Flight attendants (FA) who flew before the ban on smoking in commercial aircraft (implemented progressively from 1988 to 2000) present with many of the co-morbidities of chronic obstructive pulmonary disease (COPD) but their pulmonary dysfunction differs from the standard definition of this disorder. At the time that they were hired, they were required to be in good health. They were exposed to second-hand cigarette smoke (SHCS) in the course of their often vigorous work activities, as well as to a wide range of other stressors (including interpersonal stress, aviation-associated threats and emergencies, disrupted diurnal rhythms and polluted air (Crawford and Holcomb, 1991; Grajewski et al., 2003; Repace, 2004; Sutton et al., 2007). Recent studies (Whelan et al., 2003; Arjomandi et al., 2009, 2012; McNeely et al., 2014) demonstrate significant rates of abnormal pulmonary function in this population: air trapping, reduced flow at mid-volume, reduced exercise tolerance, and significantly elevated rates of chronic bronchitis and sinusitis. There are also increased rates of cardiac disease, depression and anxiety (Pinkerton et al., 2012), sleep disturbances, skin and reproductive cancers and hearing loss (Czura, 2005; McNeely et al., 2014). Many of these symptoms are also co-morbidities of COPD (Shrikrishna and Hopkinson, 2012; Vestbo et al., 2013a).

The magnitude and nature of these pulmonary abnormalities do not meet the standard criteria of COPD using the Forced Expiratory Volume in one second (FEV1) to Forced Vital Capacity (FVC) ratio (Vestbo et al., 2013b) (now recognized as an unreliable criterion at the individual level). In addition, Han et al. (2010) have proposed that there is significant heterogeneity in the clinical symptoms of COPD and that a broader definition may be warranted to characterize its differing phenotypes. COPD is thought to be irreversible; and this and the hypothesized effects of cigarette smoke (CS)on the autonomic nervous system may in part account for the lasting nature of tobacco smoke exposure.

The nature of the morbidities experienced by FA exposed to SHCS is poorly defined. The significance and etiology of the clusters of pathology mentioned above is not clear. Apart from the screening studies noted above, little has been published that serves to define or to suggest treatments for these specific patterns of symptoms. Treatment tends to be directed at the primary symptom, and the possible significance of clusters of co-occurring symptoms is rarely examined.

Our work with FA has led us to examine the role of autonomic nervous system (ANS) activation in this population of first responders, whose job is “to perform vital crewmember functions onboard air carrier aircraft, including emergency functions for aircraft evacuations, firefighting, first aid, and response to security threats” (Ballough, 2004). In addition to its role as a principal cause of COPD (Moritsugu, 2007), cigarette smoke (CS) has been shown to affect the ANS via its influence on fear learning (Gould, 2006). Exposure to CS at the time of stressful experience potentiates fear conditioning (Davis et al., 2006; Cohen et al., 2007; Kutlu and Gould, 2015), and makes long-term post-traumatic stress and autonomic imbalance more likely (van der Velden et al., 2008). This imbalance of the ANS negatively affects the immune, cardiovascular, respiratory and musculoskeletal systems and may reinforce the pathological changes triggered directly in the lungs by CS. The possible role of autonomic disturbance induced by workplace exposure to SHCS in causing pulmonary, immune or affective disorders has largely been ignored, and as far as we know, no therapeutic approaches specifically aimed at such conditions have been proposed. Our selection of tests and of the MM/qigong intervention are informed by this perspective.

Qigong, a traditional Chinese health practice, has been used in China for hundreds of years in treating people with respiratory, autonomic and immune dysfunction (Johnson, 2002). Qigong includes a large range of practices using specific postures, movements, breathing patterns and visualizations to treat disease and improve health. Many of these practices have a broad positive influence on a range of pathologies (Lee, 2015; Dong and Bergren, 2016), and tend to restore physiological functions to normal range (Cohen, 1999). Qigong, Tai Chi and Hatha Yoga, have been proposed to constitute a novel category of exercise, “meditative movement” (MM) (Larkey et al., 2009). Specific MM practices have been shown to be an effective intervention for COPD (Ng et al., 2011; Liu et al., 2012; Chan et al., 2013) and to be equivalent or superior to conventional pulmonary rehabilitation in benefiting several of its symptoms (Xu et al., 2010; Ng et al., 2014). MM has also been shown to benefit the immune system, both reducing chronic systemic inflammation (Jahnke et al., 2010) and increasing the effectiveness of acquired immune response to infection (Wang et al., 2012). MM practices help restore functionality to the ANS (Sun and Yan, 1992; Lee et al., 2003a,b, 2004), and also benefit depression and anxiety (Payne and Crane-Godreau, 2013) and improve quality of life in many chronic diseases (Chan and Larson, 2015; Wang et al., 2015). MM also induces positive states of mind through focusing awareness on interoceptive and proprioceptive experiences (Payne and Crane-Godreau, 2013). We believe MM could prove to be of specific benefit to the particular problems experienced by FA exposed to SHCS, especially in view of the possible linking factor of autonomic imbalance.

