The results suggest that yoga intervention helped reduce the length of hospital stay and improved wound healing by shortening the interval period for suture removal and reduced plasma TNF-alpha levels significantly as compared to the support and exercise intervention group following surgery. The results did not differ when postoperative outcomes were analyzed in only the mastectomy group. Regression analysis on postoperative outcomes showed intervention to affect the duration of drain retention and hospital stay and TNF-alpha levels.

The interval for suture removal was explained by surgery complications and the type of surgery.

Sustained elevated levels of TNF-alpha at four weeks following surgery is indicative of stress and delayed wound healing. This could be attributed to stress-induced increase in cortisol levels in the pre- and immediate postoperative period damping inflammation or infections in the postoperative period.  The results showed that our intervention was effective in reducing postoperative complications. The type of surgery (mastectomy vs breast conservation) did not affect postoperative outcomes in our study which is consistent with earlier findings.

In our study, there was a decrease of TNF-alpha levels following four weeks of surgery in the intervention group indicating that our intervention helped reduce risk factors in the pre- / postoperative period that could have delayed wound healing. However, an earlier study with relaxation training for breast cancer patients undergoing radiotherapy did not show any significant reductions in TNF-alpha levels. This could be because radiation therapy has been shown to induce proinflammatory cytokine levels in earlier studies irrespective of stress levels possibly confounding the effects of stress reduction intervention.

Stress response can have an impact on wound healing as it is regulated by inflammatory mediators. Depression and stress are associated with enhanced production of proinflammatory cytokines.  Results from animal models show that stress stimulates proinflammatory genes, delays wound contraction and myofibroblast differentiation leading to delayed wound closure by > 25% and decreases immune / inflammatory responses required for bacterial clearance leading to infection in the host.  Given the substantial contribution of stress for wound repair and immune changes associated with distress, even small alterations in anxiety could have substantial clinical implications both directly through physiological mechanisms and indirectly through increased pain and decreased compliance.

Our findings are consistent with earlier studies using behavioral and relaxation approaches to improve postoperative outcomes. A variety of hypnotic-relaxation interventions appear to shorten hospital stays, decrease pain and promote faster recovery following surgery.  For example, 241 patients undergoing a stressful medical procedure were randomized to receive perioperative standard care, structured attention or self-hypnotic relaxation. The self-hypnotic relaxation patients showed lower pain and anxiety, lower use of self-administered pain medication, shorter procedure times and less hemodynamic instability than the other two groups.  Others such as relaxation with guided imagery and exercise have demonstrated stress-relieving outcomes closely associated with wound healing.

We propose several mechanisms for action for our yoga intervention. The internal awareness and relaxation associated with these practices are known to alter perceptible thoughts and emotions and reduce reactivity to stressful situations or stimuli thereby altering stress responses and reducing distress.  The effects could be attributed primarily to the reduction in distress in the immediate postoperative period that could have buffered the effects of stress hormones, facilitated recruitment of inflammatory cells at the wound site and reduced the rate of infections and the sustained elevated levels of proinflammatory cytokines at a later period. Secondly, various yogic breathing practices are known to increase oxygen consumption that could hasten wound repair.

 Lastly, our intervention also helped facilitate compliance to treatment such as the use of breathing exercises that are known to reduce risks of pulmonary infections following surgery.  We compared our intervention with standard physiotherapy rehabilitation and supportive therapy that has earlier shown beneficial

effects with respect to postoperative and wound healing outcomes.  This study supports the hypotheses that adding an active stress reduction component (yoga-based relaxation and breathing exercises) to these interventions would hasten the recovery process.

Our study however, has several limitations: i) We assessed only one cytokine TNF-alpha in our study; other cytokines and biological determinants (IL-1, IL-10, transforming growth factor beta (TGF ß ), fibroblast growth factor (FGF), vascular endothelial growth factor (VEGF)) that are known to impact wound healing at various stages and intervals following surgery were not assessed.  ii) We assessed plasma TNF-alpha levels only after four weeks of surgery and not immediately afterwards; an immediate postoperative TNF-alpha level would have helped determine if changes were occurring due to perioperative or postoperative factors. Also, assessing these levels immediately following surgery and after subsequent intervals thereafter, would have provided more valuable information on the dynamics of wound healing. iii) The results of this study are also limited by the fact that there was 29% attrition in the study on follow- up assessments, further affecting the power of our study. iv) Differences in the preoperative nutritional status of the subjects could have confounded the observations in this study. However, it may be pointed out that although a detailed nutritional assessment of each subject was not done, all subjects were evaluated for any clinical signs of malnutrition and hemoglobin levels before surgery. All subjects were moderately built and did not show any clinical signs of malnutrition or abnormal hemoglobin levels. We did not assess serum albumin levels as none of the subjects showed any clinical signs of malnutrition. This measure could have been used to assess differences in nutritional status across subjects.

However, our results support the beneficial effects of yoga intervention on both postsurgery outcomes and wound healing. Further studies on the dynamics of wound repair processes using more relevant and advanced biological determinants of stress and wound repair are warranted. Future studies should evaluate the mechanisms of action of yoga intervention through possible neuroendocrine and immune markers.