Thiazide-Induced Hypokalemia or Hypomagnesemia and Cardiac Arrhythmias

Many physicians believe that thiazide-induced reduction of serum K levels represents a significant depletion of total body K, which predisposes patients to increased cardiac arrhythmias. The intracellular concentration of K, which is many times greater than the extracellular concentration, is maintained by an Na⁺, K⁺ adenosine phosphatase pump [1]. Diffusion plays no important role. Thus, reduction of extracellular K levels need not significantly influence intracellular K levels. Numerous studies reviewed by Kassirer and Harrington [2] indicated that the reduction in intracellular K concentration after administration of thiazide diuretics is not biologically significant and comprises only 5% of total body K. Loss of body K is not continuous but occurs primarily during the first 2 to 3 days of treatment, after which urinary excretion comes back into balance with intake of K [3]. Therefore, total body K will be initially but not continuously drained.

Another misconception is that a reduction in serum K levels increases the electrical instability of the heart. The ratio of concentrations of various ions, including K⁺, between the extracellular and intracellular spaces influences the electrical excitability of the cell. As expressed by the Nernst equation, a reduction of the ratio of K from outside to inside the cell (which occurs with thiazide use) increases the negativity of the cell's resting membrane potential [1].

Thiazide-induced depletion of Mg has also been proposed as a cause of cardiac arrhythmias. Serum Mg determinations are not reliable indicators of intracellular changes [4]. In addition, thiazides should only slightly affect Mg excretion because they do not act at the loop of Henle, which is the major site of reabsorption of Mg [5]. Furosemide does act at the loop of Henle and can, particularly in large parenteral doses, aggravate hypomagnesemia, especially in patients with congestive heart failure [6]. The investigators who did this study [6] and other investigators have reported that treatment with Mg, particularly in patients receiving furosemide, is beneficial in severe myocardial infarction. In patients with uncomplicated hypertension, however, intracellular Mg or K levels are not significantly reduced during thiazide treatment [7-9]. Other investigators who reported a greater loss of intracellular K and Mg [10, 11] studied patients with congestive heart failure, which is associated with intracellular deficits of both K and Mg, even in the absence of diuretic treatment [12, 13].

Effects of Thiazides on Cardiac Arrhythmias

Two early electrocardiographic monitoring studies found that thiazides increased cardiac arrhythmias. One study was flawed because the monitoring period was too brief [14]. The other study failed to consider the large day-to-day variability in the frequency of arrhythmic events [15]. The investigators selected only patients with baseline ventricular premature beats of less than 6/h, which is at the low end of the daily fluctuation; the number of premature beats was likely to increase in the second or postdrug monitoring because of regression toward the mean.

Better-controlled trials by Papademetriou and colleagues and others [16-22] failed to show any increase in cardiac arrhythmias after thiazide treatment during 24 to 48 hours of electrocardiographic monitoring in hypertensive patients either at rest [16-22] or during exercise [20] or in patients with left ventricular hypertrophy [16]. Normalization of hypokalemia also had no effect on arrhythmic activity [21]. One study [9] found increased arrhythmias only in a few patients with serum K levels less than 3.1. mmol/L, whereas another found no increase in arrhythmias with K levels less than 3.0 mmol/L [22].

These studies were relatively short term and included patients without evidence of ischemic heart disease. In contrast, some of the long-term clinical trials (described below) suggest that diuretic-induced hypokalemia may increase the risk for coronary heart disease events unless it is controlled by low doses of diuretics plus K-sparing diuretics.

Thiazide-Induced Hypokalemia in Acute Myocardial Infarction

Plasma catecholamine levels are elevated in acute myocardial infarction, especially during the early stages. Catecholamines reduce serum K levels and increase the hypokalemia associated with thiazide diuretics [23]. Could this aggravated hypokalemia cause fatal arrhythmias during acute myocardial infarction? During acute infarction, administration of K sufficient to increase the plasma level to normal failed to prevent arrhythmias or reduce mortality [24, 25]. Nevertheless, it seems prudent to correct thiazide-induced hypokalemia in this condition.