Our aim was to determine the effects of a novel form of MM/Qigong training by pre-and post-testing of pulmonary, autonomic and immune function in a cohort of FA previously exposed to occupational SHCS.

Outcome measures were selected to monitor a wide range of health measures. Further technical details about these measures are available in a previously published protocol paper (Payne et al., 2016).

We determined the required sample size for statistical significance based on a predicted difference of 46 meters in the 6MWT and an assessment of previous similar studies. Using this criterion, 20 participants in each group were needed to achieve a statistical power of 80% at a significance level of 5%. To compensate for anticipated drop-out, we recruited a total of 26 participants into the Selection group. Descriptive statistics, including mean, percent change and p-values, were used to describe and summarize baseline data. Handling missing data: In cases where data for a subject on one outcome measure is missing, we have chosen to eliminate that subject from the calculations for that measure. Significance levels for changes in measures from pre- to post-intervention test were determined using a one-tailed t-test. Comparisons between completers and non-completers were handled using a 2-tailed t-test. Categorical changes between these two groups were evaluated using a Chi-squared contingency table. Raw data and statistical calculations made are available (Payne and Crane-Godreau, 2016f).

Our results indicate improved autonomic function, improved exercise tolerance and cardio-pulmonary function, and reduced systemic inflammation. Autonomic dysfunction, reduced functional ability, and increased systemic inflammation are all co-morbidities of COPD. These results are in line with other studies of the effects of Qigong, specifically on respiratory disease (Ng et al., 2011) as well as on other health factors (Jahnke et al., 2010), supporting the possibility that Qigong may be a useful therapeutic intervention for our target population. Elevated blood pressure, systemic inflammation, autonomic imbalance and reduced functional ability are all strong risk factors for a variety of pathologies, especially in the elderly, suggesting that Qigong practice may have preventive and ameliorative health effects in FA exposed to SHCS and possibly in other populations.

Our intervention was effective in the primary outcomes, the 6MWT and hs-CRP. There was a 13% increase in the distance walked in the 6MWT. This indicates that, despite no significant change in FEV1 or FVC, participants had a substantial and significant increase in functional ability. This outcome was not unique to this study (Chan et al., 2011; Ng et al., 2011). The intervention itself contained no form of aerobic exercise or resistance training; we speculate that the change is due to increased efficiency of body use and movement patterns and a more responsive autonomic nervous system. Other studies have shown increases in strength and bone density due to Qigong practice apparently not due to conventional exercise factors (Lin, 2007; Wayne et al., 2007).

There was a more than 50% decrease in average hs-CRP. 10 of 14 participants improved, 2 stayed the same at low levels, and two increased. All 4 participants with levels over the clinical norm (3.0) reduced substantially, 3 to below the clinical threshold. These results are likely of clinical relevance. In a cohort of women, average age 68, systemic inflammation would be expected to increase with advancing age, and is a risk factor for the development of many diseases of the elderly, including cardiovascular disease, COPD, diabetes, and cancer (Il'yasova et al., 2005; Franceschi et al., 2007; Guarner and Rubio-Ruiz, 2014). Inflammation is associated with autonomic dysfunction (Black, 2003; McGovern and Mazzone, 2014; Mirakaj et al., 2014), so this result is compatible with the hypothesis that the reduction in hs-CRP may be due in part to improved autonomic function. Studies of qigong have shown positive effects on inflammation as well as AD (Lee et al., 2005; Jahnke et al., 2010), and there are both speculative (Pavlov and Tracey, 2012) and well-established (Felten, 1999) links between the autonomic and immune systems which could provide a mechanism for these results.

Secondary outcome measures provided an opportunity to look at the impact of qigong on conditions co-morbid with COPD. These measures included blood pressure, resting and post-6MWT heart rate, blood oxygen saturation, the COPD Assessment Test (CAT), spirometry, Borg Dyspnea Inventory, the COMPASS 31, the Zung Inventories of Anxiety and Depression, the Multidimensional Assessment of Interoceptive Awareness (MAIA), and Vitamin D.

