Loop Diuretics for Chronic Renal Insufficiency: A Continuous Infusion Is More Efficacious Than Bolus Therapy David W. Rudy, MD; James R. Voelker, MD; Paula K. Greene, RN; Francisco A. Esparza, MS; and D. Craig Brater, MD
• Objective: To test the hypothesis that a continuous, low-dose infusion of a loop diuretic is more efficacious and better tolerated than conventional intermittent bolus therapy in patients with severe chronic renal insufficiency (CRI). • Design: Randomized, crossover clinical trial with subjects serving as their own controls. • Setting: The General Clinical Research Center of Indiana University Hospital, Indianapolis, Indiana. • Patients: Eight adult volunteers with severe stable CRI (mean creatinine clearance, 0.28 mlVs; range, 0.15 to 0.47 mL7s) were recruited from the outpatient nephrology clinics of Indiana University Medical Center. • interventions: On admission, diuretic drugs were withdrawn and patients were equilibrated on an 80 mmol/d sodium, 60 mmol/d potassium metabolic diet. Patients were randomized to receive a 12-mg intravenous dose of bumetanide given either as two 6-mg bolus doses separated by 6 hours or as the same total dose administered as a 12-hour continuous infusion. When sodium balance was re-established, each patient was crossed over to the alternative study limb. All patients completed both phases of the study. • Measurements and Results: Comparable amounts of bumetanide appeared in the urine during the study period (infusion, 912 ± 428 jig; bolus, 944 ± 421 |xg; difference, 32 jig; 95% CI, - 16 ftg to 80 p,g, P = 0.16). The continuous infusion resulted in significantly greater net sodium excretion (infusion, 236 ± 77 mmol; bolus, 188 ± 50 mmol; difference, 48 mmol; CI, 16 mmol to 80 mmol, P- 0.01). No patient had episodes of druginduced myalgias during the continuous infusion compared with 3 of 8 patients with bolus therapy. • Conclusions: In patients with severe CRI, a continuous intravenous infusion of bumetanide is more effective and less toxic than conventional intermittent bolus therapy. Continuous administration will probably be useful in patients with severe CRI who have not achieved an adequate natriuresis or who show evidence of drug toxicity with standard diuretic dosing regimens. A similar benefit may occur in selected diuretic-resistant patients with cardiac or hepatic disease, and studies in these patients seem warranted.
Annals of Internal Medicine. 1991;115:360-366. From Indiana University Hospital, Indianapolis, Indiana. For current author addresses, see end of text. 360
Large doses of loop diuretics are typically administered to patients with severe chronic renal insufficiency (CRI) to circumvent diuretic resistance. Although a natriuretic response can be elicited in such patients, it remains subnormal because of the reduced filtration capacity of the diseased kidney. Although increasing the dosing frequency to several times per day can be used to enhance cumulative response, net sodium output frequently remains inadequate to meet clinical needs. One potential problem with administering repetitive large doses of a loop diuretic is the development of acute tolerance to the drug (1-4). Although never systematically examined in CRI, this phenomenon occurs in normal subjects receiving even small doses of diuretics. Acute diuretic tolerance is presumably caused by compensatory renal sodium retention because its appearance has been shown to occur in response to excessive sodium excretion and to correlate with both the rate and magnitude of natriuresis (2-4). Compensatory renal sodium retention can occur not only during the time of diuretic activity but can also continue well after the drug effect has subsided (1). The positive sodium balance caused by this adaptive response can further antagonize net diuretic effect. Because patients with severe CRI already have a reduced excretory capacity for electrolytes, any decrease in sodium elimination either during or after drug administration can diminish diuretic effectiveness enough to preclude their clinical usefulness. Because conventional loop diuretic dosing regimens in severe CRI often result in therapeutic failure, it is important to identify alternative dosing schemes that are more effective. A dosing pattern that produces a continuous but submaximal response may be one such strategy. This method would provide a constant delivery of drug to the active site so that a persistent natriuretic effect would occur. If compensatory sodium retention does occur in CRI, it is also possible that this problem could be minimized by a constant low-dose infusion, thus further enhancing the drug's overall therapeutic effectiveness. We would also expect this treatment modality to be safer because high and potentially toxic serum concentrations associated with intermittent bolus therapy would be avoided. A reduced but effective dose given by a continuous intravenous infusion might therefore result in greater cumulative sodium excretion and could be better tolerated than conventional repetitive high-dose administration. To test this hypothesis in patients with severe CRI, we compared the pharmacokinetic and dynamic determinants of response to
© 1991 American College of Physicians
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 03/06/2017
Table 1. Demographic Data Patient
Sex, Age
Weight
y
kg
Creatinine Clearance mLls
Diagnosis
1
F, 55
82.6
0.15
2
M, 52
73.1
0.38
3
M, 30
72.8
0.28
4 5 6
M, 55 M, 24 M, 31
102.2 91.2 58.4
0.47 0.30 0.27
7
M, 37
86.4
0.27
Polycystic kidney disease, hypertension Focal segmental glomerular sclerosis, hypertension, gout, coronary artery disease Diabetes mellitus, hypertension, type IV renal tubular acidosis Diabetes mellitus, hypertension Hypertension Diabetes mellitus, hypertension, type IV renal tubular acidosis Hypertension
8
F, 43
61.2
0.17
Hypertension, diabetes mellitus
intravenous bumetanide given as a continuous infusion and by conventional intermittent bolus therapy. Methods Eight adult patients with severe stable CRI participated in the study (Table 1). The patients were selected so that their concurrent medical conditions and treatment regimens would not affect diuretic responsiveness. The study protocol was approved by the Institutional Review Board of Indiana University. Experimental Protocol After obtaining informed consent, patients were hospitalized at the General Clinical Research Center. All diuretics were withdrawn; other medications were continued as previously prescribed. Patients were placed on an 80-mmol sodium and 60-mmol potassium per day metabolic diet. Sodium balance was documented to occur by day 6 on the basis of stable daily weights and 24-hour urinary electrolyte excretion. The study was done in a randomized crossover fashion comparing in each patient the natriuretic response to two bumetanide treatment strategies. Bumetanide was chosen for study because the more rapid elimination of this drug relative to furosemide in patients with severe CRI enabled us to conduct the study in a practical amount of time (5). After sodium balance was established, the patients were randomized to receive 12 mg of intravenous bumetanide administered either as two 6-mg bolus doses infused over 5 minutes and separated by 6 hours, or a 1-mg bolus loading dose given over 3 minutes followed by a 0.912 mg/h infusion for 12 hours. The bumetanide infusion was prepared by diluting 12 mg of bumetanide in 500 mL of 5% dextrose in water and was infused at a rate of 38 mL/h. Because we wanted to simulate the clinical situation, we incorporated more than one bolus dose. Further, we felt that an adequate comparison of regimens required large bolus doses that would reach the upper aspect of the dose-response curve and that the infusion dose should fall approximately midway on the steep portion of this curve. On the basis of previous data from our laboratory, we expected the respective doses to achieve these goals (5). The dosing interval of 6 hours for bolus therapy was chosen because this time represents the most frequent dosing interval used in clinical situations (5-8). A loading dose was used with the infusion to attain a steady-state plasma concentration rapidly. To ensure frequent spontaneous voiding, the patients drank water, 10 mL/kg body weight, 1 hour before drug dosing. Patients then drank 150 mL of water per hour for the next 14 hours. During the infusion this amount was reduced to 100 mL/h to compensate for the volume of intravenous 5% dex-
Medications
