author is an experienced yoga therapist and yoga teacher. We conformed to the Helsinki Declaration concerning human rights and informed consent, and followed standard procedures concerning treatment of humans.

Funding

This work was carried out without funds from any funding agency.