Cardiovascular measures improved substantially. There was a significant decrease of 7% in average resting systolic blood pressure. 11 out of 14 participants showed reduced systolic blood pressure, and the numbers of subjects above the clinical threshold of 140 mgHg reduced from 4 to 1. There was also an almost significant 12% drop in average pulse pressure. Although the decrease in diastolic pressure did not reach significance, in all three blood pressure-related measures (systolic, diastolic and pulse pressures), the number of participants above accepted clinical range diminished substantially, a result which may be meaningful in clinical terms. Also, resting heart rate decreased significantly by 6%. Since the intervention did not include any forms of conventional physical exercise known to decrease blood pressure or resting heart rate, we speculate that improved autonomic cardiovascular control may be responsible. This hypothesis is supported by the fact that there was a substantial (42%) increase in the change of heart rate after the 6MWT. Although this change did not quite reach significance, it is evidence of a more dynamic cardiovascular autonomic control.

In comparing the graphs of the changes in the 6MWT, hs-CRP and systolic blood pressure, we note that in the measures having a clinical threshold, the change was toward a normal clinical level rather than an across-the-board change. In systolic blood pressure and hs-CRP, values in the normal range showed relatively little change whereas values outside clinical range moved into normal range, with the amount of change approximately proportional to the prior deviation from normal. However, the 6MWT by contrast showed a fairly uniform amount of change regardless of baseline value. Qigong practice claims to normalize physiological variables rather than simply to reduce or increase them in a linear fashion (Cohen, 1999); these results are consistent with this claim.

In addition to objective measures, we included questionnaires to evaluate self-reported autonomic symptoms, affective state and bodily awareness. These are important for evaluating quality of life and changes in cognitive and affective functioning.

The COMPASS 31, a questionnaire evaluating general autonomic function, dropped from 15 to 11.5. This 24% change was significant (p ≤ 0.05), and supports the hypothesis of improved autonomic function as a result of the intervention.

The Zung Anxiety score dropped from 29.4 to 26.9 (p ≤ 0.05), a significant 8% drop. Since anxiety is associated with autonomic imbalance (increased sympathetic and reduced parasympathetic activity), this result too is compatible with improved autonomic function, as well as being important for better health and quality of life.

The MAIA is designed to measure the degree of awareness a person has of their bodily feelings, and the comfort they have with these feelings. This is hypothesized to relate to the capacity for self-regulation (Farb, 2015; Payne et al., 2015). This capacity includes the ability of the prefrontal cortex, insula and anterior cingulate gyrus to modulate arousal in the limbic system and hypothalamus (Critchley et al., 2003; Craig, 2008). The significant 8% increase in the MAIA score is consistent with the hypothesis that our intervention acted in part by improving autonomic functioning via increased top-down control.

The mechanisms for the long-term pathological effects of SHCS exposure are only partly understood. Our results suggest an overlooked role for the ANS. Cigarette smoke (CS), whether inhaled directly or second-hand, is known to be a principle cause of COPD and its comorbidities as well as other respiratory system disorders (Moritsugu, 2007). In COPD, inflammation, mucous secretion and plasma extravasation are in excess, and antigen responsiveness is inhibited, all together giving rise to pathology. The pulmonary pathology of COPD is accompanied by systemic abnormalities including inflammation and neuro-humoral activation (Andreas et al., 2005), as well as multiple co-morbidities, including cardiovascular, immune, musculoskeletal and affective disorders (Sin et al., 2006; Fabbri et al., 2008). Although only one of the FA recruited for the present study had COPD as defined by the GOLD standard FEV1/FVC<0.70 (Vestbo et al., 2013b), many of them manifested pulmonary dysfunction (as previously noted by Arjomandi) as well as other symptoms that are recognized as co-morbidities of COPD (Vestbo et al., 2013b). Both core pathology and the co-morbidities of COPD are known to be associated with chronic dysfunction of the autonomic nervous system (ANS) (van Gestel and Steier, 2010). In addition, it is recognized that AD adversely influences the immune system (Czura, 2005), predisposing to inflammation, infections and cancer, all of which are associated with COPD (Libert, 2003; Czura, 2005; Pavlov and Tracey, 2015).