Clonidine, aspirin (low dose), hydralazine, propranolol Prazosin, furosemide, colchicine, diltiazem, aspirin (low dose) Clonidine, insulin, sodium bicarbonate Insulin, labetalol Verapamil, enalapril, furosemide Enalapril, insulin, furosemide Metoprolol, minoxidil, clonidine, furosemide, calcium carbonate Cimetidine, furosemide, metaclopromide, clonidine, calcium carbonate, nifedipine, insulin
trose in water. One hour after bumetanide was initially administered, patients were served breakfast and given their prescribed doses of other medications. Patients were continued on the metabolic diet throughout the day and had free access to distilled water. Patients remained supine during the diuretic phase of the study except to void. In order to simulate the clinical use of diuretics and to determine if acute diuretic tolerance occurs in patients with CRI, urinary losses of solutes were not replaced during the study period. To characterize pharmacokinetic-dynamic parameters during bolus therapy, serum samples were obtained for measurement of electrolytes, creatinine, and bumetanide concentrations at - 35, 0, 5, 15, 30, 45, 60, 90, 120, 180, 240, 360, 365, 375, 395, 425, 485, 605, 725, 845, and 1440 minutes. For the continuous infusion regimen, samples were obtained at - 35, 0, 120, 300, 420, 600, 720, 725, 735, 750, 780, 840, and 1440 minutes. Urine samples for measurements of volume, sodium, creatinine, and drug concentrations were obtained for each study day by spontaneous voiding at - 60, 0, 30, 60, 120, 180, 240, 300, 360, 390, 420, 480, 540, 600, 660, 720, 780, 840, 960, and 1440 minutes. To facilitate the return to sodium balance before the second study limb, on the day after the first study, all patients received intravenous 0.9% saline over 12 hours at a rate calculated to replace net sodium loss from the diuretic. Patients were maintained on the metabolic diet and monitored as before. Once sodium balance was re-established (after approximately 3 days), the second limb of the protocol was conducted. Laboratory Determinations Sodium concentrations were measured by flame photometry (Model 943; Instrumentation Laboratory, Lexington, Massachusetts) and creatinine concentrations by autoanalyzer using the Jaffe reaction (Technicon Autoanalyzer II, Elmsford, New York). Bumetanide was measured by a modification of a highpressure liquid chromatography method previously described (9). Briefly, aliquots of urine or serum samples plus piretanide as an internal standard were adjusted to pH 1.7 with 1.0 M phosphate buffer. Dichloromethane was added, the sample vortexed, and the aqueous phase removed by aspiration. The remaining organic layer was dried at low heat under a stream of nitrogen. The samples were reconstituted in mobile phase consisting of 60:40 acetonitrile and 0.01M sodium acetate adjusted to pH 3.3. The samples were injected onto a C8 5-ju,m column (Beckman, San Ramon, California) and eluted at a flow rate of one mL/min. Drug was detected by fluorescence spectroscopy (Hitachi model 650-15, Tokyo, Japan) using excitation and emission wavelengths of 338 nm and 440 nm, respectively. A standard curve was determined for each assay and the correlation coefficient (r2) always exceeded 0.990 over the concentration range of 5 to 2000 /xg/mL. The coefficients of vari-
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 03/06/2017
361
Table 2. Pharmacokinetics of Bumetanide in Patients with Chronic Renal Insufficiency Variable
Bolus
Half-life, min Systemic clearance, mLlmin Renal clearance, mLlmin Nonrenal clearance, mLlmin Bumetanide excretion, 0 to 14 hours /Jig fe,* % Maximum serum drug concentration,/x#/L
82 202 15.1 187
Infusion
±30 76 ± 24 ± 49 196 ±43 ± 6.1 14.4 ± 5.8 ± 51 179 ± 46
944 ± 421 912 ± 428 8.0 ± 3.6 7.9 ± 3.7 1118 ± 526 155 ± 76
P Value > > > >
0.2 0.2 0.2 0.2
0.2 > 0.2 0.001
* fe = fraction of administered dose excreted in the urine. ation for urine and serum samples were less than 10% and 5%, respectively.
Data Analysis The pharmacokinetic parameters for bumetanide were determined using noncompartmental methods (10). Systemic clearance was calculated as dose divided by the area under the curve (AUC) of the serum concentration versus time profile extrapolated to infinity. The AUC was determined by a combination of trapezoidal and log-trapezoidal methods. Renal clearance was calculated as the amount of unchanged drug excreted into the urine over 24 hours divided by the serum AUC to 24 hours. Nonrenal clearance was considered the difference between systemic clearance and renal clearance. Half-life was estimated as 0.693 divided by the terminal elimination rate constant determined from the serum concentration versus the time curve. The fraction of the dose excreted unchanged in the urine was the amount of drug measured in the urine to 24 hours divided by the 12-mg dose. The drug effect during the two treatments was assessed by net cumulative sodium excretion and by the relation between the urinary diuretic excretion rate and the natriuretic response. The latter method was used because the site of action of bumetanide as well as other loop diuretics is the lumen (urinary) side of the nephron (5, 11, 12), and previous studies have shown that this relationship accurately quantifies the pharmacodynamics of response to loop diuretics (6, 7, 13). Bumetanide's natriuretic efficiency both within and between treatments was calculated as net cumulative sodium divided by bumetanide excretion within a given interval. Statistical analyses of paired data were made using the twotailed Student /-test for paired data, with a P value less than 0.05 considered significant. Data are expressed as the mean ± standard deviation.
cumulative amounts of drug to be delivered to the urinary site of action (infusion, 912 ± 428 /xg; bolus, 944 ± 42 teg) (Table 2). For each treatment, roughly 8% of the administered dose was excreted unchanged in the urine during the 14 hours that the drug was in the body. Although identical amounts of drug were delivered to the urinary site of action during the 14-hour study period, natriuresis (assessed as either total Na + excretion or the increment of Na + excreted over baseline) was 25% greater with the continuous infusion (Table 3). This difference cannot be attributed to dissimilar prestudy conditions because sodium balance was measured before each treatment, and the baseline sodium excretions were similar before each phase of study (Table 3). Figure 2 depicts the relation between urinary bumetanide and response for the two treatments. The doseresponse curves for the bolus study show that the maximal response with the second bolus was less than after the first dose. Of note, the upper plateau of response was not clearly defined with either 6-mg bolus. In contrast to the bolus dosing scheme, bumetanide excretion rates during the infusion coincided with the lower aspect of the steep portion of the dose-response curves derived from bolus data. Therefore, the relation between urinary bumetanide excretion and response is
Results Pharmacokinetic parameters did not differ between the two treatment groups except that a predictably higher mean peak serum bumetanide concentration was obtained during intermittent bolus therapy (Table 2). Moreover, these parameters were similar to those reported previously in similar patient samples (5, 14-16). As we designed, the two dosing regimens resulted in very different serum concentration versus time profiles (Figure 1, top). This difference is also reflected in the relation between urinary drug excretion rate and time (Figure 1, bottom). During bolus therapy, most of the bumetanide was delivered into the urine during the time interval immediately after each dose. In contrast, the rate of bumetanide excretion remained constant during the infusion limb. Despite the different patterns of urinary drug excretion, both treatments caused similar 362
Figure 1. Serum and urinary bumetanide parameters versus time. Top panel. Serum bumetanide concentration-time profiles for the intermittent bolus doses and the continuous infusion. Bottom panel. Urinary excretion rate of bumetanide versus time for the two treatment groups. Values are means + SD.
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 03/06/2017
Table 3. Response to Bumetanide in Patients with Chronic Renal Insufficiency Variable
Bolus
Baseline sodium excretion, mmol/24 h 70 ± 30 Sodium excretion, 0 to 14 hour total, mmol* 229 ± 62 Sodium excretion, 0 to 14 hour net, mmoHi 188 ± 50 Urine volume, 0 to 14 hour, mL 3341 ± 603 Efficiency, mmoll^g 0.25 ± 0.15
Infusion 66 ± 27
P Value > 0.2
274 ± 77
0.001
236 ± 77
0.01
3734 ± 726 0.33 ± 0.21
0.001 0.01
* Total sodium excretion = mmol sodium excreted during the study interval. t Net sodium excretion = total sodium excretion minus baseline sodium excretion during the study interval.
