Sequential Nephron Blockade Breaks Resistance to Diuretics in Edematous States

Abstract

Diuretic therapy in edematous diseases often yields an inadequate natriuretic response ("diuretic resistance"). To study the functional changes in patients with congestive heart failure, liver cirrhosis with ascites, and nephrotic syndrome, characterized by a reduced effective arterial blood volume (EABV), different diuretic strategies were studied. It was shown that monotherapy with hydrochlorothiazide or furosemide was followed by an inadequate natriuretic response. Correlation of diuretic response with pretreatment fractional sodium excretion of the patient revealed a clear-cut interdependency: Those patients were resistant whose FE_Na⁺ was greatly below normal (<0.2%). In addition, it was found that the coad-ministration of the carboanhydrase inhibitor acetazolamide to diuretic therapy was very effective. We therefore conclude that an increase in proximal-tubular Na⁺ reabsorption is the major ("pharmacodynamic") determinant for diuretic resistance in edematous diseases with functional "underfilling" of the vascular tree. This alteration of the kidney can easily be overcome by coadministration of a carboanhydrase inhibitor (e.g., acetazolamide).

Diuretics have long been the cornerstone for the treatment of edema. This is because the kidney-the selective target organ-plays the key role in the regulation of salt and water homeostasis. It is a common clinical phenomenon, however, that patients with edematous disorders frequently manifest a reduced (i.e., smaller than expected) natriuretic response to diuretics that cannot be increased by increasing the dose of diuretic. This inadequate response to diuretics (diuretic resistance) manifest in patients with congestive heart failure, liver cirrhosis, and nephrotic syndrome has long been a matter of research and debate ^(1-6).

The mechanism(s) responsible for diuretic resistance certainly differ in the different clinical conditions and are assumed to be very complex. Principally they may be divided into "pharmacokinetic" and "pharmacodynamic" mechanisms once the patient's noncompliance is ruled out. The pharmacokinetic determinants of renal response to diuretics are a function of both the total amount of "free" diuretic (not bound to protein) reaching the site of action and the diuretic's time course of delivery into the urine, which may be changed because of alterations in volume of distribution, bioavailability, or protein binding in edematous states ^(1,3,5,6).

For the assessment of the pharmacodynamic response to a diuretic, the correlation of the response to the amount of free diuretic in tubular fluid (and then in final urine) and the amount of filtered Na⁺ load reaching the nephron segment in question has to be considered. As shown recently ⁽⁷⁾, the latter determines the effectiveness or intrinsic activity of a diuretic in edematous states. In this context, vascular "underfilling" ^(8-11) was reported to increase Na⁺ reabsorption in the proximal tubule, leaving only a decreased Na⁺ load to the distal segments of the nephron. This results in reduced effectiveness of diuretics (loop diuretics, thiazides). The reabsorption of Na⁺ was also claimed to be increased in the distal tubule in some edematous disorders ^(12-17). This alteration may therefore attenuate the effectiveness of diuretics that act in the loop of Henle. A third pharmacodynamic alteration may be the result of a change in the sensitivity of the renal tubuar cells to diuretics through neurohumoral factors activated in the edematous disorder ^(1,3,18,19).

Therefore the scope of this study was to evaluate the functional state of the nephron in edematous states by studying dose-response relations of different diuretics alone and in pathopysiologically grounded combinations. The purpose was to show the clinician that by simply combining different classes of diuretics rather than increasing the dosage of conventional monotherapy, the effectiveness of diuretic treatment may be improved with fewer side effects.

MATERIALS AND METHODS

This (ethically approved) trial was performed on patients who had given their written informed consent to the study. They consisted of three groups: (a) nine patients with congestive heart failure (CHF), (b) nine patients with liver cirrhosis and ascites (LC), and (c) six patients with nephrotic syndrome (NS).

The patients with CHF were characterized by their clinical symptoms and cardiac index (CI; via Swan-Ganz catheterization). Liver cirrhosis was diagnosed by clinical signs and by sonography, blood chemistry, or laparoscopy. Nephrotic syndrome was defined by proteinuria (>3 g/day) and hypoalbuminemia (<30 g/L). They all had edema. Underlying histologic changes were studied on kidney biposies. A control group was composed of six healthy volunteers. Before the study, low plasma albumin (<30 g/L) was substituted. Only patients whose plasma potassium was normal (5 mM > K⁺ > 3.5 mM) entered the study. Patients with hyponatremia (plasma Na⁺ <130 mM) and elevated plasma creatinine >1.2 mg/dl were excluded from the study. The demographic and clinical data of the patients are given in Table 1.

