- Email: [email protected]
A Randomized Controlled Trial of Fludrocortisone for the Treatment of Hyperkalemia in Hemodialysis Patients
Dialysis Therapies
A Randomized Controlled Trial of Fludrocortisone for the Treatment of Hyperkalemia in Hemodialysis Patients Mohammed O. Kaisar, MD, Kathryn J. Wiggins, MD, Joanne M. Sturtevant, BPharm, BSc, Carmel M. Hawley, MD, Scott B. Campbell, MD, PhD, Nicole M. Isbel, MD, David W. Mudge, MD, Andrew Bofinger, MD, PhD, James J.B. Petrie, MD, and David W. Johnson, MD, PhD ● Background: Previous small uncontrolled studies suggested that fludrocortisone may significantly decrease serum potassium concentrations in hemodialysis patients, possibly through enhancement of colonic potassium secretion. The aim of this study is to evaluate the effect of oral fludrocortisone on serum potassium concentrations in hyperkalemic hemodialysis patients in an open-label randomized controlled trial. Methods: Thirty-seven hemodialysis patients with predialysis hyperkalemia were randomly allocated to administration of either oral fludrocortisone (0.1 mg/d; n ⴝ 18) or no treatment (control; n ⴝ 19) for 3 months. The primary outcome measure was midweek predialysis serum potassium concentration, which was measured monthly during the trial. Prospective power calculations indicated that the study had an 80% probability of detecting a decrease in serum potassium levels of 0.7 mEq/L (0.7 mmol/L). Results: Baseline patient characteristics were similar, except for slightly longer total weekly dialysis hours in the fludrocortisone group (13.0 ⴞ 1.3 versus 12.1 ⴞ 1.0; P ⴝ 0.02). At the end of the study period, no significant changes in serum potassium concentrations were observed between the fludrocortisone and control groups (4.8 ⴞ 0.5 versus 5.2 ⴞ 0.7 mEq/L [mmol/L], respectively; P ⴝ 0.10). Similar results were obtained when changes in serum potassium levels over time were examined between the 2 arms by using repeated-measures analysis of variance, with or without adjustment for total weekly dialysis hours. Secondary outcomes, including predialysis mean arterial pressure, interdialytic weight gain, serum sodium level, and hospitalization for hyperkalemia, were not significantly different between groups. There were no observed adverse events. Conclusion: Administering fludrocortisone to hyperkalemic hemodialysis patients is safe and well tolerated, but does not achieve clinically important decreases in serum potassium levels. Am J Kidney Dis 47:809-814. © 2006 by the National Kidney Foundation, Inc. INDEX WORDS: Dialysis; fludrocortisone acetate; chronic kidney failure; potassium.
H
YPERKALEMIA IS a frequent and sometimes life-threatening problem in hemodialysis patients. Severe hyperkalemia (serum potassium concentration ⬎ 6.0 mEq/L [⬎6.0 mmol/L]) has been reported to occur in 10% of such patients immediately before their dialysis treatment.1 In a recent audit at our institution, severe hyperkalemia (potassium ⬎ 6.0 mEq/L) was noted in 12% of all patients, and potentially life-threatening hyperkalemia (potassium ⬎ 7.0 mEq/L) was observed in 7% of all patients (David W. Johnson and Mohammed O. Kaisar, unpublished observations). Factors predisposing these individuals to serious elevations in plasma potassium levels include oligoanuria, chronic metabolic acidosis, relative mineralocorticoid deficiency, suboptimal adherence to dietary potassium restriction, and medications (eg, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, -blockers, nonsteroidal antiinflammatory drugs, and unfractionated heparin).2 Unfortunately, currently available therapies for hyperkalemia are of limited effectiveness. Shifting potassium into cells through the use of bicarbonate and adrenergic agents has little or no effect.3 Low-potassium dialysate is helpful in
augmenting potassium removal, but has been shown to increase the risk for intradialytic cardiac arrest.4-6 Ion-exchange resins (Resonium A [sodium polystyrene sulfonate; Sanofi-Synthelabo, Sydney, New South Wales, Australia]) are expensive and unpalatable and have been associated with colonic necrosis when administered rectally.7 Many hemodialysis patients have considerable difficulty complying with dietary
From the Department of Renal Medicine, University of Queensland at Princess Alexandra Hospital, Brisbane, QLD, Australia. Received November 10, 2005; accepted in revised form January 25, 2006. Originally published online as doi:10.1053/j.ajkd.2006.01.014 on March 17, 2006. Support: None. Potential conflicts of interest: None. Address reprint requests to David W. Johnson, MD, PhD, Department of Renal Medicine, Level 2, Ambulatory Renal and Transplant Services Bldg, Princess Alexandra Hospital, Ipswich Road, Woolloongabba, Brisbane, QLD 4102, Australia. E-mail: [email protected] © 2006 by the National Kidney Foundation, Inc. 0272-6386/06/4705-0010$32.00/0 doi:10.1053/j.ajkd.2006.01.014
American Journal of Kidney Diseases, Vol 47, No 5 (May), 2006: pp 809-814
809
810
KAISAR ET AL
potassium restrictions and show little, if any, response to diuretics.2 During the last 2 decades, several pretest/ posttest case studies showed that the potent mineralocorticoid fludrocortisone acetate is capable of safely decreasing serum potassium concentrations in patients with end-stage renal failure,8-10 possibly by enhancement of gastrointestinal (predominantly colonic) potassium secretion.10 However, all 3 studies were uncontrolled, involved small patient numbers (range, 5 to 21 patients), and potentially were limited significantly by cointervention bias. For example, results may have been explained by covert changes in other treatment variables (eg, greater compliance with dietary potassium restriction) as a result of participation in a study of hyperkalemia (Hawthorn effect). Consequently, before recommendations for a change in clinical practice can be made, it is critically important to undertake a randomized controlled trial of fludrocortisone as a treatment for hyperkalemia in patients with end-stage renal failure. The primary aim of the present prospective randomized study is to compare the effects of oral fludrocortisone (0.1 mg/d) versus no treatment on midweek predialysis serum potassium concentrations in hemodialysis patients. METHODS
Study Subjects All adult patients receiving in-center hemodialysis at the Princess Alexandra Hospital (Brisbane, Australia) or Logan Hospital (Logan City, Australia) were invited to participate in the study if they met the following inclusion criteria: (1) 18 years or older, (2) midweek predialysis potassium concentration greater than 4.5 mEq/L (⬎4.5 mmol/L; the upper reference limit for the laboratory) and less than 7.0 mEq/L (⬍7.0 mmol/L) during the 3 months before study commencement, and (3) able to give informed consent. Exclusion criteria included a history of psychological illness or condition that interfered with the ability to understand or comply with the requirements of the study, pregnancy or breastfeeding, or known hypersensitivity to fludrocortisone. Informed consent was obtained from all patients before inclusion in the trial, and the study protocol was reviewed and approved by the Princess Alexandra Hospital Human Research Ethics Committee.
