Advantages of Diuretics EDWARD D. FREIS, M.D. Washington, D. C. Our recent Veterans Administration Cooperative Study-involving 663 hypertensive male patients-showed that a diuretic was more effective than a beta blocker in controlling hypertension in most respects. In addition, in this study we failed to find evidence that thiazide-induced hypokalemia was associated with increased evi- dence of cardiac arrhythmias in patients without overt heart dis- ease. We are concerned that the current desire to avoid hypokale- mia at all costs may result in the prescription of ineffective dosages of diuretics for the treatment of hypertension. When chlorothiazide became available for clinical trials in 1957, we knew that it was a breakthrough drug [I]. It was a medication that seemed to control blood pressure as well as the strictest no-salt diet, but it was much better tolerated. The diuretics were important not only because they were antihypertensive in themselves, but also because they enhanced the an- tihypertensive activity of other drugs. Thiazide diuretics, therefore, soon became the favored step-one agent and have largely retained that posi- tion ever since. From the Veterans Administration Medical Center and the Department of Medicine, Georgetown Uni- versity School of Medicine, Washington, D.C. Re- quests for reprints should be addressed to Dr. Edward D. Freis, VA Medical Center, 50 Irving St. N.W., Washington, DC. 20422. Hemodynamic studies revealed that thiazides caused a modest but definite reduction in plasma volume and extracellular fluid volume; fur- thermore, this reduction seemed to be involved in the antihypertensive effect of the drug [2]. The early fall in blood pressure was associated with a reduced cardiac output and relatively unchanged total peripheral resist- ance [3]. After several weeks, however, cardiac output returned to normal and total peripheral resistance decreased. This late homeostatic adjust- ment does not seem to be due to a direct vasodilator effect of the drug. The mechanism is unknown but may involve poorly understood autoregu- latory reactions. Interestingly, these hemodynamic reactions are the op- posite of the changes that have been described in the development of volume loading types of experimentally induced hypertension. In any event, the hallmark of the antihypertensive action of diuretics is a reduc- tion in volume that lasts for as long as the diuretic is given [4,5]. The two main characteristics of interest regarding any drug are thera- peutic effectiveness and toxicity. Until recently there was no question about the position of the diuretics as the step-one treatment for hyperten- sion. In recent years, however, their premier position has been chal- lenged. There has been increasing concern that thiazide-induced hypo- kalemia may contribute to cardiac arrhythmias and even to sudden death. On the other hand, beta blockers have been somewhat effective in reduo ing the incidence of sudden death in patients who have already sustained a myocardial infarct. Therefore, some physicians-particularly in Scandi- navia-have been using a beta blocker for step-one treatment and add- October 5, 1994 The American Journal of Medicine 107 NEW CONCEPTS IN HYPERTENSION THERAPY SYMPOSIUM-FREIS ing thiazide second. Furthermore, the thiazide is usually added in smaller doses than in the past to minimize the possibility of hypokalemia developing. What is the relative effectiveness of thiazides versus beta blockers as antihypertensive agents? That question was the subject of a recent VA Cooperative Study, which included 683 male patients with diastolic blood pressure averaging 95 to 114 mm Hg [6]. They were randomly as- signed double-blind to receive either propranolol titrated from 40 to 320 mg twice a day or hydrochlorothiazide 25 to 100 mg twice a day. After six months of treatment, hydrochlorothiazide low- ered blood pressure by an average of 17.5113.1 mm Hg as compared with a lowering of 8.3111.3 mm Hg by proprano- loI. White patients responded better than black patients to propranolol and vice versa. Other indications of a some- what greater effectiveness of the thiazide as compared with propranolol over the long term were as follows: (1) with thiazide, a diastolic blood pressure <90 mm Hg was achieved in 66 percent of the patients compared with 53 percent with propranolol; (2) among those taking the diu- retic, fewer patients required termination for high blood pressure; (3) the need to titrate to high doses with hydro- chlorothiazide was