Metabolic impact of adding a thiazide diuretic to captopril

Thiazide diuretics are among the anti-hypertensive agents that demonstrate effectiveness in improving cardiovascular morbidity and mortality. However, there has been concern regarding adverse metabolic effects of thiazide diuretics on insulin sensitivity in those pre-disposed to diabetes, such as those individuals that are overweight and have hypertension (1,2). In this regard, population-based studies suggest incident or new-onset diabetes increased in groups randomized to thiazide type diuretic treatment (3). Although the exact mechanism for thiazide-induced impairments in glucose metabolism has yet to be determined, groups have traditionally explored alterations in insulin metabolic signaling mechanisms as it relates to 1) reductions in glucose disposal and/or 2) impairments in pancreatic beta cell insulin secretion (Fig 1). In contrast, therapies directed at reducing renin-angiotensin-aldosterone system (RAAS) have been observed to improve glucose metabolism (4).

Figure 1.

Obesity and hypertension predispose individuals to insulin resistance and diabetes through activation of the renin-angiotensin-aldosterone system (RAAS) and systemic and local tissue inflamation and oxidative stress. Use of thiazide-type diuretics enhance activation of the RAAS and impair insulin sensitivity through potential actions on skeletal muscle glucose disposal, pancreatic beta cell insulin secretion, and heaptic glucose production. Recent data would suggest inhibition of the RAAS with ACE inhibitors (ACEi) or angiotensin receptor blockers (ARB) attenuate the negative metabolic impact of thiazide diuretics through direct actions on early phase pancreatic beta cell insulin secretion and hepatic insulin sensitivity.

Diuretic induced natiuresis stimulates the RAAS and it has been suggested that the addition of an ACE inhibitor (ACEi) or an angiotensin receptor blocker (ARB) to a thiazide diuretic would negate the adverse metabolic effects of thiazide diuretics on insulin sensitivity through attenuation of diuretic-induced hypokalemia, oxidative stress, and inflammation (5,6). In this issue of Hypertension, McHenry et al (7) build upon their previous work (8) exploring the addition of an ACEi captopril to bendrofluazide and utilizing hyperinsulinemic, euglycemic clamps to evaluate insulin sensitivity. The authors utilized a prospective randomized, cross-over design with a 6 week wash-in followed by 12 weeks of treatment in each arm with clamp studies to evaluate glucose disposal and endogenous glucose production as a marker of hepatic insulin sensitivity. In the author’s previous report (8), individuals were up-titrated to 100 mg per day of captopril and then randomly assigned to either 5 mg of bendrofluazide or placebo. Clamp studies in a cohort of 15 mildly hypertensive (newly diagnosed or currently on <2 anti-hypertensive agents), overweight (BMI ~29 kg/m²) individuals showed no difference when comparing the addition of captopril to high dose bendrofluazide (e.g. 5 mg) to captopril alone. However, the authors observed the combination increased post-absorptive endogenous glucose production as a marker for hepatic insulin resistance.

In the current study (7), the authors utilize a similar paradigm of mildly hypertensive, overweight (e.g BMI ~29kg/m²) individuals but this time randomize to 1.25 mg of bendrofluazide instead of the 5 mg and as a blood pressure (BP) control doxazosin was used when systolic BP was >160 mmHg and diastolic BP >95 mm Hg. Only 12 out of 23 enrolled patients completed the study protocol but notably, there were no differences in BP, glycemic, or lipid measures between the two treatment groups. Similarly, there were no differences on clamp studies between the two groups with either exogenous glucose infusion (glucose disposal) or endogenous glucose production (hepatic insulin sensitivity). There are similarities between the two reports that adding bendrofluazide, at any dose, to ACEi had a neutral impact on glucose disposal. However, the addition of 5 mg of bendrofluazide compared to 1.25 mg to ACEi increased hepatic insulin resistance (8).

Data from these two mechanistic applications of systemic insulin sensitivity in humans can be viewed in one of three ways; 1) addition of an ACEi does truly negate the adverse metabolic impact of thiazides in the treatment of hypertension, 2) thiazides may not influence glucose disposal at all, or 3) thiazides influence insulin sensitivity through a mechanism other than glucose disposal such as through hepatic or even pancreatic beta cell insulin resistance. Recent data from our group (6) suggests the addition of an ARB to hydrochlorothiazide in obese, hypertensive individuals prevented glycemic excursions following a glucose load. Approximately 400 individuals were randomized to either an ARB/hydrochlorothiazide or amlodipine/hydrochlorothiazide combination and titrated to goal BP with an oral glucose tolerance test (OGTT) administered at baseline and after four months. Those allocated to the ARB/hydrochlorothiazide arm had substantially lower fasting glucose and two hour glucose and post-prandial insulin responses, suggesting the combination improves pancreatic beta cell insulin secretion. Another recent report from a Japanese cohort of insulin resistant, hypertensive individuals also suggests that treatment with ARB improves early phase insulin secretion (9).

While our understanding of the mechanisms of thiazide-induced abnormalities on glucose metabolism are evolving, it is becoming evident that there is an impact on systemic and local tissue inflammatory responses in the liver and pancreas (1,2) (Fig 1). In light of recent data (6), results from the current study would suggest the addition of RAAS inhibition to thiazide type diuretics may not impact glucose disposal but may improve hepatic or pancreatic insulin sensitivity. The reasons for this are unclear. Whether RAAS inhibition acts through direct reductions in angiotensin II and aldosterone’s adverse effects (e.g. inflammation and oxidative stress) on pancreatic beta cells or through preventing reductions in circulating potassium and magnesium, the precise protective mechanism by which RAAS inhibition improves islet insulin secretion remain to be elucidated (10).

In summary, the use of thiazides has been clearly linked to new-onset diabetes in clinical trials and impairments in insulin sensitivity, especially in those at increased risk such as overweight, hypertensive individuals, as seen in the current study (7). While these observational results from a relatively small cohort should be viewed cautiously, the use of mechanistic clamp studies in larger cohorts will be necessary to help elucidate the pancreatic beta cell and insulin sensitive tissue effects of diuretics and/or RAAS blockers. The finding in the current study that the addition of an ACEi to low dose bendrofluazide had no impact on measures of insulin sensitivity, should help guide clinicians in choosing an anti-hypertensive regimen in at-risk populations. This being said, previous work results from the current group (8) and others (1,2) would continue to suggest the use of higher dose thiazide-type diuretics alone or in combination should be avoided in these same at-risk populations.