Study Protocol The study was a prospective, open-label, randomized, controlled trial. Patients were randomly assigned to administration of either fludrocortisone acetate (Bristol-MyersSquibb, Sydney, Australia; 0.1 mg/d) or no treatment for 3 months. Randomization was performed by using sequen-
tially numbered opaque sealed envelopes. The sequence of interventions was obtained from a computer-generated random number list with randomization blocks of 6. At the time of inclusion in the study, demographic and clinical data were recorded. All patients received standard advice regarding dietary potassium restriction from a dedicated renal dietician. Dialysate potassium concentration was not changed during the trial. Dosages of medications that potentially could alter serum potassium concentration (eg, diuretics, -blockers, angiotensin-converting enzyme inhibitors, angiotensin receptor antagonists) also were not modified. All patients were followed up for the entire 3-month study period. Serum potassium, sodium, chloride, urea, creatinine, and albumin were measured monthly. Predialysis and postdialysis body weight and predialysis blood pressure were recorded at every dialysis session. Medical reviews were performed weekly, and compliance was confirmed by tablet counting. The primary outcome measure was midweek predialysis serum potassium concentration at the end of the 3-month study period. Secondary outcome measures included relative changes in serum potassium concentrations during the study period, sodium polystyrene sulfonate powder prescription, hospitalizations because of hyperkalemia, average interdialytic weight gain, mean predialysis blood pressure, and serum sodium, chloride, and bicarbonate concentrations.
Statistical Analysis Normality of data was assessed by means of the Kolmogorov-Smirnov test using Lilliefors correction. Results are expressed as mean ⫾ SD, frequency (percentage), or median and interquartile range, depending on data type. Comparisons of demographic and clinical characteristics were performed by using Student t-test, chi-square test, or MannWhitney U test, depending on data distribution. Differences in midweek serum potassium concentrations between the treatment and control groups at individual times were assessed by using unpaired t-test. Changes in serum potassium concentrations between the 2 groups over time additionally were evaluated by means of repeated-measures analysis of variance. The association between randomization group and change in serum potassium levels between the beginning and end of the study was assessed further by using multiple linear regression. Differences in proportions of categorical secondary outcome events (eg, hospitalization rates or sodium polystyrene sulfonate prescription) between the control and treatment groups were evaluated by means of chi-square or Fisher exact test, as appropriate. All data were analyzed on an intention-to-treat basis using the statistical software package SPSS, version 13.0 (SPSS Inc, Chicago, IL). Prospective power calculations indicated the study had at least an 80% probability of detecting a clinically significant decrease in serum potassium levels of 0.7 mEq/L (0.7 mmol/L) if 18 patients were recruited into each of the 2 arms of the study, assuming ␣ ⫽ 0.05, 0.5-mEq/L population SD, 10% dropout rate, 20% Hawthorn effect, 20% noncompliance rate, and no interim analyses. P less than 0.05 is considered significant.
FLUDROCORTISONE THERAPY FOR HYPERKALEMIA
811
Fig 1. Numbers of patients who entered the study, were assigned to a study group, and completed the protocol.
