Increased Secretion of Potassium in the Rectum of Humans With Chronic Renal Failure

Increased Secretion of Potassium in the Rectum of Humans With Chronic Renal Failure Rodolfo S. Martin, MD, Sergio Panese, MD, Mirta Virginillo, MD, Marisa Gimenez, MD, Maria Litardo, Elvira Arrizurieta, MD, and John P. Hayslett, MD • Previous studies have shown that the intestinal excretion of potassium increases in patients with renal failure and serves to guard against severe potassium retention. It is not known, however, whether the rise in intestinal potassium excretion occurs because of an increase in intestinal potassium secretion or a reduction in potassium absorption. Therefore, studies were performed to evaluate the rate of potassium secretion in the human rectum of controls and subjects with renal insufficiency, using a dialysis bag technique. The results demonstrate that net potassium secretion was increased in subjects with renal failure ( - 5.2 ± 0.9 J.lEq x min-1) compared with the control value of - 2.0 ± 0.4 J.lEq (P < .05). This change in intestinal secretion of potassium was shown to be independent of the passive effects of plasma potassium. The rise in potassium secretion, however, correlated directly with an increase in transepithelial potential difference (lumen-negative). Although plasma aldosterone levels were higher in patients than in controls, the scatter of data precludes an assessment of the role of aldosterone in the mechanism of the rise in potassium secretion. These data suggest that augmented intestinal potassium excretion in patients with chronic renal insufficiency is caused by increased net potassium secretion in the large intestine, and highlight the role of the intestine in maintaining potassium balance. © 1986 by the National Kidney Foundation, Inc. INDEX WORDS: Potassium secretion; human rectum; potassium adaptation; dialysis bag technique; renal insufficiency.

E

ARLIER STUDIES demonstrated that intestinal excretion of potassium (K) increases in severe renal insufficiency, and may play an important role in reducing retention of K. 1 The fraction of dietary K recovered in feces rose from 12 % in controls to approximately 35 % in patients with low rates of creatinine clearance. In these studies, however, it was not determined whether the increase in fecal K reflected an adaptive increase in intestinal secretion of K or reduced absorption of dietary K. In subsequent studies performed in a model of renal insufficiency in the rat, Bastl et al demonstrated a marked increase in colonic secretion of K.2 Moreover, this change in electrolyte transport was accompanied by an increase in the lumen-negative potential and in activity of Na-KATPase; these alterations in the colonic epithelium were similar to the changes that have also been demonstrated in animals with normal renal function fed a high K diet. 3 The present study was performed to determine whether an adaptive increase in intestinal K secretion occurs in humans with severe renal insufficiency. MATERIALS AND METHODS

Selection of Subjects Patients undergoing evaluation of functional gastrointestinal (GI) and hemorrhoidal disorders were invited to participate as

control subjects. Individuals with chronic renal failure were evaluated during periods of apparent stability of renal function, and while euvolemic by clinical criteria. All subjects received routine preparation for sigmoidoscopy, which included an enema the evening before and on the morning of the day of study. Informed consent was obtained from all subjects.

Determination of Electrolyte Transport and Electrical Potential Difference in Rectum Net electrolyte movement and transmucosal potential difference in the rectum were measured using a modification of the dialysis bag technique introduced by Edmonds. 4 Dialysis tubing (Fisher Scientific Products, Pittsburgh, Penn, cat no. 8667 A), 3.2 cm in diameter and with an initial length of approximately 14 cm, was closed at one end with double-O silk, and, after soaking, was turned inside out. This dialysis membrane permitted diffusion of substances with a molecular weight < 12,000. The bag was attached to three pieces of polyethylene (PE) tubing, 46 cm in length, held together by double-O silk as

From the Instituto de Investigaciones Medicas, University of Buenos Aires, the Children's Hospital of the City of Buenos Aires, Argentina, and Yale University, New Haven, Conn. Supported by the National Council for Scientific and Technical Research (Argentina), the University of Buenos Aires, and US Public Health Service Grant AM 1806/. Address reprint requests to Rodolfo S. Martin, MD, Instituto de Investigaciones Medicas, Donato Alvarez 3150, 1427 Buenos Aires, Argentina. © 1986 by the National Kidney Foundation, Inc. 0272-6386/86/0802-0005$03.00/0

