|
INTRODUCTION
Renal sodium and water retention is a hallmark of congestive heart failure (CHF).1,2 Thus, the use of diuretics is a standard component of the therapy for patients with CHF. Since there are approximately 5 million heart failure patients in this country and 500,000 newly diagnosed patients every year, expertise in the use of diuretics in CHF is extremely important.3 This expertise will be an ever-increasing need by physicians as the age of Americans and the prevalence of CHF continues to increase. The present chapter will discuss the pathophysiology of sodium and water retention in CHF, the various diuretics used in CHF including their mechanisms of action and side effects, and the potential effect of diuretics on renal function in patients with CHF.
PATHOPHYSIOLOGY OF SODIUM AND WATER RETENTION IN CONGESTIVE HEART FAILURE
The renal retention of sodium and water in the patient with CHF presents several dilemmas, which have been difficult to understand. Early proposals suggested that a decrease in total blood volume was the initiator of sodium and water retention by the kidney in patients with heart failure. In fact, based on this belief, sodium chloride administration was even recommended for treatment of cardiac failure. However, when the methodology to accurately measure total blood volume in patients with heart failure became available, blood volumes were found to be increased.4 Thus, the undefined and enigmatic term “effective blood volume” was proposed to be decreased in heart failure with resultant triggering of the intact kidney to retain sodium and water. Extrarenal reflexes are incriminated in CHF since the kidney no longer retains sodium and water after a successful heart transplant.
Search for the effective blood volume led to the proposal that a decrease in cardiac output provided the afferent stimulus for renal sodium and water retention in heart failure. However, there were other circumstances in which the normal kidney retains excess amounts of sodium and water in the presence of an increase in cardiac output. Foremost is high-output cardiac failure, which may occur with beriberi, thyrotoxicosis, and cirrhosis, and lead to renal sodium and water retention.
On this background, we proposed that body fluid volume was regulated by the normal kidney in response to the absolute or relative filling of the arterial vascular tree.5,6 There are estimates that total blood volume is distributed 85% on the venous side and 15% on the arterial side of the circulation. In this context, total blood volume could be increased if the expansion was predominantly on the venous side of the circulation, yet the kidney could be retaining sodium and water primarily due to underfilling of the arterial circulation. Based on this, we proposed that the arterial circulation could be underfilled either in an absolute manner by a decrease in cardiac output or relatively underfilled by arterial vasodilation, or a combination thereof.5,6
There are several baroreceptors capable of sensing an underfilling of the arterial circulation, due to either a decrease in cardiac output or arterial vasodilation. The baroreceptors in the carotid sinus and aortic arch respond to stretch. Under normal arterial filling, there is a tonic inhibition of adrenergic outflow from the central nervous system, which is mediated via the vagus and glossopharyngeal nerves from these arterial baroreceptors. There are also receptors in the afferent arterioles of the glomerulus, which respond to stretch as well as b-adrenergic stimulation.7 Thus, patients with severe autonomic insufficiency could still respond to arterial underfilling via the renal receptors. The ventricular receptors have been less well studied, but may be important in elderly cardiac failure patients with diastolic dysfunction but normal left ventricular ejection fractions. In this regard, it has been estimated that as many as 50% of elderly patients with heart failure are due to diastolic, rather than systolic dysfunction. There are also studies showing that impaired baroreceptor sensitivity occurs in patients with cardiac failure, and, therefore, could also be a factor in the increase in sympathetic tone associated with CHF.8 The mortality in CHF correlates directly with plasma norepinephrine, and the sympathetic nervous system (SNS) is a known potent stimulator of the renin-angiotensin-aldosterone system (RAAS).9,10
In addition to the arterial baroreceptors, there are receptors on the low pressure side of the circulation. Specifically, in normal subjects, an increase in atrial pressure is known to (1) decrease arginine vasopressin (AVP) concentrations and cause a solute-free water diuresis, (2) increase atrial natriuretic peptide (ANP) and cause a natriuresis, and (3) decrease renal sympathetic tone.11–13 In contrast, in subjects with CHF the increase in atrial pressure is associated with sodium and water retention and renal vasoconstriction. These observations suggest that the low-pressure receptors are either impaired in cardiac failure patients or are overridden by arterial baroreceptor-mediated increased activity of the SNS and RAAS and the nonosmotic release of AVP.
