Acute Effect of Furosemide on Glomerular Filtration Rate in Diastolic Dysfunction

Abstract

We sought to evaluate the acute effect of furosemide on glomerular filtration rate (GFR) in subjects with diastolic dysfunction. An equal number of subjects with documented diastolic dysfunction (DD) and healthy volunteers (controls) were enrolled and underwent a baseline GFR measurement via plasma clearance of technetium-99m-diethylenetriaminepentaacetic acid. Within three to seven days of the baseline, study subjects were scheduled for a second GFR study, which was performed immediately after administration of furosemide (20 mg orally and 20 mg intravenously). There were eight healthy volunteers (8 males with a mean age 42 ± 7.8 years; 6 white, 2 Asian) and eight subjects with diastolic dysfunction (7 males, 1 female, with a mean age 64.5 ± 9.3 years; 7 whites, 1 African-American). There was a significant post-furosemide decline in GFR in the healthy volunteers, baseline vs. post-furosemide 131.6 ± 19.8 vs. 117 ± 18.2 mL/min, respectively (p = 0.03), and the patients with DD, baseline vs. post-furosemide 117.5 ± 22.3 vs. 92 ± 21.7 mL/min, respectively (p = 0.0002). A strong trend was detected, though not statistically significant, of greater GFR decline in subjects with DD compared to the healthy volunteers, 25.5 ± 9.9 vs. 14.6 ± 15.6 mL/min, respectively (p = 0.12). To conclude, acute administration of furosemide might potentially cause a greater decline in GFR in subjects with diastolic dysfunction.

Keywords:

• loop diuretics
• glomerular filtration rate
• diastolic dysfunction

INTRODUCTION

Loop diuretics, mostly furosemide, continue to find widespread use in a variety of fluid overload states. So named because of their predominant site of action in the nephron, they produce natriuresis and diuresis primarily by blocking sodium chloride reabsorption at the loop of Henle.[1]

In spite of having been part of the clinical armamentarium for decades and affecting salt and water balance via a renal mechanism, their effect on glomerular filtration rate (GFR) is not completely defined. Experimental studies are conflicting in this regard.[2–9] Some have shown no effect, some have demonstrated a decrease in GFR after administration of a loop diuretic, while others have shown an increase in GFR. Human data regarding this issue are scant and inconsistent. In one study that enrolled nine hypertensive subjects with normal renal function, as measured by blood urea nitrogen and creatinine, Loon et al. found a non-significant increase in GFR after acute administration of furosemide.[10] Another study detected a fall in GFR after furosemide administration in twelve patients with congestive heart failure.[11] It is conceivable that the GFR response to loop diuretics might be different in subjects with heart failure due to a variety of factors, such as hormonal activity, decreased effective renal perfusion, and/or salt and water retention. The effect of loop diuretics on GFR in instances of isolated diastolic dysfunction has not been studied. The treatment of symptomatic diastolic dysfunction, as outlined in the American College of Cardiology/American Heart Association Task Force Heart Failure, consists of judicious use of pre-load reducing agents such as diuretics.[12]

We sought to evaluate the acute effect of furosemide on GFR in subjects with diastolic dysfunction.

METHODS

An equal number of subjects with documented diastolic cardiac dysfunction and healthy volunteers without a history of cardiac or renal disease were evaluated. The inclusion criteria consisted of preserved renal function defined as calculated creatinine clearance greater than 60 mL/min and normal left ventricular systolic function. An additional inclusion criterion for the healthy volunteers included the absence of diastolic dysfunction. All subjects had cardiac function, left ventricular ejection fraction, and the presence or absence of diastolic dysfunction, assessed by means of two-dimensional echocardiography. Creatinine clearance was calculated using serum creatinine, age, sex, and weight.[13] The exclusion criteria included the presence of edema, states of decreased effective arterial volume (such as cirrhosis of liver and nephrosis), current diuretic use, allergy or adverse reaction to furosemide, and left ventricular ejection fraction less than 50%. The presence of edema can result in inaccuracy of radionuclide GFR due to increase in the volume of distribution of the radiotracer.[14]

