- Email: [email protected]
Thiazide-induced disturbances in carbohydrate, lipid, and potassium metabolism
Thiazide-induced disturbances in carbohydrate, lipid, and potassium metabolism Long-term thiazide diuretic therapy for hypertension is associated with disturbances in carbohydrate, lipid, and potassium metabolism that theoretically may have serious adverse effects. It appears that diuretic-induced hypokalemia interferes with production of insulin, producing mild elevations of blood glucose in nondiabetic patients. The insulinopenia worsens glucose metabolism in prediabetic and type II diabetic patients. Increases in low-density lipoprotein cholesterol% triglycerides, and the low-density lipoprotein/high-density lipoprotein cholesterol ratio are frequently seen following thiazide treatment of hypertension. These changes are more pronounced in younger patients. Decrements of serum potassium of 0.6 mEq/L are commonly observed with diuretic therapy. Usually, patients remain asymptomatic and no potassium replacement is necessary. In patients with underlying heart disease, however, alterations in potassium metabolism may produce increased frequency and complexity of ventricular ectopic activity. All these metabolic disturbances appear to be, in part, dose related, and there is currently no evidence that they have clinical significance. (AM HEART J 106:245, 1983.)
Eliseo Perez-Stable,
M.D., and Potoula
V. Caralis, M.D. Miami,
The results of two Veterans Administration studies published in 1967 and 19701y2 demonstrated the effectiveness of pharmacotherapy in decreasing morbidity and mortality in hypertensive male patients with diastolic pressure above 105 mm Hg. In 1979 the Hypertension Detection and Follow-up Program (HDFP),3*4 a 5-year cooperative study involving more than 10,000 patients, published its findings that drug therapy produced a 20 % decrease in mortality in patients with diastolic pressure of 90 to 104 mm Hg. This finding is important because more than 70% of the hypertensive population has this level of blood pressure, and about 60% of the deaths attributable to high blood pressure occur in this subgroup of patients. As a result, there seems to be a strong tendency in the United States to administer drug therapy to all patients with diastolic pressure over 90 mm Hge5 Furthermore, pharmacotherapeutic intervention to prevent early hypertension in children at high risk of developing hypertension has recently been suggested.6 Some investigators,7v8 however, have questioned certain methodologic aspects of the HDFP study, particularly the absence of a control group, and have recommended a more conservative approach to drug use in the
From the Medical Medical Center, Reprint Miami
requests: VA Medical
and Research Services, Miami and the Department of Medicine, Eliseo Center,
Perez-Stable, 1201 N.W.
Veterans Administration University of Miami.
M.D., Chief, 16th St., Miami,
Medical Service, FL 33125.
Flu.
management of mild hypertension in adults and a much more cautious approach in children.g The thiazide diuretics have been the backbone of all the long-range studies evaluating antihypertensive pharmacotherapy. In general, diuretic therapy is considered to be well tolerated and its complications rather mild and more than offset by the beneficial effects of lowering elevated blood pressure. However, from the beginning of diuretic therapy in clinical use, it was observed that this treatment could induce metabolic disturbances. This article will review the current evidence on the incidence and mechanisms of these disturbances, as well as the clinical implications of long-term thiazide diuretic therapy in hypertension. The importance of understanding the clinical implications was emphasized recently by Oliver, lo who noted that there is a world of difference between removing a risk factor such as cigarette smoking and adding a new, unknown risk such as prescribing a drug. DISTURBANCES
IN CARBOHYDRATE
METABOLISM
Shortly after the introduction of thiazide diuretic therapy for hypertension,” it was recognized that a disturbance of carbohydrate tolerance could occur as a consequence of its use. This phenomenon has been found in nondiabetic, prediabetic, and diabetic patients.‘2-22 The extensive report by the European Working Party on Hypertension in the Elderly observed that diuretic therapy (EWPHE)23 impaired glucose tolerance within 2 years in elderly 245
246
Perez-Stable
and Caralis
patients, although the impairment may take more than 5 years to develop in younger individuals. Berghmd and Andersonz4 did not observe glucose intolerance during their 6-year follow-up study in middle-aged hypertensive patients. However, the degree of hypokalemia was insignificant in these patients, who were also younger than those in the EWPHE study. Both of these factors may explain the discrepancy. Although clinical diabetes is likely to be induced by thiaxide diuretics only in individuals who already have borderline diabetes, it seems that an impairment of glucose tolerance “across the board” is associated with long-term oral thiaxide therapy. The mechanisms by which thiazides induce a disturbance of carbohydrate metabolism have not been completely elucidated. Mechanisms that have been postulated include decreased insulin secretion by the beta cells of the pancreas,25-27 decreased tissue sensitivity to insulin, increased insulin output*5 accelerating the development of insulin depletion in the prediabetic state, and a deficiency in the enteropancreatic insulin axis. There is evidence both in favor of and against a primary defect in tissue sensitivity to insulin and decreased beta-cell insulin production.28-30 Two causes have been implicated in the impairment of carbohydrate metabolism: (1) the direct effect of the thiaxide molecule at any or all loci in glucose metabolism (mentioned previously) and (2) diuretic-induced potassium depletion. It is well established that potassium depletion is associated with impaired glucose tolerance,31-35 which can be reversed by normalizing the potassium deficiency. With the use of the “glucose clamp” technique in normal subjects with exchange resininduced potassium depletion, Rowe et al.36 were able to demonstrate glucose intolerance that was associated with a decrease in immtmoreactive insulin (IRI) and normal tissue glucose utilization. These findings suggest that the hypokalemia-induced hyperglycemia is the result of an insulinopenia, rather than a deficiency in the insulin receptors. More recently, Helderman et al.37 published an elegant study using the glucose clamp technique in which they demonstrated that thiazide diuretics have no direct effect on any aspect of glucose homeostasis and implied that the carbohydrate disturbance results from diuretic-induced hypokalemia. The defect was found to reside at the level of beta-cell responsiveness to glucose stimulus, as has been shown for other models of hypokalemia.% In two subjects who received total replacement of potassium losses, glucose metabolism remained nor-
American
July, 1083 Heart Journal
mal. The clinical implication is that glucose intolerance can be prevented by rigorous maintenance of normal potassium balance. Hypokalemia appears to interfere with the conversion of proinsulin to insulin at the level of the beta cells. Generally the conversion is largely complete, and in normal subjects only about 5% of the beta-cell secretory product consists of proinsulin. Since proinsulin has only about 10% of the biologic activity of insulin, a situation in which the conversion of proinsulin to insulin is incomplete (resulting in an elevated level of IRI) could be interpreted as a hyperinsulinemic state if the assay used to determine IRI does not differentiate between proinsulin and insulin. This may explain some of the discrepancies in the literature. There are several important clinical implications of the thiazide-induced disturbance of carbohydrate metabolism. Clinical diabetes is likely to be induced only in patients who already have borderline diabetes, In patients whose diabetes is controlled with either diet alone or with diet and oral hypoglycemic agents, diuretic therapy may decrease glucose tolerance and induce clinically significant hyperglycemia. In insulin-dependent diabetic patients (type I) who are producing no effective insulin, these complications should not be observed. The glucose intolerance appears to be the result of the diuretic-induced hypokalemia, indicating that potassium supplementation or the use of a potassium-sparing diuretic, such as spironolactone, triamterene, or amiloride, will usually reverse the abnormality. Finally, the balance between an increased cardiovascular risk because of a small rise in blood glucose levels and the decreased risk brought about by the hypotensive effect of the diuretics remains to be determined. However, it seems that the risk of mild hyperglycemia is more than offset by the control of the hypertension. DISTURBANCES
IN LIPID METABOLISM
As early as 1964, Schoenfield and Goldberger% reported that serum cholesterol levels increased in patients treated with a thiaxide diuretic and decreased when the drug was discontinued. In 1974, Johnson et al.3g suggested that thiaxide diuretics produced an elevation in plasma triglycerides, but it was not until 1976 that Ames and Hill4oT41 reported increments in serum cholesterol of 11 mg/dl and in serum triglycerides of 34 mg/dl in patients treated with chlorthalidone and hydrochlorothiaxide. These investigators also made the important observation that the increases in blood lipids were more pronounced in subjects who had the lowest levels of
Volume Number
106 1, Part 2
cholesterol. Helgeland et al.42 did not observe hypercholesterolemia with the use of hydrochlorothiazide, but they did see a 9% increase in triglycerides. In a cooperative study of the Veterans Administration and the National Heart, Lung and Blood Institute,43 more than 1000 patients were randomly assigned to two treatment groups in a double-blind trial. After 1 year of treatment, the group that received chlorthalidone had increases of 10 mg/dl in total cholesterol levels, 9.8 mg/dl in triglyceride levels, and 12.6 mg/dl in low-density lipoprotein (LDL) cholesterol levels above the changes observed in the placebo group. There was no difference in high-density lipoprotein (HDL) changes between the two groups. The total cholesterol data demonstrated that the magnitude of the increase was inversely related to baseline concentration. Subjects younger than 30 years of age who were treated with chlorthalidone had an average increase of about 18 mg/dl, twice the increase found in older subjects. However, this increase was apparently related to age only because baseline cholesterol level was related to age, with the younger subjects having the lowest baseline concentration and hence the largest increases. In a more recent investigation, Grimm et al.t4 by means of a double-blind trial, confirmed the previous clinical observation that in patients with mild hypertension, chlorthalidone and hydrochlorothiazide produce elevations in blood lipoprotein levels. Compared with baseline, the plasma cholesterol levels increased 6% and 8%) and the triglyceride levels increased 17% and 15% during treatment with hydrochlorothiazide and chlorthalidone, respectively. A cholesterol-lowering diet largely prevents this increase. The daily dose of chlorthalidone was 100 mg. In a smaller number of patients, the use of 50 mg suggested that the degree to which the diuretic affects blood lipid levels may be dose related, although this was not statistically confirmed. Analysis of the data accumulated to date, particularly from the investigations of Goldman et al.43 and Grimm et a1.44consistently indicates that the use of the oral diuretics chlorthalidone and hydrochlorothiazide for the treatment of mild hypertension is associated with increases in the plasma concentrations of total cholesterol, triglycerides, and LDL cholesterol, but not HDL cholesterol. The mechanism of diuretic-induced hyperlipidemia is unknown. The role of the volume contraction initiated by oral diuretics, producing a hemoconcentration of lipoproteins, is probably minimal The degree of hemoconcentration as judged by changes in hematocrit and albumin have been inconsistent.
