DISCUSSION

The fasting blood glucose and HbAlc assays show that glucose homeostasis improved in the yoga group, compared to the controls, over the 12-week period of the trial. The three reductions of medication in the yoga group, with no reductions in the control group, strengthen this conclusion.

Our yoga 'package' included several different components, known to help with glucose control in NIDDM, including exercise, diet, relaxation and counseling. The effect of exercise in increasing glucose utilization would have been small, since the yoga employed is very gentle and the average heart rate over a yoga session is about the same as the resting level; exercise therapy for diabetes involves at least half an hour of active exercise with heart rate well above the resting level.

Surwit and Feingloss8 reported that relaxation training leads improved glucose tolerance in NIDDM patients without affecting insulin sensitivity or glucose-stimulated insulin secretary activity. This could be mediated by decreases in sympathetic and adrenal cortical activity. Yoga presumably has similar effects but may also act in other ways, since it includes postural and breathing exercises in addition to simple relaxation.

It might be argued that attention, alone, could have caused the observed effects, since the control group did not have periods of attention, matching those of the yoga classes. It would be difficult to rule out such a possibility, since attention is a complex process, and the quality of attention provided by yoga instructor undoubtedly contributes to the effectiveness of the therapy. To study this scientifically will require considerable sophistication, since attention in a yoga class cannot readily be dissociated from the practices being taught; controls for attention cannot be administered like placebo tablets in drug trials. However, this methodological difficulty does not detract from the evidence that yoga can benefit people with NIDDM. The potentials of yoga as a therapy for NIDDM must stand upon comparison with other therapies (including those based primarily on attention). Before this can critically be accomplished, yoga therapy for NIDDM must be further characterized.

The absence of adverse effect is consistent with observations by one of us (RN), at her yoga therapy clinic, on hundreds of cases of diabetes. It might be thought that yoga could exacerbate the tendency to neuropathic arthropathy in people with diabetes but, contrary to popular misconceptions, therapeutic yoga uses very gentle exercises (with body awareness), and avoids putting any strain on the musculoskeletal system. Indeed, yoga may offer a viable alternative to exercise therapy in cases of arthropathy, and might actually benefit the condition.

Yoga therapy for NIDDM should now be studied to (a) optimize it, (b) determine the extent to which the effects are due to exercise, relaxation, 'attention placebo' or other factors and (c) compare its efficacy and range of applicability to those of other behavioral intervention.

Robin Monro PhD
Joyce Powar B.
Yoga Biomedical Trust
Cambridge
Anil Coumar MBBS.
P Dandona FRCP
Royal Free Hospital
London
R Nagarathna MRCP
(Vivekananda Kendra Yoga Research Foundation)

ACKNOWLEDGEMENT

We thank members of the Vivekananda Foundation for providing training in the yoga methods used.