American Journal of Kidney Diseases, Vol VIII, No 2 (August), 1986: pp 105-110

105

MARTIN ET AL

106

Dialysis Membrane

PE90

1------11.0 cm------1

Sample Removal shown in Fig I. The volume of the bag fabricated in this way was approximately 20 mL, and its length was 11.0 to 11.5 cm. 1\vo pieces of PE 90 tubing were used for filling the closed bag with solution and withdrawing samples, respectively, while one piece of PE 360 tubing, filled with 0.15 mollL NaCI in 4% agar, was used as the exploring electrode for estimation of transmural potential difference. At the time of study, the bag was inserted under direct vision through a sigmoidoscope, and then the sigmoidoscope was removed, leaving the empty bag lying in the rectum distal to the sigmoid flexure. The position of the bag was similar in all subjects, and the distal end of the bag was 4 cm from the rectal annulus. The bag was filled after insertion with Ringer's solution, containing polyethylene glycol (PEG 4000) as a marker for net water movement. The composition of the solution in mmollL was Na, 140; Cl, 120; K, 5; and HC0 3 , 25; the concentration of PEG 4000 was 2 g x L- I. Samples of I to 2 mL were removed for analysis at the beginning of the study and at the end of three consecutive 20-minute collection periods. After sample collection, the dead space in the tubing was filled with normal Ringer's solution. At the termination of the procedure, the bag was emptied and removed by gentle traction. To determine the transmural potential difference, the exploring electrode was connected to a DC millivoltmeter (8022 B Digital Multimeter, John Fluke Mfg Co, Inc, Washington, DC) via a silver-chloride half-cell. The indifferent electrode consisted of intracath tubing (Critikon,14-gauge) placed in an antecubical vein. This electrode was filled with 0.15 mol/L NaCI and was also connected to the millivoltmeter via a silverchloride half-cell and to an intravenous (IV) bottle containing isotonic saline, which ran at a slow rate during the procedure. Both half-cells were sterilized by gas autoclaving before use. Electrical symmetry between exploring and indifferent electrodes was tested in vitro before and after each experiment and agreed within 1 mV. Measurements of transmural potential difference were made at the midpoint of each collection period, and polarity was expressed with the indifferent electrode connected to ground.

Calculations The use of a bag made from dialysis membrane to estimate net electrolyte movement across the wall of human rectum is

Fig. 1. Construction of the dialysis bag and set-up for measuring the potential difference across the wall of human rectum.

based on the assumption that the intestinal segment assumes the shape of the fluid-filled bag, and its lining epithelium is separated from the membrane by a thin layer of fluid. Since the membrane is highly permeable to water and electrolytes, changes in concentration within the lumen of the bag reflect changes in mucosal fluid outside of the membrane, and can, therefore, be used to estimate net movement of water and electrolytes across the rectal epithelium. The calculation of net movement of water and electrolytes was based on formulas proposed by Edmonds.' The rate of net transport of water, in,.L x min-I, was estimated as:

x

VI

1

The term V I is the volume of test solution at the beginning of the collection period, and PEG I and PEG2 are the concentrations of PEG at the beginning and end of the same period. Time in minutes is given by t. Net movement of electrolytes, in I'Eq x min-I, was calculated from the formula: Jinet

(VI

x

II) - (V2

x 12)

where II and 12 are the concentrations of the electrolyte, in mEq/L, at the beginning and end of the collection period, and V2 is the volume of the test solution at the end of the collection period. The value of colonic clearance of K (CJ was calculated as Ck = J::" ... Pk> where Pk is plasma K concentration, and is expressed as mL x min-I.

Chemical Determinations and Statistical Methods The concentrations of sodium and K were determined by flame photometry, with an internal standard, and PEG was measured in duplicate by the method of Hyden. S Plasma aldosterone was determined by radioimmunoassay (RIA), as previously described. 6 All data are expressed as mean ± SEM and analyzed for statistical significance by Student's t test.

107

COLONIC POTASSIUM IN RENAL FAILURE

Table 1.

Comparison of Individual Data from Control Subjects and Patients with Renal Insufficiency,

J:

JHOH

Control Subjects

Without diarrhea RJ SN QG With diarrhea 8M RS MF FD Mean ± SEM Chronic renal failure RE GV CN W MJ DiMJ SS Mean ± SEM

oot

IJ
J~et

p.

PN.

