Effect of 12 weeks of yoga training on the somatization, psychological symptoms, and stress-related biomarkers of healthy women

A national survey of yoga practitioners reported that the percentage of females attending yoga training was 84.2% in the USA [17]. Because yoga is also far more popular for women than men in Japan, it seemed that it would be difficult to recruit men for a yoga study. Therefore, we recruited 39 healthy, adult women who had no experience with yoga. The participants were recruited by posters, flyers, and the Internet from ten yoga-training centers and from halls where yoga lessons were only a part of the lesson program. The recruitment sites were in Fukuoka, Kumamoto, and Kagoshima Prefectures of Japan. The following exclusion criteria were applied: (i) age < 20 years and > 50 years; (ii) taking medication including supplements in the month prior to the experiment; (iii) having an illness; and (iv) having a past history of significant physical or mental illness. All participants received detailed information on the purpose of the study and provided written informed consent. The participants who agreed were then handed questionnaires, a paper cup, a tube with a screw cap, and a self-addressed return parcel before and after the 12 weeks of yoga training. Each participant collected urine samples and, at the same time, answered the POMS and SCL-90-R questionnaires at home. The urine, frozen at home by the participant, and the questionnaires were sent as soon as possible via a parcel delivery service, which used a freezer van to keep the samples frozen at -18°C (-0.4°F). Participants who completed the yoga- training program received 2,000 yen (about $20). The study was approved by the Institutional Review Board of Kyushu University.

Yoga classes were conducted one day a week for about one hour each session for 12 weeks. All instructors of the yoga classes are certificated. The participants were requested to attend at least 10 of the 12 weekly yoga sessions and to practice on their own at home at least twice a week for over 30 minutes during this 12-week period. If they could not attend 10 of the 12 classes or the requested home practice, they were dropped from the study. The self-reported home practice activity time was confirmed at a classroom session. Some yoga classes were closed classes, for only the participants of our study, and others were not. Of the various types of yoga, we chose cyclic meditation yoga as our yoga intervention because of the availability of scientific studies on this form of yoga (reviewed in [18]). At the beginning of the cyclic meditation yoga training, the instructors emphasized practicing slowly with awareness and relaxation. The cyclic meditation yoga consisted of the following practices;

Although the participants sometimes practiced breathing exercises (Bhramari Pranayama) or meditation (Om meditation), they mainly practiced cyclic meditation yoga.

Urine samples were collected from all participants for the quantification of biopyrrin, 8-OHdG, cortisol, and creatinine. The participants were asked to avoid vigorous exercise and heavy psychological stress for 24 hours prior to urine collection. Urine was collected between 6:00 and 9:00 am, and 2 ml of urine from each participant was stored at -80°C (-112°F) until analysis. The urinary biopyrrin, cortisol, and 8-OHdG concentrations were measured by enzyme-linked immunosorbent assay kits (Shino-Test, Tokyo, Japan, Oxford Biomedical Research, Inc. MI, and Nikken Seil Co., Ltd, Shizuoka, Japan, respectively). The urine creatinine concentration was analyzed using the Accuras Auto-Cre diagnosis kit (Shino-Test, Tokyo, Japan) and biopyrrin, 8-OHdG, and cortisol concentrations were corrected based on the urine creatinine concentration. One participant was excluded from the analysis of biopyrrin, 8-OHdG, and cortisol concentrations because her urine was sent to us without having been frozen.

A questionnaire about demographic characteristics, the POMS questionnaire (Educational and Industrial Testing Service, San Diego, CA), and the SCL-90-R questionnaire were given to each of the participants. The questionnaire about demographic characteristics included questions about age, race, and education. The POMS questionnaire assesses six mood subscales: tension-anxiety, depression, anger-hostility, vigor, fatigue, and confusion. High vigor scores reflect a good mood or emotion, and low scores on the other subscales reflect a good mood or emotion. Yokoyama et al. previously translated the 65-item scale of the POMS into Japanese and demonstrated the reliability and validity of this Japanese version of the POMS [19]. The Japanese version of the POMS (Kaneko Shobo Co., Tokyo, Japan) was used for the present study. The test taker rates his/her mood over the past seven days on a 5-point scale ranging from “not at all” to “extremely”. The SCL-90-R is a validated and reliable questionnaire that is sensitive to changes in psychological distress. The validity and reliability of the Japanese version of SCL-90-R has been confirmed [20]. It is a 90-item self-report symptom inventory and consists of nine symptom dimensions, somatization, obsessive-compulsive, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, and psychoticism. The test taker rates how much each of 90 problems had distressed or bothered them in the past seven days on a 5-point scale ranging from “not at all” to “extremely. Somatization, anxiety, depression, and hostility were chosen for study because it has been demonstrated that the practice of yoga improves the mental state by lowering the levels of anxiety, depression and anger-hostility [4, 5, 21, 22] and because we wanted to ensure that the practice of yoga improves somatic symptoms. These questionnaires were chosen because of their sensitivity to change through therapeutic intervention in about 12 weeks [10, 22]. Of the participants, the data of three were excluded from the SCL-90-R data analysis because they did not fill out the back page of the questionnaire.