RESULTS
Numbers of patients who entered the study, were assigned to a study group, and completed the protocol are shown in Fig 1. Thirty-seven of 41 patients who entered the study completed the study protocol and were included in the intentionto-treat analysis (18 patients, fludrocortisone group; 19 patients, control group). Baseline demographics of the 2 groups were similar, except for longer dialysis hours in the fludrocortisone cohort (Table 1). At the end of the 3-month study period, midweek predialysis serum potassium concentrations were not significantly different between the fludrocortisone and control groups (4.8 ⫾ 0.5 versus 5.2 ⫾ 0.7 mEq/L [4.8 ⫾ 0.5 versus 5.2 ⫾ 0.7 mmol/L],
respectively; P ⫽ 0.10; Fig 2). Secondary analysis of the primary outcome measure by using repeated-measures analysis of variance again did not show a significant change in serum potassium levels between the 2 arms during the 3-month period (P ⫽ 0.21). Adjustment for number of dialysis hours per week did not alter results (P ⫽ 0.17). Using paired t-tests, serum potassium concentrations did not change significantly during the study period in either group. Similar results were observed in post hoc analyses in which patients were subdivided into high- and low-potassium groups, using a predialysis serum potassium concentration of 5.5 mEq/L as an arbitrary cutoff value (data not shown). Serum potassium level response to fludrocortisone therapy did
812
KAISAR ET AL Table 1. Baseline Characteristics of the Study Population
Characteristic
Control (n ⫽ 19)
Fludrocortisone (n ⫽ 18)
Age (y) 64.9 ⫾ 13.9 56.1 ⫾ 16.9 Male sex 10 (53) 14 (78) White 18 (95) 18 (100) 28.4 ⫾ 10.3 28.6 ⫾ 7.0 Body mass index (kg/m2) Diabetes mellitus 6 (32) 7 (39) Ischemic heart disease 11 (58) 9 (50) Cerebrovascular disease 3 (16) 3 (17) Peripheral vascular disease 4 (21) 2 (11) Smoker Never 8 (42) 7 (39) Former 8 (42) 10 (56) Current 3 (16) 1 (6) Potassium-active drugs ACE inhibitor/angiotensin receptor blocker 7 (37) 7 (39) -Blocker 6 (32) 8 (44) Diuretic 4 (21) 3 (17) Cyclosporine 1 (5) 0 (0) Urea reduction ratio 0.72 ⫾ 0.05 0.68 ⫾ 0.07 Dialysis duration (h/wk) 12.1 ⫾ 1.0 13.0 ⫾ 1.3* Recirculation (%) 7.9 ⫾ 1.5 7.6 ⫾ 1.8 Access type Fistula 14 (74) 13 (72) Graft 3 (16) 5 (28) Catheter 2 (10) 0 (0) Blood flow (mL/min) 267 ⫾ 33 270 ⫾ 27 Dialysate potassium (mEq/L) 2 (1-2) 2 (2-2) Urine output (mL/d) 0 (0-200) 75 (0-237.5) Serum sodium (mEq/L) 138 ⫾ 3 140 ⫾ 2 Serum potassium (mEq/L) 5.3 ⫾ 0.7 5.1 ⫾ 0.5 Serum chloride (mEq/L) 97 ⫾ 3 99 ⫾ 1 Serum bicarbonate (mEq/L) 23 ⫾ 3 23 ⫾ 2 Serum glucose (mg/dL) 131.5 ⫾ 39.6 122.5 ⫾ 66.7 Serum albumin (g/dL) 3.93 ⫾ 0.29 4.11 ⫾ 0.32 Serum urea nitrogen (mg/dL) 65.0 ⫾ 14.0 67.8 ⫾ 13.2 Interdialytic weight gain (kg) 1.7 (1.2-2.1) 1.7 (1.2-2.3) Mean arterial pressure (mm Hg) 95 (87-100) 97 (88-109) NOTE. Results expressed as mean ⫾ SD, number (percent), or median (interquartile range). To convert potassium, sodium, chloride, and bicarbonate in mEq/L to mmol/L, multiply by 1; glucose in mg/dL to mmol/L, multiply by 0.05551; albumin in g/dL to g/L, multiply by 10; urea nitrogen in mg/dL to mmol/L, multiply by 0.357. Abbreviation: ACE, angiotensin-converting enzyme. *P ⬍ 0.05 versus controls.
not correlate significantly with predialysis serum potassium concentration (r2 ⫽ 0.18; P ⫽ 0.08). No difference was observed in 3-month serum potassium concentrations between the fludro-
cortisone and control groups in the subset of patients administered angiotensin-converting enzyme inhibitors or angiotensin receptor blockers (P ⫽ 0.95). Analyses of secondary outcome measures, including serum sodium, chloride, and bicarbonate levels; average interdialytic weight gain; and mean predialysis blood pressure, did not show significant differences between the 2 groups (Table 2). No patient was prescribed sodium polystyrene sulfonate before or throughout the study period, and no patient was hospitalized for hyperkalemia. A single patient in the control group died of sepsis late in the study period. Fludrocortisone was tolerated well by all subjects, and there were no observed adverse events attributable to this agent. DISCUSSION
Results of the present study show that administering fludrocortisone to hyperkalemic hemodialysis patients is safe and well tolerated, but does not achieve decreases in serum potassium levels that are either statistically significant or clinically important. To our knowledge, this is the only randomized controlled trial of fludrocortisone administration for the treatment of hyperkalemia in hemodialysis patients. Results of this study differ from those of 3 previous uncontrolled studies, which reported that fludrocortisone therapy decreased serum potassium concentrations in dialysis patients.8-10 Nyman et al8 used fludrocortisone to treat symptomatic hypotension in 5 anuric dialysis patients and observed a decrease in serum potassium levels from 4.7 to 4.0 mEq/L (4.7 to 4.0 mmol/L) during a treatment period of 3 to 9 months. Similarly, Singhal et al9 found that fludrocortisone (0.1 to 0.3 mg/d) decreased serum potassium levels by 0.7 mEq/L during a 3- to 6-month period in 21 hemodialysis subjects. Finally, Furuya et al10 studied 15 hemodialysis patients administered fludrocortisone in increasing dosages (0, 0.05, 0.1, 0.15, and 0.2 mg/d) during 5 consecutive 4-week periods. A concomitant decrease in salivary sodium-potassium ratio suggested that fludrocortisone decreased serum potassium levels in patients with end-stage renal failure by enhancement of gastrointestinal (predominantly colonic) potassium secretion through augmented sodium, potassium-adenosine triphos-
FLUDROCORTISONE THERAPY FOR HYPERKALEMIA
813
Fig 2. Midweek predialysis serum potassium concentrations in the fludrocortisone group (n ⴝ 18) and controls (n ⴝ 19). There were no statistically significant differences between the 2 groups at any time. To convert potassium in mmol/L to mEq/L, multiply by 1.