TABLE 1: Demographic data of the patients

The participants were kept on a standard diet, receiving a defined fluid (1.5 L/day) and salt intake (6 g NaCl; 100 mmol Na⁺/day) 3 days before and throughout the trial. Routine blood chemistry and urine analysis were performed at entry into the study. Diuretic medication was withheld for ≥4 days before the study (washout). Medication for CHF [except angiotensin-converting enzyme (ACE) inhibitors and β-blockers), medication for LC (except β-blockers), and medications for NS were allowed to be continued at constant dosage throughout the study. At the end of 4 days' diuretic washout, the fractional excretion of sodium (FE_Na⁺) was determined in each patient. Then 100 mg spironolactone was given as a basic treatment to avoid K⁺ wasting ^(1,20,21).

Study protocol

The format of the protocol was previously reported ⁽²²⁾. To evaluate the acetazolamide effectiveness on the test diuretics, hydrochlorothiazide (HCTZ), furosemide (FU), and acetazolamide (AA), the diuresis and urinary excretion of Na⁺ and body weight control were examined for 24 h at 12-h intervals.

The effect of single- (25 mg) and double-dose treatment with HCTZ was compared with the effect of single (40 mg) and double doses of the loop diuretic FU. Then the effect of AA, 250 mg, coadministered with the single-dose thiazide or loop diuretic was evaluated.

After diuretic washout and assessment of FE_Na⁺, every patient received 100 mg spironolactone 2 days before and during the study. Then they were randomly divided into two groups assigned to either 25-mg HCTZ or 40-mg FU treatment. The following day, they received the double dose of either diuretic. For the next day, the patients were on a placebo interval and were then crossed over to receive first a single and next day a double dose of thiazide or loop diuretic. The next 2 days, they were given AA with a single dose of the test diuretic given the day before. On the next day, the test diuretic was changed back and again given with AA. The results of the respective diuretic treatment obtained before and after cross-over were pooled. The mean values ± SEM are given. The results were correlated with the pretreatment FE_Na⁺ data of the patients by simple correlation analysis.

Laboratory techniques

Measurements of urinary Na⁺ and K⁺ were performed by using a Zeiss automatic electrolyte analyzer, FL 6. Samples for Na⁺ and K⁺ were diluted in 1:500 with a La₂O₃ solution. The fractional sodium excretion, FE_Na⁺ (%), represents the amount of Na⁺ excretion (mmol/time) as a percentage of filtered load [plasma Na⁺ concentration × glomerular filtration rate (GFR).

GFR was taken from creatinine clearance. Measurement of inulin clearance related to body surface would have given exact GFR data, because creatinine clearance tends increasingly to overestimate GFR as kidney function is reduced. However, when comparing different diuretic regimens in an individual, this consistent error is less significant.

RESULTS

Diuretic treatment of patients with CHF

The administration of 25 mg HCTZ to patients with CHF who were receiving spironolactone treatment yielded only a moderate response compared with the controls, which was only slightly increased by doubling the dose (Fig. 1). The data scattered much more than did the control data. Practically the same held true for the treatment with FU, 40 and then 80 mg, given orally (Fig. 1).

FIG. 1: Sodium excretion per day after the administration of hydrochlorothiazide (HCTZ), furosemide (FU), and acetazolamide (AA) compared with pretreatment data (pre) [on "basic" spironolactone (100 mg/day) treatment throughout the study] in congestive heart failure, liver cirrhosis with ascites, and nephrotic syndrome. In the headline, the excretion data of healthy controls are given.

When correlating the effectiveness of either diuretic treatment with pretreatment FE_Na⁺ data, it became obvious that the diuretic respose of the patient depended on the pretreatment FE_Na⁺(Figs. 2 and 3). The lower the patient's FE_Na⁺, the smaller was the natriuretic response to diuretics (r = 0.77 − 0.90). The pretreatment FE_Na⁺ in turn was significantly correlated (p < 0.01) with the CI of the patients (see Table 1). Patients with a high New York Heart Association (NYHA) rating and a low CI (<2 L/min/m²) had a FE_Na⁺ <0.2% and were in turn resistant to diuretic treatment.

FIG. 2: Left: Sodium excretion per day after the administration of 50 mg hydrochlorothiazide (HCTZ) in relation to the patients' pretreatment fractional sodium excretion, FE_Na⁺. Right: The increment of sodium excretion achieved by acetazolamide, 250 mg, coadministered with 25 mg HCTZ in relation to the patients' pretreatment FE_Na⁺. Co, healthy controls.

FIG. 3: The same experiment as depicted in Fig. 2. Instead of hydrochlorothiazide (HCTZ), furosemide (FU) is administered. Left: 80 mg FU. Right: 40 mg FU and 250 mg acetazolamide (AA).