Thiazide
carbohydrate,
lipid,
and K+ effects
247
In addition, the duration of hyperlipidemia is maintained for at least 16 months, which argues against the role of plasma volume changes, because the plasma volume reduction induced by diuretics tends to decrease with time. Thiazide diuretics could interfere with the production, release, or action of insulin, which increases lipoprotein lipase activity. Triglycerides in lipoprotein of very low density are hydrolyzed by lipoprotein lipase in the capillary endothelial cells. A decrease in lipoprotein lipase activity would elevate lipoprotein levels in a manner consistent with the observed changes. Further research is needed on this and other potential mechanisms. Potassium depletion interferes with the conversion of proinsulin to insulin, creating a biologically insulinopenic state. It would be interesting to determine whether maintaining a normal potassium balance plays a role in the prevention of thiazide-induced disturbances of lipid metabolism. What are the clinical implications of this undesirable effect of diuretics on lipid metabolism? First, it is reassuring to know that the lipid effects of diuretic drugs do not negate the beneficial impact of these drugs on cerebrovascular disease.45s46 This is not surprising, since plasma lipoprotein levels are known to be less important predictors of cerebrovascular accidents than is blood pressure. With respect to coronary heart disease, however, prospective epidemiologic studies and other evidence suggest that such alterations in lipid transport, if uncorrected, might accelerate atherogenesis. Thus any reduction of coronary risk conferred by the hypotensive effect of pharmacotherapy might be counteracted by its effects on lipoprotein metabolism. This is only speculation, but the matter deserves intensive investigation and close attention by the clinician. The evidence is clear that in patients with diastolic pressure greater than 100 mm Hg, the benefits of lowering the pressure overcome any complication produced by hyperlipidemia. The question of whether treatment is beneficial for patients with mildly elevated blood pressure (diastolic of 90 to 99 mm Hg) and for adolescents with hypertension remains controversial. The results of the HDFP, claiming a 20% reduction in mortality, is a strong argument in favor of pharmacotherapy for mild hypertension regardless of the metabolic effects of diuretic drugs. However, this study has been criticized for its methodology, and disagreement on the treatment of mild hypertension remains. The clinician who decides to initiate drug therapy should be aware that dietary fat restriction can help prevent the hyperlipidemic effects of diuretics. In addition, plasma lipoproteins should be measured
248
Perez-Stable
and Caralis
American
several weeks after initiation of therapy to determine whether the patient has an increased level of lipids. The best alternative to thiazides for singledrug therapy of hypertension is a beta blocker; unfortunately, however, these drugs produce a similar hyperlipidemic effect.47 Hydralazine has been shown to have a hypocholesterolemic effect,48 but its use as single-drug therapy is not effective. It is interesting that the diuretic indapamide has incorporated in its structure a moiety similar to hydralazine. Prazosin, clonidine, converting enzyme inhibitors, and methyldopa do not seem to induce an increase in plasma lipids, but none has yet been shown to be as effective as single-drug therapy in long-range studies. DISTURBANCES
IN POTASSIUM
METABOLISM
The long-term administration of thiazide diuretics for the control of hypertension is associated with an increase in the delivery of sodium to the distal tubules and an activation of the renin-angiotensinaldosterone system. The inevitable consequence of these two physiologic events is an increase in the urinary losses of potassium. Usually, the degree of hypokalemia is mild and does not interfere with the continuous action of these drugs. The reduction in serum potassium levels in hypertensive patients receiving thiazide therapy ranges from 0.3 to 1.2 mEq/L. The average decrease is 0.6 mEq/L, with the result that 50% of patients have serum potassium ‘concentrations of less than 3.5 mEq/L and only 7% have levels below 3.0 mEq/L. The average decrease is less with furosemide (0.3 mEq/L), with only 5% of patients having values below 3.5 mEq/L/g-51 Longacting diuretics, such as chlorthalidone, induce a more pronounced degree of hypokalemia. The impact of thiazide therapy on total body potassium levels is relatively mild. Several investigators have shown that the thiazide-induced deficit is around 200 mEq but may be up to 400 mEq with long-term chlorthalidone administration.52p53 An increase in sodium and a decrease in potassium in the diet augment the total body potassium depletion.54 The literature has been somewhat confounded by studies that have used relatively high drug doses for the study of diuretic-induced metabolic abnormalities. There are data on mild hypertension to suggest that therapeutic efficacy need not be compromised by lower diuretic doses but may result in fewer adverse effects.52s55 The clinical significance of these relatively mild changes in potassium metabolism during diuretic therapy of hypertension is controversial. Many practicing physicians prescribe potassium supplementa-
July, 1983 Heart Journal
tion or potassium-sparing agents routinely to hypertensive patients treated with thiazide diuretics. The cost of this prescribing routine to patients has been estimated at more than 250 million dollars per year.56 Kassirer and Harringto@ and others57*58 have pointed out that routine potassium supplementation for patients on diuretic therapy is not only expensive but may also be burdensome, unnecessary, and potentially dangerous, except for patients receiving digitalis concomitantly. In these patients, deficiencies of both potassium and chloride may initiate fatal arrhythmias, and the maintenance of normokalemia should be pursued energetically.5vv60 The arrhythmogenic effect of hypokalemia has also been demonstrated in patients with acute myocardial infarction, during episodes of myocardial ischemia without infarction, and during surgical procedures under general anesthesia.61-64 An increased incidence of ventricular tachycardia and fibrillation in patients with acute myocardial infarction and serum potassium levels less than 3.5 mEq/L has been well documented (40 % vs 20% )?l Hulting65 found a fivefold increase of early ventricular fibrillation in all patients entering a coronary care unit with serum potassium levels less than 3.9 mEq/L, regardless of whether or not a myocardial infarction had occurred. These and similar studies have been criticized because they failed to establish adequately the temporal relationship between the serum potassium level and the observation of ventricular ectopy and to exclude other possible causes of ectopy.57 However, it seems that a mild degree of hypokalemia may favor serious ventricular arrhythmias in the presence of acute myocardial ischemia. Perhaps even more important is the recent publication of three reports claiming diuretic-induced ventricular ectopy in hypertensive patients in an ambulatory setting. Hollifield and Slaton% demonstrated an increased frequency of premature ventricular contractions in hypertensive patients on long-term diuretic therapy both at rest and during dynamic exercise. Holland et al.67 found a similar increased incidence of ventricular arrhythmias in hypertensive patients on long-term diuretic therapy and related the arrhythmias to the reduction of serum potassium concentration. Caralis et al.@ also studied ambulatory hypertensive patients and were able to identify a subgroup of patients who developed increased frequency and severity of ectopic ventricular activity after 4 to 6 weeks of oral diuretic therapy. Ectopic ventricular activity reverted to baseline after serum potassium levels were normalized, in spite of the maintenance of the diuretic. Changes in serum concentrations of potassium, mag-
Volume Number
106 1, Part 2