Our body fluid volume hypothesis is depicted in Fig. 9-1 for arterial underfilling secondary to a decrease in cardiac output and in Fig. 9-2 for arterial vasodilation. While this hypothesis provides a potential sequence of events whereby normal kidneys retain excess sodium and water in association with cardiac failure, several dilemmas exist with respect to the efferent arm of this body fluid volume regulation.14 With renal vasoconstriction, a fall in glomerular filtration rate (GFR) could contribute to the sodium and water retention in cardiac failure, but it is well established that sodium retention occurs in CHF prior to a decrease in GFR. Thus, enhanced tubular reabsorption must be involved in the sodium retention with cardiac failure. The role of increased aldosterone in this increased tubular reabsorption therefore emerged as a potential factor. However, several problems arose. Specifically, all patients with cardiac failure do not have elevated plasma aldosterone concentrations and large exogenous doses of aldosterone to normal individuals do not cause edema. This is because of the so-called aldosterone escape phenomenon from the hormone’s sodium-retaining effect, in which urinary sodium excretion returns to intake levels in spite of continued aldosterone administration. Moreover, patients with CHF do not demonstrate the aldosterone escape phenomenon. Patients with cardiac failure have also been shown to be resistant to the natriuretic response to increasing doses of ANP.15 There is evidence that the effects of increased angiotensin II and adrenergic activity to increase proximal tubule sodium reabsorption in cardiac failure limits sodium delivery to the collecting duct sites of aldosterone and ANP action (Fig. 9-3). We have proposed that this diminished distal sodium delivery is a major factor in the failure of the CHF patient to escape from the sodium-retaining effect of aldosterone and to respond normally to ANP.16
Experimental results indicate that aldosterone not only increases Na-K-ATPase activity as well as the expression and membrane trafficking of the collecting duct epithelial sodium channel (ENaC) but may also increase the NaCl cotransporter (thiazide sensitive cotransporter) in the distal convoluted and connecting tubules. There is also evidence that angiotensin II, independent of aldosterone, is an important factor in ENaC activity. There is substantial evidence that aldosterone increases cardiac fibrosis and angiotensin II as a major factor in cardiac remodeling.17,18 Thus, angiotensin-converting enzyme (ACE) inhibitors and angiotensin receptor blockers (ARBs) as well as non-natriuretic doses of mineralocorticoid antagonists (spironolactone and eplerenone) have been shown to increase survival in cardiac patients (vide infra).
The renal sodium and water retention in patients with heart failure may have substantial deleterious effects in addition to causing pulmonary congestion. The resultant increased cardiac preload is associated with myocardial wall stress, cardiac dilatation, and ventricular hypertrophy, which are factors known to be associated with increased cardiac morbidity and mortality in CHF. Thus, in addition to ACE inhibition, ARBs, and low-dose mineralocorticoid antagonists, diuretics need to be considered in the therapeutic armamentarium in cardiac failure patients. Figure 9-4 shows a deleterious vicious cycle, which may occur in cardiac failure patients and where several proven and unproven interventions, including diuretics, may intervene.19
Use of Loop Diuretics in Congestive Heart Failure
The loop diuretics available in the United States include furosemide, bumetanide, ethacrynic acid, and torsemide.20 The half-lives of these agents are short and range from 1 hour with bumetanide to 3–4 hours for torsemide. After oral administration of loop diuretics, the peak serum concentration occurs within 0.5–2 hours, with furosemide being somewhat slower than bumetanide and torsemide. The predominant effect of loop diuretics is to inhibit the electroneutral Na-K-2Cl cotransporter at the apical surface of the thick ascending limb cells whereby up to 25% of filtered Na and Cl can be excreted.21 The transepithelial voltage of the thick ascending limb is oriented normally with lumen positive and this accounts for absorption of Na, Ca, and Mg via the paracellular pathway.22 In fact, this paracellular pathway accounts for approximately 50% of Na transport by the thick ascending limb. Loop diuretics inhibit both the transcellular and paracellular pathways in the thick ascending limb for Na transport, and increase Ca and Mg excretion primarily by inhibiting the paracellular pathway secondary to abolishing the lumen positive transepithelial voltage.