phatase activity. The apparent disparity in results between these studies and the present investigation may have occurred because the former investigations were uncontrolled and therefore potentially subject to cointervention bias. Moreover, the previous studies involved highly selected patients from single centers and specifically prohibited medications that might affect potassium metabolism, such that the effect of fludrocortisone on serum potassium concentrations in these patients may have overestimated the effect that would be observed in unselected hemodialysis populations. Alternatively, it is possible that the fludrocortisone dose used in our study (0.1 mg/d) was Table 2. Secondary Outcome Measures at the End of the 3-Month Study Period
Outcome Measure
Control (n ⫽ 19)
Fludrocortisone (n ⫽ 18)
Serum sodium (mEq/L) 137 ⫾ 3 138 ⫾ 4 Serum chloride (mEq/L) 96 ⫾ 3 97 ⫾ 2 Serum bicarbonate (mEq/L) 24 ⫾ 3 24 ⫾ 3 Serum glucose (mg/dL) 142.3 ⫾ 59.4 120.7 ⫾ 43.2 Interdialytic weight gain (kg) 1.6 (1.3-2.1) 1.8 (1.4-2.1) Mean arterial pressure (mm Hg) 91 (86-105) 102 (89-109)
P
0.28 0.29 0.63 0.21 0.49 0.22
NOTE. Results expressed as mean ⫾ SD or median (interquartile range).To convert sodium, chloride, and bicarbonate in mEq/L to mmol/L, multiply by 1; glucose in mg/dL to mmol/L, multiply by 0.05551.
insufficient to achieve an important decrease in serum potassium levels. Nevertheless, this dose was the most commonly reported safe and effective fludrocortisone dose used in previous studies of potassium-level lowering in dialysis patients.8-10 In the dose-ranging study by Furuya et al,10 0.15 mg/d of fludrocortisone decreased mean serum potassium levels from 5.6 to 4.9 mEq/L (5.6 to 4.9 mmol/L). Fludrocortisone doses as low as 0.05 mg/d were still significantly effective, whereas doses greater than 0.15 mg/d did not result in an additional decrease in serum potassium levels. The investigators concluded that a daily fludrocortisone dose of “0.05 and/or 0.10 mg should be sufficient to ameliorate hyperkalemia.”10 Moreover, doses greater than 0.2 mg/d have been associated with significant increases in blood pressure.8,10 Thus, although it cannot be excluded that a dose greater than 0.1 mg/d would have been equally safe and more effective, it does not seem likely. There were several other potential limitations of our study. The open-label design may have introduced bias. This was minimized by providing standard dietary advice to all patients and leaving dialysate potassium concentrations and dosages of potassium-active medications unchanged throughout the study period. Stable serum potassium concentrations over time in the control group also argued against a significant Hawthorn effect. Nevertheless, participants in the fludrocortisone arm conceivably could have increased their dietary potassium intake in response to the knowledge that they were being
814
administered a potentially effective potassiumlowering medication. Our study also examined midweek blood samples and it therefore is possible that different results may have been obtained if blood samples were drawn for serum potassium after the 3-day break because colonic potassium excretion may become a more important determinant of serum potassium levels during longer interdialytic intervals. We attempted to indirectly assess this possibility by dividing fludrocortisone-treated patients into groups with high and low predialysis potassium levels, but did not find appreciable differences in response. There also was a trend toward more men in the fludrocortisone group (P ⫽ 0.11), which potentially may have mitigated the effect of fludrocortisone on serum potassium levels because men generally have greater muscle mass and hence greater potassium stores. Furthermore, based on the prospective assumptions of a final sample size of 16 patients per group (after a 10% dropout rate) and an estimated Hawthorn effect of 20%, our study did not have sufficient power to detect a fludrocortisone-induced decrease in serum potassium concentration less than 0.7 mEq/L (⬍0.7 mmol/L). However, based on our actual final sample size of 18 patients in the fludrocortisone group and 19 patients in the control group (for whom a Hawthorn effect was not apparent), the study had 80% power to detect a fludrocortisone-induced decrease in serum potassium levels of 0.47 mEq/L. It is doubtful that a smaller change in serum potassium levels would be clinically important. In conclusion, results of this study suggest that administering fludrocortisone to hyperkalemic hemodialysis patients is safe and well tolerated, but does not achieve decreases in serum potassium levels that are either statistically significant or clinically important. Standard use of fludrocor-
KAISAR ET AL
tisone for the treatment of the hyperkalemia in hemodialysis patients currently cannot be recommended. ACKNOWLEDGMENT The authors thank the nursing staff of the Princess Alexandra Hospital Haemodialysis Unit for their invaluable assistance.
REFERENCES 1. Tzamaloukas AH, Avasthi PS: Temporal profile of serum potassium concentration in nondiabetic and diabetic outpatients on chronic dialysis. Am J Nephrol 7:101-109, 1987 2. Imbriano LJ, Durham JH, Maesaka JK: Treating interdialytic hyperkalemia with fludrocortisone. Semin Dial 16: 5-7, 2003 3. Blumberg A, Weidmann P, Shaw S, Gnadinger M: Effect of various therapeutic approaches on plasma potassium and major regulating factors in terminal renal failure. Am J Med 85:507-512, 1988 4. Morrison G, Michelson EL, Brown S, Morganroth J: Mechanism and prevention of cardiac arrhythmias in chronic hemodialysis patients. Kidney Int 17:811-819, 1980 5. Hou S, McElroy PA, Nootens J, Beach M: Safety and efficacy of low-potassium dialysate. Am J Kidney Dis 13:137143, 1989 6. Karnik JA, Young BS, Lew NL, et al: Cardiac arrest and sudden death in dialysis units. Kidney Int 60:350-357, 2001 7. Scott TR, Graham SM, Schweitzer EJ, Bartlett ST: Colonic necrosis following sodium polystyrene sulfonate (Kayexalate)-sorbitol enema in a renal transplant patient. Report of a case and review of the literature. Dis Colon Rectum 36:607-609, 1993 8. Nyman N, Mulgaonkar S, Walcer R, Viscuso R, Jacobs MG: Effects of a mineralocorticoid on extra-renal potassium homeostasis in anuric patients. Kidney Int 23:155A, 1983 (abstr) 9. Singhal PC, Desroches L, Mattana J, Abramovici M, Wagner JD, Maesaka JK: Mineralocorticoid therapy lowers serum potassium in patients with end-stage renal disease. Am J Nephrol 13:138-141, 1993 10. Furuya R, Kumagai H, Sakao T, Maruyama Y, Hishida A: Potassium-lowering effect of mineralocorticoid therapy in patients undergoing hemodialysis. Nephron 92:576-581, 2002
