The effect of thiazide therapy on glucose, insulin and cholesterol metabolism and of glucose on potassium: results of a cross-sectional study in patients from the Hypertension Detection and Follow-up Program.

Fasting and one hour post-glucose load blood samples were obtained from 497 participants in the Hypertension Detection and Follow-up Program (HDFP), 79.8% of whom were on antihypertensive therapy at the time of their five-year examination. Major findings include a positive correlation between glucose/insulin ratio and serum potassium (P = 0.0014) and a weaker negative correlation between fasting insulin and serum potassium (P = 0.004). These data are compatible with a primary effect of hypokalaemia producing insulin 'resistance'. In addition, the glucose load was followed by a mean reduction in serum potassium of 0.135 +/- 0.525 meq/l (P less than 0.001). Twenty percent of participants experienced a drop of more than 0.5 meq/l. Cholesterol was associated with the fasting glucose/insulin ratio (P less than 0.032). The results are compatible with the hypothesis that prevention of hypokalaemia may prevent certain metabolic effects attributed to thiazide.