Furosemide has been shown to stimulate prostaglandin E2 production from thick ascending limb cells.23 Loop diuretics reduce renal vascular resistance and increase renal blood, an effect that depends at least in part on prostaglandins. Inhibition of prostaglandin synthesis with nonsteroidal anti-inflammatory drugs (NSAIDs) has been shown to decrease the diuretic potency of loop diuretics.24 Prostaglandins may also be the primary mediator of the venodilatory effect of loop diuretics.25 This venodilatory effect and the resultant increased splanchnic compliance no doubt accounts for the effect of loop diuretics to decrease cardiac preload and pulmonary edema prior to any effect on urinary electrolyte excretion.26
In patients with chronic CHF, however, loop diuretic administration may acutely increase cardiac afterload, left ventricular end-diastolic pressure, and worsen pulmonary edema.27 This response is probably due to the effect of loop diuretics to stimulate the renin-angiotensin system and secondarily the SNS system. Macula densa cells are in the cortical thick ascending limb and mediate renin secretion and tubuloglomerular feedback (TGF). An increase in NaCl transport across the apical membrane of the macula densa cells via the Na-K-2Cl cotransporter activates TGF and increases renin secretion in association with constriction of the glomerular afferent arteriole and a resultant decrease in GFR. Loop diuretics block this NaCl transport in the macula densa cells and therefore inhibit TGF.28 This may be the reason that GFR is maintained during loop diuretic administration even in the presence of a decrease in extracellular fluid volume (ECFV). The effect of loop diuretics to increase prostaglandin E2 and nitric oxide (neuronal nitric oxide synthase [NOS] is highly expressed in the macula densa) has been proposed to account for adenosine 3’5’ cyclic monophosphate (cyclic AMP)-mediated renin secretion.29 In this regard, inhibition of prostaglandins with NSAIDs has been shown to block the effect of loop diuretics to increase renin secretion.30 In contrast, any effect of nitric oxide to induce renin secretion appears only to be permissive, since loop diuretics still increase renin secretion in experimental knockout models of NOS.31 The effect of loop diuretics to stimulate the RAAS theoretically can have adverse effects on cardiac function (Fig. 9-5).
The potential beneficial effects of loop diuretics in CHF patients are shown in Fig. 9-6. The resultant negative sodium and water balance will not only improve pulmonary congestion, but the associated decrease in ventricular filling pressure can diminish ventricular dilatation. Myocardial function then may improve secondary to a decrease in mitral insufficiency and endomyocardial ischemia as cardiac dilatation is ameliorated. In this setting of improved myocardial function, renal function may actually improve during loop diuretic therapy. Excess diuresis in CHF, however, may further decrease cardiac output and worsen renal function. This particularly occurs in the presence of ACE inhibitor or ARB therapy, which predisposes to a fall in glomerular filtration pressure, and thus decreased GFR, by blocking the vasoconstrictor effect of angiotensin II on the glomerular efferent arteriole.32 Nevertheless, since angiotensin II and aldosterone are known to mediate cardiac remodeling and fibrosis respectively, and ACE inhibitors and ARBs have been shown to decrease mortality in CHF patients, their use is indispensable, even if some decrease in renal function occurs.17,18 Decreasing the loop diuretic dose and lessening the degree of sodium restriction should be considered if blood urea nitrogen (BUN) and serum creatinine concentration are rising during diuretic therapy in a CHF patient.
As discussed above, loop diuretics activate the RAAS by blocking the Na-K-2Cl cotransporter in the macula densa independent of any effect on ECFV. Moreover, this effect on RAAS can be accentuated by ECFV depletion if the diuretic-inducing renal excretion rate exceeds the estimated 12–14 mL/min mobilization of interstitial fluid into the circulation in CHF patients.33 Ultrafiltration has been used in CHF, particularly in the setting of diuretic resistance. It has several potential advantages over loop therapy.34 The RAAS activation in CHF need not be accentuated with ultrafiltration if the rate of fluid removal does not exceed the rate of interstitial fluid mobilization. Also, for the same volume of negative fluid balance, more sodium is removed with ultrafiltration than diuretic-induced fluid loss. This is because the diuresis with loop diuretics is always hypotonic, in contrast to the isotonic fluid removal with ultrafiltration. In this regard, it is the amount of sodium removal that determines the effect on ECFV. Ultrafiltration is invasive and more expensive than diuretic therapy. However, with newer ultrafiltration instrumentation and the use of peripheral access, the potential for diminished hospitalization may make this approach cost-effective.
