|Home | About | Journals | Submit | Contact Us | Français|
1. To review the current controversies related to the use of thiazide diuretics as first-line treatment of hypertension
2. To discuss the causal roles for hyperuricemia and hypokalemia on the adverse consequences of thiazide usage
Thiazides significantly reduce morbidity and mortality in hypertensive subjects. However, there remains debate about thiazide usage as first line treatment of hypertension. This negative impact of thiazides may be partially attributed to the ability of thiazides to exacerbate features of metabolic syndrome and/or increase the risk for developing diabetes. Several clinical trials suggest that thiazide-induced hyperuricemia and hypokalemia may account for some of these negative effects. Thiazide treatment is also associated with a decline of renal function in spite of a lowering blood pressure. In this review, we discuss the clinical and experimental evidence supporting a potential role of hyperuricemia and hypokalemia on the development of renal injury and worsening of the metabolic syndrome..
Hyperuricemia and hypokalemia may have pivotal roles in the exacerbation of the metabolic syndrome in response to thiazides. We propose that controlling serum uric acid and serum potassium could improve thiazide efficacy and also reduce its risk for inducing metabolic syndrome or diabetes.
Thiazides, the distal convoluted tubule diuretics, were introduced as antihypertensive drugs in 1957 . Thiazides are often favored as first-line monotherapy or as a part of combined antihypertensive treatment for hypertension. Several prospective randomized trials have documented the benefit of thiazides in reducing cardiovascular morbidity and mortality in subjects with hypertension [2-6]. Additionally, the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC7) recommends thiazide diuretics as the first-line therapy for most patients with uncomplicated hypertension . However, because thiazides can cause a variety of adverse metabolic derangements, some have challenged this recommendation [8••-12]. The purpose of this review is to discuss the current controversies on diuretic use [8••, 13••]. In addition, we discuss recent studies suggesting a role for hyperuricemia and hypokalemia in mediating metabolic abnormalities and renal disease induced by thiazides.
Thiazides primarily inhibit sodium reabsorption at the renal distal convoluted tubule, thereby reducing plasma volume and lowering blood pressure (BP). Chronic thiazide use also reduces peripheral vascular resistance which also contributes to the lowering of BP [14-16]. Hydrochlorothiazide (HCTZ) is the most commonly used thiazide in clinical practice, although others include chlorthalidone, quinethazone, metolazone and bendrofluazide. Chlorthalidone has a longer half-life (40-60 hours) compared to HCTZ (3.2-13.1 hours)  and thus may be more effective at lowering systolic BP at night [18, 19]. Additionally, most clinical trials demonstrating the benefits of thiazides in reduction of cardiovascular end points used chlorthalidone rather than HCTZ. Therefore, the substitution of HCTZ with chlorthalidone in clinical practice has been encouraged [8••, 20-22].
Thiazides are particularly beneficial in controlling hypertension in certain groups, including African Americans, females and the elderly [23, 24]. In addition, other predictors for lowering BP in response to thiazides include shorter duration of hypertension, a higher baseline BP level, a lower baseline plasma renin activity, a lower baseline urinary aldosterone excretion, and a relatively higher urinary sodium excretion in response to HCTZ .
Because of the effects of thiazides to lower urinary calcium excretion, thiazides are also used to prevent recurrence of calcium nephrolithiasis, especially in hypercalciuric subjects . Thiazide treatment is associated with a positive calcium balance with an increase in cortical bone mineral density and therefore has been used to treat osteoporosis and prevent fractures in postmenopausal women [26-28]. Additionally, thiazides directly stimulate osteoblast differentiation and bone mineral formation, further slowing the course of osteoporosis .