Coadministration of AA to thiazide or loop diuretic treatment was particularly effective in those patients who had responded poorly to a double dose of a thiazide or loop diuretic (i.e., in those who had a low FE_Na⁺). This is illustrated on the right side of Figs. 2 and 3, which show an inverse relation between the increment in natriuresis (given in percentage of that with single-dose diuretic) and the patients' pretreatment FE_Na⁺ data (r = 0.75 − 0.85). In other words, the low responders with reduced FE_Na⁺ benefited more from low-dose combination therapy than from high-dose monotherapy.

Diuretic treatment of patients with cirrhotic ascites

The results obtained in cirrhotic patients were very much the same as those obtained in patients with CHF. Again the effectiveness of treatment with thiazides or loop diuretics was only moderate, and the data also scattered widely (Fig. 1). Those patients with cirrhosis who had a low pretreatment FE_Na⁺ and a weak response to diuretic monotherapy had a significantly greater excretory increment when AA was coadministered (Figs. 2 and 3).

Diuretic treatment of patients with nephrotic syndrome

Despite the small number of patients with NS who entered the study, it could be shown that (consistent with CHF and LC) coadministration of AA with the thiazide or the loop diuretic was predominantly effective in those patients characterized by a low FE_Na⁺(Figs. 2 and 3) and who had first shown an only small response to a single and double dose of HCTZ or FU (Fig. 1).

DISCUSSION

The purpose of the study was to aid the clinician simply and effectively to find a diuretic regimen with fewer adverse effects for the treatment of the edematous states.

In agreement with earlier observations in heart ⁽²³⁾ and liver disease ⁽²⁴⁾, those edematous patients were found to be resistant to diuretics (i.e., thiazides and loop diuretics), whose pretreatment FE_Na⁺ data had been below normal. It was these particular patients-irrespective of the underlying disease-who manifested the greatest benefit by the coadministration with AA, known to block Na⁺ reabsorption in the proximal renal tubule. Patients with a FE_Na⁺ >0.35% showed no significant increment in the response to additional AA, yet these patients had an increased diuresis with the loop diuretic and thiazide combination rather than with increasing dosage of either drug alone. This combination therapy was shown earlier to be very effective in these disease entities ^(12-17).

For the first time it was shown that patients with NS had a benefit from AA coadministration when they presented with a lower than normal FE_Na⁺ before diuretic treatment and when these patients manifested diuretic resistance to both thiazides and loop diuretics. By making use of a diuretic of very high (FU) and of very low protein binding (HCTZ), the well-known attenuation of the effectiveness of a diuretic through intratubular protein binding ⁽⁶⁾ could be ruled out as the predominant mechanism of diuretic resistance in our nephrotic patients. Rather, changes of renal tubular sodium handling have to be assumed to be the main disorder common in all the three disease entities, resulting in Na⁺ retention and edema formation.

The pathogenesis of Na⁺ and water retention in edematous states such as CHF, LC, and NS was analyzed by Schrier ⁽⁸⁾ (see Fig. 4). These diseases appear to have disturbances of volume regulation as a hallmark of their clinical syndrome. The effective arterial blood volume (EABV), a term amplified by Seldin in 1975 ⁽²⁵⁾, is considered to be reduced in these edematous states ^(8-11). The concept of EABV does not necessarily reflect the entire intravascular volume but rather a critical compartment in the arterial side ⁽⁹⁾. Thus arterial blood volume is better correlated with renal sodium regulation than is total blood volume. EABV behaves as if it were determined by two forces. The first is the filling of the arterial tree, a function of venous return and left ventricular performance. The second is the magnitude of arteriolar runoff and the size and compliance of the vessels. A reduction in EABV, therefore, can mean an actual diminution of arterial volume (as in CHF) or excessive peripheral runoff despite increased arterial filling (as in the AV shunts and vasodilatation; e.g., liver cirrhosis) or a dissociation between the vascular and interstitial compartments caused by diminished plasma oncotic pressure (e.g., NS) ⁽⁹⁾. Despite their complex pathophysiologic characteristics, the common quality of all these edematous states appears to be a shrinkage of EABV in a setting in which the retention of salt and water ensues, which manifests clinically in a low FE_Na⁺ or low luminal urine sodium concentration, respectively.

