After 12 weeks of yoga training, all negative subscale scores (tension-anxiety, depression, anger-hostility, fatigue, and confusion) from the POMS and somatization, anxiety, depression, and hostility from the SCL-90-R were significantly decreased compared with those before starting yoga training. This is the first study to demonstrate the effect of yoga training on the somatization symptoms of healthy women who had almost normal somatization scores on the SCL-90-R (Before yoga: 6.86 ± 6.69, After yoga: 3.64 ± 4.04 in this study and in a Japanese community sample: 7.56 ± 5.64 in a previous study ; these figures were calculated as the somatization scores in this paper and the previous paper multiplied by 12, because the number of questions about somatization in SCL-90-R are 12). Our findings suggest that regular yoga training reduces not only the level of psychological symptoms, but also somatization symptoms.
Somatization symptoms include medically unexplained physical symptoms such as headache, dizziness, chest pain, lower back pain, and nausea. Some of these physical symptoms have been shown to improve with yoga training. For example, it has been reported that the practice of yoga improved the symptoms of patients with clinically diagnosed migraine  and chemotherapy-induced nausea . However, these symptoms are not classified as somatization symptoms because they can be explained medically. Additionally, there is evidence of lower back pain improvement with yoga training [25, 26]. However, low back pain is only one of the 12 somatization score items on the SCL-90-R. A few studies have demonstrated that mindfulness training improves somatization symptoms [9, 10]. Meditation is a major component of mindfulness. However, the participants of our study practiced a combination of classical postures, breathing exercises, and meditation. Therefore, our findings show that this combination of yoga practices also has the potential to improve somatization symptoms.
Yoga has been reported to be effective with respect to negative psychological symptoms, such as anxiety [3, 27–29], anger/hostility , and depression [29–31]. Our previous study also suggested that long-term yoga training can improve negative psychological symptoms such as anxiety and anger . Our current results using the POMS and SCL-90-R questionnaires confirm these previous findings that negative psychological symptoms, such as anxiety, anger/hostility, depression, and confusion, are improved by yoga.
As for the relation between mood and somatization, previous study showed that only anxiety, not depression nor anger, had a direct effect on the somatic symptoms of anxiety disorder patients, whereas anxiety and depression, not anger, had a direct effect on the somatic symptoms of somatoform disorder patients . However, in healthy people there were no significant correlations between the changes in mood (tension-anxiety, depression, anger-hostility, vigor, fatigue, and confusion in POMS) and the changes in somatization. Our results suggested that anxiety, depression, and anger are not intimately linked to somatic symptoms in healthy people.
Contrary to our expectation, the level of urinary 8-OHdG after 12 weeks of yoga training showed a significant increase compared with that before starting yoga training. In Figure 3A, the urinary 8-OHdG level of the participants before starting yoga was significantly lower than that of the control group from our previous study  (p < 0.01; Mann–Whitney U-test). Before being corrected based on the urine creatinine concentration, the average urinary 8-OHdG level of the participants before starting yoga (7.0 ng/mL) was lower than the reference value for women (10.3 ng/mL) described in the document for the ELISA kit we used that appears on the manufacturing company’s website (http://www.jaica.com/guidance_8ohdg/index.html) Japanese. It is unclear why the urinary 8-OHdG level before starting yoga was lower than that of the control group. It is possible that the yoga-training group had an atypical background in terms of their daily activity. A previous study suggested that a heavy burden of work in addition to daily domestic roles increases urinary excretion levels of 8-OHdG . We speculate that the participants in the yoga-training group, in contrast with the others, had sufficient time to attend and continue the 12 weeks of training sessions. Most of them may have been working-women without daily domestic roles. Viewed in this light, through the practice of yoga, their activity levels may have increased and approached the activity levels of the long-term yoga group. Further studies are needed to elucidate the relationship between urinary 8-OHdG and psychosomatic distress that include the activity level.
The analysis of urinary biopyrrin in our previous cross-sectional study of yoga indicated no significant differences in the level of urinary biopyrrin between the control and long-term yoga groups . The current results also showed no significant change in urinary biopyrrin levels after 12 weeks of yoga training. This may be because it is difficult for healthy people to reduce their concentration of urinary biopyrrin by yoga training because their urinary excretion of biopyrrins is already stable at a lower concentration. Our results suggest that biopyrrin is suitable for studies of clinical populations, but not suitable for studies of healthy people.
Cortisol is an accepted, objective stress-related biomarker, because dysregulation of the level of cortisol is related to pathologies associated with stress-related symptoms, such as anxiety, depression, and negative affect [34–36]. However, a previous cross-sectional study of meditation and our previous cross-sectional study of yoga demonstrated that the baseline cortisol level of long-term practitioners was almost the same as that of a control group [6, 37]. Our present results also showed no significant change in urinary cortisol level by 12 weeks of yoga training. Although it has been reported that chronic stress causes a high basal cortisol level, the basal cortisol level of most healthy participants may be stable at a lower concentration. Another possibility is that we did not get the accurate peak cortisol level of all participants because we did not take into consideration differences in the participant’s sleep/wake schedules, although urine was collected between 6:00 and 9:00 am. Circadian rhythm should be tracked leading up to the urine collection to more accurately measure the cortisol level.
This study has some limitations. The first is the small sample size. However, we were able to demonstrate significant differences in somatization and psychological scores because there were more participants than the calculated sample size necessary to insure significance. The second limitation is that there is a lack of a comparison group. By using the previous data as a reference, we were able to show that the data of the yoga-training group before yoga training were almost the same as the data of the control group, except for the level of 8-OHdG. Also, the data of the yoga-training group after yoga training were almost the same as the data in the long-term yoga group. These data indicate that the changes of somatization and psychological symptoms in the yoga-training group compared with the “true” control group would have remained significant if an RCT has been used. Despite these limitations, the present findings suggest that yoga training can reduce the somatization score and scores related to mental health indicators, such as anxiety, depression, anger, and fatigue. Yoga training may affect both physical and mental well-being and be useful for preventing somatization and mental disorders. Further studies that measure psychosomatic symptoms and stress-related biomarkers of patients with somatization using yoga intervention compared with controls without yoga intervention, taking into consideration differences in their sleep/wake schedules using methods such as 24 hr urine collection, and/or sleep/wake logs, are needed to verify the effect of yoga.