In addition to overzealous use and activation of the RAAS, there are other potential adverse effects of loop diuretic use. Urinary potassium and magnesium losses can lead to deleterious effects on cardiac function secondary to hypokalemia and hypomagnesemia. In addition to the inhibition of Na-K-2Cl cotransporter, the effect of loop diuretics to increase potassium delivery to the collecting duct site of aldosteronemediated potassium secretion is also involved. Hypokalemia also contributes to loop diuretic-related metabolic alkalosis by stimulating ammonium production.35 Loop diuretics can also cause excess urinary calcium losses, which can potentially decrease plasma-ionized Ca and worsen cardiac function in a patient with CHF. Loop diuretics may cause hyponatremia, however this is less frequent than with distal convoluted tubule (DCT) diuretics.36 This may be because loop, but not DCT, diuretics impair urinary concentration. Ototoxicity, sometimes irreversible, may also occur with loop diuretic therapy. This complication has occurred with the rapid infusion of high doses of a loop diuretic in patients with impaired renal function.37 Thus, it is recommended not to infuse loop diuretics faster than 4 mg/min. With metabolic alkalosis secondary to loop diuretics, the usual method of saline infusion to decrease proximal tubule bicarbonate reabsorption and correct metabolic alkalosis obviously cannot be used in patients with CHF. In this setting, a carbonic anhydrase inhibitor, such as acetazolamide, can be used to increase urinary bicarbonate excretion and correct metabolic alkalosis in CHF when saline infusion is not indicated.38 An increase in urinary pH >7 indicates bicarbonaturia is present. It is important to remember that bicarbonaturia increases urinary potassium loss and thus can worsen hypokalemia if potassium replacement is not instituted.
Thiazide Diuretics
Thiazide diuretics are benzothiadiazide derivatives.39 Other structurally related diuretics are quinazolinones (e.g., metolazone) and benzophenone (e.g., chlorthalidone). These diuretics inhibit Na and Cl transport along the DCT and thus have been termed “DCT diuretics.” Acute administration of these diuretics is associated with inhibition of carbonic anhydrase with increased Na, K, Cl, bicarbonate (HCO3), phosphate, and urate. However, chronically these DCT diuretics primarily increase urinary Na, Cl, and K excretion. In fact, as ECFV decreases with DCT diuretic use, increased uric acid reabsorption may lead to hyperuricemia and precipitate gout in predisposed patients. DCT diuretics also decrease calcium excretion and thus are used to treat nephrolithiasis. Increased urinary Mg excretion and hypomagnesemia may also occur with DCT diuretics.
The molecular site of action of the DCT diuretics has been identified as a transport protein entitled thiazide-sensitive cotransporter (TSC) or sodium chloride cotransporter (NCC).40 This transport protein has been identified in all mammalian species that have been examined.41
Since there is less NaCl reabsorption in the DCT than the thick ascending limb of Henle’s loop, these DCT diuretics are less potent than loop diuretics in patients with CHF. However, the addition of a DCT diuretic to a CHF patient receiving a loop diuretic may dramatically increase NaCl excretion. Similar to loop diuretics, DCT diuretics are highly protein bound in the circulation and thus reach the tubular lumen primarily by the organic anion secretory pathway in the proximal tubule rather than by glomerular filtration.42 In contrast to loop diuretics, the TGF mechanism is not blocked by DCT diuretics since their nephron site of action is beyond the macula densa.43
Therapy with DCT diuretics is associated with more hyponatremia and hypokalemia than loop diuretics.44 With respect to the hyponatremia, the DCT diuretics impair solute-free water excretion but, in contrast to loop diuretics, which impair urinary concentration, DCT diuretics do not impair solute-free water reabsorption. Because of the effect of loop diuretics to impair the countercurrent concentrating mechanism, these agents can lead to the excretion of hypotonic urine even in the presence of antidiuretic hormone (i.e., AVP).45 There is also evidence that furosemide may attenuate the action of AVP at the nephron level.46 Since patients with CHF exhibit a nonosmotic stimulation of AVP and are generally receiving loop diuretics, the addition of a DCT diuretic may increase the occurrence of hyponatremia.47
