Methods We recruited 21 volunteers with NIDDM through the Royal Free Hospital outpatient clinic and the local media, excluding those with end stage liver or kidney disease, or congestive cardiac failure. Thirteen were taking medication for NIDDM, the remainder were on diet control alone. Only one had practiced yoga previously (not recently). Volunteers were allocated to yoga and control groups without stratification by drawing lots. Table 1gives biodata for two groups.
Table 1 Biodata of patients
The control group continued with normal medical treatment only. The yoga group in addition to normal treatment was offered five yoga classes per week at the hospital. Most patients attended one or two classes per week and practiced one or more times per week at home. The 90-minute classes utilized a standard sequence of simple postural, breathing and relaxation exercises, taking appropriate precautions for patients with hypertension or lower back pain6. Fasting blood glucose (FBG) and glycated haemoglobin (HbAlc)7 were assayed before randomisation and after 12 weeks of yoga. Covariate analysis was carried out with BMDP Statistical Software. A simple questionnaire on subjective responses to yoga was administered by post to the patients in the yoga group (100% response).