Diuretics and Cholesterol

Over the short term, thiazides modestly increase cholesterol levels, with an average increase of less than 0.3 mmol/L [26, 27]. Long-term clinical trials indicate, however, that serum cholesterol levels are elevated only during the first 6 to 12 months of diuretic treatment and then decrease to pretreatment levels [28-33].

Diuretics, Plasma Glucose, and Insulin Resistance

It has long been suspected that thiazides increase plasma glucose levels, at least over the short term. Several clinical trials have provided controlled data on long-term effects. After 5 years of diuretic treatment, the Multifactorial Primary Prevention of Cardiovascular Disease trial [33] found that, in the thiazide group, the glucose level increased by a mean of 0.3 mmol/L over the baseline level 1 hour after a glucose load, whereas the placebo group had a mean increase of 0.7 mmol/L. In the Heart Attack Primary Prevention in Hypertension trial [34], 6000 patients were studied, and the incidence of diabetes after 45 months was 6.9/1000 patient-years in diuretic recipients and 6.1/1000 patient-years in the group receiving ß-blockers. In the Medical Research Council trial, there was no significant long-term change in plasma glucose levels in patients receiving thiazides compared with those receiving placebo [35]. In a 10-year controlled trial, Berglund and colleagues [36] found no evidence that low doses of thiazides increased the incidence of diabetes. Only the Hypertension Detection and Follow-up Program trial reported an increase in diabetes with thiazides compared with other drugs [37]. Therefore, the consensus of the trials indicates that long-term thiazide therapy does not increase the incidence of diabetes.

Hyperinsulinism and insulin resistance are frequently present in hypertension, diabetes, and obesity. Some experimental work has led to the hypothesis that hyperinsulinism and insulin resistance may be a risk factor in coronary heart disease. Although diuretics have been associated with a decrease in insulin sensitivity and an elevation of plasma insulin levels [38], a direct relation of these changes to coronary heart disease has not yet been shown in humans. The current evidence about the importance of hyperinsulinism in aggravating coronary heart disease is inconclusive and requires further investigation.

Effects of Thiazides on Coronary Heart Disease-Related Morbidity and Mortality

In all of the randomized trials on morbidity and mortality in hypertension, diuretics have been used as the primary treatment. ß-blockers were also used in a few of the trials. Newer drugs, such as angiotensin-converting enzyme inhibitors and calcium channel blockers, have not yet been tested in long-term trials. A meta-analysis of 14 randomized trials primarily involving diuretics indicated a 42% reduction in stroke and a 14% decrease in the complications of treatment-related coronary artery disease [39].

ß-blockers have been compared with diuretics in four of the larger trials. Three trials used large doses of diuretics without potassium supplementation [34, 40, 41]. In these trials, ß-blockers were no more effective in reducing coronary heart disease events, including sudden death, than were diuretics. In the remaining trial, a ß-blocker was compared with low-dose thiazide plus a potassium-sparing diuretic [42]. The thiazide regimen was associated with a 44% reduction in coronary heart disease, whereas the ß-blocker was associated with a 5% reduction. No reason is apparent for the poor performance of the ß-blocker in preventing coronary heart disease-related morbidity and mortality.

Why diuretics reduce coronary heart disease events less than they reduce stroke has been a subject of speculation. Several hypotheses have been advanced: 1) Treatment periods have not been long enough to significantly influence the more slowly progressive atherosclerosis of coronary heart disease; 2) hypertension is only one of several other important risk factors in coronary heart disease [for example, a diet high in fat and cholesterol and cigarette smoking], and therefore blood pressure reduction alone may not be sufficient to effectively retard progression; and 3) most important for this discussion, the adverse effects of diuretics on serum cholesterol, glucose, and potassium levels may counteract the beneficial effects of blood pressure reduction.