A Randomized Controlled Trial of Fludrocortisone for the Treatment of Hyperkalemia in Hemodialysis Patients Mohammed O. Kaisar, MD, Kathryn J. Wiggins, MD, Joanne M. Sturtevant, BPharm, BSc, Carmel M. Hawley, MD, Scott B. Campbell, MD, PhD, Nicole M. Isbel, MD, David W. Mudge, MD, Andrew Bofinger, MD, PhD, James J.B. Petrie, MD, and David W. Johnson, MD, PhD ● Background: Previous small uncontrolled studies suggested that fludrocortisone may significantly decrease serum potassium concentrations in hemodialysis patients, possibly through enhancement of colonic potassium secretion. The aim of this study is to evaluate the effect of oral fludrocortisone on serum potassium concentrations in hyperkalemic hemodialysis patients in an open-label randomized controlled trial. Methods: Thirty-seven hemodialysis patients with predialysis hyperkalemia were randomly allocated to administration of either oral fludrocortisone (0.1 mg/d; n ⴝ 18) or no treatment (control; n ⴝ 19) for 3 months. The primary outcome measure was midweek predialysis serum potassium concentration, which was measured monthly during the trial. Prospective power calculations indicated that the study had an 80% probability of detecting a decrease in serum potassium levels of 0.7 mEq/L (0.7 mmol/L). Results: Baseline patient characteristics were similar, except for slightly longer total weekly dialysis hours in the fludrocortisone group (13.0 ⴞ 1.3 versus 12.1 ⴞ 1.0; P ⴝ 0.02). At the end of the study period, no significant changes in serum potassium concentrations were observed between the fludrocortisone and control groups (4.8 ⴞ 0.5 versus 5.2 ⴞ 0.7 mEq/L [mmol/L], respectively; P ⴝ 0.10). Similar results were obtained when changes in serum potassium levels over time were examined between the 2 arms by using repeated-measures analysis of variance, with or without adjustment for total weekly dialysis hours. Secondary outcomes, including predialysis mean arterial pressure, interdialytic weight gain, serum sodium level, and hospitalization for hyperkalemia, were not significantly different between groups. There were no observed adverse events. Conclusion: Administering fludrocortisone to hyperkalemic hemodialysis patients is safe and well tolerated, but does not achieve clinically important decreases in serum potassium levels. Am J Kidney Dis 47:809-814. © 2006 by the National Kidney Foundation, Inc. INDEX WORDS: Dialysis; fludrocortisone acetate; chronic kidney failure; potassium.
H
YPERKALEMIA IS a frequent and sometimes life-threatening problem in hemodialysis patients. Severe hyperkalemia (serum potassium concentration ⬎ 6.0 mEq/L [⬎6.0 mmol/L]) has been reported to occur in 10% of such patients immediately before their dialysis treatment.1 In a recent audit at our institution, severe hyperkalemia (potassium ⬎ 6.0 mEq/L) was noted in 12% of all patients, and potentially life-threatening hyperkalemia (potassium ⬎ 7.0 mEq/L) was observed in 7% of all patients (David W. Johnson and Mohammed O. Kaisar, unpublished observations). Factors predisposing these individuals to serious elevations in plasma potassium levels include oligoanuria, chronic metabolic acidosis, relative mineralocorticoid deficiency, suboptimal adherence to dietary potassium restriction, and medications (eg, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, -blockers, nonsteroidal antiinflammatory drugs, and unfractionated heparin).2 Unfortunately, currently available therapies for hyperkalemia are of limited effectiveness. Shifting potassium into cells through the use of bicarbonate and adrenergic agents has little or no effect.3 Low-potassium dialysate is helpful in
augmenting potassium removal, but has been shown to increase the risk for intradialytic cardiac arrest.4-6 Ion-exchange resins (Resonium A [sodium polystyrene sulfonate; Sanofi-Synthelabo, Sydney, New South Wales, Australia]) are expensive and unpalatable and have been associated with colonic necrosis when administered rectally.7 Many hemodialysis patients have considerable difficulty complying with dietary
From the Department of Renal Medicine, University of Queensland at Princess Alexandra Hospital, Brisbane, QLD, Australia. Received November 10, 2005; accepted in revised form January 25, 2006. Originally published online as doi:10.1053/j.ajkd.2006.01.014 on March 17, 2006. Support: None. Potential conflicts of interest: None. Address reprint requests to David W. Johnson, MD, PhD, Department of Renal Medicine, Level 2, Ambulatory Renal and Transplant Services Bldg, Princess Alexandra Hospital, Ipswich Road, Woolloongabba, Brisbane, QLD 4102, Australia. E-mail: [email protected] © 2006 by the National Kidney Foundation, Inc. 0272-6386/06/4705-0010$32.00/0 doi:10.1053/j.ajkd.2006.01.014
American Journal of Kidney Diseases, Vol 47, No 5 (May), 2006: pp 809-814
809
810
KAISAR ET AL
potassium restrictions and show little, if any, response to diuretics.2 During the last 2 decades, several pretest/ posttest case studies showed that the potent mineralocorticoid fludrocortisone acetate is capable of safely decreasing serum potassium concentrations in patients with end-stage renal failure,8-10 possibly by enhancement of gastrointestinal (predominantly colonic) potassium secretion.10 However, all 3 studies were uncontrolled, involved small patient numbers (range, 5 to 21 patients), and potentially were limited significantly by cointervention bias. For example, results may have been explained by covert changes in other treatment variables (eg, greater compliance with dietary potassium restriction) as a result of participation in a study of hyperkalemia (Hawthorn effect). Consequently, before recommendations for a change in clinical practice can be made, it is critically important to undertake a randomized controlled trial of fludrocortisone as a treatment for hyperkalemia in patients with end-stage renal failure. The primary aim of the present prospective randomized study is to compare the effects of oral fludrocortisone (0.1 mg/d) versus no treatment on midweek predialysis serum potassium concentrations in hemodialysis patients. METHODS
Study Subjects All adult patients receiving in-center hemodialysis at the Princess Alexandra Hospital (Brisbane, Australia) or Logan Hospital (Logan City, Australia) were invited to participate in the study if they met the following inclusion criteria: (1) 18 years or older, (2) midweek predialysis potassium concentration greater than 4.5 mEq/L (⬎4.5 mmol/L; the upper reference limit for the laboratory) and less than 7.0 mEq/L (⬍7.0 mmol/L) during the 3 months before study commencement, and (3) able to give informed consent. Exclusion criteria included a history of psychological illness or condition that interfered with the ability to understand or comply with the requirements of the study, pregnancy or breastfeeding, or known hypersensitivity to fludrocortisone. Informed consent was obtained from all patients before inclusion in the trial, and the study protocol was reviewed and approved by the Princess Alexandra Hospital Human Research Ethics Committee.