nesium, and calcium and in intra-red blood cell concentrations of potassium were not correlated statistically with increased ventricular ectopy. However, in patients over 60 years of age with clinical, ECG, or radiologic evidence of organic heart disease, even a mild degree of hypokalemia induced a significant increase in ventricular arrhythmias. It may be that the underlying heart disease renders the myocardium more susceptible to the arrhythmogenic effect of hypokalemia. It is well established that in the setting of acute myocardial infarction a clear association exists between more complex forms of ventricular ectopy and the occurrence of ventricular tachycardia and sudden death.6g-73 The ability to sustain reentrant circuits in areas of ischemic and infarcted muscle has been given as the best explanation for the occurrence of ventricular fibrillation and sudden death.74s75 More recently, it has been documented that ventricular ectopy is related to survival in patients with coronary artery disease outside the postinfarction period. 76*77In these patients the abolition of complex grades of ectopic ventricular activity has been shown to prevent the recurrence of potentially fatal arrhythmias. Potassium may reduce repetitive discharges associated with increased susceptibility or may stabilize the cell membrane, thus reducing the ability of the heart to maintain reentry loops. In patients with coronary heart disease, even a mild degree of potassium depletion might uncover the tendency for some of these patients to have fatal arrhythmias. The report of the Multiple Risk Factor Intervention Trial,78 with an average follow-up of 7 years, is interesting because, of the patients who had abnormal baseline ECGs, the greatest increase in mortality caused by coronary heart disease occurred in hypertensive men who were intensively treated with diuretics (29.2 % vs 17.7 % ). It can only be speculated at this time whether pharmacotherapy was responsible for the increased mortality. The possibility that chronic potassium deficits may contribute to increased mortality by facilitating serious ventricular arrhythmias warrants further investigation. In conclusion, we consider that for the majority of hypertensive patients who do not have clinical, ECG, or radiologic evidence of coronary heart disease, routine potassium salts supplementation or the use of potassium-sparing agents should not be recommended. However, in hypertensive patients who have evidence of coronary heart disease, particularly if ectopic ventricular activity is detected, the prevention of potassium depletion might be indicated
Thiazide carbohydrate,
lipid, and K+ effects
249
during long-term oral diuretic therapy. In these circumstances, beta blockers may be a better option than diuretics to control the hypertension. CONCLUSION
Thiazide diuretics have been the principal agents used in chemotherapy for hypertension for the past 25 years. They have been found to be effective, safe, and well tolerated, and they produce a minimal number of complications. However, there is evidence that long-term use of thiazides may be associated with disturbances in carbohydrate, lipid, and potassium metabolism that theoretically could induce serious adverse effects in the long term. The presence of potassium deficiency seems to be essential for the abnormality in carbohydrate metabolism and is probably also important in producing hyperlipidemia. Despite these theoretic considerations, there is no evidence that these metabolic changes have significant clinical consequences. In patients with mild hypertension, low doses of diuretics may minimize these abnormalities without producing a loss in effectiveness. It seems that clinicians ought to be particularly cautious in using thiazide diuretics in patients with mild hypertension (diastolic pressure less than 100 mm Hg), adolescents with hypertension, and patients with organic heart disease associated with ventricular ectopy. REFERENCES
1.
2.
3.
4.
5.
6.
7. 8.
Veterans Administration Cooperative Study Group on Antihypertensive Agents: Effects of treatment on morbidity in hypertension. I. Results in patients with diastolic blood pressures averaging 115 through 129 mm Hg. JAMA 202:1028, 1967. Veterans Administration Cooperative Study Group on Antihypertensive Agents: Effects of treatment on morbidity in hypertension. II. Results in patients with diastolic blood pressure averaging 90 through 114 mm Hg. JAMA 213:1143, 1970. Hypertension Detection and Follow-up Program Cooperative Group: Five-year findings of the hypertension detection and follow-up program. I. Reduction in mortality of persons with high blood pressure, including mild hypertension. JAMA 242:2562, 1979. Hypertension Detection and Follow-up Program Cooperative Group: Five-year findings of the hypertension detection and follow-up program. II. Mortality by race, sex and age. JAMA 242~2572, 1979. The Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure: The 1980 report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med 140:1280, 1980. Berenson GS, Voors AW, Webber LS, et ah A model of intervention for prevention of early essential hypertension in the 1980s. Hypertension 541, 1983. Freis ED: Should mild hypertension be treated? N Engl J Med 307:306, 1982. Kaplan NM: Whom to treat: The dilemma of mild hypertension. AM HEART J 101:867, 1981.
July,
250
Perez-Stable
and Caralis
9. Jesse MJ: “Early” essential hypertension, prevention, intervention. Hypertension 5:54, 1983. 10. Oliver MF Risks of correcting the risks of coronary disease and stroke with drugs. N Enel J Med 306:297. 1982. 11. Wilkins RW: New drugs for ihe treatment of ‘hypertension. Ann Intern Med 50:1, 1959. 12. Ferguson MJ: Diuretic drugs and diabetes mellitus. Am J Cardiol 7:568, 1961. 13. Hauman RL, Weller JM: Hyperglycemic effect of chlorothiazide. Clin Res 9:180A, 1961. 14. Hollis WC: Aggravation of diabetes mellitus during treatment with chlorothiazide. JAMA 176:947. 1961. 15. Shapiro AP, Benedek TG, Small JE: Effect of thiazides on carbohydrate metabolism in patients with hypertension. N Engl J Med 265:1028, 1961. 16. Zatuchni J, Kordas F: The diabetogenic effect of thiazide diuretics. Am J Cardiol 7:565, 1961. 17. Runyan JW: Influence of thiazide diuretics on carbohydrate metabolism in patients with mild diabetes. N Engl J Med 267:541, 1962. 18. Lewis PH, Kohner EM, Petrie A, et al: Deterioration of glucose tolerance in hypertensive patients on prolonged diuretic treatment. Lancet 1:564, 1976. 19. Weller JN, Borondy PE: Effects of benzothiadiazine drugs on carbohydrate metabolism. Metabolism 14:708, 1965. 20. Wolf FW, Pamley WW, White K, et al: Drug-induced diabetes. JAMA 185:568, 1973. 21. Breckenridge A, Welborn TA, Dollery CT, et al: Glucose tolerance in hypertensive patients on long-term diuretic therapy. Lancet 1:61, 1967. 22. Murphy MB, Kohner E, Lewis PJ, et al: Glucose intolerance in hypertensive patients treated with diuretics: A fourteenyear follow-up. Lancet 2:1293, 1982. 23. Amery A, Berthaux P, Bulpitt C, et al: Glucose intolerance during diuretic therapy: Results of trial by the European Working Party on Hypertension in the Elderly. Lancet 1:681, 1978. 24. Berglund G, Anderson 0: Beta-blockers or diuretics in hypertension? A six-year follow-up of blood pressure and metabolic side effects. Lancet 1:744, 1981. 25. Selzer HS, Allen EW: Inhibition of insulin secretion in diazoxide diabetes. Diabetes 14:439A, 1965. 26. Fajans SS, Floyd JC, Knopf RF, et al: Benzothiadiazine suppression of insulin release from normal and abnormal islet cell-tissue in man. J Clin Invest 45:481, 1966. 27. Hicks BH. Ward JD. Jarrett RJ. et al: A controlled studv of clopamine, clorexolone and hydrochlorothiazide in diabetes. Metabolism 22:101, 1973. 28. Frerichs J, Creutzfeld W: Insulin release from pancreas of the rat, the rabbit and miniature pig in vitro. Diabetologia 1:80A, 1965. 29. Beardwood DM, Alden JS, Graham CA, et al: Evidence for peripheral action of chlorothiazide in normal man. Metabolism 14:561, 1965. 30. Barnett CA, Whitney JE: The effect of diazoxide and chlorothiazide on glucose uptake in vitro. Metabolism 15:88, 1966. 31. Gorden P: Glucose intolerance with hypokalemia. Diabetes 22:544, 1973. 32. Sagild U, Andersen V, Andreasen PB: Glucose tolerance and insulin responsiveness in experimental potassium depletion. Acta Med Stand 169:243, 1961. 33. Gardner LI, Talbot NB, Cook CD, et al: The effect of potassium deficiency on carbohydrate metabolism. J Lab Clin Med 35:592, 1950. 34. Fuhrman FA: Glycogen, glucose tolerance and tissue metabolism in potassium-deficient rats. Am J Physiol 167:314, 1951. 35. Rapoport MI, Hurd HF: Thiazide-induced glucose intolerance treated with potassium. Arch Intern Med 113:405, 1964.