There are reasons why DCT diuretics may be associated with more hypokalemia than loop diuretics. First of all, DCT diuretics have a longer duration of action than loop diuretics and this is particularly true of chlorthalidone and metolazone.48 Both DCT and loop diuretics increase tubule flow rate in the connecting tubule and collecting duct, the main sites of potassium secretion. Thus, in the presence of hyperaldosteronism associated with CHF, both diuretics will increase urinary potassium excretion. Thus, resultant hypokalemia is a particularly adverse side effect when a DCT diuretic is added to loop diuretic therapy in a patient with cardiac failure. There is also evidence that high luminal concentrations of calcium, which occur with loop diuretics but not with DCT diuretics, inhibit the functional activity of ENaC and thus decrease potassium excretion.49 Hypomagnesemia, which occurs with both DCT and loop diuretics, has also been shown to increase urinary K excretion, and magnesium replacement may diminish K losses.50 Taken together, diuretic-related hypokalemia and hypomagnesemia may predispose to cardiac arrhythmias, particularly in patients with CHF receiving cardiac glycosides, for example, digoxin. Potassium replacement may also obviate the hyperglycemic effect of DCT by normalizing pancreatic insulin release.51 In the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) the effect of the DCT diuretic chlorthalidone on total cholesterol versus placebo was only an increase of 2.2 mg/dL.52
Cortical Collecting and Connecting Tubule Diuretics
Diuretics whose action is localized in the cortical collecting and/or connecting tubule have been referred to as potassium-sparing diuretics.53 Filtered potassium is mostly reabsorbed in the proximal portions of the nephron so that the rate of urinary potassium excretion is primarily determined by potassium secretion in the connecting tubule and cortical collecting duct.54 These are the sites where the potassium-sparing diuretics inhibit potassium secretion. The apical membrane of connecting tubule and principal cells of the collecting tubule express Na and K channels.55 The mechanism for sodium reabsorption in these sites is through conductive ENaC. The basolateral Na-K-ATPase by extruding cellular sodium creates a low intracellular sodium reabsorption, which provides an electrochemical gradient for sodium entry through the sodium channels.56 Thus, sodium entry into the cell creates a lumen-negative transepithelial potential difference, which provides the electrochemical driving force, along with the Na-K-ATPasegenerated high intracellular potassium, for potassium secretion. Amiloride and triamterene decrease potassium secretion by blocking these sodium conductance channels, thereby decreasing the electrochemical gradient for potassium excretion.57 This effect of these two diuretics occurs independent of aldosterone action. The sodium channel is comprised of three homologous subunits (a, b, g-ENaC), which in the mouse are expressed in the apical membrane during a sodium-retaining state.58 In mammals, aldosterone has been shown to increase the abundance of the a-subunit of ENaC and to redistribute all three subunits to the apical region of the principal cells in the collecting duct.59 The mechanism of action of mineralocorticoid receptor blockers, namely spironolactone, is by blocking the nuclear localization of the mineralocorticoid receptors.60 The natriuretic effect of spironolactone is modest, 1–2% of filtered sodium, and is dependent on the presence of aldosterone.61 Spironolactone is, therefore, ineffective in adrenalectomized animals and patients with Addison’s disease. The peak response of spironolactone may occur as late as 48 hours and the effect wanes over a period of 48–72 hours after stopping the diuretic.62 Eplerenone is another competitive aldosterone antagonist, which appears to be 50–70% as potent as spironolactone, but does not have the estrogenic side effects of spironolactone (e.g., gynecomastia, breast tenderness, menstrual irregularities, decreased libido, and impotence).63
The use of the potassium-sparing diuretics as an adjunct to loop diuretics, so as to prevent hypokalemia, is quite effective. These potassium-sparing diuretics should, however, not be used in conjunction with potassium supplements, since this agent’s main side effect is hyperkalemia. The hyperkalemia of mineralocorticoid receptor antagonists is most likely to occur in elderly patients, patients with diabetes, and patients with reduced kidney function, all of which can occur in association with CHF.64 Since CHF patients are receiving ACE inhibitors, ARBs, and b-blockers, all of which have been shown to improve survival, but also predispose to potassium retention, the use of potassium-sparing diuretics must be monitored carefully in this setting. The potassium-losing effect of loop diuretics may, however, attenuate this potential adverse side effect of hyperkalemia during use of potassium-sparing diuretics in the CHF patients. By blocking hydrogen ion secretion, as well as potassium secretion, amiloride, triamterene, and mineralocorticoid, antagonists may predispose to, or worsen, hyperchloremic metabolic acidosis.65
