Antihypertensive Therapy and New On-set of Diabetes Mellitus: The Double-Edge Sword
In this issue of the Journal, Park et al1 prospectively analyzed the incidence of on-set of diabetes mellitus in hypertensive patients receiving either selective or nonselective β-blockers over 60-month time span. This is an important question faced by all health care professionals that attempt to reduce morbidity and mortality in population of patients who suffer from high blood pressure. It is also a complex question. However, the outcomes from the analytics revealed a number of interesting outcomes that require careful thought.
Based on the evidence presented, it seems that nonselective β-blockers significantly increased the incidence of diabetes mellitus, whereas the selective β-blockers did not seem to do so. One of the early indications that treatment of hypertensive patients with β-blockers could lead to diabetes was a report by Bengtsson et al.2 The evidence from the latter investigation indicated that treatment with either β-blockers and/or thiazides resulted in the development of diabetes.2 Further evidence was provided by other reports.3–5 It is interesting that the evidence provided from the analysis by Park et al1 indicates that carvedilol but not metoprolol caused significant increase in the new on-set of diabetes mellitus in hypertensive patients. The data from the Carvedilol Or Metoprolol European Trial (COMET) in patients with chronic heart failure indicate that metoprolol rather than carvedilol was the β-blocker that increased the incidence of diabetes mellitus as assessed over 5 years.6 It is possible that the differential outcome is due to ethnicity of the population, as the patients in COMET were of predominantly of European decent, whereas those in the Park et al1 analysis are Asians. It is also possible that the differences in the outcome are related to the underlying nature of the pathophysiology of the cohort investigated.
One other important outcome from the analysis by Park et al1 was when the data for patients taking selective and nonselective β-blockers were combined, the dilution in analysis seems to suggest β-blockers as a class which do not significantly lead to the new on-set diabetes mellitus in matched hypertensive patients. This emphasizes the importance of stratification of mixed and nonhomogenous data sets. Hence, it would have also been helpful to determine sex differences in the new on-set of diabetes mellitus with this class of antihypertensive in this population. Furthermore, it seems that body mass index (BMI) was not a determining factor in the development of diabetes in patients treated with β-blockers. BMI has been identified as an independent risk factor in the development of diabetes mellitus in hypertensive patients taking antihypertensive drugs including β-blockers.5,7,8 It is not readily clear why BMI was not a factor in this particular analysis,1 but race may have an influence.
The evidence from the analysis by Park et al1 also seems to reveal that the only diuretic that led to the new on-set of diabetes was furosemide, and that the thiazide diuretics did not seem to readily cause the new on-set of diabetes mellitus in this cohort. This is not entirely surprising, as it has been previously suggested that thiazide-induced diabetes mellitus is likely linked to dose, and a lower dose of such a class of drugs would not significantly lead to the new on-set of diabetes.9 It is possible that the database1 contained patients who were predominately being treated with lower doses of the thiazides.
The differential effects of angiotensin converting enzyme (ACE) inhibitors and angiotensin receptor blockers in causing new on-set of diabetes mellitus were also noted in the analysis.