Thiazide-induced dysglycemia: it’s time to take notice

Figure 1. Glucose levels measured over time of two patients taking hydrochlorothiazide and hydrochlorothiazide plus trandolapril.

Almost a half century ago, Wolff et al. wrote [1]:

“During the course of studies of hypertensive patients, including a double blind study, observations were made concerning the diabetogenic activity of benzothiadiazine derivatives. This activity was observed to occur in nonobese patients without a family history of diabetes as well as in obese hypertensives with such a family history. These observations suggest that benzothiadiazines should not be given to young or middle-aged hypertensives in whom there is a lengthy life expectancy and for whom alternative hypotensive therapy is feasible.”

Despite these insightful recommendations, benzothiadiazine derivatives (most often hydrochlorothiazide [HCTZ]) are the first-line recommended therapy for hypertension, regardless of glycemic status [2]. Epidemiologic data, as well as data from randomized controlled trials, have associated a new diagnosis of diabetes with hypertension treatment that contains a thiazide diuretic, including chlorthalidone [3,4], HCTZ [5,6] and bendroflumethiazide [7]. There is controversy over the long-term significance of diuretic-induced diabetes [8], primarily related to the benefit of blood pressure lowering versus the hazard of dysglycemia. Recent data from the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) indicate that there is no increased risk for cardiovascular outcomes in those with chlorthalidone-induced hyperglycemia or diabetes [9]. However, this may, in fact, be due to a lack of power to detect a difference in outcomes in this subset of patients, since ALLHAT was not designed a priori for this comparison [10] and less than a quarter of the patients had a fasting glucose measurement during follow-up, with just over 3 years follow-up. In hypertensive patients followed for 15 years, diabetes associated with diuretics is linked to significant cardiovascular risk, as for diabetes from other causes [5].

Although half a century ago, the mechanism of thiazide-induced glucose elevation was not well understood, the inverse relationship between potassium and glucose levels, in which lower potassium is associated with higher glucose, is now well documented [11]. There is also an association between hypokalemia and impaired insulin secretion, which may partially explain this inverse relationship [12] but is unlikely to be the only mechanism. Although it has been suggested that maintenance of potassium homeostasis by supplementation or with concomitant use of an angiotensin-converting enzyme (ACE) inhibitor or angiotensin II receptor blockers may reduce or prevent the dysglycemia associated with thiazide diuretics, this has not been prospectively evaluated [11,13]. We have observed that ACE inhibition is protective against diabetes in hypertensive coronary artery disease patients taking a calcium antagonist but not in those taking a β-blocker plus a thiazide diuretic, even at low or moderate doses [14]. As it is not known if potassium supplementation modifies the increased risk for thiazide-induced dysglycemia or diabetes, the National Heart Lung and Blood Institute recently convened a working group to develop a clinical trial specifically designed to shed light on the association and correlation between glucose-, potassium- and thiazide-induced dysglycemia and diabetes [12].

Prediabetes is characterized by either impaired fasting glucose (fasting plasma glucose: 100–125 mg/dl) or impaired glucose tolerance (2-h plasma glucose: 140–199 mg/dl after a 75-g glucose load) [15]. Prediabetes is a component of the metabolic syndrome, which is a constellation of risk factors including abdominal adiposity, hyperglycemia, hypertension and dyslipidemia [16]. Cardiovascular disease risk in patients with metabolic syndrome is increased two- to fourfold compared with those without metabolic syndrome [17,18]. Internists and cardiologist alike should be aware that the use of agents that negatively impact glucose homeostasis in hypertensive patients, especially in those who also have prediabetes or metabolic syndrome, may hasten diabetes development. This is especially relevant when these patients are relatively young, have a lengthy life expectancy and, perhaps, may be most susceptible to the worsening glucose tolerance often associated with thiazide diuretics. These patients are likely to receive antihypertensive therapy for decades, and the implications of thiazide-induced diabetes may not be fully understood in the relatively short-term clinical trial data that are available to date [19].

