Clinical Science (1992) 83, 567-574 (Printed in Great Britain)
567
Effects of acute NaCI, KCI and KHCO, loads on renal electrolyte excretion in humans Marjolijn V A N BUREN, Ton J. RABELINK, Herman J. M.
VAN
RIJN and Hein A. KOOMANS
Department of Nephrology and Hypertension, University Hospital Utrecht, Utrecht, The Netherlands (Received 30 Apri1/26 June 1992; accepted 15 July 1992)
1. Potassium salts increase sodium excretion in humans. To define the role of the potassium ion in this effect, we compared the effects of equimolar single oral loads of 100mmol of NaCl and KCI on renal electrolyte excretion in seven healthy subjects. In a second group (0=7), we infused equimolar loads of NaCl or KCI (0.75mmol/kg in 2 h). 2. In both experiments the KCI load quickly increased plasma potassium and aldosterone concentrations and potassium and sodium excretion to a maximum by 2h after the load, whereas the NaCl load had no such effect. 3. In a third group (n=7) we compared the effects of single oral loads of KCI and KHCO, (1mmol/kg), to assess the role of the anion in the natriuretic effect of potassium salts. 4. KC1 and KHCO, transiently stimulated urinary excretion of potassium and sodium in an identical manner. also followed the changes in acid excretion me. Whereas both KCI and KHCO, loading decreased acid excretion, this effect was greater after KHCO, loading. Interestingly, acid excretion did not decrease further after the first collection hour after the potassium load, although the plasma potassium concentration was still increasing. 6. From these data we conclude (1) that increased excretion of sodium, potassium and chloride and decreased excretion of protons after administration of potassium salts are the specific effects of the potassium component; (2) that potassium also appears to ndary, indirect effects on proton excretion, nism of which remains to be clarified.
INTRODUCTION
An acute potassium load in humans causes a rapid increase in sodium and potassium excretion [1,2]. This effect of sodium excretion is well known, and in the past potassium salts have been used as natriuretic agents [3-61. However, the mechanism of this natriuresis is still unknown and its relation to the kaliuresis is unclear. It may be due to the
accompanying anion load, since the excretion of these anions, generally chloride, will require simultaneous excretion of cations. Since chloride is a mainly extracellular anion, the natriuresis may form the unspecific consequence of the associated extracellular volume expansion. However, potassium may also increase sodium excretion in a more specific manner, by decreasing tubular sodium reabsorption. If this is the case, the natriuretic effect of potassium could even be greater than that of sodium. Also, if the potassium ion (i.e. the rise in plasma potassium concentration) is crucial for this effect, the accompanying anion (chloride or bicarbonate) may have relatively little influence on the magnitude of the natriuresis and the associated changes in renal electrolyte excretion. ies we have also stressed the effects: whereas after a single potassium load, potassium and sodium excretion initially increase in parallel, the natriuresis stops after a few hours, but the kaliuresis continues [l, 21. This is probably the delayed result of aldosterone stimulation, which promotes reabsorption of sodium, delivered in increased amounts to the distal nephron, in favour of potassium secretion [1, 2, 7, 81. Such time effects have not been reported for the effect of a single potassium load on acid excretion. An increase in plasma potassium concentration has been shown to directly decrease acid excretion [9, lo]. On the other hand, both increased distal sodium delivery and aldosterone release will increase distal acid excretion [l 11. Therefore, the decrease in acid excretion may be of short duration, and not in phase with the rise in plasma potassium concentration. To further analyse these issues, we studied the effects of equimolar loads of KC1 and NaC1, and of equimolar loads of KC1 and KHCO,, on renal electrolyte and acid excretion by means of clearance techniques in humans. The loads were given orally, consequent to earlier studies in our department. However, to exclude the possibility that differences in response to oral NaCl and KCl loading are due to different rates of intestinal reabsorption, we also studied the effects of intravenous NaCl and KC1
Key words: acid excretion, clearance study, potassium, sodium excretion. Correspondence: Dr M. van Buren, Department of Nephrology and Hypertension, Room F03.226, University Hospital Utrecht, PO Box 85500, 3508 GA Utrecht, The Netherlands.