American
Heart
1983
Journal
Rowe JW, Tobin JD, Rosa RM, et al: Effects of experimental potassium deficiency on glucose and insulin metabolism. Metabolism 29:498, 1980. 37. Helderman HJ, Elahi D, Andersen DK, et al: Prevention of the glucose intolerance of thiazide diuretics by maintenance of body potassium. Diabetes 32:106, 1983. 38. Schoenfield MF, Goldberger E: Hypercholesterolemia induced by thiazides: A pilot study. Curr Ther Res 6:180, 1964. 39. Johnson B, Bye C, Labrooy J, et al: The relation of antihypertensive treatment to plasma lipids and other vascular risk factors in hypertensives. Clin Sci 47:9, 1974. 40. Ames RP, Hill P: Increase in serum-lipid concentration during treatment of hypertension with chlorthalidone. Lancet 1:721, 1976. 41. Ames RP, Hill P: Elevation of serum-lipids during diuretic therapy for hypertension. Am J Med 61:748, 1976. 42. Helgeland A, Hjermann I, Holme I, et al: Serum triglycerides and serum uric acid in untreated and thiazide-treated patients with mild hypertension. Am J Med 64:34, 1978. 43. Goldman AI, Steele BW, Schnaper HE, et al: Serum lipoprotein levels during chlorthalidone therapy. __ JAMA 22:1691, 1980. 44. Grimm RH, Leon AS, Hunninghake DB, et al: Effects of thiazide diuretics on ulasma linids and linonroteins in mildlv ” hypertensive patients. Ann Intern Med Q4:7, 1981. 45. Kannell WB, Castelli WP, Gordon T, et al: Serum cholesterol lipoproteins and risk of coronary heart disease: The Framingham study. Ann Intern Med 7&l, 1971. 46. Gordon T, Castelli WP, Hjortland MC, et al: High-density lipoprotein as a protective factor against coronary heart disease: The Framingham study. Am J Med 62:707,-1977. 47. Leren P, Helgeland A. Holme I. et al: Effect of nronranolol and prazosin-on blood lipids. The Oslo study. La&et 2:4, 1980: 48. Perry HM Jr, Mills EJ: The effect of oral hydralazine on circulating human cholesterol. Am J Med Sci 243:564. 1962. 49. Healy JJ, McKenna TJ, Canning B, et al: Body composition changes in hypertensive subjects on long-term oral diuretic therapy. Br Med J 1:716, 1970. 50. Wilinson PR, Hesp R, Issler H, et al: Total body and serum potassium during prolonged thiazide therapy for essential hypertension. Lancet 1:759, 1975. 51. Leemhuis MP, Struyvenberg A: Significance of hypokalemia due to diuretics. Neth J Med 16:18, 1973. 52. Tweeddale MG, Ogilvie RI, Ruedy J: Antihypertensive and biochemical effects of chlorthalidone. Clin Pharmacol Ther 36.
22:519,
1977.
Kassirer JP, Harrington JT: Diuretics and potassium metabolism: A reassessment of the need, effectiveness, and safety of potassium therapy. Kidney Int 11:505, 1977. 54. Venkata C, Ram S, Garrett N, et al: Moderate sodium restriction and various diuretics in the treatment of hypertension: Effects of potassium wastage and blood pressure control. Arch Intern Med 141:1015, 1981. 55. Materson BJ, Oster JR, Michael UF, et al: Dose response to chlorthalidone in patients with mild hypertension. Clin Pharmacol Ther 24:192, 1978. 56. Filmyer JA: Data from the National Disease and Therapeutic Index and National Prescription Audit. Ambler, Pa, 1982, IMS American. 57. Harrington JT, Isner JM, Kassirer JP: Our national obsession with potassium. Am J Med 73:155, 1982. 58. The 1980 Report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med 140:1280, 1980. 59. Lown B, Weller JM, Wyatt N, et al: Effects of alterations of body potassium on digitalis toxicity. J Clin Invest 31:648, 53.
1952.
60. Steiness E, Olesen KH: Cardiac arrhythmias
induced
by
Volume Number
61.
62.
63.
64. 65. 66.
67. 68.
69.
106 1, Part 2
Thiazide carbohydrate,
hypokalemia and potassium loss during maintenance digoxin treatment. Br Heart J 38:168, 1976. Duke M: Thiazide-induced hypokalemia: Association with acute myocardial infarction and ventricular fibrillation. JAMA 239:43, 1978. Solomon RJ, Cole AG: Importance of potassium in patients with acute myocardial infarction. Acta Med Stand 647(suppl):87, 1981. Dyckner T, Helmers C, Lundman T, et al: Initial serum potassium level in relation to early complications and prognosis in patients with acute mvocardial infarction. Acta Med &and-197:207, 1975. Lackey E, Ross DN, Longmore DB, et al: Potassium and open-heart surgery. Lancet 1:651, 1966. Hulting J: In-hospital ventricular fibrillation and its relation to serum potassium. Acta Med &and 647(suppl):109, 1981. Hollifield JW, Slaton PE: Thiazide diuretics, hypokalemia and cardiac arrhythmias. Acta Med Stand 647(suppl):67, 1981. Holland OB, Nixon JV, Kuhnert L: Diuretic-induced ventricular ectopic activity. Am J Med 70:762, 1981. Caralis PV. Materson BJ. Perez-Stable E: Potassium and diuretic-induced ventricular arrhythmias in ambulatory hypertensive patients. Clin Res 293832A, 1981. Lown B: Sudden cardiac death: The major challenge confronting contemporary cardiology. Am J Cardiol 43:313, 1979.
lipid, and K+ effects
70. Lown B: Management of patients at high risk of sudden death. AM HEART J 103:689, 1982. 71. Moss AJ: Clinical significance of ventricular arrhythmias in patients with and without coronary artery disease. Prog Cardiovasc Dis 23:33, 1980. 72. Schulze RA, Strauss HW, Pitt B: Sudden death in the year following myocardial infarction: Relation to ventricular premature contractions in the late hospital phase and left ventricular ejection fraction. Am J Med 62:192, 1977. electrocardiography and the diagno73. Winkle RA: Ambulatory sis, evaluation and treatment of chronic ventricular arrhythmias. Prog Cardiovasc Dis 23:99, 1980. 74. Josephson ME, Soielman SR, Greenspan AM, et al: Mechanism of ventricular fibrillation in man: Observations based on electrode catheter recordings. Am J Cardiol 44:623, 1979. 75. Zipes DP, Heger JJ, Prystowsky EN: Sudden cardiac death. Am J Med 761151, 1981. 76. Ruberman W. Weinblatt E. Goldbere JD. et al: Ventricular premature complexes in prognosis YOf angina. Circulation 61:1172, 1980. 77. Hinkle LE, Carver ST, Argyros DC: The prognostic significance of ventricular premature contractions in healthy people and in people with coronary heart disease. Acta Cardiol 18(suppl):5, 1974. 78. Multiple Risk Factor Intervention Trial: Risk factor changes and mortality results. JAMA 248:1465, 1982.
Some wrong-way chemical changes during antihypertensive treatment: Comparison of indapamide and related agents Despite the beneficial therapeutic effects of antihypertensive drugs, some agents-particularly diureticsseem to go in the “wrong direction” chemically. In fact, these changes could counteract some of the benefits resulting from lowering a patient’s blood pressure. In the absence of hard evidence of the efficacy of long-term diuretic treatment of mild hypertension, we must be maximally sure that such therapy causes no harm. Thiazide and related diuretics have been associated with four distinct wrong-way chemical changes: increases in plasma concentrations of cholesterol, glucose, and uric acid, and a decrease in plasma potassium levels. The potential ramifications of such changes are well understood. The increase in circulating cholesterol, an established risk factor of myocardiai infarction and stroke, is of particular concerneach year approximately one million hypertensive patients have myocardiai infarctions. As a result, the search for safer and more effective diuretics must continue. Indapamide, a new antihypertensive drug, appears to meet these criteria. It is an effective diuretic with a considerable peripheral vasodilatory effect. Additionally, it does not appear to induce any significant change in circulating cholesterol, whereas chlorthalidone has been found to increase total cholesterol by 5%. Hydraiazine is the only antihypertensive agent that seems to lower total cholesterol levels significantly. Neither indapamide nor hydralazine appears to affect plasma glucose levels; benrothiadiazines, however, have been found to induce an increase in circulating glucose. (AM HEART J 106:251, 1983.)