Spironolactone has been shown to improve mortality in CHF (Randomized Aldactone Evaluation Study [RALES]), and eplerenone has been shown to improve mortality in patients with left ventricular dysfunction following myocardial infarction.66,67 These effects are observed at non-natriuretic doses of these mineralocorticoid receptor antagonists; however, these agents may also be useful in treating sodium retention in CHF, albeit with higher doses.68 Diuretic resistance to loop diuretics involves secondary hyperaldosteronism and upregulation of NCC and ENaC effects, which can be reversed by mineralocorticoid antagonists.69 In fact, in cirrhosis, another edematous disorder with secondary hyperaldosteronism, spironolactone has been shown to be the diuretic of choice.70
Agents that increase solute-free water excretion, so-called aquaretics, are under study by several pharmaceutical companies in phase II and III of clinical trials.71 These agents cause a water diuresis by blocking the V2 vasopressin receptor on the basolateral membrane of the collecting duct and have been shown to correct hyponatremia in cardiac failure patients.72,73 There is little effect with these agents on the hormonal status of CHF patients, since two-thirds of the solute-free water excretion is removed from the intracellular space and only one-third from the extracellular fluid (ECF) compartment. There are also AVP antagonists, which block both the V2 and V1 vascular receptors. The potential benefits of these combined V1and V2antagonists are shown in Fig. 9-7. The effect of these antagonists on morbidity and mortality in CHF patients has yet to be demonstrated. Pretreatment hyponatremia is a dire prognostic risk factor for mortality in CHF patients, but this most likely relates to the severity of the cardiac disease.74 Similarly, the hyponatremic patient with advanced cardiac failure appears to be at increased risk for cardiac arrhythmias. A primary deleterious effect of hyponatremia on cardiac function remains to be demonstrated. In vitro studies in vascular smooth muscle, however, have shown that hypo-osmolality is associated with an increase in cytosolic calcium concentration and enhanced contractile response to vasoconstrictor agents.75
REFERENCES
1. Schrier RW. Pathogenesis of sodium and water retention in high-output and low-output cardiac failure, nephrotic syndrome, cirrhosis, and pregnancy. Part I. N Engl J Med. 1988;319: 1065–1072.
1. Schrier RW. Pathogenesis of sodium and water retention in high-output and low-output cardiac failure, nephrotic syndrome, cirrhosis, and pregnancy. Part II. N Engl J Med. 1988;319:1127–1134.
2. Ghali JK, Cooper R, Ford E. Trends in hospitalization rates for heart failure in the United States, 1973-1987: evidence for increasing population prevalence. Arch Intern Med. 1990;150:769–773.
3. Schrier RW, Ecder T. Unifying hypothesis of body fluid volume regulation: implications for cardiac failure and cirrhosis. Mt Sinai J of Med. 2001;68:350–361.
4. Schrier RW. Body fluid volume regulation in health and disease: a unifying hypothesis. Ann Intern Med. 1990;113:155–159.
5. Schrier RW, Abraham WT. Hormones and hemodynamics in heart failure. N Engl J Med. 1999;341:577–585.
6. Henrich WL, Berl T, McDonald KM, et al. Angiotensin II, renal nerves, and prostaglandins in renal hemodynamics during hemorrhage. Am J Physiol. 1978;235:F46–F51.
7. Ferguson DW, Berg WJ, Roach PJ, et al. Effects of heart failure on baroreflex control of sympathetic neural activity. Am J Cardiol. 1992;69:523–531.
8. Cohn JN, Levine TB, Olivari MT, et al. Plasma norepinephrine as a guide to prognosis in patients with chronic congestive heart failure. N Engl J Med. 1984;311:819–823.
9. Brooks VL. Interactions between angiotensin II and the sympathetic nervous system in the long term control of arterial pressure. Clin Exp Pharmacol Physiol. 1997;24:83–90.
10. Henry JP, Gauer OH, Reeves JS. Evidence of atrial location of receptors in influencing urine flow. Circ Res. 1956;4:85–90.
11. Mulrow PJ, Schrier RW. Atrial Hormones and Other Natriuretic Factors. Clinical Physiology Series. Bethesda, MD: American Physiological Society; 1987.
12. Linden RJ, Kappagoda CT. Atrial Receptors. Cambridge, England: Cambridge University Press; 1982.
13. Schrier RW. An odyssey into the milieu intéreur: pondering the enigmas. J Am Soc Nephrol. 1992;2:1549–1559.
14. Cody RJ, Covit AB, Schaer GL. Sodium and water balance in chronic heart failure. J Clin Invest. 1986;77:144–52.
15. Schrier RW, Better OS. Peripheral arterial vasodilation hypothesis: implications for impairedaldosterone escape. Eur J Gstroenterol Hepatol. 1991;3:721–729.
|