Study Protocol The study was a prospective, open-label, randomized, controlled trial. Patients were randomly assigned to administration of either fludrocortisone acetate (Bristol-MyersSquibb, Sydney, Australia; 0.1 mg/d) or no treatment for 3 months. Randomization was performed by using sequen-
tially numbered opaque sealed envelopes. The sequence of interventions was obtained from a computer-generated random number list with randomization blocks of 6. At the time of inclusion in the study, demographic and clinical data were recorded. All patients received standard advice regarding dietary potassium restriction from a dedicated renal dietician. Dialysate potassium concentration was not changed during the trial. Dosages of medications that potentially could alter serum potassium concentration (eg, diuretics, -blockers, angiotensin-converting enzyme inhibitors, angiotensin receptor antagonists) also were not modified. All patients were followed up for the entire 3-month study period. Serum potassium, sodium, chloride, urea, creatinine, and albumin were measured monthly. Predialysis and postdialysis body weight and predialysis blood pressure were recorded at every dialysis session. Medical reviews were performed weekly, and compliance was confirmed by tablet counting. The primary outcome measure was midweek predialysis serum potassium concentration at the end of the 3-month study period. Secondary outcome measures included relative changes in serum potassium concentrations during the study period, sodium polystyrene sulfonate powder prescription, hospitalizations because of hyperkalemia, average interdialytic weight gain, mean predialysis blood pressure, and serum sodium, chloride, and bicarbonate concentrations.
Statistical Analysis Normality of data was assessed by means of the Kolmogorov-Smirnov test using Lilliefors correction. Results are expressed as mean ⫾ SD, frequency (percentage), or median and interquartile range, depending on data type. Comparisons of demographic and clinical characteristics were performed by using Student t-test, chi-square test, or MannWhitney U test, depending on data distribution. Differences in midweek serum potassium concentrations between the treatment and control groups at individual times were assessed by using unpaired t-test. Changes in serum potassium concentrations between the 2 groups over time additionally were evaluated by means of repeated-measures analysis of variance. The association between randomization group and change in serum potassium levels between the beginning and end of the study was assessed further by using multiple linear regression. Differences in proportions of categorical secondary outcome events (eg, hospitalization rates or sodium polystyrene sulfonate prescription) between the control and treatment groups were evaluated by means of chi-square or Fisher exact test, as appropriate. All data were analyzed on an intention-to-treat basis using the statistical software package SPSS, version 13.0 (SPSS Inc, Chicago, IL). Prospective power calculations indicated the study had at least an 80% probability of detecting a clinically significant decrease in serum potassium levels of 0.7 mEq/L (0.7 mmol/L) if 18 patients were recruited into each of the 2 arms of the study, assuming ␣ ⫽ 0.05, 0.5-mEq/L population SD, 10% dropout rate, 20% Hawthorn effect, 20% noncompliance rate, and no interim analyses. P less than 0.05 is considered significant.
FLUDROCORTISONE THERAPY FOR HYPERKALEMIA
811
Fig 1. Numbers of patients who entered the study, were assigned to a study group, and completed the protocol.
RESULTS
Numbers of patients who entered the study, were assigned to a study group, and completed the protocol are shown in Fig 1. Thirty-seven of 41 patients who entered the study completed the study protocol and were included in the intentionto-treat analysis (18 patients, fludrocortisone group; 19 patients, control group). Baseline demographics of the 2 groups were similar, except for longer dialysis hours in the fludrocortisone cohort (Table 1). At the end of the 3-month study period, midweek predialysis serum potassium concentrations were not significantly different between the fludrocortisone and control groups (4.8 ⫾ 0.5 versus 5.2 ⫾ 0.7 mEq/L [4.8 ⫾ 0.5 versus 5.2 ⫾ 0.7 mmol/L],
respectively; P ⫽ 0.10; Fig 2). Secondary analysis of the primary outcome measure by using repeated-measures analysis of variance again did not show a significant change in serum potassium levels between the 2 arms during the 3-month period (P ⫽ 0.21). Adjustment for number of dialysis hours per week did not alter results (P ⫽ 0.17). Using paired t-tests, serum potassium concentrations did not change significantly during the study period in either group. Similar results were observed in post hoc analyses in which patients were subdivided into high- and low-potassium groups, using a predialysis serum potassium concentration of 5.5 mEq/L as an arbitrary cutoff value (data not shown). Serum potassium level response to fludrocortisone therapy did
812
KAISAR ET AL Table 1. Baseline Characteristics of the Study Population
Characteristic
Control (n ⫽ 19)
Fludrocortisone (n ⫽ 18)