less; (4) escape from blood pressure control during treatment occurred less often with hydro- chlorothiazide; and (5) after withdrawal of treatment, the blood pressure remained lower over the ensuing two weeks in those previously taking the diuretic. None of these advantages of diuretics can be consid- ered major; if beta blockers can be proved effective in the primary prevention of heart attack, this would outweigh all of the considerations cited and justify the use of beta blockers as primary treatment in all patients. The same policy would apply if it could be proved that thiazides in- crease the risk of life-threatening arrhythmias. However, at present no definitive evidence exists to justify either of these assumptions. In considering the risks of hypokalemia, it is important to differentiate between patients who have overt heart dis- ease and those who do not. Obvious impairment of car- diac function, particularly congestive heart failure, may be associated with a reduced potassium content in myocar- dial cells unrelated to diuretic treatment. In patients with asymptomatic hypertension without overt heart disease, the concentrations of potassium in their myocardial cells are normal. The following discussion concerning hypoka- lemic risk pertains to patients without overt heart disease. Changes in the extracellular concentration of potassium do not reliably reflect changes in the intracellular concen- tration of potassium. Most of the studies indicate that even during long-term treatment with thiazide, losses of total body potassium remain small, averaging 5 to IO percent in most reports [7,8]. Thus, the percentage reduction of intracellular potassium is considerably less than the re- duction in extracellular potassium. Recent studies from our clinic do not provide any evi- dence that thiazide-induced hypokalemia leads to the development of increased ventricular arrhythmias [9]. Hypertensive patients with no overt heart disease were chosen for study. All had hypokalemia while taking diuret- ics, with plasma potassium concentrations of 3.2 mEq per liter or less (average 2.8 mEq per liter). Holter monitoring for 24 hours was carried out on each patient during the hypokalemia, as well as after it was corrected with potas- sium supplements and/or triamterene. There was no im- provement in ventricular ectopy after correction of hypo- kalemia. Ventricular ectopic activity improved in five patients but worsened in 10. This response is not too sur- prising since the sensitivity of myocardial cells to disturb- ances in rhythm depends upon the relative concentrations of potassium inside and outside the cell. An increase in this ratio produced either by an increase in intracellular potassium concentration or a decrease in extracellular concentration results in a more negative resting mem- brane potential, that is, the cell becomes more resistant to excitation [lo]. This change should reduce rather than in- crease the incidence of ventricular ectopy. An additional concern has been the small but definite increase in serum cholesterol that occurs with diuretic treatment. Possibly, this could increase the risk of athero- sclerosis over the long term. In the VA trial [6], however, it was found that the elevation did not persist, returning to baseline after one year of treatment. Others have found similar returns to baseline over the long term [l 1 ,121. An important consideration in the choice of initial treat- ment is the racial difference in response to diuretics as compared with beta blockers. Although diuretics are defi- nitely more effective in black people, beta blockers may be as effective or more effective than diuretics in white people. In a recent VA trial of the beta blocker nadolol versus the diuretic bendroflumethiazide, nadolol lowered blood pressure to a greater extent in white people than did the diuretic [13]. In the trial of propranolol versus hydro- chlorothiazide, the reduction of diastolic blood pressure was nearly the same with each drug in white people but not in black people, in whom the diuretic was definitely superior [6]. Are small doses of thiazide diuretics equally as effective as large doses? According to some of the recent reports in the literature they are. It is currently popular, for example, to give 25 mg hydrochlorothiazide or even less once daily and not to increase the dose beyond that point since it is believed that the dose-response curve has already pla- teaued. This does not agree with data from our VA Coop- erative Study [6]. In