generally the same in both the infusion and the bolus regimens. As a consequence, differences in pharmacodynamics cannot explain the difference in natriuretic response that occurred. The greater net sodium excretion during the continuous infusion, despite comparable amounts of bumetanide delivered to the drug's active site with both treatments, indicates that this pattern of drug delivery produced more efficient drug utilization than intermittent bolus dosing. The reduced natriuresis after the second bolus compared with that after the first and the decline in diuretic response during the continuous infusion despite constant amounts of bumetanide in the urine (Figure 3) show that acute diuretic tolerance developed during both treatments. These data indicate that compensatory renal mechanisms are invoked when even a relatively small amount of sodium is excreted. During bolus therapy, urinary excretion of sodium decreases to below the baseline level by 14 hours, indicating that net sodium retention occurs during this diuretic-free interval. In contrast, in the infusion experiment, even when the negative sodium balance was greater than with bolus therapy, continued diuretic excretion during this 12- to 14-hour period prevented net sodium retention. To explore further the magnitude of acute diuretic tolerance that developed within a treatment regimen, we quantitated drug efficiency at different time intervals (Table 4). The periods 0 to 6 hours and 6 to 14 hours were selected because these intervals represent times when equal amounts of drug are excreted. During both limbs of the study, whereas the quantity of bumetanide excreted during each interval was identical, the net sodium excretion in the second period was reduced by approximately 30%, indicating a decrease in drug efficiency and proving that acute diuretic tolerance developed with both treatment regimens. In addition to the superior efficacy with the continuous infusion, this treatment was also better tolerated. Three patients noted diffuse myalgias during bolus dosing. No patients reported symptoms during the continuous infusion, and all patients preferred this treatment regimen. Discussion Our results indicate that administering bumetanide (and probably other loop diuretics) by a continuous
infusion leads to a greater response with less risk of drug toxicity than does conventional intermittent bolus therapy. These findings are important because the goal of diuretic therapy in edematous states is to elicit effective diuresis while avoiding drug toxicity. Achieving this objective requires that a regimen be designed with considerations given to the disposition and elimination of the drug in the disease state and to the corresponding dose-response relationship (6, 7). Among the variables that can be altered when administering a diuretic are the dosing route, its magnitude, and its frequency. Continuous intravenous infusion of a loop diuretic appears to offer advantages over traditional intermittent bolus administration by minimizing the risk of extrarenal toxicity, providing a more efficient rate of drug delivery into the urine, blunting or delaying the onset of acute tolerance, and decreasing the diuretic-free interval in which compensatory sodium retention may occur. Further, a continuous intravenous infusion rate can be adjusted to allow some degree of regulation of the diuresis (17). We tested the effects of a continuous infusion in patients with severe CRI because these patients often fail to respond adequately to conventional intermittent bolus dose therapy. Lack of adequate response, in turn, may necessitate dialysis. The typical method used to overcome diuretic resistance in such patients is to administer a diuretic dose large enough to obtain a urinary drug concentration capable of inhibiting sodium transport in the thick ascending limb. Although large bolus doses of loop diuretic can be used successfully in CRI, there is a ceiling dose beyond which no further increase in diuresis is observed (18). A larger dose will only serve to increase the risk of extrarenal diuretic toxicity caused by high peak serum concentrations (19, 20). Even in situations where a safe and effective dose is found, the magnitude of natriuresis may remain inadequate because the large reduction in filtered load limits the response to loop diuretics in patients with severe
Figure 2. Bumetanide dose-response relationships. The relationship between bumetanide urinary excretion rate (dose) versus sodium excretion rate (response) is represented as the mean values per time interval for the intermittent bolus doses and continuous infusion treatments. The curves represent the best fit lines to the intermittent bolus data.
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 03/06/2017
363
CRI (5, 11, 18, 21). The standard solution to this problem is to administer boluses of the drug at frequent intervals. Our results, however, confirm that continuous infusion of a low but effective dose of a loop diuretic can increase overall response in CRI to a level above that seen with conventional therapy. Thus, continuous administration may be an important therapeutic option in appropriately selected patients. Our investigation is the first controlled study of which we are aware that directly compares conventional bolus diuretic therapy with a continuous infusion in a disease state associated with diuretic resistance. Few data exist regarding the efficacy of this form of therapy under any condition. In normal volunteers, it has been shown that a 4-hour infusion is more effective than a single bolus dose (22). Anecdotal reports have also described successful use of a continuous infusion of loop diuretics in patients with congestive heart failure (10, 23), with malignant ascites (24), and with acute renal failure occurring after cardiac surgery (25). These reports included no control groups, and the subjects were not evaluated in identical states of sodium balance. In another study in patients after cardiac surgery who had normal renal function, the diuretic response after continuous intravenous infusion was compared in a noncrossover fashion to an identical total dose given as two equal boluses (17). The two groups showed no difference in diuresis. The lack of a crossover design, possible inequalities in underlying sodium balance, and the absence of pharmacokinetic-pharmacodynamic analysis in these studies limit the ability to interpret them. We designed this study for patients to serve as their own controls and to compare each treatment during identical states of sodium balance. To simulate a clinical situation, we placed the patients on a standard low-salt diet (80 mmol of Na + /day), did not replace drug-induced volume loss during the diuretic phase of the study, and gave two bolus doses of bumetanide during that treatment limb. On the basis of our previous work and of conventional practice, we selected bolus doses of 6 mg. Although larger single doses might elicit a greater response, over one third of our patients experienced myalgias associated with drug toxicity at this dose, precluding the routine use of higher doses. This side effect Table 4. Development of Acute Diuretic Tolerance in Patients with Chronic Renal Insufficiency Method of Administration
0 to 6 Hours
6 to 14 Hours
Bolus Bumetanide excretion, ixg All ± 203 468 ± 222 Sodium excretion, net, mmol* 114 ± 43 74 ± 17 Efficiency, mmol/fig 0.29 ± 0.18 0.20 ± 0.13 Infusion Bumetanide excretion, fxg 432 ± 179 480 ± 258 Sodium excretion, net, mmol* 138 ± 56 97 ± 24 Efficiency, mmoll^g 0.38 ± 0.22 0.27 ± 0.19
P Value
> 0.2 0.03 0.001 > 0.2 0.02 0.001
* Net sodium excretion = total sodium excretion minus baseline sodium excretion.
364
has been well described in patients with renal insufficiency who were receiving large intravenous doses of bumetanide (26-28). In contrast to bolus therapy, the mean peak plasma drug concentration was six times less during the infusion, and no episodes of myalgias occurred during this treatment. Thus, simply to avoid toxicity, continuous infusion seems worth considering. Because a greater natriuresis occurred during the infusion than during bolus therapy, pharmacokinetic parameters were evaluated to determine if quantitative differences in drug disposition and elimination caused the difference. Because the total amount of drug excreted into the urine was identical in both treatment regimens, the superior efficacy of the infusion limb cannot be attributed to a difference in pharmacokinetics per se but, rather, is related to the pattern of diuretic delivery rate into the urine. The continuous infusion maintained a urinary drug delivery rate at a value that was consistently at the lower portion of the steep component of the dose-response curve. Consequently, a continuous submaximal diuresis was obtained with the infusion that, over time, produced a greater cumulative response than occurred after bolus therapy. One reason that a continuous infusion produces a greater response than bolus doses is that continuous infusion provides more efficient drug utilization. In other words, there is more sodium output relative to the excreted amount of diuretic drug. As shown in Tables 3 and 4, efficiency was greater with the infusion. Further assessment of diuretic efficiency between dosing intervals indicates that for both treatments, the efficiency was diminished in the latter half of therapy. This finding indicates that the volume loss is large enough to invoke compensatory sodium retention, producing diuretic tolerance even though patients with severe CRI have a markedly subnormal diuretic response. Although such tolerance has been previously shown in normal subjects (2), it has not been described in any disease state associated with diuretic resistance, including severe CRI. This property is an important consideration when diuretics are used clinically. Because diuretic-induced volume depletion occurs during therapy, the drug's effectiveness diminishes over time (29), and the ability to induce an effective diuresis in the volume-overloaded patient is impeded. It is possible that acute tolerance could have been completely avoided if the continuous infusion had been administered at a rate producing a submaximal diuresis less than that noted in our patients. This scenario, although speculative, might occur if the rate of plasma volume loss from the diuresis does not exceed the rate of plasma volume restoration from the fluid-expanded interstitial compartment. Further studies will be required to test this hypothesis. In addition to determining whether acute diuretic tolerance occurs in CRI, we evaluated drug-induced renal effects in the postdiuretic period. Studies in normal subjects have shown that unless dietary sodium restriction is invoked with drug therapy, compensatory renal sodium retention in the interdiuretic period can completely offset the large natriuretic effect from loop diuretics (1). These data illustrate the importance of restricting sodium intake in volume-overloaded patients receiving diuretics. The mechanism for this "braking
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 03/06/2017
that avoidance of the higher peak concentrations prevented the extrarenal toxicity observed with bolus treatment. This therapeutic modality should be considered a viable option in patients with CRI who require parenteral diuretic therapy. Its potential efficacy in other edematous states warrants further study. Presented in part at the 91st annual meeting of the American Society for Clinical Pharmacology and Therapeutics on 21 March 1990. Acknowledgments: The authors thank Dr. Steven O. Pastan for referring patients for study and Ms. Mary Pfister for preparing the manuscript. Grant Support: By National Institutes of Health grants R01 DK37994 and M01 RR750-16. Dr. Voelker is the recipient of a Clinical Associate Physician Award, M01 RR750-16. Dr. Rudy is a recipient of a Pharmaceutical Manufacturers Association Foundation Fellowship for Careers in Clinical Pharmacology Award. Roche Laboratories supplied bumetanide and Hoechst-Roussel supplied piretanide. Requests for Reprints: David W. Rudy, MD, Indiana University School of Medicine, Wishard Memorial Hospital, OPW 320, 1001 West 10th Street, Indianapolis, IN 46202. Current Author Addresses: Drs. Rudy, Voelker, and Brater, and Mrs. Greene and Mr. Esparza: Indiana University School of Medicine, Clinical Pharmacology Division, Wishard Memorial Hospital, OPW 320, 1001 West 10th Street, Indianapolis, IN 46202.