H. Mitchell
Perry, Jr., M.D. St. Louis, MO.
From the Hypertension Division, Washington University School of Meditine, and the St. Louis Veterans Administration Medical Center. Reprint requests: H. Mitchell Perry, M.D., Director, Hypertension Division, Washington University School of Medicine, 915 N. Grand Blvd., Bldg. No.
3, St. Louis,
MO
63106.
Lowering moderately pressure is one of the maneuvers available decrease disease and
or severely elevated blood most beneficial therapeutic today.’ It can significantly delay death. Nonetheless, 251
Eliseo Perez-Stable,
M.D., and Potoula
V. Caralis, M.D. Miami,
The results of two Veterans Administration studies published in 1967 and 19701y2 demonstrated the effectiveness of pharmacotherapy in decreasing morbidity and mortality in hypertensive male patients with diastolic pressure above 105 mm Hg. In 1979 the Hypertension Detection and Follow-up Program (HDFP),3*4 a 5-year cooperative study involving more than 10,000 patients, published its findings that drug therapy produced a 20 % decrease in mortality in patients with diastolic pressure of 90 to 104 mm Hg. This finding is important because more than 70% of the hypertensive population has this level of blood pressure, and about 60% of the deaths attributable to high blood pressure occur in this subgroup of patients. As a result, there seems to be a strong tendency in the United States to administer drug therapy to all patients with diastolic pressure over 90 mm Hge5 Furthermore, pharmacotherapeutic intervention to prevent early hypertension in children at high risk of developing hypertension has recently been suggested.6 Some investigators,7v8 however, have questioned certain methodologic aspects of the HDFP study, particularly the absence of a control group, and have recommended a more conservative approach to drug use in the
From the Medical Medical Center, Reprint Miami
requests: VA Medical
and Research Services, Miami and the Department of Medicine, Eliseo Center,
Perez-Stable, 1201 N.W.
Veterans Administration University of Miami.
M.D., Chief, 16th St., Miami,
Medical Service, FL 33125.
Flu.
management of mild hypertension in adults and a much more cautious approach in children.g The thiazide diuretics have been the backbone of all the long-range studies evaluating antihypertensive pharmacotherapy. In general, diuretic therapy is considered to be well tolerated and its complications rather mild and more than offset by the beneficial effects of lowering elevated blood pressure. However, from the beginning of diuretic therapy in clinical use, it was observed that this treatment could induce metabolic disturbances. This article will review the current evidence on the incidence and mechanisms of these disturbances, as well as the clinical implications of long-term thiazide diuretic therapy in hypertension. The importance of understanding the clinical implications was emphasized recently by Oliver, lo who noted that there is a world of difference between removing a risk factor such as cigarette smoking and adding a new, unknown risk such as prescribing a drug. DISTURBANCES
IN CARBOHYDRATE
METABOLISM
Shortly after the introduction of thiazide diuretic therapy for hypertension,” it was recognized that a disturbance of carbohydrate tolerance could occur as a consequence of its use. This phenomenon has been found in nondiabetic, prediabetic, and diabetic patients.‘2-22 The extensive report by the European Working Party on Hypertension in the Elderly observed that diuretic therapy (EWPHE)23 impaired glucose tolerance within 2 years in elderly 245
246
Perez-Stable
and Caralis
patients, although the impairment may take more than 5 years to develop in younger individuals. Berghmd and Andersonz4 did not observe glucose intolerance during their 6-year follow-up study in middle-aged hypertensive patients. However, the degree of hypokalemia was insignificant in these patients, who were also younger than those in the EWPHE study. Both of these factors may explain the discrepancy. Although clinical diabetes is likely to be induced by thiaxide diuretics only in individuals who already have borderline diabetes, it seems that an impairment of glucose tolerance “across the board” is associated with long-term oral thiaxide therapy. The mechanisms by which thiazides induce a disturbance of carbohydrate metabolism have not been completely elucidated. Mechanisms that have been postulated include decreased insulin secretion by the beta cells of the pancreas,25-27 decreased tissue sensitivity to insulin, increased insulin output*5 accelerating the development of insulin depletion in the prediabetic state, and a deficiency in the enteropancreatic insulin axis. There is evidence both in favor of and against a primary defect in tissue sensitivity to insulin and decreased beta-cell insulin production.28-30 Two causes have been implicated in the impairment of carbohydrate metabolism: (1) the direct effect of the thiaxide molecule at any or all loci in glucose metabolism (mentioned previously) and (2) diuretic-induced potassium depletion. It is well established that potassium depletion is associated with impaired glucose tolerance,31-35 which can be reversed by normalizing the potassium deficiency. With the use of the “glucose clamp” technique in normal subjects with exchange resininduced potassium depletion, Rowe et al.36 were able to demonstrate glucose intolerance that was associated with a decrease in immtmoreactive insulin (IRI) and normal tissue glucose utilization. These findings suggest that the hypokalemia-induced hyperglycemia is the result of an insulinopenia, rather than a deficiency in the insulin receptors. More recently, Helderman et al.37 published an elegant study using the glucose clamp technique in which they demonstrated that thiazide diuretics have no direct effect on any aspect of glucose homeostasis and implied that the carbohydrate disturbance results from diuretic-induced hypokalemia. The defect was found to reside at the level of beta-cell responsiveness to glucose stimulus, as has been shown for other models of hypokalemia.% In two subjects who received total replacement of potassium losses, glucose metabolism remained nor-
American
July, 1083 Heart Journal
mal. The clinical implication is that glucose intolerance can be prevented by rigorous maintenance of normal potassium balance. Hypokalemia appears to interfere with the conversion of proinsulin to insulin at the level of the beta cells. Generally the conversion is largely complete, and in normal subjects only about 5% of the beta-cell secretory product consists of proinsulin. Since proinsulin has only about 10% of the biologic activity of insulin, a situation in which the conversion of proinsulin to insulin is incomplete (resulting in an elevated level of IRI) could be interpreted as a hyperinsulinemic state if the assay used to determine IRI does not differentiate between proinsulin and insulin. This may explain some of the discrepancies in the literature. There are several important clinical implications of the thiazide-induced disturbance of carbohydrate metabolism. Clinical diabetes is likely to be induced only in patients who already have borderline diabetes, In patients whose diabetes is controlled with either diet alone or with diet and oral hypoglycemic agents, diuretic therapy may decrease glucose tolerance and induce clinically significant hyperglycemia. In insulin-dependent diabetic patients (type I) who are producing no effective insulin, these complications should not be observed. The glucose intolerance appears to be the result of the diuretic-induced hypokalemia, indicating that potassium supplementation or the use of a potassium-sparing diuretic, such as spironolactone, triamterene, or amiloride, will usually reverse the abnormality. Finally, the balance between an increased cardiovascular risk because of a small rise in blood glucose levels and the decreased risk brought about by the hypotensive effect of the diuretics remains to be determined. However, it seems that the risk of mild hyperglycemia is more than offset by the control of the hypertension. DISTURBANCES
IN LIPID METABOLISM
As early as 1964, Schoenfield and Goldberger% reported that serum cholesterol levels increased in patients treated with a thiaxide diuretic and decreased when the drug was discontinued. In 1974, Johnson et al.3g suggested that thiaxide diuretics produced an elevation in plasma triglycerides, but it was not until 1976 that Ames and Hill4oT41 reported increments in serum cholesterol of 11 mg/dl and in serum triglycerides of 34 mg/dl in patients treated with chlorthalidone and hydrochlorothiaxide. These investigators also made the important observation that the increases in blood lipids were more pronounced in subjects who had the lowest levels of
Volume Number
106 1, Part 2
cholesterol. Helgeland et al.42 did not observe hypercholesterolemia with the use of hydrochlorothiazide, but they did see a 9% increase in triglycerides. In a cooperative study of the Veterans Administration and the National Heart, Lung and Blood Institute,43 more than 1000 patients were randomly assigned to two treatment groups in a double-blind trial. After 1 year of treatment, the group that received chlorthalidone had increases of 10 mg/dl in total cholesterol levels, 9.8 mg/dl in triglyceride levels, and 12.6 mg/dl in low-density lipoprotein (LDL) cholesterol levels above the changes observed in the placebo group. There was no difference in high-density lipoprotein (HDL) changes between the two groups. The total cholesterol data demonstrated that the magnitude of the increase was inversely related to baseline concentration. Subjects younger than 30 years of age who were treated with chlorthalidone had an average increase of about 18 mg/dl, twice the increase found in older subjects. However, this increase was apparently related to age only because baseline cholesterol level was related to age, with the younger subjects having the lowest baseline concentration and hence the largest increases. In a more recent investigation, Grimm et al.t4 by means of a double-blind trial, confirmed the previous clinical observation that in patients with mild hypertension, chlorthalidone and hydrochlorothiazide produce elevations in blood lipoprotein levels. Compared with baseline, the plasma cholesterol levels increased 6% and 8%) and the triglyceride levels increased 17% and 15% during treatment with hydrochlorothiazide and chlorthalidone, respectively. A cholesterol-lowering diet largely prevents this increase. The daily dose of chlorthalidone was 100 mg. In a smaller number of patients, the use of 50 mg suggested that the degree to which the diuretic affects blood lipid levels may be dose related, although this was not statistically confirmed. Analysis of the data accumulated to date, particularly from the investigations of Goldman et al.43 and Grimm et a1.44consistently indicates that the use of the oral diuretics chlorthalidone and hydrochlorothiazide for the treatment of mild hypertension is associated with increases in the plasma concentrations of total cholesterol, triglycerides, and LDL cholesterol, but not HDL cholesterol. The mechanism of diuretic-induced hyperlipidemia is unknown. The role of the volume contraction initiated by oral diuretics, producing a hemoconcentration of lipoproteins, is probably minimal The degree of hemoconcentration as judged by changes in hematocrit and albumin have been inconsistent.