PD (mV)

IJ
(mEq x L-')

P", (mg x dL-')

P Aldosterone (ng x dL-')

1.5 0.8 1.1

20.7 14.5 3.3

5.6 3.4 1.5

-1.7 -1.2 -1.8

-28

134

-62

147

3.4 4.9 3.7

42.6 33.3 15.0 23.1 21.8 ± 4.9

9.7 4.4 14.2 4.5 6.2 ± 1.6

-3.7 -1.1 -3.1 -1.3 -2.0 ± 0.4

-33 -37 -20 -36 ± 8.0

141 149 148 149 144.7 ± 2.5

3.2 3.7 3.7 2.5 3.6 ± 0.3

0.8 0.8 0.7 0.8 0.9 ± 0.1

27.5 17.0 7.5 9.5 12.3 ± 3.5

34.4 22.5 3.4 40.0 8.3 9.7 -1.7 16.7 ± 6.0

8.4 6.8 4.9 1.1 6.0 2.9 4.9 5.0 ± 0.9

-4.1 -5.5 -5.7 -3.4 -3.5 -3.6 -10.3 -5.2 ± 0.9'

137 132 135 140 143 135 141 137.6 ± 1.5

4.0 5.2 5.1 4.0 3.5 4.8 5.8 4.6 ± 0.3'

9.0 3.3 5.0 8.3 6.4 3.6 3.6 5.6 ± 0.9'

43.0 34.0 5.8 22.0 16.5

-27 -70 -56

-51.0

± 16.0

7.0 5.0

24.3

± 6.5

Note: ' indicates P < .05 or less compared to control. Abbreviations: P, plasma; PO, potential difference.

tion of 16.7 ± 6.0 p.L x min-I, and 5.0 ± 0.9 p.Eq x min-I, respectively, in patients with chronic renal failure were not different from control. However, as shown in Table I and Fig 2, the rate of net K secretion of 5.2 ± 0.9 p.Eq x min-I was more than 2.5-fold higher (P < .05) than control. The average rise in luminal K was 4.9 ± 1.1 mEq x L- I. When J~t was plotted against the reciprocal of plasma creatinine, used as an index of glomerular filtration rate, K secretion increased markedly in subjects with low rates of filtration «30 mL x min-I), as shown in Fig 3. This was in agreement with the data of Hayes et all that showed that the fraction of dietary K excreted in feces was increased in subjects with a low filtration rate.

RESULTS

The results of transport studies in individual control subjects and patients with chronic renal insufficiency (CRI) are shown in Table 1. Since some control subjects had a history of diarrhea, they are identified separately, but are considered together with subjects without a history of diarrhea in calculation of mean control values. The mean rates of net water and sodium absorption were 21.8 ± 4.9 p.L x min-I and 6.2 ± 1.6 p.Eq x min-I, respectively, and the rate of net K secretion was 2.0 ± 0.4 p.Eq X min-I in the control group. In these experiments, the concentration of K in the luminal solution rose a mean of 2.4 ± 0.7 mEq x L- t •

The mean rates of net water and sodium absorp~

p
6

iii

1.2

p
~

b

li

1.0 ~

o z

4

0.8 ... c:

I-ILI ILl ::l.. 3

ILl'

... .!:

511.

11.

E

O'e

Q~

0::

0.6

_


UI eX

U

Z

0.4 ~­

32

U

9

U

8 tiZ

~E

~..:

u-

1LIl;

Fig 2. Mean J~ (JlEq x min-1) and mean colonic clearance of potassium (mUmin). Ck • In control and experimental subjects. Groups with and without history of bowel disease were included together under the heading Normal Renal Function.

~

:;)

iii

0.2

~

.J

OL..--'-_....L.o

NORMAL CHRONIC RENAL RENAL FUNCTION FAILURE

NORMAL CHRONIC RENAL RENAL FUNCTION FAILURE

o

8

108

~ iii

MARTIN ET AL

levels of Pk on the K-diffusion gradient across the paracellular shunt is an unlikely explanation for the increased rate of K secretion in experimental subjects. Since recent studies in the rat colon, however, demonstrate that chronic K loading stimulates the mechanism of active K secretion, by an aldosterone-independent process,9 it is possible that adaptive epithelial changes may have occurred as a result of higher K levels in patients with renal insufficiency. Since the transmural potential difference was measured in some control and experimental subjects, the effect of potential difference on K secretion was also assessed. Figure 4 demonstrates a highly significant correlation between J~t and potential difference (y = O.04x ± 0.25, r = .78, P < .005). A higher luminal negativity could be expected to enhance K movement through the paracellular shunt and across the apical cell membrane. It seems probable, therefore, that the increase in K secretion in experimental subjects was influenced, at least in part, by a rise in the electrical driving force. Lastly, since rectal mucosa in humans is a target site for the action of aldosterone,lo plasma aldosterone levels were measured in subjects with CRI and in controls. Although the level of24.3 ± 6.5 ng x dL- I was higher in the experimental group than in controls (12.3 ± 3.5 ng x dL- I), as shown in Table 1, the difference was not statistically significant.