Figure 3. Net urinary sodium excretion rate versus time. Diuretic response to bumetanide over time is represented by the urinary sodium excretion rate above baseline. Values less than zero indicate net sodium retention. Top panel. Response following intermittent bolus doses of drug. Bottom panel. Response during continuous drug infusion. Values are means + SD.
phenomenon" is presumably the same as the cause of acute tolerance, neither of which has been precisely defined. Theoretically, the continuous inhibition of sodium transport in the thick ascending limb with a loop diuretic could blunt the effect. This process began to appear at hour 14 with bolus therapy as shown by the sodium excretion rates falling below baseline, that is, net sodium retention (Figure 3). Despite the greater overall sodium excretion that had occurred with the continuous infusion at this time, natriuresis continued because the diuretic was still being eliminated renally. These findings document that the braking phenomenon can occur in patients with severe CRI and that it can be blunted by continuous infusion of a loop diuretic. Patients with CRI often require repetitive intravenous bolus doses of loop diuretics to elicit a satisfactory diuretic response and thus are at increased risk for drug toxicity. We have shown that a continuous, low-dose intravenous infusion of the loop diuretic, bumetanide, produced a greater natriuresis than the same amount administered as two equally divided bolus doses. The infusion was accompanied by significantly lower peak serum drug concentrations. We presume
References 1. Wilcox CS, Mitch WE, Kelly RA, Skorecki K, Meyer TW, Friedman PA, et al. Response to furosemide. 1. Effects of salt intake and renal compensation. J Lab Clin Med. 1983;102:450-8. 2. Hammarlund MM, Odlind B, Paalzow LK. Acute tolerance to furosemide diuresis in humans. Pharmacokinetic-pharmacodynamic modeling. J Pharmacol Exp Ther. 1985;233:447-53. 3. Cook JA, Smith DE. Development of acute tolerance to bumetanide: bolus injection studies. Pharm Res. 1987;4:379-84. 4. Cook JA, Smith DE. Development of acute tolerance to bumetanide: constant-rate infusion studies. Pharm Res. 1988;5:86-91. 5. Voelker JR, Cartwright-Brown D, Anderson S, Leinfelder J, Sica DA, Kokko JP, et al. Comparison of loop diuretics in patients with chronic renal insufficiency. Kidney Int. 1987;32:572-8. 6. Brater DC. Resistance to diuretics: emphasis on a pharmacological perspective. Drugs. 1981;22:477-94. 7. Brater DC. Pharmacodynamic considerations in the use of diuretics. Annu Rev Pharmacol Toxicol. 1983;23:45-62. 8. Brater DC. Disposition and response to bumetanide and furosemide. Am J Cardiol. 1986;57:20A-25A. 9. Brater DC, Day B, Burdette A, Anderson S. Bumetanide and furosemide in heart failure. Kidney Int. 1984;26:183-9. 10. Gibaldi M, Perrier D. Pharmacokinetics. 2d ed. New York: Marcel Dekker; 1982:45-111. 11. Odlind B, Beermann B, Lindstrom B. Coupling between renal tubular secretion and effect of bumetanide. Clin Pharmacol Ther. 1983; 34:805-9. 12. Odlind B. Relation between renal tubular secretion and effects of five loop diuretics. J Pharmacol Exp Ther. 1979;211:238-44. 13. Smith D. Loop diuretics: effects in an eliminating organ. In: Smith RB, Kroboth PD, Juhl RP, eds. Pharmacokinetics and Pharmacodynamics, v 2. Current Problems, Potential Solutions. Cincinnati: H. Whitney Books; 1988:107-38. 14. Marcantonio LA, Auld WH, Murdoch WR, Purohit R, Skellern GG, Howes CA. The pharmacokinetics and pharmacodynamics of the diuretic bumetanide in hepatic and renal disease. Br J Clin Pharmacol. 1983;15:245-52. 15. Lau HS, Hyneck ML, Berrardi RR, Swartz RD, Smith DE. Kinetics, dynamics, and bioavailability of bumetanide in healthy subjects and patients with chronic renal failure. Clin Pharmacol Ther. 1986;39: 635-45. 16. Pentikainen PJ, Pasternack A, Lampainen E, Neuvonen PJ, Penttila A. Bumetanide kinetics in renal failure. Clin Pharmacol Ther. 1985; 37:582-8. 17. Copeland JG, Campbell DW, Plachetka JR, Salomon NW, Larson DF. Diuresis with continuous infusion of furosemide after cardiac surgery. Am J Surg. 1983;146:796-9. 18. Brater DC, Anderson SA, Brown-Cartwright D. Response to furosemide in chronic renal insufficiency: rationale for limited doses. Clin Pharmacol Ther. 1986;40:134-9. 19. Gallagher KL, Jones JK. Furosemide-induced ototoxicity. Ann Intern Med. 1979;91:744-5. 20. Schwartz GH, David DS, Riggio RR, Stenzel KH, Rubin AL. Ototoxicity induced by furosemide. N Engl J Med. 1970;282:1413-4.
1 September 1991 • Annals of Internal Medicine
• Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 03/06/2017
365
21. Brater DC, Voelker JR. Use of diuretics in patients with renal disease. In: Brenner BM, Stein JH, eds. Contemporary Issues in Nephrology: Pharmacotherapy in Renal Disease and Hypertension. New York: Churchill Livingstone; 1987:115-47. 22. Mclnnes GT, Thomson AH, Rapeport WG, Cook A. Loop diuretics—infusion or bolus? Clin Sci. 1982;63:33P-34P. 23. Lawson DH, Gray JM, Henry DA, Tillstone WJ. Continuous infusion of frusemide in refractory oedema. Br Med J. 1978;2:476. 24. Amiel SA, Blackburn AM, Rubens RD. Intravenous infusion of frusemide as treatment for ascites in malignant disease. Br Med J [Clin Res]. 1984;283:1041. 25. Krasna MJ, Scott GE, Scholz PM, Spotnitz AJ, Mackenzie JW, Penn F. Postoperative enhancement of urinary output in patients with
366
1 September 1991 • Annals
of Internal
Medicine
26.
27.
28. 29.
acute renal failure using continuous furosemide therapy. Chest. 1986;89:294-5. Berg KJ, Tromsdal A, Wideroe TE. Diuretic action of bumetanide in advanced chronic renal insufficiency. Eur J Clin Pharmacol. 1976; 9:265-75. Johnson CA, Grant KL, Madalon MG. Severe musculoskeletal syndrome in patients with renal failure and hypoalbuminemia receiving bumetanide. Clin Pharm. 1987;6:735-41. Cuthbert MF. Reports of adverse reactions with bumetanide. Postgrad Med J. 1975;51(Suppl 6):51-2. Grantham JJ, Chonko AM. The physiologic basis and clinical use of diuretics. In: Brenner BM, Stein JH, eds. Contemporary Issues in Nephrology: Sodium and Water Homeostasis. New York: Churchill Livingstone; 1978:178-211.
• Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 03/06/2017
• Objective: To test the hypothesis that a continuous, low-dose infusion of a loop diuretic is more efficacious and better tolerated than conventional intermittent bolus therapy in patients with severe chronic renal insufficiency (CRI). • Design: Randomized, crossover clinical trial with subjects serving as their own controls. • Setting: The General Clinical Research Center of Indiana University Hospital, Indianapolis, Indiana. • Patients: Eight adult volunteers with severe stable CRI (mean creatinine clearance, 0.28 mlVs; range, 0.15 to 0.47 mL7s) were recruited from the outpatient nephrology clinics of Indiana University Medical Center. • interventions: On admission, diuretic drugs were withdrawn and patients were equilibrated on an 80 mmol/d sodium, 60 mmol/d potassium metabolic diet. Patients were randomized to receive a 12-mg intravenous dose of bumetanide given either as two 6-mg bolus doses separated by 6 hours or as the same total dose administered as a 12-hour continuous infusion. When sodium balance was re-established, each patient was crossed over to the alternative study limb. All patients completed both phases of the study. • Measurements and Results: Comparable amounts of bumetanide appeared in the urine during the study period (infusion, 912 ± 428 jig; bolus, 944 ± 421 |xg; difference, 32 jig; 95% CI, - 16 ftg to 80 p,g, P = 0.16). The continuous infusion resulted in significantly greater net sodium excretion (infusion, 236 ± 77 mmol; bolus, 188 ± 50 mmol; difference, 48 mmol; CI, 16 mmol to 80 mmol, P- 0.01). No patient had episodes of druginduced myalgias during the continuous infusion compared with 3 of 8 patients with bolus therapy. • Conclusions: In patients with severe CRI, a continuous intravenous infusion of bumetanide is more effective and less toxic than conventional intermittent bolus therapy. Continuous administration will probably be useful in patients with severe CRI who have not achieved an adequate natriuresis or who show evidence of drug toxicity with standard diuretic dosing regimens. A similar benefit may occur in selected diuretic-resistant patients with cardiac or hepatic disease, and studies in these patients seem warranted.
Annals of Internal Medicine. 1991;115:360-366. From Indiana University Hospital, Indianapolis, Indiana. For current author addresses, see end of text. 360
Large doses of loop diuretics are typically administered to patients with severe chronic renal insufficiency (CRI) to circumvent diuretic resistance. Although a natriuretic response can be elicited in such patients, it remains subnormal because of the reduced filtration capacity of the diseased kidney. Although increasing the dosing frequency to several times per day can be used to enhance cumulative response, net sodium output frequently remains inadequate to meet clinical needs. One potential problem with administering repetitive large doses of a loop diuretic is the development of acute tolerance to the drug (1-4). Although never systematically examined in CRI, this phenomenon occurs in normal subjects receiving even small doses of diuretics. Acute diuretic tolerance is presumably caused by compensatory renal sodium retention because its appearance has been shown to occur in response to excessive sodium excretion and to correlate with both the rate and magnitude of natriuresis (2-4). Compensatory renal sodium retention can occur not only during the time of diuretic activity but can also continue well after the drug effect has subsided (1). The positive sodium balance caused by this adaptive response can further antagonize net diuretic effect. Because patients with severe CRI already have a reduced excretory capacity for electrolytes, any decrease in sodium elimination either during or after drug administration can diminish diuretic effectiveness enough to preclude their clinical usefulness. Because conventional loop diuretic dosing regimens in severe CRI often result in therapeutic failure, it is important to identify alternative dosing schemes that are more effective. A dosing pattern that produces a continuous but submaximal response may be one such strategy. This method would provide a constant delivery of drug to the active site so that a persistent natriuretic effect would occur. If compensatory sodium retention does occur in CRI, it is also possible that this problem could be minimized by a constant low-dose infusion, thus further enhancing the drug's overall therapeutic effectiveness. We would also expect this treatment modality to be safer because high and potentially toxic serum concentrations associated with intermittent bolus therapy would be avoided. A reduced but effective dose given by a continuous intravenous infusion might therefore result in greater cumulative sodium excretion and could be better tolerated than conventional repetitive high-dose administration. To test this hypothesis in patients with severe CRI, we compared the pharmacokinetic and dynamic determinants of response to
© 1991 American College of Physicians
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 03/06/2017
Table 1. Demographic Data Patient
Sex, Age
Weight
y
kg
Creatinine Clearance mLls
Diagnosis
1
F, 55
82.6
0.15
2
M, 52
73.1
0.38
3
M, 30
72.8
0.28
4 5 6
M, 55 M, 24 M, 31
102.2 91.2 58.4
0.47 0.30 0.27
7
M, 37
86.4
0.27
Polycystic kidney disease, hypertension Focal segmental glomerular sclerosis, hypertension, gout, coronary artery disease Diabetes mellitus, hypertension, type IV renal tubular acidosis Diabetes mellitus, hypertension Hypertension Diabetes mellitus, hypertension, type IV renal tubular acidosis Hypertension
8
F, 43
61.2
0.17
Hypertension, diabetes mellitus
intravenous bumetanide given as a continuous infusion and by conventional intermittent bolus therapy. Methods Eight adult patients with severe stable CRI participated in the study (Table 1). The patients were selected so that their concurrent medical conditions and treatment regimens would not affect diuretic responsiveness. The study protocol was approved by the Institutional Review Board of Indiana University. Experimental Protocol After obtaining informed consent, patients were hospitalized at the General Clinical Research Center. All diuretics were withdrawn; other medications were continued as previously prescribed. Patients were placed on an 80-mmol sodium and 60-mmol potassium per day metabolic diet. Sodium balance was documented to occur by day 6 on the basis of stable daily weights and 24-hour urinary electrolyte excretion. The study was done in a randomized crossover fashion comparing in each patient the natriuretic response to two bumetanide treatment strategies. Bumetanide was chosen for study because the more rapid elimination of this drug relative to furosemide in patients with severe CRI enabled us to conduct the study in a practical amount of time (5). After sodium balance was established, the patients were randomized to receive 12 mg of intravenous bumetanide administered either as two 6-mg bolus doses infused over 5 minutes and separated by 6 hours, or a 1-mg bolus loading dose given over 3 minutes followed by a 0.912 mg/h infusion for 12 hours. The bumetanide infusion was prepared by diluting 12 mg of bumetanide in 500 mL of 5% dextrose in water and was infused at a rate of 38 mL/h. Because we wanted to simulate the clinical situation, we incorporated more than one bolus dose. Further, we felt that an adequate comparison of regimens required large bolus doses that would reach the upper aspect of the dose-response curve and that the infusion dose should fall approximately midway on the steep portion of this curve. On the basis of previous data from our laboratory, we expected the respective doses to achieve these goals (5). The dosing interval of 6 hours for bolus therapy was chosen because this time represents the most frequent dosing interval used in clinical situations (5-8). A loading dose was used with the infusion to attain a steady-state plasma concentration rapidly. To ensure frequent spontaneous voiding, the patients drank water, 10 mL/kg body weight, 1 hour before drug dosing. Patients then drank 150 mL of water per hour for the next 14 hours. During the infusion this amount was reduced to 100 mL/h to compensate for the volume of intravenous 5% dex-
Medications
Clonidine, aspirin (low dose), hydralazine, propranolol Prazosin, furosemide, colchicine, diltiazem, aspirin (low dose) Clonidine, insulin, sodium bicarbonate Insulin, labetalol Verapamil, enalapril, furosemide Enalapril, insulin, furosemide Metoprolol, minoxidil, clonidine, furosemide, calcium carbonate Cimetidine, furosemide, metaclopromide, clonidine, calcium carbonate, nifedipine, insulin