Age (y) 64.9 ⫾ 13.9 56.1 ⫾ 16.9 Male sex 10 (53) 14 (78) White 18 (95) 18 (100) 28.4 ⫾ 10.3 28.6 ⫾ 7.0 Body mass index (kg/m2) Diabetes mellitus 6 (32) 7 (39) Ischemic heart disease 11 (58) 9 (50) Cerebrovascular disease 3 (16) 3 (17) Peripheral vascular disease 4 (21) 2 (11) Smoker Never 8 (42) 7 (39) Former 8 (42) 10 (56) Current 3 (16) 1 (6) Potassium-active drugs ACE inhibitor/angiotensin receptor blocker 7 (37) 7 (39) -Blocker 6 (32) 8 (44) Diuretic 4 (21) 3 (17) Cyclosporine 1 (5) 0 (0) Urea reduction ratio 0.72 ⫾ 0.05 0.68 ⫾ 0.07 Dialysis duration (h/wk) 12.1 ⫾ 1.0 13.0 ⫾ 1.3* Recirculation (%) 7.9 ⫾ 1.5 7.6 ⫾ 1.8 Access type Fistula 14 (74) 13 (72) Graft 3 (16) 5 (28) Catheter 2 (10) 0 (0) Blood flow (mL/min) 267 ⫾ 33 270 ⫾ 27 Dialysate potassium (mEq/L) 2 (1-2) 2 (2-2) Urine output (mL/d) 0 (0-200) 75 (0-237.5) Serum sodium (mEq/L) 138 ⫾ 3 140 ⫾ 2 Serum potassium (mEq/L) 5.3 ⫾ 0.7 5.1 ⫾ 0.5 Serum chloride (mEq/L) 97 ⫾ 3 99 ⫾ 1 Serum bicarbonate (mEq/L) 23 ⫾ 3 23 ⫾ 2 Serum glucose (mg/dL) 131.5 ⫾ 39.6 122.5 ⫾ 66.7 Serum albumin (g/dL) 3.93 ⫾ 0.29 4.11 ⫾ 0.32 Serum urea nitrogen (mg/dL) 65.0 ⫾ 14.0 67.8 ⫾ 13.2 Interdialytic weight gain (kg) 1.7 (1.2-2.1) 1.7 (1.2-2.3) Mean arterial pressure (mm Hg) 95 (87-100) 97 (88-109) NOTE. Results expressed as mean ⫾ SD, number (percent), or median (interquartile range). To convert potassium, sodium, chloride, and bicarbonate in mEq/L to mmol/L, multiply by 1; glucose in mg/dL to mmol/L, multiply by 0.05551; albumin in g/dL to g/L, multiply by 10; urea nitrogen in mg/dL to mmol/L, multiply by 0.357. Abbreviation: ACE, angiotensin-converting enzyme. *P ⬍ 0.05 versus controls.
not correlate significantly with predialysis serum potassium concentration (r2 ⫽ 0.18; P ⫽ 0.08). No difference was observed in 3-month serum potassium concentrations between the fludro-
cortisone and control groups in the subset of patients administered angiotensin-converting enzyme inhibitors or angiotensin receptor blockers (P ⫽ 0.95). Analyses of secondary outcome measures, including serum sodium, chloride, and bicarbonate levels; average interdialytic weight gain; and mean predialysis blood pressure, did not show significant differences between the 2 groups (Table 2). No patient was prescribed sodium polystyrene sulfonate before or throughout the study period, and no patient was hospitalized for hyperkalemia. A single patient in the control group died of sepsis late in the study period. Fludrocortisone was tolerated well by all subjects, and there were no observed adverse events attributable to this agent. DISCUSSION
Results of the present study show that administering fludrocortisone to hyperkalemic hemodialysis patients is safe and well tolerated, but does not achieve decreases in serum potassium levels that are either statistically significant or clinically important. To our knowledge, this is the only randomized controlled trial of fludrocortisone administration for the treatment of hyperkalemia in hemodialysis patients. Results of this study differ from those of 3 previous uncontrolled studies, which reported that fludrocortisone therapy decreased serum potassium concentrations in dialysis patients.8-10 Nyman et al8 used fludrocortisone to treat symptomatic hypotension in 5 anuric dialysis patients and observed a decrease in serum potassium levels from 4.7 to 4.0 mEq/L (4.7 to 4.0 mmol/L) during a treatment period of 3 to 9 months. Similarly, Singhal et al9 found that fludrocortisone (0.1 to 0.3 mg/d) decreased serum potassium levels by 0.7 mEq/L during a 3- to 6-month period in 21 hemodialysis subjects. Finally, Furuya et al10 studied 15 hemodialysis patients administered fludrocortisone in increasing dosages (0, 0.05, 0.1, 0.15, and 0.2 mg/d) during 5 consecutive 4-week periods. A concomitant decrease in salivary sodium-potassium ratio suggested that fludrocortisone decreased serum potassium levels in patients with end-stage renal failure by enhancement of gastrointestinal (predominantly colonic) potassium secretion through augmented sodium, potassium-adenosine triphos-
FLUDROCORTISONE THERAPY FOR HYPERKALEMIA
813
Fig 2. Midweek predialysis serum potassium concentrations in the fludrocortisone group (n ⴝ 18) and controls (n ⴝ 19). There were no statistically significant differences between the 2 groups at any time. To convert potassium in mmol/L to mEq/L, multiply by 1.
phatase activity. The apparent disparity in results between these studies and the present investigation may have occurred because the former investigations were uncontrolled and therefore potentially subject to cointervention bias. Moreover, the previous studies involved highly selected patients from single centers and specifically prohibited medications that might affect potassium metabolism, such that the effect of fludrocortisone on serum potassium concentrations in these patients may have overestimated the effect that would be observed in unselected hemodialysis populations. Alternatively, it is possible that the fludrocortisone dose used in our study (0.1 mg/d) was Table 2. Secondary Outcome Measures at the End of the 3-Month Study Period
Outcome Measure
Control (n ⫽ 19)
Fludrocortisone (n ⫽ 18)
Serum sodium (mEq/L) 137 ⫾ 3 138 ⫾ 4 Serum chloride (mEq/L) 96 ⫾ 3 97 ⫾ 2 Serum bicarbonate (mEq/L) 24 ⫾ 3 24 ⫾ 3 Serum glucose (mg/dL) 142.3 ⫾ 59.4 120.7 ⫾ 43.2 Interdialytic weight gain (kg) 1.6 (1.3-2.1) 1.8 (1.4-2.1) Mean arterial pressure (mm Hg) 91 (86-105) 102 (89-109)
P
0.28 0.29 0.63 0.21 0.49 0.22
NOTE. Results expressed as mean ⫾ SD or median (interquartile range).To convert sodium, chloride, and bicarbonate in mEq/L to mmol/L, multiply by 1; glucose in mg/dL to mmol/L, multiply by 0.05551.