Emerging evidence from human epidemiology and animal testing suggests that the presence of nicotine within the body may alter the susceptibility of the nervous system to long-term dysfunction. CS exposure during stress has been shown to strengthen learned fear responses (Davis et al., 2006; Elias et al., 2010; Kutlu and Gould, 2015). In mouse model studies, long-term fear memory is strengthened. In human epidemiological studies, there appears to be an increased likelihood of developing chronic or post-traumatic stress with consequent autonomic dysfunction (van der Velden et al., 2008). ANS dysfunction negatively affects the immune, cardiovascular, respiratory, and musculoskeletal systems. It seems likely that this may reinforce the pathological pulmonary changes triggered by cigarette smoke through its direct effect on lung tissue. In this model the effects of CS on fear learning and consequent ANS imbalance are an intrinsic part of the etiology of COPD and COPD-related symptoms. This concept is a distinct break from most previous literature, which has documented self-medication through nicotine use in those who already have affective disorders. However, there is evidence that nicotine exposure precedes and exacerbates the effects of exposure to stress or trauma (Elias et al., 2010). Since FA are first responders and are at risk for high-stress exposure, we believe that the hypothesis that systemic co-morbidities of COPD may be exacerbated by cigarette smoke exposure at or close to the time of stressful or traumatic events deserves further consideration.

This study addressed a specific population, FA exposed to second-hand cigarette smoke. This, and the requirement to attend classes, effectively limited the population to females over 49 with reasonable functional ability. The generalizability of these results to the non-FA population, males, younger people, and those with more severe pathology remains to be established; however given the results of other qigong studies with different populations, it appears likely that similar results would be obtained.

However, in comparing our results with those of other studies of MM, we note that the intervention we developed differs from other MM practices in that it emphasizes integrating the principles of MM practice with daily life, rather than setting aside specific times for practice as is done in most studies (Ng and Tsang, 2009b). We believe there are both advantages and disadvantages to this approach, involving both effectiveness and compliance.

As we have said elsewhere (Payne and Crane-Godreau, 2013), we believe that precise and complete descriptions of qigong interventions are essential in order to enable comparisons between different studies. Details of this intervention are available as supplementary materials to this paper, and a full description and analysis of the exercise selection process, the final choice of exercises, and issues of participant compliance will be published separately. Compliance in particular remains a significant issue, and we plan future studies to explore ways of increasing both compliance and effectiveness of this form of intervention.

The large percentage of participants failing to complete the study is a possible source of bias. This study was not designed to accommodate people with significant levels of affective disturbance. In some cases, affective disturbance only became apparent to us after participants had been consented and had begun to undergo testing. Improved screening is indicated in future similar studies. Life-style was claimed as the reason for non-completion of this study by non-Caucasian participants; we are exploring how to effectively reach a more diverse population.

In this study, we did not use a control group. Since our purpose was to determine which MM exercises are most effective and best tolerated, use of a control group would not have been appropriate. This however diminishes the power of results from the pre-and post-intervention tests as it is not possible determine with certainty whether results are due to attention, seasonal or other factors. However, we are currently proceeding with the RCT stage of the study (Phase II), using an attention/health education control group, which will provide a well-controlled validation of the effects of the selected MM exercises. It is however possible that positive results from the testing in Phase I, followed by negative results from Phase II, might reflect the ineffectiveness of digital delivery rather than the ineffectiveness of the intervention. In this case an RCT testing in-person delivery would be indicated.

The 6MWT is known to be influenced by motivation (Southard and Gallagher, 2013). Every precaution has been taken by the researchers to standardize the administration of the test. Nonetheless, in the Selection stage, subjects may be more highly motivated in the post-test situation, and this could skew results. As mentioned above, the use of a control group in the RCT stage will provide a check on these results.

We are currently conducting Phase II of this study, a randomized controlled trial investigating the effectiveness of digital delivery of these practices. Results from this will indicate whether we need to further refine our method of delivery, or whether we should explore ways of making this material immediately available to FA. In Phase II we are using wearable Holter EKG monitors to analyze heart rate variability. These measures will allow us to test further the hypothesis that the benefits of these practices are due to improved autonomic function, as well as to examine the relative contributions of sympathetic and parasympathetic factors to the reduction of systemic inflammation.

PP and MC in part receive compensation from teaching and consulting relating to Meditative Movement and MC from consulting on flight attendant health concerns; DZ is financially compensated by the lab that provides the biomarker analysis for the study. SF and JL declare no conflict of interest.