trose in water. One hour after bumetanide was initially administered, patients were served breakfast and given their prescribed doses of other medications. Patients were continued on the metabolic diet throughout the day and had free access to distilled water. Patients remained supine during the diuretic phase of the study except to void. In order to simulate the clinical use of diuretics and to determine if acute diuretic tolerance occurs in patients with CRI, urinary losses of solutes were not replaced during the study period. To characterize pharmacokinetic-dynamic parameters during bolus therapy, serum samples were obtained for measurement of electrolytes, creatinine, and bumetanide concentrations at - 35, 0, 5, 15, 30, 45, 60, 90, 120, 180, 240, 360, 365, 375, 395, 425, 485, 605, 725, 845, and 1440 minutes. For the continuous infusion regimen, samples were obtained at - 35, 0, 120, 300, 420, 600, 720, 725, 735, 750, 780, 840, and 1440 minutes. Urine samples for measurements of volume, sodium, creatinine, and drug concentrations were obtained for each study day by spontaneous voiding at - 60, 0, 30, 60, 120, 180, 240, 300, 360, 390, 420, 480, 540, 600, 660, 720, 780, 840, 960, and 1440 minutes. To facilitate the return to sodium balance before the second study limb, on the day after the first study, all patients received intravenous 0.9% saline over 12 hours at a rate calculated to replace net sodium loss from the diuretic. Patients were maintained on the metabolic diet and monitored as before. Once sodium balance was re-established (after approximately 3 days), the second limb of the protocol was conducted. Laboratory Determinations Sodium concentrations were measured by flame photometry (Model 943; Instrumentation Laboratory, Lexington, Massachusetts) and creatinine concentrations by autoanalyzer using the Jaffe reaction (Technicon Autoanalyzer II, Elmsford, New York). Bumetanide was measured by a modification of a highpressure liquid chromatography method previously described (9). Briefly, aliquots of urine or serum samples plus piretanide as an internal standard were adjusted to pH 1.7 with 1.0 M phosphate buffer. Dichloromethane was added, the sample vortexed, and the aqueous phase removed by aspiration. The remaining organic layer was dried at low heat under a stream of nitrogen. The samples were reconstituted in mobile phase consisting of 60:40 acetonitrile and 0.01M sodium acetate adjusted to pH 3.3. The samples were injected onto a C8 5-ju,m column (Beckman, San Ramon, California) and eluted at a flow rate of one mL/min. Drug was detected by fluorescence spectroscopy (Hitachi model 650-15, Tokyo, Japan) using excitation and emission wavelengths of 338 nm and 440 nm, respectively. A standard curve was determined for each assay and the correlation coefficient (r2) always exceeded 0.990 over the concentration range of 5 to 2000 /xg/mL. The coefficients of vari-
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 03/06/2017
361
Table 2. Pharmacokinetics of Bumetanide in Patients with Chronic Renal Insufficiency Variable
Bolus
Half-life, min Systemic clearance, mLlmin Renal clearance, mLlmin Nonrenal clearance, mLlmin Bumetanide excretion, 0 to 14 hours /Jig fe,* % Maximum serum drug concentration,/x#/L
82 202 15.1 187
Infusion
±30 76 ± 24 ± 49 196 ±43 ± 6.1 14.4 ± 5.8 ± 51 179 ± 46
944 ± 421 912 ± 428 8.0 ± 3.6 7.9 ± 3.7 1118 ± 526 155 ± 76
P Value > > > >
0.2 0.2 0.2 0.2
0.2 > 0.2 0.001
* fe = fraction of administered dose excreted in the urine. ation for urine and serum samples were less than 10% and 5%, respectively.
Data Analysis The pharmacokinetic parameters for bumetanide were determined using noncompartmental methods (10). Systemic clearance was calculated as dose divided by the area under the curve (AUC) of the serum concentration versus time profile extrapolated to infinity. The AUC was determined by a combination of trapezoidal and log-trapezoidal methods. Renal clearance was calculated as the amount of unchanged drug excreted into the urine over 24 hours divided by the serum AUC to 24 hours. Nonrenal clearance was considered the difference between systemic clearance and renal clearance. Half-life was estimated as 0.693 divided by the terminal elimination rate constant determined from the serum concentration versus the time curve. The fraction of the dose excreted unchanged in the urine was the amount of drug measured in the urine to 24 hours divided by the 12-mg dose. The drug effect during the two treatments was assessed by net cumulative sodium excretion and by the relation between the urinary diuretic excretion rate and the natriuretic response. The latter method was used because the site of action of bumetanide as well as other loop diuretics is the lumen (urinary) side of the nephron (5, 11, 12), and previous studies have shown that this relationship accurately quantifies the pharmacodynamics of response to loop diuretics (6, 7, 13). Bumetanide's natriuretic efficiency both within and between treatments was calculated as net cumulative sodium divided by bumetanide excretion within a given interval. Statistical analyses of paired data were made using the twotailed Student /-test for paired data, with a P value less than 0.05 considered significant. Data are expressed as the mean ± standard deviation.
cumulative amounts of drug to be delivered to the urinary site of action (infusion, 912 ± 428 /xg; bolus, 944 ± 42 teg) (Table 2). For each treatment, roughly 8% of the administered dose was excreted unchanged in the urine during the 14 hours that the drug was in the body. Although identical amounts of drug were delivered to the urinary site of action during the 14-hour study period, natriuresis (assessed as either total Na + excretion or the increment of Na + excreted over baseline) was 25% greater with the continuous infusion (Table 3). This difference cannot be attributed to dissimilar prestudy conditions because sodium balance was measured before each treatment, and the baseline sodium excretions were similar before each phase of study (Table 3). Figure 2 depicts the relation between urinary bumetanide and response for the two treatments. The doseresponse curves for the bolus study show that the maximal response with the second bolus was less than after the first dose. Of note, the upper plateau of response was not clearly defined with either 6-mg bolus. In contrast to the bolus dosing scheme, bumetanide excretion rates during the infusion coincided with the lower aspect of the steep portion of the dose-response curves derived from bolus data. Therefore, the relation between urinary bumetanide excretion and response is
Results Pharmacokinetic parameters did not differ between the two treatment groups except that a predictably higher mean peak serum bumetanide concentration was obtained during intermittent bolus therapy (Table 2). Moreover, these parameters were similar to those reported previously in similar patient samples (5, 14-16). As we designed, the two dosing regimens resulted in very different serum concentration versus time profiles (Figure 1, top). This difference is also reflected in the relation between urinary drug excretion rate and time (Figure 1, bottom). During bolus therapy, most of the bumetanide was delivered into the urine during the time interval immediately after each dose. In contrast, the rate of bumetanide excretion remained constant during the infusion limb. Despite the different patterns of urinary drug excretion, both treatments caused similar 362
Figure 1. Serum and urinary bumetanide parameters versus time. Top panel. Serum bumetanide concentration-time profiles for the intermittent bolus doses and the continuous infusion. Bottom panel. Urinary excretion rate of bumetanide versus time for the two treatment groups. Values are means + SD.
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 03/06/2017
Table 3. Response to Bumetanide in Patients with Chronic Renal Insufficiency Variable
Bolus
Baseline sodium excretion, mmol/24 h 70 ± 30 Sodium excretion, 0 to 14 hour total, mmol* 229 ± 62 Sodium excretion, 0 to 14 hour net, mmoHi 188 ± 50 Urine volume, 0 to 14 hour, mL 3341 ± 603 Efficiency, mmoll^g 0.25 ± 0.15
Infusion 66 ± 27
P Value > 0.2
274 ± 77
0.001
236 ± 77
0.01
3734 ± 726 0.33 ± 0.21
0.001 0.01
* Total sodium excretion = mmol sodium excreted during the study interval. t Net sodium excretion = total sodium excretion minus baseline sodium excretion during the study interval.