FIG. 4: Top: Pathogenesis of sodium retention in edematous diseases modified after Schrier ⁽⁸⁾ and Seldin ⁽⁹⁾. EABV, effective arterial blood volume; RAA system, renin-angiotensin-aldosterone system; GFR, glomerular filtration rate; RPF, renal plasma flow; FF, filtration fraction; ADH, antidiuretic hormone. Bottom: Relation between urinary delivery of diuretic and natriuretic response in normal controls, in patients with chronic renal failure, and in edematous states with reduced effective arterial blood volume (EABV). To be independent of individual GFRs, urinary diuretic excretion (μg/min) is divided by each patient's GFR (ml/min) yielding fractional diuretic excretion μg/ml) corresponding to tubular disposition of the diuretic. Natriuretic response is given by fractional Na⁺ excretion. The "plateau" of the curve represents the intrinsic activity of the diuretic in the respective diseases. This, in turn, is set by the Na⁺ load at the tubular site of the diuretic's action. Schematically modified after refs. ^{30, 31, 33, and 34}.

When EABV is consistently contracted, at least four efferent pathways for volume control are activated ⁽⁹⁾: the sympathetic system, the renin-angiotensin-aldosterone system, the nonosmotic release of antidiuretic hormone (ADH), and the renal prostaglandin production.

As for the renal mechanisms for sodium retention, the primary epithelial side of action of renal nerves appears to be the proximal tubule, where α-adrenergic stimulation enhances Na⁺ reabsorption ^(9,26,27). Linked to enhanced adrenergic nerve activity, an increased renin secretion and angiotensin II formation finally increases the efferent arteriolar resistance disproportionately ^(28,29). As a consequence, net proximal tubular Na⁺ transport is increased. In addition, angiotensin stimulates aldosterone secretion from the adrenal cortex. Aldosterone, in turn, increases Na⁺ reabsorption in the distal nephron. However, because of greater (≤85% of filtrate) than normal (65-70% of filtrate) reabsorption of sodium by the proximal tubule, sodium delivery to the distal tubule is decreased. This change in renal handling of sodium is proposed as the cause for failure of edematous patients to "escape" from the sodium-retaining effect of aldosterone ⁽¹¹⁾.

The common inherent disorder, despite pharmacokinetically and dynamically grounded differences of Na⁺ handling in the three disease entities, is a lower than normal ratio between natriuresis and free diuretic in the urine ^(23,30-32). As first shown by Brater et al. ⁽³³⁾ with a modified Hill plot, patients with CHF could be pharmacologically characterized by a downward and rightward shift in the concentration-response relation (Fig. 4). The same pattern in the dose-response curve was then portrayed by Keller et al. for LC ⁽³⁰⁾ and NS ⁽³¹⁾.

Clinically, these curves suggest that the diuretic response is much lower than normal and cannot be increased by increasing the dosage. However, as shown by Keller et al. ⁽³⁰⁾ in cirrhotic patients with ascites, undesired side effects such as K⁺ excretion dissociated from Na⁺ reabsorption may be increased further by higher dosing.

The mechanism(s) underlying the downward and rightward shift of the dose-response relation in the three edematous states has not been tested. In advanced renal failure, the opposite occurs: Brater et al. ⁽³⁴⁾ reported an upward shift of the dosage-response curve, once the low GFR data of the patients were normalized to be comparable with the normal GFR data of the other patients (see Fig. 4). The response given as FE_Na⁺ related to urinary recovery or disposition, respectively, of diuretic was relatively increased (i.e., the "plateau" of the patients with renal failure was higher than normal in contrast to what was seen in edematous states with reduced EABV). Interestingly, this upward shift of the dose-response curve was also seen with thiazides studied in renal failure ⁽³⁵⁾. The high FE_Na⁺ in renal failure is associated with an elevated Na⁺ load in the thick ascending limb of Henle's loop and the distal nephron after reduced Na⁺ reabsorption (Na⁺ rejection) by a number of uremic factors ⁽³⁶⁾. On the other hand, the low FE_Na⁺ in contracted EABV is associated with a predominant increase in proximal-tubular Na⁺ reabsorption ^(8-11). That is, the plateau of the (normalized) dose-response curve represents the intrinsic activity of the diuretic under the clinical condition. The plateau is set by the load or availability, respectively, of sodium at the tubular site of action of the diuretic.

The message of these data is that a diuretic with a proximal tubular site of action given in addition to loop or thiazide diuretics may break the resistance to these diuretics, predictable by the low FE_Na⁺. This can best be reconciled with the previously outlined predominant increase in proximal tubular Na⁺ transport. In conclusion, simple monitoring of the patients' urine sodium before diuretic treatment may enable the clinician to establish a diuretic regimen that meets the renal alterations that take place in edematous states. A combination of low-dose diuretics with different sites of renal action (sequential nephron blockade) rather than conventional high-dose monotherapy provides more effectiveness and safety for the patient.