12

!II

~ oQ.

'z""

10

0.5:

E

•

8

Q~

!;i't6 It: _

fd.c:><

• •

!II

~:; 4

• •

Z

9

•

2

8 !;i z

o

0.3

0.6

• 0.9

1.2

1.5

II PLASMA CREATININE, mg/dl

Fig 3. Inverse correlation between J~et and the reciprocal of simultaneous plasma creatinine, used as an index of glomerular filtration rate.

Since acute increases in Pk have been shown to stimulate K secretion in experimental animals, in both the large intestine and distal convolution of the kidney,7.8 it was of interest that the chronic ambient K concentration in patients with CRI of 4.6 ± 0.3 mEq x L- I was significantly higher than the level in controls of 3.6 ± 0.3 (P < .05). To account for this difference in Pk levels between groups, the flux of potassium (J~et) was factored by Pk and is referred to as Ck , since previous studies demonstrated a linear relationship between these parameters during acute changes in Pk in the range of concentrations observed in these experiments.7.8 As shown in Fig 2, Ck in subjects with renal insufficiency (1.16 ± 0.13 mL x min-I) was significantly (P < .05) higher than the control value of 0.57 ± 0.14. These results indicate, therefore, that the effect of modestly elevated

DISCUSSION

Previous studies have shown that the human rectum and colon are capable of potassium secretion. ll . 12 The major finding in the present study is the demonstration that K secretion in the rectum

6 • (G.V.I

5 LL

o c 4

~!

iii Lff It: U

::L

•

3

•

w_-•

VI ~~

u., z-

2

9

.,..,-.

•

8 ~

•

w

z

o

-20

-40

COLONIC TRANSEPITHELIAL PO,

-60 mV

-80

Fig 4. The correlation of J~ and potential difference in control and experimental subjects. Patient GV departed from the whole group and was not included in calculation of the regression line.

109

COLONIC POTASSIUM IN RENAL FAILURE

increased markedly in patients with CRI. These results obtained in humans agree with studies performed in an experimental model of CRI in which net K secretion rose significantly along the length of the large intestine compared to control.2 The rise in rectal, and probably colonic, K secretion may play an important role in mitigating hyperkalemia in severe chronic renal failure in humans. As glomerular filtration is reduced, adaptive changes in the renal collecting duct system result in higher rates of K secretion in individual surviving nephrons\3 and, thereby, maintain normal or near normal Pk levels until filtration is reduced to approximately 25 to 30 mL x min-I. At lower rates of glomerular filtration, Pk levels rise, owing to apparent saturation of the renal secretory process for K. An alternate mechanism to provide a means to eliminate K, therefore, would guard against lethal hyperkalemia. The study of Hayes et all and the present study suggest that an adaptive increase in intestinal excretion of K, due to K secretion in the large intestine, is induced in patients with glomerular filtration rates of < 30 mL X min-I. The clinical importance of this adaptive change in intestinal transport may be substantial, since Hayes et all showed that approximately one third of dietary K was excreted in fecal material of patients with CRI, compared to approximately 12 % in controls. An increase in K secretion could involve either changes in the passive forces that affect the movement of K across the colonic mucosa or stimulation of active secretion. Regarding passive forces, experimental studies have shown that acute elevations of Pk result in marked increases in K secretion in the distal convolution and large intestine,7.8 owing to a rise in the concentration gradient of K across the paracellular shunt and/or transient activation of the K pump in the basolateral cell membrane. The present study shows that the increase in K secretion in patients with renal failure was not explained solely on this basis. Alternatively, chronic oral-K loading or induction of chronic hyperaldosteronism in the rat have been reported to induce adaptive epithelial changes in the large intestine that result in increased net K secretion. 14 The adaptive change is characterized by stimulation of active K secretion in association with an increase in the Na-K-ATPase activity, amplification of the area of the basolateral cell membrane, and a rise in transmural potential difference. Com-