The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) was a large randomized double-blind study which included a total of 42,418 hypertensive subjects initially treated with four different drug classes (thiazide; chlorthalidone, calcium channel blocker (CCB); amlodipine, angiotensin- converting enzyme inhibitor (ACEI); lisinopril and α-blocker; doxazosin) and followed for a mean of 4.9 years. ALLHAT demonstrated similar primary outcomes (coronary heart disease mortality or non fatal myocardial infarction) for all four drugs. Nevertheless, when secondary outcomes were analyzed, chlorthalidone-based treatment was superior in the prevention of heart failure to the other three drugs and was also associated with less stroke rates compared with lisinopril and doxazosin . Additionally, a post hoc analysis of ALLHAT showed that amlodipine, lisinopril and chlorthalidone were not different in the rate of development of end stage renal disease (ESRD) or reduced (>50%) GFR . Similarly, a subgroup analysis of hypertensive patients with metabolic syndrome could not demonstrate differences in outcome regardless of whether subjects were initially treated with CCB, α-blocker, or ACEI compared with thiazide-type diuretics, particularly in black participants . These findings have led many authorities to conclude that thiazides should be preferred for the initial treatment of uncomplicated hypertension .
However, some critics have challenged this conclusion. For example, it has been argued that the superior benefit of thiazides to reduce stroke and heart failure might be due to the fact that the chlorthalidone treated group had lower systolic BP (2-3 mmHg) than the lisinopril treated group. In addition, despite lower BP in the chlorthalidone group, the morbidity and mortality of coronary heart disease were similar with those treated with ACEI, suggesting the possibility of a superior effect of the ACEI over thiazides on coronary heart disease [8••]. Using chlorthalidone as one of the initial drugs rather than the add on drug for the ACEI or CCB may have also prevented the optimal use of the ACEI or CCB in hypervolemic subjects needing diuretics [9•]. Because the combination of antihypertensive drugs used in ALLHAT is not common in clinical practice, the interpretation of the results should be cautious [8••]. Finally, the average age of participants at entry in ALLHAT was 67 years; therefore the results of ALLHAT may be not applicable to the younger hypertensive population [8••].
While there is no doubt that thiazides reduce mortality in hypertensive subjects, thiazides also may cause a number of metabolic abnormalities that increase cardiovascular risk, including hyperlipidemia, insulin resistance, new onset diabetes mellitus, hypokalemia, hyperuricemia and stimulation of the renin-angiotensin-aldosterone system (RAAS) [9•, 11, 12, 32]. While these effects can be lessened by using low doses of thiazides, they still occur to an alarming degree [33, 34]. In the ALLHAT study, for example, a significant increase in the development of type 2 diabetes was observed in the thiazide-treated group compared to the other treatment groups (11.6% vs. 9.8% and 8.1%, for 4-year incidence in chlorthalidone, amlodipine and lisinopril groups, respectively; P<0.05). Some have argued that these metabolic effects did not translate into a greater frequency of cardiovascular events, and that thiazides as initial therapy for hypertension offers possibly superior cardiovascular disease outcomes in older hypertensive adults even with the metabolic syndrome, as compared with treatment with CCB and ACEI . However, if one also considers that 1) antihypertensive management is usually life-long; that 2) hypertension and metabolic syndrome are rampant (involving 31% and 27% of the US population, respectively); and that 3) increasing numbers of adolescents are developing obesity and hypertension, then one must be concerned about the long term consequences of thiazide therapy. In fact, it has been also shown that longer treatment with thiazides (up to16 years) is significantly associated with an increased incidence of new onset diabetes mellitus . In this study, thiazide use increased the incidence of cardiovascular events to the same extent as pre-existing diabetes in treated hypertensive subjects .
Additional unfavorable effects of thiazides include lack of improvement in insulin resistance in obese hypertensive patients , the development of atherosclerosis in subjects [37-39], the activation of the RAAS, and an increase in serum plasminogen activator inhibitor type I (PAI-l) activity and homocysteine levels [38-43].