A meta-analysis of the clinical trials found that long-term treatment with diuretics was associated with an average increase in total cholesterol levels of about 1% [39], which is too small to explain the shortfall in the reduction of coronary heart disease events. Also, as I have previously indicated, long-term treatment with thiazide diuretics did not increase plasma glucose levels or the incidence of diabetes. However, recent evidence suggests that diuretic-induced hypokalemia may be an important factor in the smaller reduction of coronary heart disease compared with that for stroke.

Most of the clinical trials used large doses of diuretics without adjunctive potassium-sparing drugs [33-35, 37-49]. In four of the trials, however, smaller daily doses of hydrochlorothiazide (12.5 and 25 mg) or chlorthalidone (12.5 mg) plus a potassium-sparing diuretic were prescribed [46, 42, 50, 51].

In these four trials, the number of sudden deaths was reduced by 30% [42], 30% [50], and 66% [51], and "total coronary deaths" were reduced by 12% [46]. Except for the last trial, these reductions were considerably greater than those in the trials that used high-dose thiazide therapy by itself.

In addition, a recently published case–control study [52] evaluated the treatment records of 114 hypertensive patients who had primary cardiac arrest from 1977 to 1990. The risk for sudden death was greater in the patients who received high doses of diuretics, considerably lower in those receiving small doses, and lower still in those treated with low-dose thiazides combined with potassium-sparing diuretics.

Although these studies indicate that small doses of diuretics are more effective than large doses in preventing sudden death, the evidence is not entirely convincing. Retrospective case-report studies [52] may be biased. For example, it is possible that the patients who have severe and complicated hypertension, and therefore the highest risk for sudden death, received large doses of diuretics, whereas the patients with mild hypertension received small doses. There are also problems with the prospective, controlled trials that used low-dose diuretics plus potassium-sparing drugs. Although three of four trials reported a considerable reduction in the number of sudden deaths, the remaining trial [46], which had the highest incidence of sudden death, found no significant difference in this event between patients receiving drugs and those receiving placebo. Only elderly patients were included in the four studies, and the results may not be applicable to younger patients. Because diuretics and potassium-sparing drugs were administered together, it is not possible to determine the relative effectiveness of each component separately. Finally, although ß-blockers have antiarrhythmic as well as antihypertensive properties, they were no more effective in reducing the number of sudden deaths than were high-dose diuretics. To settle this question, controlled trials are needed in patients of various ages who are randomly assigned to receive low and high doses of diuretics with and without potassium-sparing drugs.

Recommendations

Top Recommendations Author & Article Info References

It is currently advisable in clinical practice to follow the recommendations of the Joint National Committee on Detection, Evaluations, and Treatment of High Blood Pressure [53], which recommends beginning treatment with a low-dose diuretic. In addition to this recommendation, a potassium-sparing diuretic may also be added for patients who require increased doses or patients who show signs or symptoms of ischemic heart disease. Serum K levels should be determined before and at periodic intervals after treatment with diuretics has been initiated.

The following are the wholesale prices per 100 tablets for several representative antihypertensive drugs [54]. Thiazide diuretics are the least expensive: The price for generic hydrochlorothiazide ranges from $1 to $5, depending on the manufacturer. The combination of 25 mg of hydrochlorothiazide plus 50 mg of triamterene (generic preparations) costs $10 to $15. Generic propranolol in 40-mg tablets is reasonably priced at $5 to $10. On the other hand, generic 50-mg atenolol costs $63 dollars; the cost of diltiazem SR at 60 mg ranges from $37 to $56 depending on the manufacturer. The angiotensin-converting enzyme inhibitor lisinopril, 10 mg, costs $78—many times more than the cost of hydrochlorothiazide.

In selected patients, however, drugs other than diuretics should probably be given as primary therapy. These include ß-blockers in patients with previous myocardial infarction and angiotensin-converting enzyme inhibitors in hypertensive patients with diabetes, especially diabetic nephropathy, and in patients with congestive heart failure. In most patients, however, low-dose diuretics plus a potassium-sparing diuretic when indicated are safe, effective, and far less costly than most other antihypertensive drugs. To paraphrase Mark Twain, it may be said of diuretics that the reports of their demise have been greatly exaggerated.