parable changes in colonic function have also been reported by Bastl et al 2 in an experimental model of renal insufficiency, although in a recent report by Hene et al, 15 an increase in Na-K-ATPase activity was not found in the rectal mucosa of patients with renal insufficiency. An effort was made in the present study to determine the cause for increased K secretion in patients with renal insufficiency. It seems likely that the observed elevation in potential difference in patients played at least a partial role in the process. To determine, however, whether hyperaldosteronism was responsible for the changes in K secretion and transmural potential difference, plasma aldosterone levels were measured. The human rectum is a target site for the action of aldosterone,IO and elevated plasma aldosterone levels have been demonstrated in patients with advanced renal failure.l 6 . 17 Moreover, inhibition of the peripheral action of aldosterone in these patients leads to a rise in P k .18 Although the mean plasma aldosterone level was higher in patients than in controls in the present study, the difference was not statistically significant, perhaps owing to the scatter of data and relatively limited number of subjects. The question whether hyperaldosteronism plays a major role in modulating K secretion in chronic renal failure will require further studies in a larger number of patients. In summary, the present study shows that K secretion is significantly increased in the human rectum of patients with CRI. This process may provide an important means to prevent excessive K retention in patients with severe renal insufficiency. Since increased K secretion was not accounted for by elevated levels of P k , the mechanism for intestinal K adaptation appears to involve an alteration in the properties of epithelia that include, at least in part, a rise in luminal negativity. ACKNOWLEDGMENT The authors thank members of the Departments of Nephrology and Gastroenterology of the University of Buenos Aires for technical assistance and Dr Ignacio L. Reisin for helpful suggestions.

REFERENCES 1. Hayes CP Jr, McLeod MI, Robinson R: An extrarenal mechanism for the maintenance of potassium balance in severe chronic renal failure. Trans Assoc Am Physicians 80:207-216, 1967 2. Bastl C, Hayslett Jp, Binder HJ: Increased large intestinal

110

secretion of potassium in renal insufficiency. Kidney Int 12 :916, 1977 3. Fisher KA, Binder HJ, Hayslett JP: Potassium secretion by colonic mucosal cells after potassium adaptation. Am J Physiol 231:987-994, 1976 4. Edmonds CJ: Absorption of sodium and water by human rectum measured by a dialysis method. Gut 12:356-362, 1971 5. Hyden S: A turbidimetric method for the determination of higher polyethylene glycols in biological materials. Ann R Agric Coli Sweden 22:139-145, 1955 6. Ito T, Woo J, Haning R, et al: A radioimmunoassay for aldosterone in human peripheral plasma including the comparison of alternate techniques. J Clin Endocrinol Metab 34: 106111,1972 7. Hayslett Jp, Halevy J, Pace PE, et al: Demonstration of net potassium absorption in mammalian colon. Am J Physiol 242:G209-G214, 1982 8. Stanton BA, Giebisch GH: Potassium transport by the distal tubule: Effects of potassium loading. Am J Physiol 243:F487-F493, 1982 9. Foster ES, Jones WJ, Hayslett Jp, et al: Role of aldosterone and dietary potassium in potassium adaptation in the distal colon of the rat. Gastroenterology 88:41-46,1985 10. Edmonds CJ, Godfrey RC: Measurement of electrical potentials of the human rectum and pelvic colon in normal and aldosterone-treated patients. Gut 11:330-337, 1970

MARTIN ET AL

11. Salas-Coli CA, Kermode JC, Edmonds CJ: Potassium transport across the distal colon in man. Clin Sci 51:287-296, 1976 12. Levitan M, Ingelfinger FJ: Effect of D-aldosterone on salt and water absorption from the intact human colon. J Clin Invest 44:801-808, 1965 13. Schon DA, Silva P, Hayslett JP: Mechanism of potassium excretion in renal insufficiency. Am J Physiol 227: 13231330, 1974 14. HayslettJp, Binder JH: Mechanism of potassium adaptation. Am J Physiol 243(12):FI03-FI12, 1982 15. Hene RJ, Boer P, Koomans HA, et al: Sodium potassium ATPase activity in human rectal mucosa with and without renal insufficiency. Am J Kidney Dis 5: 177-181, 1985 16. Hene RJ, Boer P, Koomans HA, et al: Plasma aldosterone concentrations in chronic renal disease. Kidney Int 21 :98101, 1982 17. Hene RJ, Boer P, RoofC, et al: Relation between plasma aldosterone concentration and renal handling of sodium and potassium, in particular in patients with chronic renal failure. Nephron 37:94-99, 1984 18. Ber! T, Katz FH, Henrich WL, et al: Role of aldosterone in the control of sodium excretion in patients with advanced chronic renal failure. Kidney Int 14:228-235, 1978