generally the same in both the infusion and the bolus regimens. As a consequence, differences in pharmacodynamics cannot explain the difference in natriuretic response that occurred. The greater net sodium excretion during the continuous infusion, despite comparable amounts of bumetanide delivered to the drug's active site with both treatments, indicates that this pattern of drug delivery produced more efficient drug utilization than intermittent bolus dosing. The reduced natriuresis after the second bolus compared with that after the first and the decline in diuretic response during the continuous infusion despite constant amounts of bumetanide in the urine (Figure 3) show that acute diuretic tolerance developed during both treatments. These data indicate that compensatory renal mechanisms are invoked when even a relatively small amount of sodium is excreted. During bolus therapy, urinary excretion of sodium decreases to below the baseline level by 14 hours, indicating that net sodium retention occurs during this diuretic-free interval. In contrast, in the infusion experiment, even when the negative sodium balance was greater than with bolus therapy, continued diuretic excretion during this 12- to 14-hour period prevented net sodium retention. To explore further the magnitude of acute diuretic tolerance that developed within a treatment regimen, we quantitated drug efficiency at different time intervals (Table 4). The periods 0 to 6 hours and 6 to 14 hours were selected because these intervals represent times when equal amounts of drug are excreted. During both limbs of the study, whereas the quantity of bumetanide excreted during each interval was identical, the net sodium excretion in the second period was reduced by approximately 30%, indicating a decrease in drug efficiency and proving that acute diuretic tolerance developed with both treatment regimens. In addition to the superior efficacy with the continuous infusion, this treatment was also better tolerated. Three patients noted diffuse myalgias during bolus dosing. No patients reported symptoms during the continuous infusion, and all patients preferred this treatment regimen. Discussion Our results indicate that administering bumetanide (and probably other loop diuretics) by a continuous
infusion leads to a greater response with less risk of drug toxicity than does conventional intermittent bolus therapy. These findings are important because the goal of diuretic therapy in edematous states is to elicit effective diuresis while avoiding drug toxicity. Achieving this objective requires that a regimen be designed with considerations given to the disposition and elimination of the drug in the disease state and to the corresponding dose-response relationship (6, 7). Among the variables that can be altered when administering a diuretic are the dosing route, its magnitude, and its frequency. Continuous intravenous infusion of a loop diuretic appears to offer advantages over traditional intermittent bolus administration by minimizing the risk of extrarenal toxicity, providing a more efficient rate of drug delivery into the urine, blunting or delaying the onset of acute tolerance, and decreasing the diuretic-free interval in which compensatory sodium retention may occur. Further, a continuous intravenous infusion rate can be adjusted to allow some degree of regulation of the diuresis (17). We tested the effects of a continuous infusion in patients with severe CRI because these patients often fail to respond adequately to conventional intermittent bolus dose therapy. Lack of adequate response, in turn, may necessitate dialysis. The typical method used to overcome diuretic resistance in such patients is to administer a diuretic dose large enough to obtain a urinary drug concentration capable of inhibiting sodium transport in the thick ascending limb. Although large bolus doses of loop diuretic can be used successfully in CRI, there is a ceiling dose beyond which no further increase in diuresis is observed (18). A larger dose will only serve to increase the risk of extrarenal diuretic toxicity caused by high peak serum concentrations (19, 20). Even in situations where a safe and effective dose is found, the magnitude of natriuresis may remain inadequate because the large reduction in filtered load limits the response to loop diuretics in patients with severe
Figure 2. Bumetanide dose-response relationships. The relationship between bumetanide urinary excretion rate (dose) versus sodium excretion rate (response) is represented as the mean values per time interval for the intermittent bolus doses and continuous infusion treatments. The curves represent the best fit lines to the intermittent bolus data.
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 03/06/2017
363
CRI (5, 11, 18, 21). The standard solution to this problem is to administer boluses of the drug at frequent intervals. Our results, however, confirm that continuous infusion of a low but effective dose of a loop diuretic can increase overall response in CRI to a level above that seen with conventional therapy. Thus, continuous administration may be an important therapeutic option in appropriately selected patients. Our investigation is the first controlled study of which we are aware that directly compares conventional bolus diuretic therapy with a continuous infusion in a disease state associated with diuretic resistance. Few data exist regarding the efficacy of this form of therapy under any condition. In normal volunteers, it has been shown that a 4-hour infusion is more effective than a single bolus dose (22). Anecdotal reports have also described successful use of a continuous infusion of loop diuretics in patients with congestive heart failure (10, 23), with malignant ascites (24), and with acute renal failure occurring after cardiac surgery (25). These reports included no control groups, and the subjects were not evaluated in identical states of sodium balance. In another study in patients after cardiac surgery who had normal renal function, the diuretic response after continuous intravenous infusion was compared in a noncrossover fashion to an identical total dose given as two equal boluses (17). The two groups showed no difference in diuresis. The lack of a crossover design, possible inequalities in underlying sodium balance, and the absence of pharmacokinetic-pharmacodynamic analysis in these studies limit the ability to interpret them. We designed this study for patients to serve as their own controls and to compare each treatment during identical states of sodium balance. To simulate a clinical situation, we placed the patients on a standard low-salt diet (80 mmol of Na + /day), did not replace drug-induced volume loss during the diuretic phase of the study, and gave two bolus doses of bumetanide during that treatment limb. On the basis of our previous work and of conventional practice, we selected bolus doses of 6 mg. Although larger single doses might elicit a greater response, over one third of our patients experienced myalgias associated with drug toxicity at this dose, precluding the routine use of higher doses. This side effect Table 4. Development of Acute Diuretic Tolerance in Patients with Chronic Renal Insufficiency Method of Administration
0 to 6 Hours
6 to 14 Hours
Bolus Bumetanide excretion, ixg All ± 203 468 ± 222 Sodium excretion, net, mmol* 114 ± 43 74 ± 17 Efficiency, mmol/fig 0.29 ± 0.18 0.20 ± 0.13 Infusion Bumetanide excretion, fxg 432 ± 179 480 ± 258 Sodium excretion, net, mmol* 138 ± 56 97 ± 24 Efficiency, mmoll^g 0.38 ± 0.22 0.27 ± 0.19
P Value
> 0.2 0.03 0.001 > 0.2 0.02 0.001
* Net sodium excretion = total sodium excretion minus baseline sodium excretion.
364
has been well described in patients with renal insufficiency who were receiving large intravenous doses of bumetanide (26-28). In contrast to bolus therapy, the mean peak plasma drug concentration was six times less during the infusion, and no episodes of myalgias occurred during this treatment. Thus, simply to avoid toxicity, continuous infusion seems worth considering. Because a greater natriuresis occurred during the infusion than during bolus therapy, pharmacokinetic parameters were evaluated to determine if quantitative differences in drug disposition and elimination caused the difference. Because the total amount of drug excreted into the urine was identical in both treatment regimens, the superior efficacy of the infusion limb cannot be attributed to a difference in pharmacokinetics per se but, rather, is related to the pattern of diuretic delivery rate into the urine. The continuous infusion maintained a urinary drug delivery rate at a value that was consistently at the lower portion of the steep component of the dose-response curve. Consequently, a continuous submaximal diuresis was obtained with the infusion that, over time, produced a greater cumulative response than occurred after bolus therapy. One reason that a continuous infusion produces a greater response than bolus doses is that continuous infusion provides more efficient drug utilization. In other words, there is more sodium output relative to the excreted amount of diuretic drug. As shown in Tables 3 and 4, efficiency was greater with the infusion. Further assessment of diuretic efficiency between dosing intervals indicates that for both treatments, the efficiency was diminished in the latter half of therapy. This finding indicates that the volume loss is large enough to invoke compensatory sodium retention, producing diuretic tolerance even though patients with severe CRI have a markedly subnormal diuretic response. Although such tolerance has been previously shown in normal subjects (2), it has not been described in any disease state associated with diuretic resistance, including severe CRI. This property is an important consideration when diuretics are used clinically. Because diuretic-induced volume depletion occurs during therapy, the drug's effectiveness diminishes over time (29), and the ability to induce an effective diuresis in the volume-overloaded patient is impeded. It is possible that acute tolerance could have been completely avoided if the continuous infusion had been administered at a rate producing a submaximal diuresis less than that noted in our patients. This scenario, although speculative, might occur if the rate of plasma volume loss from the diuresis does not exceed the rate of plasma volume restoration from the fluid-expanded interstitial compartment. Further studies will be required to test this hypothesis. In addition to determining whether acute diuretic tolerance occurs in CRI, we evaluated drug-induced renal effects in the postdiuretic period. Studies in normal subjects have shown that unless dietary sodium restriction is invoked with drug therapy, compensatory renal sodium retention in the interdiuretic period can completely offset the large natriuretic effect from loop diuretics (1). These data illustrate the importance of restricting sodium intake in volume-overloaded patients receiving diuretics. The mechanism for this "braking
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 03/06/2017
that avoidance of the higher peak concentrations prevented the extrarenal toxicity observed with bolus treatment. This therapeutic modality should be considered a viable option in patients with CRI who require parenteral diuretic therapy. Its potential efficacy in other edematous states warrants further study. Presented in part at the 91st annual meeting of the American Society for Clinical Pharmacology and Therapeutics on 21 March 1990. Acknowledgments: The authors thank Dr. Steven O. Pastan for referring patients for study and Ms. Mary Pfister for preparing the manuscript. Grant Support: By National Institutes of Health grants R01 DK37994 and M01 RR750-16. Dr. Voelker is the recipient of a Clinical Associate Physician Award, M01 RR750-16. Dr. Rudy is a recipient of a Pharmaceutical Manufacturers Association Foundation Fellowship for Careers in Clinical Pharmacology Award. Roche Laboratories supplied bumetanide and Hoechst-Roussel supplied piretanide. Requests for Reprints: David W. Rudy, MD, Indiana University School of Medicine, Wishard Memorial Hospital, OPW 320, 1001 West 10th Street, Indianapolis, IN 46202. Current Author Addresses: Drs. Rudy, Voelker, and Brater, and Mrs. Greene and Mr. Esparza: Indiana University School of Medicine, Clinical Pharmacology Division, Wishard Memorial Hospital, OPW 320, 1001 West 10th Street, Indianapolis, IN 46202.