568
M. van Buren et al.
loading. Our specific questions were (1) whether the natriuretic effect of an acute KCl load is the consequence of the potassium or the chloride ion; (2) whether the acute effect of KCl and KHCOJ on sodium and potassium excretion would be quantitatively different; (3) how changes in plasma potassium concentration and acid excretion are related in time. METHODS
Studies were carried out in 21 healthy subjects using the protocols described below. Informed consent was obtained from all participants, and the studies were approved by the Hospital Ethical Committee for Studies in Humans. Study I: KCI versus NaCl orally
This study was performed to answer the question as to whether equimolar oral loads of KC1 and NaCl would have different effects on renal sodium and potassium excretion. To this effect, seven healthy subjects (three males, four females, age range 21-27 years) underwent three clearance experiments on separate days, one serving as a timecontrol experiment, one to assess the effects of a KCl load and one to assess the effects of a NaCl load. Since the effects of these loads on electrolyte excretion might be symmetrical (NaCl favouring sodium excretion over potassium excretion as much as KCl would favour potassium excretion over sodium excretion), we aimed at comparable baseline sodium and potassium excretions. Therefore, the subjects were given a diet containing equal amounts of sodium and potassium (both 100mmol daily) throughout the study period. This diet was supplied by the metabolic ward, where the subjects paid daily visits to bring their 24 h urine collections. The first clearance study was performed after at least 4 days on this diet, and the second and third clearance studies after intervals of at least 2 days each. The clearance studies were performed in a quiet room in the metabolic ward, with the subjects in the supine position. Hourly urine collections for the determination of sodium, potassium and chloride excretion were made for 8 h (from 08.00 to 16.00 hours). Blood samples for the determinations of plasma electrolyte concentrations were taken at the midpoint of each hour, and for the determination of plasma hormone levels at the midpoint of every second hour. Throughout the study the subjects drank of 200ml of water/h to ensure adequate urine flow. The first clearance study served as a timecontrol experiment. In the other two clearance studies either 75mmol of NaCl or 75mmol of KCl (random order) were given orally after the first 2 collection hours. On the day of the clearance experiments the subjects did not have breakfast, and consumed their lunch and dinner after the clearance studies.
Study II: KCI versus NaCl intravenously
A second group of seven subjects (three males, four females, age range 19-24 years) also underwent one time-control clearance experiment and two test experiments. These two test experiments concerned a NaCl load versus a KCl load intravenously to exclude the possibility that any differences in response to either load in study I were due to different rates of intestinal reabsorption of the loads. The results of study I (see below) did not show the presumed 'symmetrical' effect of NaCl and KCl on electrolyte excretion alluded to above, but instead showed significant effects after the KC1 load only. We therefore decided to abandon the 100mmol Na+/100mmol K + diet, and instead to prescribe a 150mmol Na+/100mmol K + diet, which would more closely resemble the normal sodium intake of the subjects. This diet was taken throughout the study, and the first clearance experiment (timecontrol study) was performed after at least 4 days on this diet. The two test experiments (random order) were performed after an interval of at least 2 days. Clearance studies were performed as described above, except that the hourly urine collections were made for 6h, from 09.00 until 15.00 hours (this adaptation compared with the protocol of study I was made since the changes in electrolyte excretion aimed at had occurred within this time period). In the test experiments either NaCl C0.05 mol/l NaCl in 1% (w/v) glucose] or KCl c0.05 mol/l KCl in 1% (w/v) glucose] was infused over 2 h, from 10.00 until 12.00 hours. The infusion rate was adjusted individually ( 15m12 h - kg - ) in order to administer 0.75mmol of Na+ or K + 2h-' kg-'. By following this design of a 2 h infusion, we tried to generate roughly similar changes in body NaCl and KCl load, and in plasma potassium concentration, as obtained after oral loading. In the time-control experiment 5% (w/v) glucose was infused at a rate of 3 m12 h-' kg-', and additional water was given to match the fluid load of the NaCl and KCl infusions. The KCl infusions were given directly into the subclavian vein by a catheter introduced through a cubital vein; the other infusions were performed via a cubital vein. On the study days the meals were taken as in study I.
'
'
Study 111: NaCl versus KHC03 orally
A third group of seven subjects (four males, three females, age range 21-26 years) underwent one timecontrol and two test clearance experiments. The two test experiments used an oral KCl load and a KHCOJ load to establish the specificity of potassium instead of the accompanying anion in the natriuretic and anti-aciduretic effects of a potassium load. The subjects consumed a diet containing 150mmol of Naf and 100mmol of K+/day, and underwent three clearance experiments of 6 h each.
K+ load and electrolyte excretion
The first clearance study served as a time-control study. In the two test experiments either KCl or KHC03 (1mmol/kg, random order) was taken orally after the first 60min collection period. In addition to the determination of sodium, potassium and chloride, these samples were also assessed for bicarbonate, ammonium, titratable acid and pH. These measurements were made immediately after urine collection. Laboratory methods
Plasma and urine samples were analysed for sodium and potassium by standard flame photometry. Chloride was measured by an autoanalyser. Plasma bicarbonate concentration and venous plasma pH were determined by a ABL-3 radiometer immediately after the blood samples were taken. Urinary bicarbonate, ammonia and titratable acid concentrations were assessed by potentiometric titration, and urinary pH was determined with a Radiometer PHM 60 pH-meter. Plasma aldosterone concentration, plasma renin activity and plasma atrial natriuretic peptide concentration were determined by standard radioimmunoassays.