Cases in point

Cases from our hypertension research clinic illustrate the impact of metabolically active drugs in metabolically sensitive patients. Patient 1 (a 62-year-old woman) and Patient 2 (a 53-year-old woman) have hypertension and metabolic syndrome according to criteria outlined by the National Cholesterol Education Panel [16]. Following a 75-g glucose load, their fasting glucose, 1-h and 2-h glucose tolerance were assessed at baseline, after 6 weeks of treatment with HCTZ daily and after 6 weeks of HCTZ and trandolapril daily (Figure 1). Potassium and blood pressure values were also assessed at each time point (Table 1).

At baseline, Patient 1 had normal fasting glucose (glucose < 100 mg/dl) while Patient 2 had impaired fasting glucose (glucose 100–125 mg/dl), but neither patient had prediabetes (2-h glucose 140–199 mg/dl) (Figure 1). Although both patients achieved considerable blood pressure reduction during the 12-week period, after just 6 weeks of HCTZ (25 mg daily), Patient 1 had a 2-h glucose diagnostic of prediabetes, while Patient 2 had a 2-h glucose diagnostic of diabetes (glucose > 200 mg/dl). With continuation of HCTZ (25 mg) and addition of the ACE inhibitor trandolapril (4 mg daily) for an additional 6 weeks, the 2-h glucose of Patient 1 returned to normal (glucose < 139 mg/dl) while in Patient 2, 2-h glucose remained diagnostic of diabetes. Importantly, in both of these patients, fasting glucose (0-h value) was not altered from baseline after HCTZ treatment, either alone or in combination with trandolapril, yet there were clear increases in the 2-h glucose values, suggesting that monitoring fasting glucose alone may be insufficient to fully assess the dysglycemic effect of HCTZ. Neither patient had significant diuretic-induced hypokalemia, nor was the observed modest reduction in potassium associated with an increase in fasting glucose.

These two cases demonstrate the heightened susceptibility to the dysmetabolic effects of HCTZ in patients with metabolic syndrome over a short period of treatment. Patient 1 would probably progress to diabetes with continued exposure to HCTZ over the long term [5]. Although these data are limited, they suggest that further study of the impact of thiazide diuretics in patients with metabolic syndrome is warranted and, until more data are available, caution should be exercised if long-term thiazide diuretic use is planned. A 2-h oral glucose tolerance test should be used to routinely monitor the dysglycemic effects, rather than fasting glucose alone, in patients at risk for developing diabetes, including those of Hispanic ethnicity and black race, and those with increased BMI, left ventricular hypertrophy, history of stroke and elevated blood pressure [14,20,21].

The prevalence of prediabetes and metabolic syndrome continues to increase, driven largely by obesity and physical inactivity. The cardiovascular consequences associated with these conditions have recently led to important changes in hypertension treatment guidelines. The European Society of Hypertension/European Society of Cardiology no longer endorse thiazide diuretics in hypertensive patients with diabetes [22] and the American Association of Clinical Endocrinologists recommends thiazide diuretics only at low dosage and only with adequate potassium replacement in patients with diabetes [23]. Additionally, the American College of Endocrinology and the American Association of Clinical Endocrinologists recommend an annual glucose tolerance test in patients with prediabetes and even more careful monitoring in patients with prediabetes and metabolic syndrome. The National Institute for Health and Clinical Excellence, together with the British Hypertension Society, recommend use of renin–angiotensin system inhibitors as first line in younger patients, with diuretics reserved for the elderly or black patients of any age [24]. While thiazide diuretics remain an inexpensive antihypertensive choice, the long-term costs of diabetes that may result far outweigh the short-term medication cost savings.

Table 1. Blood pressure and potassium measurements of two patients at baseline, after taking HCTZ and HCTZ plus trandolapril.

Measurement	Baseline	Week 6 after HCTZ 25 mg daily	Week 12 after HCTZ 25 mg + trandolapril 4 mg daily
Patient 1
BP (mmHg)	146/95	134/81	123/72
Potassium (mEq/dl)	4.0	3.7	3.5
Patient 2
BP (mmHg)	148/97	136/87	125/84
Potassium (mEq/dl)	4.2	3.9	3.9

BP: Blood pressure; HCTZ: Hydrochlorothiazide.

Financial & competing interests disclosure

This work is supported in part by NIH grants 1K23HL086558 and M01RR000082. The author has no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this editorial manuscript.