Glomerular filtration rate was measured using technetium-99m-diethylenetriaminepentaacetic acid (^99mTc-DTPA) via the plasma clearance method as described previously.[14] A patient dose and standard dose of 5 mCi of ^99mTc-DTPA were prepared in each case. The patient dose was injected at the start of the test and the standard used for dose calibration. The ^99mTc-DTPA standard was diluted 1:10,000 using deionized water. Patients were asked to return to the nuclear medicine suite one and three hours after the injection for sampling of 10 ml venous blood from another vein (other than that used for injection) in a lavender topped (EDTA) tube. Plasma samples were obtained after 10 minutes' centrifugation. The radioactivity in the standards and plasma was counted in a standard gamma scintillation counter. The GFR derivation consisted of the mean of the one-sample (i.e., three-hour sample) and two-sample (i.e., one- and three-hour samples) calculation methods.[14] Sampling between 60 and 240 minutes after injection is considered adequate for GFR measurement via the plasma clearance method.[15]

All subjects underwent a baseline GFR measurement. Within three to seven days of the baseline measurement, the subjects were scheduled for a second GFR study. The timing of the second test was determined thus to eliminate interference due to residual radioactivity from the first study day. At the same time, a short time interval between the two tests was intended to avoid occurrence of a significant change in the clinical and physiologic state of the study subject. On the second study day immediately prior to the injection of the radionuclide, subjects were administered furosemide 20 mg orally and 20 mg intravenously. The protocol was so designed to ensure a continuous diuretic effect over the three-hour time period of the test. Intravenous furosemide acts rapidly and achieves a peak effect within 30 minutes, and the effect dissipates after two hours.[16] Oral furosemide achieves a peak effect within one to two hours and has a duration of action of six to eight hours. The intent was to employ an effective but not supra-therapeutic dose of the diuretic. No water load was given on either of the study days because the intent was to study inherent diuretic effect not altered by extraneous factors. The GFR measurement by plasma clearance method obviated the need to achieve a high urine flow rate.

Subjects with diastolic dysfunction were recruited based on impaired diastolic relaxation reported on echocardiography performed as part of usual clinical care if they satisfied other study criteria. All healthy volunteers underwent echocardiographic studies to determine eligibility in the study.

Statistical Analysis

Paired and unpaired t-test was used for analysis. Results are expressed as mean ± standard deviation (SD).

The protocols were approved by the Institutional Review Board (IRB) of the University of Missouri-Columbia and Research and Development Committee of the Harry S. Truman Memorial Veterans' Hospital. All subjects provided informed written consent.

RESULTS

There were eight healthy volunteers (8 males with a mean age 42 ± 7.8 years; 6 white, 2 Asian). There were eight subjects with diastolic dysfunction (7 males, 1 female with a mean age 64.5 ± 9.3 years; 7 whites, 1 African-American). Five subjects with diastolic dysfunction had a history of hypertension.

There was a significant post-furosemide decline in GFR in the healthy volunteers, baseline vs. post-furosemide 131.6 ± 19.8 vs. 117 ± 18.2 mL/min, respectively (p = 0.03; see Figure 1). Similarly, there was a significant drop in GFR in the patients with diastolic dysfunction, baseline vs. post-furosemide 117.5 ± 22.3 vs. 92 ± 21.7 mL/min, respectively (p = 0.0002; see Figure 1).

Figure 1. Decline in glomerular filtration rate before and after furosemide (mean ± SD).

A strong trend was detected of greater GFR decline in subjects with diastolic dysfunction compared to the healthy volunteers, 25.5 ± 9.9 vs 14.6 ± 15.6 mL/min, respectively (p = 0.12; see Figure 2).

Figure 2. Comparative post-furosemide decline in glomerular filtration rate in healthy volunteers and subjects with diastolic dysfunction (DD) (mean ± SD).

There was no statistically significant difference between the mean blood pressure before and three hours after furosemide, 90 ± 6.6 vs. 86.3 ± 6.3 mmHg, respectively (n = 11; p = 0.19).

One subject (healthy volunteer) inadvertently missed the one-hour sample collection time during the baseline study, and in this instance, GFR was derived solely using the one-sample method.[14] One subject could not come for the second GFR study per the protocol-defined period (within three to seven days of the first study) but had the test performed on the ninth day, which was not felt to be of any meaningful significance. No subjects experienced dizziness or any overt symptoms suggestive of hypovolemia.

DISCUSSION

Our data suggest that the acute administration of furosemide to humans results in a decline in fall in glomerular filtration rate, which might be exaggerated in subjects with diastolic dysfunction.