Statistical analyses were performed using a statistical software package (PASW Statistics 18, version 18.0.0 for Windows; SPSS Inc., Chicago, IL, USA). The Kruskal-Wallis test was used to compare age and education among the yoga-training, control, and long-term yoga groups. Distribution of the subscale scores of POMS and biopyrrin, 8-OHdG, and cortisol concentrations was analyzed using the Kolmogolov-Smirnov test, and we found that the tension-anxiety, depression, anger-hostility, fatigue, and confusion scores of POMS and biopyrrin, 8-OHdG, and cortisol concentrations were not normally distributed. In addition, the anxiety and hostility subscales of SCL-90-R in Japanese community samples had a floor effect (average – 1SD < 0) [20]. Therefore, we chose to run non-parametric statistics. The Wilcoxon signed-rank test was utilized to compare the biopyrrin, 8-OHdG, and cortisol concentrations, which were corrected based on the urine creatinine concentration and the scores of questionnaires before and after 12 weeks of yoga training. The Mann–Whitney U-test was used to compare the biopyrrin, 8-OHdG, and cortisol concentrations and the scores of questionnaires between the control group and the yoga-training group before yoga training and between the long-term yoga group and the yoga-training group after yoga training. Spearman rank correlation was used to test the relationship between the somatization score and the POMS scores or between the changes of subscale scores in POMS and SCL-90-R and the changes of stress-related urinary biomarkers (8-OHdG, biopyrrin, and cortisol). Differences of p < 0.05 were considered to be statistically significant.

We estimated that a sample size of 21 would allow us to detect significant differences in the before and after yoga training somatization scores from the SCL-90-R with 80% power (α = 0.05) based on the mean and standard deviation of the somatization scores reported in a previous study of mindfulness [10]. The sample size was calculated using the Power and Sample Size Calculation Software version 3.0 for Windows (Vanderbilt University, Nashville, TN, USA).

The data of 24 of the 39 participants, those who completed at least 10 of the 12 weekly yoga sessions and practiced on their own at home at least twice a week for over 30 minutes during the 12 weeks, were available for analysis (yoga-training group). There were no significant differences in the subscale scores of POMS and SCL-90-R and the stress-related urinary biomarkers before starting yoga training between those who dropped out before the end of the yoga training and those able to complete the 12 weeks of yoga training (data not shown). The demographic data of the yoga-training group, including age, race, and education, are shown in Table 1. As reference data, the demographic data of the long-term yoga group (with more than two years of experience with yoga) and the control group (without experience with yoga) in our previous study [6] are also shown in Table 1. No significant differences were found among the three groups.

The results of the subscale scores of the POMS questionnaire are shown in Figure 1. To add reference for the current results, the subscale scores of the control (without experience with yoga) and long-term yoga (with more than two years of experience with yoga) groups from our previous cross-sectional study are shown on the left and right sides, respectively [6]. All negative subscale scores, tension-anxiety (p = 0.022; Figure 1A), depression (p = 0.010; Figure 1B), anger-hostility (p = 0.020; Figure 1C), fatigue (p = 0.001; Figure 1E), and confusion (p = 0.004; Figure 1F) of the POMS after the 12 weeks of yoga training were significantly decreased compared with those before starting yoga training. There was a trend toward an increase of the vigor score after the 12 weeks of yoga training (p = 0.083; Figure 1D). There were no significant differences in any of the negative subscale scores of the POMS questionnaire between the control group and the yoga-training group before yoga training and between the long-term yoga group and the yoga-training group after yoga training.

Scores for somatization (p = 0.006), depression (p = 0.002), anxiety (p = 0.002), and hostility (p = 0.007) from SCL-90-R were significantly decreased after 12 weeks of yoga training compared with those before starting yoga training (Figure 2A-D).

Therefore, we investigated if changes in mood were related to changes in somatization. However, we did not find any significant correlation between the changes in mood (tension-anxiety, depression, anger-hostility, vigor, fatigue, and confusion in POMS) and the changes in somatization (data not shown). These results suggest that the change of somatization does not have a direct relation with the change of mood.

The stress-related biomarkers urine concentration of 8-OHdG, biopyrrin, and cortisol are shown in Figure 3. As a reference for the current results, the urine concentration of the control group is shown on the left and that of the long-term yoga group is shown on the right [6]. Contrary to our expectation, the urinary 8-OHdG concentration after 12 weeks of yoga training showed a significant increase compared with that before the start of yoga training (Figure 3A). The urinary 8-OHdG concentration of the yoga-training group before yoga training was significantly lower than that of the control group from our previous study (Figure 3A). No significant changes were observed in the levels of urinary biopyrrin and cortisol after the 12 weeks of yoga training (Figure 3B, C). There were no significant differences in the urinary biopyrrin and cortisol levels between the control group and the yoga-training group before yoga training and between the long-term yoga group and the yoga-training group after yoga training.

We investigated if the changes of the subscale scores of the POMS and SCL-90-R are related to the changes of stress-related urinary biomarkers, such as 8-OHdG, biopyrrin, and cortisol; however, we did not find any significant correlations (data not shown).