Thiazide
carbohydrate,
lipid,
and K+ effects
247
In addition, the duration of hyperlipidemia is maintained for at least 16 months, which argues against the role of plasma volume changes, because the plasma volume reduction induced by diuretics tends to decrease with time. Thiazide diuretics could interfere with the production, release, or action of insulin, which increases lipoprotein lipase activity. Triglycerides in lipoprotein of very low density are hydrolyzed by lipoprotein lipase in the capillary endothelial cells. A decrease in lipoprotein lipase activity would elevate lipoprotein levels in a manner consistent with the observed changes. Further research is needed on this and other potential mechanisms. Potassium depletion interferes with the conversion of proinsulin to insulin, creating a biologically insulinopenic state. It would be interesting to determine whether maintaining a normal potassium balance plays a role in the prevention of thiazide-induced disturbances of lipid metabolism. What are the clinical implications of this undesirable effect of diuretics on lipid metabolism? First, it is reassuring to know that the lipid effects of diuretic drugs do not negate the beneficial impact of these drugs on cerebrovascular disease.45s46 This is not surprising, since plasma lipoprotein levels are known to be less important predictors of cerebrovascular accidents than is blood pressure. With respect to coronary heart disease, however, prospective epidemiologic studies and other evidence suggest that such alterations in lipid transport, if uncorrected, might accelerate atherogenesis. Thus any reduction of coronary risk conferred by the hypotensive effect of pharmacotherapy might be counteracted by its effects on lipoprotein metabolism. This is only speculation, but the matter deserves intensive investigation and close attention by the clinician. The evidence is clear that in patients with diastolic pressure greater than 100 mm Hg, the benefits of lowering the pressure overcome any complication produced by hyperlipidemia. The question of whether treatment is beneficial for patients with mildly elevated blood pressure (diastolic of 90 to 99 mm Hg) and for adolescents with hypertension remains controversial. The results of the HDFP, claiming a 20% reduction in mortality, is a strong argument in favor of pharmacotherapy for mild hypertension regardless of the metabolic effects of diuretic drugs. However, this study has been criticized for its methodology, and disagreement on the treatment of mild hypertension remains. The clinician who decides to initiate drug therapy should be aware that dietary fat restriction can help prevent the hyperlipidemic effects of diuretics. In addition, plasma lipoproteins should be measured
248
Perez-Stable
and Caralis
American
several weeks after initiation of therapy to determine whether the patient has an increased level of lipids. The best alternative to thiazides for singledrug therapy of hypertension is a beta blocker; unfortunately, however, these drugs produce a similar hyperlipidemic effect.47 Hydralazine has been shown to have a hypocholesterolemic effect,48 but its use as single-drug therapy is not effective. It is interesting that the diuretic indapamide has incorporated in its structure a moiety similar to hydralazine. Prazosin, clonidine, converting enzyme inhibitors, and methyldopa do not seem to induce an increase in plasma lipids, but none has yet been shown to be as effective as single-drug therapy in long-range studies. DISTURBANCES
IN POTASSIUM
METABOLISM
The long-term administration of thiazide diuretics for the control of hypertension is associated with an increase in the delivery of sodium to the distal tubules and an activation of the renin-angiotensinaldosterone system. The inevitable consequence of these two physiologic events is an increase in the urinary losses of potassium. Usually, the degree of hypokalemia is mild and does not interfere with the continuous action of these drugs. The reduction in serum potassium levels in hypertensive patients receiving thiazide therapy ranges from 0.3 to 1.2 mEq/L. The average decrease is 0.6 mEq/L, with the result that 50% of patients have serum potassium ‘concentrations of less than 3.5 mEq/L and only 7% have levels below 3.0 mEq/L. The average decrease is less with furosemide (0.3 mEq/L), with only 5% of patients having values below 3.5 mEq/L/g-51 Longacting diuretics, such as chlorthalidone, induce a more pronounced degree of hypokalemia. The impact of thiazide therapy on total body potassium levels is relatively mild. Several investigators have shown that the thiazide-induced deficit is around 200 mEq but may be up to 400 mEq with long-term chlorthalidone administration.52p53 An increase in sodium and a decrease in potassium in the diet augment the total body potassium depletion.54 The literature has been somewhat confounded by studies that have used relatively high drug doses for the study of diuretic-induced metabolic abnormalities. There are data on mild hypertension to suggest that therapeutic efficacy need not be compromised by lower diuretic doses but may result in fewer adverse effects.52s55 The clinical significance of these relatively mild changes in potassium metabolism during diuretic therapy of hypertension is controversial. Many practicing physicians prescribe potassium supplementa-
July, 1983 Heart Journal
tion or potassium-sparing agents routinely to hypertensive patients treated with thiazide diuretics. The cost of this prescribing routine to patients has been estimated at more than 250 million dollars per year.56 Kassirer and Harringto@ and others57*58 have pointed out that routine potassium supplementation for patients on diuretic therapy is not only expensive but may also be burdensome, unnecessary, and potentially dangerous, except for patients receiving digitalis concomitantly. In these patients, deficiencies of both potassium and chloride may initiate fatal arrhythmias, and the maintenance of normokalemia should be pursued energetically.5vv60 The arrhythmogenic effect of hypokalemia has also been demonstrated in patients with acute myocardial infarction, during episodes of myocardial ischemia without infarction, and during surgical procedures under general anesthesia.61-64 An increased incidence of ventricular tachycardia and fibrillation in patients with acute myocardial infarction and serum potassium levels less than 3.5 mEq/L has been well documented (40 % vs 20% )?l Hulting65 found a fivefold increase of early ventricular fibrillation in all patients entering a coronary care unit with serum potassium levels less than 3.9 mEq/L, regardless of whether or not a myocardial infarction had occurred. These and similar studies have been criticized because they failed to establish adequately the temporal relationship between the serum potassium level and the observation of ventricular ectopy and to exclude other possible causes of ectopy.57 However, it seems that a mild degree of hypokalemia may favor serious ventricular arrhythmias in the presence of acute myocardial ischemia. Perhaps even more important is the recent publication of three reports claiming diuretic-induced ventricular ectopy in hypertensive patients in an ambulatory setting. Hollifield and Slaton% demonstrated an increased frequency of premature ventricular contractions in hypertensive patients on long-term diuretic therapy both at rest and during dynamic exercise. Holland et al.67 found a similar increased incidence of ventricular arrhythmias in hypertensive patients on long-term diuretic therapy and related the arrhythmias to the reduction of serum potassium concentration. Caralis et al.@ also studied ambulatory hypertensive patients and were able to identify a subgroup of patients who developed increased frequency and severity of ectopic ventricular activity after 4 to 6 weeks of oral diuretic therapy. Ectopic ventricular activity reverted to baseline after serum potassium levels were normalized, in spite of the maintenance of the diuretic. Changes in serum concentrations of potassium, mag-
Volume Number
106 1, Part 2