insufficient to achieve an important decrease in serum potassium levels. Nevertheless, this dose was the most commonly reported safe and effective fludrocortisone dose used in previous studies of potassium-level lowering in dialysis patients.8-10 In the dose-ranging study by Furuya et al,10 0.15 mg/d of fludrocortisone decreased mean serum potassium levels from 5.6 to 4.9 mEq/L (5.6 to 4.9 mmol/L). Fludrocortisone doses as low as 0.05 mg/d were still significantly effective, whereas doses greater than 0.15 mg/d did not result in an additional decrease in serum potassium levels. The investigators concluded that a daily fludrocortisone dose of “0.05 and/or 0.10 mg should be sufficient to ameliorate hyperkalemia.”10 Moreover, doses greater than 0.2 mg/d have been associated with significant increases in blood pressure.8,10 Thus, although it cannot be excluded that a dose greater than 0.1 mg/d would have been equally safe and more effective, it does not seem likely. There were several other potential limitations of our study. The open-label design may have introduced bias. This was minimized by providing standard dietary advice to all patients and leaving dialysate potassium concentrations and dosages of potassium-active medications unchanged throughout the study period. Stable serum potassium concentrations over time in the control group also argued against a significant Hawthorn effect. Nevertheless, participants in the fludrocortisone arm conceivably could have increased their dietary potassium intake in response to the knowledge that they were being
814
administered a potentially effective potassiumlowering medication. Our study also examined midweek blood samples and it therefore is possible that different results may have been obtained if blood samples were drawn for serum potassium after the 3-day break because colonic potassium excretion may become a more important determinant of serum potassium levels during longer interdialytic intervals. We attempted to indirectly assess this possibility by dividing fludrocortisone-treated patients into groups with high and low predialysis potassium levels, but did not find appreciable differences in response. There also was a trend toward more men in the fludrocortisone group (P ⫽ 0.11), which potentially may have mitigated the effect of fludrocortisone on serum potassium levels because men generally have greater muscle mass and hence greater potassium stores. Furthermore, based on the prospective assumptions of a final sample size of 16 patients per group (after a 10% dropout rate) and an estimated Hawthorn effect of 20%, our study did not have sufficient power to detect a fludrocortisone-induced decrease in serum potassium concentration less than 0.7 mEq/L (⬍0.7 mmol/L). However, based on our actual final sample size of 18 patients in the fludrocortisone group and 19 patients in the control group (for whom a Hawthorn effect was not apparent), the study had 80% power to detect a fludrocortisone-induced decrease in serum potassium levels of 0.47 mEq/L. It is doubtful that a smaller change in serum potassium levels would be clinically important. In conclusion, results of this study suggest that administering fludrocortisone to hyperkalemic hemodialysis patients is safe and well tolerated, but does not achieve decreases in serum potassium levels that are either statistically significant or clinically important. Standard use of fludrocor-
KAISAR ET AL
tisone for the treatment of the hyperkalemia in hemodialysis patients currently cannot be recommended. ACKNOWLEDGMENT The authors thank the nursing staff of the Princess Alexandra Hospital Haemodialysis Unit for their invaluable assistance.
REFERENCES 1. Tzamaloukas AH, Avasthi PS: Temporal profile of serum potassium concentration in nondiabetic and diabetic outpatients on chronic dialysis. Am J Nephrol 7:101-109, 1987 2. Imbriano LJ, Durham JH, Maesaka JK: Treating interdialytic hyperkalemia with fludrocortisone. Semin Dial 16: 5-7, 2003 3. Blumberg A, Weidmann P, Shaw S, Gnadinger M: Effect of various therapeutic approaches on plasma potassium and major regulating factors in terminal renal failure. Am J Med 85:507-512, 1988 4. Morrison G, Michelson EL, Brown S, Morganroth J: Mechanism and prevention of cardiac arrhythmias in chronic hemodialysis patients. Kidney Int 17:811-819, 1980 5. Hou S, McElroy PA, Nootens J, Beach M: Safety and efficacy of low-potassium dialysate. Am J Kidney Dis 13:137143, 1989 6. Karnik JA, Young BS, Lew NL, et al: Cardiac arrest and sudden death in dialysis units. Kidney Int 60:350-357, 2001 7. Scott TR, Graham SM, Schweitzer EJ, Bartlett ST: Colonic necrosis following sodium polystyrene sulfonate (Kayexalate)-sorbitol enema in a renal transplant patient. Report of a case and review of the literature. Dis Colon Rectum 36:607-609, 1993 8. Nyman N, Mulgaonkar S, Walcer R, Viscuso R, Jacobs MG: Effects of a mineralocorticoid on extra-renal potassium homeostasis in anuric patients. Kidney Int 23:155A, 1983 (abstr) 9. Singhal PC, Desroches L, Mattana J, Abramovici M, Wagner JD, Maesaka JK: Mineralocorticoid therapy lowers serum potassium in patients with end-stage renal disease. Am J Nephrol 13:138-141, 1993 10. Furuya R, Kumagai H, Sakao T, Maruyama Y, Hishida A: Potassium-lowering effect of mineralocorticoid therapy in patients undergoing hemodialysis. Nephron 92:576-581, 2002