Hypokalemia is a well-known complication of thiazides. While it is more common with higher doses of HCTZ (100 mg/d), hypokalemia is still observed with low doses (12.5 to 25 mg/d). For example, in ALLHAT 12.7% of subjects treated with low dose chlorthalidone developed hypokalemia . In turn, hypokalemia may have several significant cardiovascular consequences. For example, hypokalemia increases the risk for sudden death and cardiovascular events in hypertensive patients [44-46]. Hypokalemia may also counter the beneficial cardiovascular effect of thiazides in elderly patients with isolated systolic hypertension . Additionally, hypokalemia may also reduce the effectiveness of thiazides in lowering BP, and potassium supplementation has been shown to augment the BP-lowering effects of thiazides . Furthermore, the known effect of low dose thiazides to increase the risk for diabetes  may be due in part to hypokalemia. For example, Zillich et al, have evaluated 59 clinical trials which were conducted between 1966 and 2004, and observed that thiazide-induced hypokalamia is commonly associated with increased blood glucose .
Thiazides also induce hyperuricemia, even at low dose . The elevation in mean serum uric acid caused by low dose of diuretics (12.5mg hydrochlorothiazide, or 1.25 mg of bendrofluazide) may occur without any significant changes in serum potassium concentration [51-53], thereby suggesting that alterations in serum uric acid may be more sensitive to the action of diuretics than serum potassium. Recent studies suggest that thiazide induced hyperuricemia can be associated with increased cardiovascular risk. For example, in the WORKSITE study in which 7978 mild-moderate hypertensive subjects were followed on treatment for 20 years, the risk for cardiovascular events was markedly increased in those who had high serum uric acid levels . Similarly, in the Systolic Hypertension in Elderly Program (SHEP) trial, the reduction in coronary artery disease events observed with chlorthalidone was completely abrogated in those subjects whose uric acid increased more than 1 mg/dl during treatment . These studies emphasize the potential importance of hyperuricemia and hypokalemia in increasing the risk for cardiovascular events in subjects on thiazide treatment.
One interesting complication of thiazide usage is the increased risk for developing features of the metabolic syndrome. Thiazides are associated with an increase in serum triglycerides and in fasting glucose and insulin resistance which may be related with potassium depletion [55-58]. Since thiazides are commonly administered to subjects with metabolic syndrome, identifying ways to prevent exacerbation of these metabolic derangements could theoretically improve the efficacy and benefit of diuretic therapy.
Recently it has been shown that endothelial dysfunction itself can cause features of the metabolic syndrome. A rise in blood glucose after eating stimulates insulin which then acts on endothelial cells to release nitric oxide (NO) that then causes systemic vasodilation and facilitates glucose uptake into peripheral muscle and fat . Hence, one third of insulin’s action is thought to act by stimulating the NO-mediated vasodilation. Mice lacking endothelial NO synthase can not produce NO and develop all characteristics of the metabolic syndrome. In this regard, both hypokalemia [60••, 61] and hyperuricemia [62, 63] can cause endothelial dysfunction and low NO levels. This led us to hypothesize that thiazide induced hypokalemia or hyperuricemia might provide the mechanism by which thiazides might cause metabolic syndrome.
To test this hypothesis, we first administered HCTZ to rats with metabolic syndrome induced by a high fructose diet [64••]. As expected, we found that HCTZ treatment worsened hypertriglycemia, insulin resistance, hyperglycemia, hypokalemia and hyperuricemia in spite of lowering BP [64••]. We next examined the effect of lowering serum uric acid (using allopurinol, a xanthine oxidase inhibitor) or of correcting the hypokalemia (with supplemental potassium). Interestingly, both correcting the hyperuricemia with allopurinol and improving the hypokalemia with potassium supplementation could improve the insulin resistance and hypertriglyceridemia along with a further reduction of BP [64••].
HCTZ has been reported to cause endothelial dysfunction in animal and human experiments [65-69]. We confirmed this finding and also showed that thiazide usage reduced urinary nitrite/nitrate excretion. More importantly, we found that correction of hyperuricemia, or of hypokalemia, could largely reverse the endothelial dysfunction as measured by the vasodilatory response of aortic segments in response to acetylcholine [60••, 62]. These studies suggest that correction of hypokalemia and hyperuricemia might be able to improve the blood pressure in response to thiazides and to reduce the risk for metabolic syndrome, cardiovascular complications and diabetes associated with thiazide use. The proposed mechanism by which hypokalemian and hyperuricemia may account for thiazide-associated exacerbation of the metabolic syndrome is shown in Figure 1.