Investigators have previously studied the effect of furosemide on GFR in subjects with heart failure. Dikshit et al. studied the acute effect of furosemide on GFR in subjects with left ventricular failure following acute myocardial infarction.[17] Furosemide was administered intravenously in dose ranging from 0.5 to 1 mg per kg. Data regarding baseline renal function were not presented, which is pertinent with respect to the therapeutic effect of furosemide. Furosemide acts intraluminally primarily at the thick ascending limb of the loop of Henle after excretion into the luminal fluid via the proximal tubular organic acid pathway.[18] Therefore, decreased concentration of the active drug reaches the site of action in the face of decreased tubular function associated with renal insufficiency. Thus, the efficacy of a given dose is dissimilar in subjects with varying renal function, and higher doses are required in the presence of renal failure.[19] GFR was measured via urinary clearance of inulin. A significant increase in GFR was detected after 15 and 30 minutes. However, subsequently GFR progressively declined such that after two hours, it decreased to levels 15 to 20 mL/min below baseline (actual figures not presented), though apparently statistical significance was not detected. No further measurements were obtained after this time period.

Gottlieb et al. administered intravenous furosemide in doses ranging from 20 to 200 mg to 12 males, age range 42 to 73 years, with congestive heart failure, and measured GFR by the plasma clearance method.[11] The baseline serum creatinine was described as being less than 2.5 mg/dL. A significant drop in normalized GFR from a mean pre-furosemide GFR of 84 mL/min to 63 mL/min was detected. However, the inconsistent dose employed in subjects with potentially differing renal function makes interpretation difficult. Conceivably, the results could be biased by a significant number of subjects receiving large (potentially overt hypovolemia generating) doses. On the other hand, a particular elderly subject with significant renal insufficiency—for example, a serum creatinine value of 2.4 mg/dL and poor muscle mass, indicative of advanced chronic kidney disease—could have received an ineffective dose of 20 mg.

In another study, a significant decline in creatinine clearance was detected after intravenous furosemide administration to edematous subjects with symptomatic heart failure (ejection fraction ≤ 40%).[20]

The present study enrolled subjects with uniformly preserved renal function and utilized a consistent dose of furosemide. GFR was measured over a three-hour period, the entire period of maximal furosemide effect. We intended to study the effect of loop diuretics on a subgroup of subjects with diastolic dysfunction (DD) and preserved systolic function, which represents nearly one-third of instances of heart failure.[12] The primary pathophysiologic problem in DD represents decreased ventricular compliance and/or impaired relaxation, which results in an inappropriate rise in pressure with relatively small changes in volume.[21] Signs and symptoms of pulmonary venous congestion occur in spite of a normal or above-normal ejection fraction and symptomatic relief requires pre-load reduction, such as with diuretics. However, caution is necessary, as there is significant dependency of the cardiac output on the elevated filling pressure in DD, as outlined in the American College of Cardiology/American Heart Association Task Force Heart Failure guidelines.[12] The effect of diuretics on GFR in such circumstances has not been studied.

We found an impressive trend of greater decline in GFR in subjects with DD compared to healthy volunteers, which suggests that such subjects may be more prone to a decline in GFR after the acute administration of loop diuretics.

There are limitations to the present study. First, in view of the small sample size, the results can only be considered preliminary and a prelude for further investigations. Further, the subjects with diastolic dysfunction were significantly older than the volunteers, and thus could be more prone to a decline in GFR after diuretic administration. In addition, though there was no significant fall in mean blood pressure in either group and none of the subjects had symptoms of hypovolemia, one cannot rule out that volume changes might have played a role in the decline in GFR. Last, it must be emphasized that we studied the acute (three-hour) effect of furosemide on GFR, and the findings cannot necessarily be extrapolated to an extended time-period or chronic use.

To conclude, there is a significant fall in GFR after acute administration of furosemide to subjects with diastolic dysfunction. There is a need for additional research in this arena.

ACKNOWLEDGMENTS

This work was carried out with the use of resources and facilities of the Harry S. Truman Memorial Veterans' Hospital. The Missouri Foundation for Medical Research provided partial financial support for this work. Partial results were presented at annual meetings of the American Society of Nephrology (2001 and 2003) published in abstract form in the abstract editions of the Journal of the American Society of Nephrology (J Am Soc Nephrol. 2001;12:518A; J Am Soc Nephrol. 2003:768A). The authors are indebted to, and sincerely thank, all of the subjects who participated in this research, donated their time, and lent their body to this non-therapeutic protocol without which this work would not have been possible. The authors acknowledge Patricia Stone for technical assistance in the performance of the radionuclide glomerular filtration rate studies, Rex Freelon for the performance of two-dimensional echocardiography, John Hewett, Ph.D., for statistical consultation, and Harold Moore, M.A., for the creation of the figures.