569
Table I.Steady-state data on the days before the clearance studies. Values are means &$EM. Abbreviations: UN,+v, 24h urinary sodium excretion; UK+V, 24h urinary potassium excretion. UN,+V (mmol/24 h)
uK+v (mmoll24h)
Body wt. (kg)
Study I Control Oral KCI Oral NaCl
107.2 f9.3 II I.3 II.l 99.4 k 12.8
*
87.4k6.9 79.4k 12.0 83. I k0.9
65.7k3.1 65.6k2.9 65.9 f 2.6
Study II Control i.v. KCI i.v. NaCl
133.6+ 12.2 121.7+19.6 136.4+ 19.0
73.4k6.2 68.2+5.1 82.7 f 3.4
67.7k3.9 67.6k4.0 61.1 f4.0
Study 111 Control Oral KCI Oral KHCO,
134.8+ 11.9 129.4+ 10.2 122.6k 15.0
75.0k9.3 12.9 +7.8 73.4k5.0
72.3k2.6 71.9 k7.0 72.1 kl.7
Ti
I
Statistical analysis and calculations
Values are given as means fSEM. Plasma aldosterone concentrations and plasma renin activity were averaged after logarithmic transformation. Statistical comparisons were only made within each study, not between the three studies. Differences in baseline data and in cumulative electrolyte excretion between control and KC1-load and NaC1- or KHC03-load experiments were analysed by oneway analysis of variance. Statistical analysis of the clearance data was performed by two-way analysis of variance for repeated measures. If the variance ratio reached statistical significance, the differences alysed by the method of the least e, using the analysis of varia square error to calculate the least significant difference. RESULTS
Appropriate balance before the clearance studies was confirmed by the 24h excretion of sodium and potassium (Table 1). Study I: KCI versus NaCl orally
The main variables measured in this study are given as hourly data in Fig. 1. Baseline sodium and potassium excretions were comparable. Immediate increments in electrolyte excretion were seen only after the KC1 load, with peak excretion rates in the second hour after the load. Therefore, detailed data for the baseline collection hour and the second hour after the load are given in Table 2 and were used for statistical analysis. Compared with the time-
t
o'-'i
Time (h)
o
I
i i i ii i Time (h)
Fig. 1. Plasma potassium and aldosterone (ALDO) concentrations and urinary potassium (UK+V) and sodium (U,,+ V) excretion during study I (KCI versus NaCl orally). Plasma concentrations: each point represents the mean value of the blood samples taken at the midpoint llection period. Urine excretion: each point represents the rate over the previous 60min. 0-0, Timecontrol study; 0-0, KCI study; A-A, NaCl study. The ingestion of the test meal is denoted by the arrows. For statistical analysis, see the text and Table 2.
control experiment, the KCl load induced significant increments in plasma potassium and aldosterone concentrations and urinary potassium excretion (P
567
Effects of acute NaCI, KCI and KHCO, loads on renal electrolyte excretion in humans Marjolijn V A N BUREN, Ton J. RABELINK, Herman J. M.
VAN
RIJN and Hein A. KOOMANS
Department of Nephrology and Hypertension, University Hospital Utrecht, Utrecht, The Netherlands (Received 30 Apri1/26 June 1992; accepted 15 July 1992)
1. Potassium salts increase sodium excretion in humans. To define the role of the potassium ion in this effect, we compared the effects of equimolar single oral loads of 100mmol of NaCl and KCI on renal electrolyte excretion in seven healthy subjects. In a second group (0=7), we infused equimolar loads of NaCl or KCI (0.75mmol/kg in 2 h). 2. In both experiments the KCI load quickly increased plasma potassium and aldosterone concentrations and potassium and sodium excretion to a maximum by 2h after the load, whereas the NaCl load had no such effect. 3. In a third group (n=7) we compared the effects of single oral loads of KCI and KHCO, (1mmol/kg), to assess the role of the anion in the natriuretic effect of potassium salts. 4. KC1 and KHCO, transiently stimulated urinary excretion of potassium and sodium in an identical manner. also followed the changes in acid excretion me. Whereas both KCI and KHCO, loading decreased acid excretion, this effect was greater after KHCO, loading. Interestingly, acid excretion did not decrease further after the first collection hour after the potassium load, although the plasma potassium concentration was still increasing. 6. From these data we conclude (1) that increased excretion of sodium, potassium and chloride and decreased excretion of protons after administration of potassium salts are the specific effects of the potassium component; (2) that potassium also appears to ndary, indirect effects on proton excretion, nism of which remains to be clarified.