Figure 3. Net urinary sodium excretion rate versus time. Diuretic response to bumetanide over time is represented by the urinary sodium excretion rate above baseline. Values less than zero indicate net sodium retention. Top panel. Response following intermittent bolus doses of drug. Bottom panel. Response during continuous drug infusion. Values are means + SD.
phenomenon" is presumably the same as the cause of acute tolerance, neither of which has been precisely defined. Theoretically, the continuous inhibition of sodium transport in the thick ascending limb with a loop diuretic could blunt the effect. This process began to appear at hour 14 with bolus therapy as shown by the sodium excretion rates falling below baseline, that is, net sodium retention (Figure 3). Despite the greater overall sodium excretion that had occurred with the continuous infusion at this time, natriuresis continued because the diuretic was still being eliminated renally. These findings document that the braking phenomenon can occur in patients with severe CRI and that it can be blunted by continuous infusion of a loop diuretic. Patients with CRI often require repetitive intravenous bolus doses of loop diuretics to elicit a satisfactory diuretic response and thus are at increased risk for drug toxicity. We have shown that a continuous, low-dose intravenous infusion of the loop diuretic, bumetanide, produced a greater natriuresis than the same amount administered as two equally divided bolus doses. The infusion was accompanied by significantly lower peak serum drug concentrations. We presume
References 1. Wilcox CS, Mitch WE, Kelly RA, Skorecki K, Meyer TW, Friedman PA, et al. Response to furosemide. 1. Effects of salt intake and renal compensation. J Lab Clin Med. 1983;102:450-8. 2. Hammarlund MM, Odlind B, Paalzow LK. Acute tolerance to furosemide diuresis in humans. Pharmacokinetic-pharmacodynamic modeling. J Pharmacol Exp Ther. 1985;233:447-53. 3. Cook JA, Smith DE. Development of acute tolerance to bumetanide: bolus injection studies. Pharm Res. 1987;4:379-84. 4. Cook JA, Smith DE. Development of acute tolerance to bumetanide: constant-rate infusion studies. Pharm Res. 1988;5:86-91. 5. Voelker JR, Cartwright-Brown D, Anderson S, Leinfelder J, Sica DA, Kokko JP, et al. Comparison of loop diuretics in patients with chronic renal insufficiency. Kidney Int. 1987;32:572-8. 6. Brater DC. Resistance to diuretics: emphasis on a pharmacological perspective. Drugs. 1981;22:477-94. 7. Brater DC. Pharmacodynamic considerations in the use of diuretics. Annu Rev Pharmacol Toxicol. 1983;23:45-62. 8. Brater DC. Disposition and response to bumetanide and furosemide. Am J Cardiol. 1986;57:20A-25A. 9. Brater DC, Day B, Burdette A, Anderson S. Bumetanide and furosemide in heart failure. Kidney Int. 1984;26:183-9. 10. Gibaldi M, Perrier D. Pharmacokinetics. 2d ed. New York: Marcel Dekker; 1982:45-111. 11. Odlind B, Beermann B, Lindstrom B. Coupling between renal tubular secretion and effect of bumetanide. Clin Pharmacol Ther. 1983; 34:805-9. 12. Odlind B. Relation between renal tubular secretion and effects of five loop diuretics. J Pharmacol Exp Ther. 1979;211:238-44. 13. Smith D. Loop diuretics: effects in an eliminating organ. In: Smith RB, Kroboth PD, Juhl RP, eds. Pharmacokinetics and Pharmacodynamics, v 2. Current Problems, Potential Solutions. Cincinnati: H. Whitney Books; 1988:107-38. 14. Marcantonio LA, Auld WH, Murdoch WR, Purohit R, Skellern GG, Howes CA. The pharmacokinetics and pharmacodynamics of the diuretic bumetanide in hepatic and renal disease. Br J Clin Pharmacol. 1983;15:245-52. 15. Lau HS, Hyneck ML, Berrardi RR, Swartz RD, Smith DE. Kinetics, dynamics, and bioavailability of bumetanide in healthy subjects and patients with chronic renal failure. Clin Pharmacol Ther. 1986;39: 635-45. 16. Pentikainen PJ, Pasternack A, Lampainen E, Neuvonen PJ, Penttila A. Bumetanide kinetics in renal failure. Clin Pharmacol Ther. 1985; 37:582-8. 17. Copeland JG, Campbell DW, Plachetka JR, Salomon NW, Larson DF. Diuresis with continuous infusion of furosemide after cardiac surgery. Am J Surg. 1983;146:796-9. 18. Brater DC, Anderson SA, Brown-Cartwright D. Response to furosemide in chronic renal insufficiency: rationale for limited doses. Clin Pharmacol Ther. 1986;40:134-9. 19. Gallagher KL, Jones JK. Furosemide-induced ototoxicity. Ann Intern Med. 1979;91:744-5. 20. Schwartz GH, David DS, Riggio RR, Stenzel KH, Rubin AL. Ototoxicity induced by furosemide. N Engl J Med. 1970;282:1413-4.
1 September 1991 • Annals of Internal Medicine
• Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 03/06/2017
365
21. Brater DC, Voelker JR. Use of diuretics in patients with renal disease. In: Brenner BM, Stein JH, eds. Contemporary Issues in Nephrology: Pharmacotherapy in Renal Disease and Hypertension. New York: Churchill Livingstone; 1987:115-47. 22. Mclnnes GT, Thomson AH, Rapeport WG, Cook A. Loop diuretics—infusion or bolus? Clin Sci. 1982;63:33P-34P. 23. Lawson DH, Gray JM, Henry DA, Tillstone WJ. Continuous infusion of frusemide in refractory oedema. Br Med J. 1978;2:476. 24. Amiel SA, Blackburn AM, Rubens RD. Intravenous infusion of frusemide as treatment for ascites in malignant disease. Br Med J [Clin Res]. 1984;283:1041. 25. Krasna MJ, Scott GE, Scholz PM, Spotnitz AJ, Mackenzie JW, Penn F. Postoperative enhancement of urinary output in patients with
366
1 September 1991 • Annals
of Internal
Medicine
26.
27.
28. 29.
acute renal failure using continuous furosemide therapy. Chest. 1986;89:294-5. Berg KJ, Tromsdal A, Wideroe TE. Diuretic action of bumetanide in advanced chronic renal insufficiency. Eur J Clin Pharmacol. 1976; 9:265-75. Johnson CA, Grant KL, Madalon MG. Severe musculoskeletal syndrome in patients with renal failure and hypoalbuminemia receiving bumetanide. Clin Pharm. 1987;6:735-41. Cuthbert MF. Reports of adverse reactions with bumetanide. Postgrad Med J. 1975;51(Suppl 6):51-2. Grantham JJ, Chonko AM. The physiologic basis and clinical use of diuretics. In: Brenner BM, Stein JH, eds. Contemporary Issues in Nephrology: Sodium and Water Homeostasis. New York: Churchill Livingstone; 1978:178-211.
• Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 03/06/2017