nesium, and calcium and in intra-red blood cell concentrations of potassium were not correlated statistically with increased ventricular ectopy. However, in patients over 60 years of age with clinical, ECG, or radiologic evidence of organic heart disease, even a mild degree of hypokalemia induced a significant increase in ventricular arrhythmias. It may be that the underlying heart disease renders the myocardium more susceptible to the arrhythmogenic effect of hypokalemia. It is well established that in the setting of acute myocardial infarction a clear association exists between more complex forms of ventricular ectopy and the occurrence of ventricular tachycardia and sudden death.6g-73 The ability to sustain reentrant circuits in areas of ischemic and infarcted muscle has been given as the best explanation for the occurrence of ventricular fibrillation and sudden death.74s75 More recently, it has been documented that ventricular ectopy is related to survival in patients with coronary artery disease outside the postinfarction period. 76*77In these patients the abolition of complex grades of ectopic ventricular activity has been shown to prevent the recurrence of potentially fatal arrhythmias. Potassium may reduce repetitive discharges associated with increased susceptibility or may stabilize the cell membrane, thus reducing the ability of the heart to maintain reentry loops. In patients with coronary heart disease, even a mild degree of potassium depletion might uncover the tendency for some of these patients to have fatal arrhythmias. The report of the Multiple Risk Factor Intervention Trial,78 with an average follow-up of 7 years, is interesting because, of the patients who had abnormal baseline ECGs, the greatest increase in mortality caused by coronary heart disease occurred in hypertensive men who were intensively treated with diuretics (29.2 % vs 17.7 % ). It can only be speculated at this time whether pharmacotherapy was responsible for the increased mortality. The possibility that chronic potassium deficits may contribute to increased mortality by facilitating serious ventricular arrhythmias warrants further investigation. In conclusion, we consider that for the majority of hypertensive patients who do not have clinical, ECG, or radiologic evidence of coronary heart disease, routine potassium salts supplementation or the use of potassium-sparing agents should not be recommended. However, in hypertensive patients who have evidence of coronary heart disease, particularly if ectopic ventricular activity is detected, the prevention of potassium depletion might be indicated
Thiazide carbohydrate,
lipid, and K+ effects
249
during long-term oral diuretic therapy. In these circumstances, beta blockers may be a better option than diuretics to control the hypertension. CONCLUSION
Thiazide diuretics have been the principal agents used in chemotherapy for hypertension for the past 25 years. They have been found to be effective, safe, and well tolerated, and they produce a minimal number of complications. However, there is evidence that long-term use of thiazides may be associated with disturbances in carbohydrate, lipid, and potassium metabolism that theoretically could induce serious adverse effects in the long term. The presence of potassium deficiency seems to be essential for the abnormality in carbohydrate metabolism and is probably also important in producing hyperlipidemia. Despite these theoretic considerations, there is no evidence that these metabolic changes have significant clinical consequences. In patients with mild hypertension, low doses of diuretics may minimize these abnormalities without producing a loss in effectiveness. It seems that clinicians ought to be particularly cautious in using thiazide diuretics in patients with mild hypertension (diastolic pressure less than 100 mm Hg), adolescents with hypertension, and patients with organic heart disease associated with ventricular ectopy. REFERENCES
1.
2.
3.
4.
5.
6.
7. 8.
Veterans Administration Cooperative Study Group on Antihypertensive Agents: Effects of treatment on morbidity in hypertension. I. Results in patients with diastolic blood pressures averaging 115 through 129 mm Hg. JAMA 202:1028, 1967. Veterans Administration Cooperative Study Group on Antihypertensive Agents: Effects of treatment on morbidity in hypertension. II. Results in patients with diastolic blood pressure averaging 90 through 114 mm Hg. JAMA 213:1143, 1970. Hypertension Detection and Follow-up Program Cooperative Group: Five-year findings of the hypertension detection and follow-up program. I. Reduction in mortality of persons with high blood pressure, including mild hypertension. JAMA 242:2562, 1979. Hypertension Detection and Follow-up Program Cooperative Group: Five-year findings of the hypertension detection and follow-up program. II. Mortality by race, sex and age. JAMA 242~2572, 1979. The Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure: The 1980 report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med 140:1280, 1980. Berenson GS, Voors AW, Webber LS, et ah A model of intervention for prevention of early essential hypertension in the 1980s. Hypertension 541, 1983. Freis ED: Should mild hypertension be treated? N Engl J Med 307:306, 1982. Kaplan NM: Whom to treat: The dilemma of mild hypertension. AM HEART J 101:867, 1981.
July,
250
Perez-Stable
and Caralis
9. Jesse MJ: “Early” essential hypertension, prevention, intervention. Hypertension 5:54, 1983. 10. Oliver MF Risks of correcting the risks of coronary disease and stroke with drugs. N Enel J Med 306:297. 1982. 11. Wilkins RW: New drugs for ihe treatment of ‘hypertension. Ann Intern Med 50:1, 1959. 12. Ferguson MJ: Diuretic drugs and diabetes mellitus. Am J Cardiol 7:568, 1961. 13. Hauman RL, Weller JM: Hyperglycemic effect of chlorothiazide. Clin Res 9:180A, 1961. 14. Hollis WC: Aggravation of diabetes mellitus during treatment with chlorothiazide. JAMA 176:947. 1961. 15. Shapiro AP, Benedek TG, Small JE: Effect of thiazides on carbohydrate metabolism in patients with hypertension. N Engl J Med 265:1028, 1961. 16. Zatuchni J, Kordas F: The diabetogenic effect of thiazide diuretics. Am J Cardiol 7:565, 1961. 17. Runyan JW: Influence of thiazide diuretics on carbohydrate metabolism in patients with mild diabetes. N Engl J Med 267:541, 1962. 18. Lewis PH, Kohner EM, Petrie A, et al: Deterioration of glucose tolerance in hypertensive patients on prolonged diuretic treatment. Lancet 1:564, 1976. 19. Weller JN, Borondy PE: Effects of benzothiadiazine drugs on carbohydrate metabolism. Metabolism 14:708, 1965. 20. Wolf FW, Pamley WW, White K, et al: Drug-induced diabetes. JAMA 185:568, 1973. 21. Breckenridge A, Welborn TA, Dollery CT, et al: Glucose tolerance in hypertensive patients on long-term diuretic therapy. Lancet 1:61, 1967. 22. Murphy MB, Kohner E, Lewis PJ, et al: Glucose intolerance in hypertensive patients treated with diuretics: A fourteenyear follow-up. Lancet 2:1293, 1982. 23. Amery A, Berthaux P, Bulpitt C, et al: Glucose intolerance during diuretic therapy: Results of trial by the European Working Party on Hypertension in the Elderly. Lancet 1:681, 1978. 24. Berglund G, Anderson 0: Beta-blockers or diuretics in hypertension? A six-year follow-up of blood pressure and metabolic side effects. Lancet 1:744, 1981. 25. Selzer HS, Allen EW: Inhibition of insulin secretion in diazoxide diabetes. Diabetes 14:439A, 1965. 26. Fajans SS, Floyd JC, Knopf RF, et al: Benzothiadiazine suppression of insulin release from normal and abnormal islet cell-tissue in man. J Clin Invest 45:481, 1966. 27. Hicks BH. Ward JD. Jarrett RJ. et al: A controlled studv of clopamine, clorexolone and hydrochlorothiazide in diabetes. Metabolism 22:101, 1973. 28. Frerichs J, Creutzfeld W: Insulin release from pancreas of the rat, the rabbit and miniature pig in vitro. Diabetologia 1:80A, 1965. 29. Beardwood DM, Alden JS, Graham CA, et al: Evidence for peripheral action of chlorothiazide in normal man. Metabolism 14:561, 1965. 30. Barnett CA, Whitney JE: The effect of diazoxide and chlorothiazide on glucose uptake in vitro. Metabolism 15:88, 1966. 31. Gorden P: Glucose intolerance with hypokalemia. Diabetes 22:544, 1973. 32. Sagild U, Andersen V, Andreasen PB: Glucose tolerance and insulin responsiveness in experimental potassium depletion. Acta Med Stand 169:243, 1961. 33. Gardner LI, Talbot NB, Cook CD, et al: The effect of potassium deficiency on carbohydrate metabolism. J Lab Clin Med 35:592, 1950. 34. Fuhrman FA: Glycogen, glucose tolerance and tissue metabolism in potassium-deficient rats. Am J Physiol 167:314, 1951. 35. Rapoport MI, Hurd HF: Thiazide-induced glucose intolerance treated with potassium. Arch Intern Med 113:405, 1964.
