RESULTS

Table 2 shows that FBG decreased in the yoga group but increased in the controls, the difference between the changes in the two groups being 0.60-(-1.20)-1.80(P<0.05; 95% confidence limits 0.27- 3.33). The corresponding changes in HbAlc were 1.56-(-0.20)=1.76(P<0.05;95% confidence limits 0.20-3.32). Table 3 shows that similar confidence limits for the changes were obtained using the final values for FBG (or HbAl) adjusted by covariate analysis for sex, age and initial FBG (or HbAl)

Since FBG reflects glucose control at the time of blood sampling, and HbAlc reflects average glucose levels over the previous (approximately) six weeks, it follows that glucose levels in the yoga group fell significantly, relative to the controls, both at a single time point, after 12 weeks of yoga, and over the preceding (approximately) six weeks. Three of the patients in the yoga group were able to reduce their dose of tablets, since their blood glucose levels fell markedly following the initiation of yoga therapy. It follows that glucose homeostasis improved in the yoga group compared to the non-intervention group.

Questionnaires, administered by post, indicated that all the patients in the yoga group, except one, wanted to continue attending yoga classes, and the majority felt better, less anxious and more in control of them selves. Two reported improved mobility. No adverse effects were observed.

Table II : Fasting blood glucose (FBG) and glycated haemoglobin (HbAlc) before and after 12 weeks of yoa

Table III : Mean changes in FBG and HbAlc (after minus before 12 weeks of yoga), adjusted by covariate analysis for sex, age and initial FBG. 95% CL = 95% confidence.