two-thirds of 312 patients, a diastolic blood pressure <90 mm Hg was attained with hydrochlo- rothiazide titrated from 25 mg twice a day to 100 mg twice a day. Of this number, goal diastolic blood pressure was achieved in 50 percent with the 25 mg twice a day dose (50 mg a day), in an additional 30 percent with 50 mg 108 October 5, 1984 The American Journal of Medicine NEW CONCEPTS IN HYPERTENSION THERAPY SYMPOSIUM-FREIS twice a day, whereas in the remaining 20 percent 100 mg twice a day was required. In a recent study by others, doses as low as 6.25 mg a day have been given [14]. It seems possible that no reduction of plasma volume or extracellular volume would result from such a small dose. A double-blind placebo-treated group would be required to be certain that the decrease in blood pressure was due to drug effects rather than to the gradual downward drift of blood pressure that occurs with repeated visits to the clinic. In summary, diuretics lower blood pressure by reducing plasma and extracellular fluid volume. This is a unique mechanism that is shared by no other antihypertensive agents. Our evidence indicates that thiazide diuretics still are the keystone of antihypertensive drug treatment not only for primary treatment, but also in enhancing antihy- pertensive effectiveness when combined with other drugs. Because of racial differences in response, beta blockers have a slight edge over diuretics for initial treatment of white patients, although diuretics are still the preferred pri- mary treatment for black patients. Also, studies from this clinic failed to find evidence that thiazide-induced hypo- kalemia was associated with an increased incidence of cardiac arrhythmias in patients without overt heart dis- ease. It is possible that, because of the current desire to avoid hypokalemia at all costs, doses of diuretics are being reduced in some cases to below the effective thera- peutic level. 1. Freis ED, Wilson IM: Potentiating effect of chlorothiazide (Diuril) in combination with antihypertensive agents. Med Ann DC. 1957; 26: 468-469. 2. Wilson IM, Freis ED: Relationship between plasma and extra- cellular fluid volume depletion and the antihypertensive effect of chlorothiazide. Circulation 1959; 20: 1028-1036. 3. Shah S, Khatri IM, Freis ED: Mechanisms of antihypertensive effect of thiazide diuretics. Am Heart J 1978; 95: 61 l-618. 4. Tarazi HC, et al: Long-term thiazide therapy in essential hyper- tension Circulation 1970; 41: 709-717. 5. Leth A: Changes in plasma and extracellular fluid volume in pa- tients with essential hypertension during long-term treatment with hydrochlorothiazide. Circulation 1970; 42: 479-485. 6. Veterans Administration Cooperative Study Group on Antihy- pertensive Agents: Comparison of propranolol and hydrochlo- rothiazide for the initial treatment of hypertension. JAMA 1982; 248: 1996-2011. 7. Wilkinson RP, Hesp R, lssler H, et al: Total body potassium dur- ing prolonged thiazide therapy for essential hypertension. Lancet 1975; I: 759-762. 8. Edmonds CS, Jasani B: Total body potassium in hypertensive patients during prolonged diuretic therapy. Lancet 1972; II: 8-12. 9. Papademetriou V, Fletcher R, Khatri IM, et al: Diuretic-induced 10. 11. 12. 13. 14. hypokalemia in uncomplicated systemic hypertension. Effect of plasma potassium correction in cardiac arrhythmias. Am J Cardiol 1983; 52: 1017-1022. Fish C: Relation of electrolyte disturbances to cardiac arrhyth- mias. Circulation 1973; 47: 408-419. Alcazar J, Ruilope L, Ladron de Guevara P, et al: Interrelation- ship between uric acid, cholesterol and triglycerides in essen- tial hypertension, abstract no. 8. Proceedings of the Ninth Meeting of the International Society of Hypertension, Feb. 20- 21, 1982, Mexico City, Mexico. Williams WR, Borhani NO, Schnaper HW, Schneider KA, Slat- koff L: The relationship between diuretics and serum choles- terol in HDFP participants. Presented at the American Col- lege of Cardiology, March 20-24, 1983, New Orleans, Louisiana. Veterans Administration Cooperative Study Group on Antihy- pertensive Agents: Efficacv of nadolol alone and combined with bendroflimethiazide and hydralazine for systemic hyper- tension. Am J Cardiol 1983; 52: 1230-1237. Andren L, Weiner L, Suensson A, et al: Enalapril with either a "very low" or "low" dose of hydrochlorothiazide is equally ef- fective in essential hypertension. A double-blind trial in 100 hypertensive patients. J Hypertens 1983; 1 (suppl 2): 384- 386. October 5, 1994 The American Journal of Medicine 109