American
Heart
1983
Journal
Rowe JW, Tobin JD, Rosa RM, et al: Effects of experimental potassium deficiency on glucose and insulin metabolism. Metabolism 29:498, 1980. 37. Helderman HJ, Elahi D, Andersen DK, et al: Prevention of the glucose intolerance of thiazide diuretics by maintenance of body potassium. Diabetes 32:106, 1983. 38. Schoenfield MF, Goldberger E: Hypercholesterolemia induced by thiazides: A pilot study. Curr Ther Res 6:180, 1964. 39. Johnson B, Bye C, Labrooy J, et al: The relation of antihypertensive treatment to plasma lipids and other vascular risk factors in hypertensives. Clin Sci 47:9, 1974. 40. Ames RP, Hill P: Increase in serum-lipid concentration during treatment of hypertension with chlorthalidone. Lancet 1:721, 1976. 41. Ames RP, Hill P: Elevation of serum-lipids during diuretic therapy for hypertension. Am J Med 61:748, 1976. 42. Helgeland A, Hjermann I, Holme I, et al: Serum triglycerides and serum uric acid in untreated and thiazide-treated patients with mild hypertension. Am J Med 64:34, 1978. 43. Goldman AI, Steele BW, Schnaper HE, et al: Serum lipoprotein levels during chlorthalidone therapy. __ JAMA 22:1691, 1980. 44. Grimm RH, Leon AS, Hunninghake DB, et al: Effects of thiazide diuretics on ulasma linids and linonroteins in mildlv ” hypertensive patients. Ann Intern Med Q4:7, 1981. 45. Kannell WB, Castelli WP, Gordon T, et al: Serum cholesterol lipoproteins and risk of coronary heart disease: The Framingham study. Ann Intern Med 7&l, 1971. 46. Gordon T, Castelli WP, Hjortland MC, et al: High-density lipoprotein as a protective factor against coronary heart disease: The Framingham study. Am J Med 62:707,-1977. 47. Leren P, Helgeland A. Holme I. et al: Effect of nronranolol and prazosin-on blood lipids. The Oslo study. La&et 2:4, 1980: 48. Perry HM Jr, Mills EJ: The effect of oral hydralazine on circulating human cholesterol. Am J Med Sci 243:564. 1962. 49. Healy JJ, McKenna TJ, Canning B, et al: Body composition changes in hypertensive subjects on long-term oral diuretic therapy. Br Med J 1:716, 1970. 50. Wilinson PR, Hesp R, Issler H, et al: Total body and serum potassium during prolonged thiazide therapy for essential hypertension. Lancet 1:759, 1975. 51. Leemhuis MP, Struyvenberg A: Significance of hypokalemia due to diuretics. Neth J Med 16:18, 1973. 52. Tweeddale MG, Ogilvie RI, Ruedy J: Antihypertensive and biochemical effects of chlorthalidone. Clin Pharmacol Ther 36.
22:519,
1977.
Kassirer JP, Harrington JT: Diuretics and potassium metabolism: A reassessment of the need, effectiveness, and safety of potassium therapy. Kidney Int 11:505, 1977. 54. Venkata C, Ram S, Garrett N, et al: Moderate sodium restriction and various diuretics in the treatment of hypertension: Effects of potassium wastage and blood pressure control. Arch Intern Med 141:1015, 1981. 55. Materson BJ, Oster JR, Michael UF, et al: Dose response to chlorthalidone in patients with mild hypertension. Clin Pharmacol Ther 24:192, 1978. 56. Filmyer JA: Data from the National Disease and Therapeutic Index and National Prescription Audit. Ambler, Pa, 1982, IMS American. 57. Harrington JT, Isner JM, Kassirer JP: Our national obsession with potassium. Am J Med 73:155, 1982. 58. The 1980 Report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med 140:1280, 1980. 59. Lown B, Weller JM, Wyatt N, et al: Effects of alterations of body potassium on digitalis toxicity. J Clin Invest 31:648, 53.
1952.
60. Steiness E, Olesen KH: Cardiac arrhythmias
induced
by
Volume Number
61.
62.
63.
64. 65. 66.
67. 68.
69.
106 1, Part 2
Thiazide carbohydrate,
hypokalemia and potassium loss during maintenance digoxin treatment. Br Heart J 38:168, 1976. Duke M: Thiazide-induced hypokalemia: Association with acute myocardial infarction and ventricular fibrillation. JAMA 239:43, 1978. Solomon RJ, Cole AG: Importance of potassium in patients with acute myocardial infarction. Acta Med Stand 647(suppl):87, 1981. Dyckner T, Helmers C, Lundman T, et al: Initial serum potassium level in relation to early complications and prognosis in patients with acute mvocardial infarction. Acta Med &and-197:207, 1975. Lackey E, Ross DN, Longmore DB, et al: Potassium and open-heart surgery. Lancet 1:651, 1966. Hulting J: In-hospital ventricular fibrillation and its relation to serum potassium. Acta Med &and 647(suppl):109, 1981. Hollifield JW, Slaton PE: Thiazide diuretics, hypokalemia and cardiac arrhythmias. Acta Med Stand 647(suppl):67, 1981. Holland OB, Nixon JV, Kuhnert L: Diuretic-induced ventricular ectopic activity. Am J Med 70:762, 1981. Caralis PV. Materson BJ. Perez-Stable E: Potassium and diuretic-induced ventricular arrhythmias in ambulatory hypertensive patients. Clin Res 293832A, 1981. Lown B: Sudden cardiac death: The major challenge confronting contemporary cardiology. Am J Cardiol 43:313, 1979.
lipid, and K+ effects
70. Lown B: Management of patients at high risk of sudden death. AM HEART J 103:689, 1982. 71. Moss AJ: Clinical significance of ventricular arrhythmias in patients with and without coronary artery disease. Prog Cardiovasc Dis 23:33, 1980. 72. Schulze RA, Strauss HW, Pitt B: Sudden death in the year following myocardial infarction: Relation to ventricular premature contractions in the late hospital phase and left ventricular ejection fraction. Am J Med 62:192, 1977. electrocardiography and the diagno73. Winkle RA: Ambulatory sis, evaluation and treatment of chronic ventricular arrhythmias. Prog Cardiovasc Dis 23:99, 1980. 74. Josephson ME, Soielman SR, Greenspan AM, et al: Mechanism of ventricular fibrillation in man: Observations based on electrode catheter recordings. Am J Cardiol 44:623, 1979. 75. Zipes DP, Heger JJ, Prystowsky EN: Sudden cardiac death. Am J Med 761151, 1981. 76. Ruberman W. Weinblatt E. Goldbere JD. et al: Ventricular premature complexes in prognosis YOf angina. Circulation 61:1172, 1980. 77. Hinkle LE, Carver ST, Argyros DC: The prognostic significance of ventricular premature contractions in healthy people and in people with coronary heart disease. Acta Cardiol 18(suppl):5, 1974. 78. Multiple Risk Factor Intervention Trial: Risk factor changes and mortality results. JAMA 248:1465, 1982.
Some wrong-way chemical changes during antihypertensive treatment: Comparison of indapamide and related agents Despite the beneficial therapeutic effects of antihypertensive drugs, some agents-particularly diureticsseem to go in the “wrong direction” chemically. In fact, these changes could counteract some of the benefits resulting from lowering a patient’s blood pressure. In the absence of hard evidence of the efficacy of long-term diuretic treatment of mild hypertension, we must be maximally sure that such therapy causes no harm. Thiazide and related diuretics have been associated with four distinct wrong-way chemical changes: increases in plasma concentrations of cholesterol, glucose, and uric acid, and a decrease in plasma potassium levels. The potential ramifications of such changes are well understood. The increase in circulating cholesterol, an established risk factor of myocardiai infarction and stroke, is of particular concerneach year approximately one million hypertensive patients have myocardiai infarctions. As a result, the search for safer and more effective diuretics must continue. Indapamide, a new antihypertensive drug, appears to meet these criteria. It is an effective diuretic with a considerable peripheral vasodilatory effect. Additionally, it does not appear to induce any significant change in circulating cholesterol, whereas chlorthalidone has been found to increase total cholesterol by 5%. Hydraiazine is the only antihypertensive agent that seems to lower total cholesterol levels significantly. Neither indapamide nor hydralazine appears to affect plasma glucose levels; benrothiadiazines, however, have been found to induce an increase in circulating glucose. (AM HEART J 106:251, 1983.)
H. Mitchell
Perry, Jr., M.D. St. Louis, MO.
From the Hypertension Division, Washington University School of Meditine, and the St. Louis Veterans Administration Medical Center. Reprint requests: H. Mitchell Perry, M.D., Director, Hypertension Division, Washington University School of Medicine, 915 N. Grand Blvd., Bldg. No.
3, St. Louis,
MO
63106.
Lowering moderately pressure is one of the maneuvers available decrease disease and
or severely elevated blood most beneficial therapeutic today.’ It can significantly delay death. Nonetheless, 251