Both clinical and population-based studies have reported that diuretics usage are not renal protective, but rather are associated with acceleration of renal disease in the population. Indeed, the usage of diuretics in the EWPHE , Syst-Eur , SHEP  and INSIGHT  studies were all statistically associated with a greater decline in renal function, mainly assessed by a rise in serum creatinine, compared to the other treatment groups. More recently Hawkins and Houston published a highly controversial paper demonstrating that thiazide usage was epidemiologically associated with the increasing incidence of ESRD in the United States. In particular, they noted a time-lag between the use of diuretics and the future development of ESRD [74, 75]. These data suggest that thiazides might cause renal injury in spite of lowering BP.
Evidence that thiazides can cause renal structural changes has now been shown in several animal models. For example, Loffing et al, have demonstrated that thiazide administration provoked apoptosis in distal tubule cells in the rat, suggesting a direct toxic effect . More recently our group found that chronic HCTZ administration to rats could cause mild renal injury as evidenced by glomerular ischemia, and medullary tubular and interstitial injuries [60••].
The mechanisms responsible for the thiazide associated renal injury remain unclear. The fact that apoptosis has been observed in the distal tubular cells suggests a potential direct toxic effect. However, a more likely mechanism could be due to the chronic effects of thiazides on metabolic abnormalities (hypokalemia, hyperuricemia, hyperglycemia) or volume (volume depletion, stimulation of the RAAS) systems. Hypokalemia, which is one of the main side effects of thiazides, is a likely candidate because it is well known that chronic hypokalemia can cause renal hypertrophy and tubulointerstitial fibrosis, especially in the outer medulla. However, in our study we compared rats with chronic HCTZ-associated hypokalemia and dietary hypokalemia, and only the HCTZ rats displayed renal injury at sacrifice [60••]. It is known that longer periods of dietary hypokalemia will cause renal injury, but this was not observed with the length of this particular study [60••]. In our study, several findings provided suggestive evidence for mechanisms, including volume depletion (as noted by hemoconcentration), increased aldosterone levels (suggesting activation of the RAAS), hypomagnesemia, and chronic hyperglycemia [60••]. Other potential mechanisms could include hyperuricemia, renal ischemia, and oxidative stress [60••].
One important mechanism to consider in HCTZ-associated renal injury is the effects of chronic hyperuricemia. Although HCTZ failed to induce fasting hyperuricemia in our study , there is no doubt that diuretics increase serum uric acid in humans. Recent experimental and clinical studies suggest that chronic elevations in uric acid may cause renal dysfunction. For example, we have reported that raising uric acid levels in rats with a uricase inhibitor can induce glomerulosclerosis, tubulointerstitial and microvascular disease [77, 78]. Raising uric acid levels can also accelerate renal injury in rats with chronic kidney disease [79, 80]. More recently, studies in humans have reported that hyperuricemia predicts the development of kidney disease  and that lowering uric acid can slow renal progression in subjects with chronic kidney disease . Thus, uric acid is also an attractive candidate to explain the observation that thiazides do not improve (and may worsen) renal function despite controlling blood pressure. The proposed mechanism for how hyperuricemia and hypokalemia may cause renal disease is shown in Figure 2.
Thiazides are useful agents for the treatment of hypertension either as monotherapy or in combination with other antihypertensive drugs. Thiazides reduce cardiovascular mortality, heart failure, and stroke in subjects with uncomplicated hypertension. However, thiazides may also carry toxicity by inducing endothelial dysfunction, worsening the metabolic syndrome, increasing the risk for diabetes, and accelerating renal disease. Clinial and experimental studies suggest that thiazide-induced hypokalemia and hyperuricemia may be responsible for some of these effects. Clinical studies are needed to determine if correction of hyperuricemia, and/or hypokalemia, can improve blood pressure control and the metabolic syndrome profile in patients on thiazides.
Dr Johnson and Dr Nakagawa are listed as inventors on several patent applications related to lowering uric acid in hypertension and cardiovascular disease.
• of special interest
•• of outstanding interest