INTRODUCTION
An acute potassium load in humans causes a rapid increase in sodium and potassium excretion [1,2]. This effect of sodium excretion is well known, and in the past potassium salts have been used as natriuretic agents [3-61. However, the mechanism of this natriuresis is still unknown and its relation to the kaliuresis is unclear. It may be due to the
accompanying anion load, since the excretion of these anions, generally chloride, will require simultaneous excretion of cations. Since chloride is a mainly extracellular anion, the natriuresis may form the unspecific consequence of the associated extracellular volume expansion. However, potassium may also increase sodium excretion in a more specific manner, by decreasing tubular sodium reabsorption. If this is the case, the natriuretic effect of potassium could even be greater than that of sodium. Also, if the potassium ion (i.e. the rise in plasma potassium concentration) is crucial for this effect, the accompanying anion (chloride or bicarbonate) may have relatively little influence on the magnitude of the natriuresis and the associated changes in renal electrolyte excretion. ies we have also stressed the effects: whereas after a single potassium load, potassium and sodium excretion initially increase in parallel, the natriuresis stops after a few hours, but the kaliuresis continues [l, 21. This is probably the delayed result of aldosterone stimulation, which promotes reabsorption of sodium, delivered in increased amounts to the distal nephron, in favour of potassium secretion [1, 2, 7, 81. Such time effects have not been reported for the effect of a single potassium load on acid excretion. An increase in plasma potassium concentration has been shown to directly decrease acid excretion [9, lo]. On the other hand, both increased distal sodium delivery and aldosterone release will increase distal acid excretion [l 11. Therefore, the decrease in acid excretion may be of short duration, and not in phase with the rise in plasma potassium concentration. To further analyse these issues, we studied the effects of equimolar loads of KC1 and NaC1, and of equimolar loads of KC1 and KHCO,, on renal electrolyte and acid excretion by means of clearance techniques in humans. The loads were given orally, consequent to earlier studies in our department. However, to exclude the possibility that differences in response to oral NaCl and KCl loading are due to different rates of intestinal reabsorption, we also studied the effects of intravenous NaCl and KC1
Key words: acid excretion, clearance study, potassium, sodium excretion. Correspondence: Dr M. van Buren, Department of Nephrology and Hypertension, Room F03.226, University Hospital Utrecht, PO Box 85500, 3508 GA Utrecht, The Netherlands.
568
M. van Buren et al.
loading. Our specific questions were (1) whether the natriuretic effect of an acute KCl load is the consequence of the potassium or the chloride ion; (2) whether the acute effect of KCl and KHCOJ on sodium and potassium excretion would be quantitatively different; (3) how changes in plasma potassium concentration and acid excretion are related in time. METHODS
Studies were carried out in 21 healthy subjects using the protocols described below. Informed consent was obtained from all participants, and the studies were approved by the Hospital Ethical Committee for Studies in Humans. Study I: KCI versus NaCl orally
This study was performed to answer the question as to whether equimolar oral loads of KC1 and NaCl would have different effects on renal sodium and potassium excretion. To this effect, seven healthy subjects (three males, four females, age range 21-27 years) underwent three clearance experiments on separate days, one serving as a timecontrol experiment, one to assess the effects of a KCl load and one to assess the effects of a NaCl load. Since the effects of these loads on electrolyte excretion might be symmetrical (NaCl favouring sodium excretion over potassium excretion as much as KCl would favour potassium excretion over sodium excretion), we aimed at comparable baseline sodium and potassium excretions. Therefore, the subjects were given a diet containing equal amounts of sodium and potassium (both 100mmol daily) throughout the study period. This diet was supplied by the metabolic ward, where the subjects paid daily visits to bring their 24 h urine collections. The first clearance study was performed after at least 4 days on this diet, and the second and third clearance studies after intervals of at least 2 days each. The clearance studies were performed in a quiet room in the metabolic ward, with the subjects in the supine position. Hourly urine collections for the determination of sodium, potassium and chloride excretion were made for 8 h (from 08.00 to 16.00 hours). Blood samples for the determinations of plasma electrolyte concentrations were taken at the midpoint of each hour, and for the determination of plasma hormone levels at the midpoint of every second hour. Throughout the study the subjects drank of 200ml of water/h to ensure adequate urine flow. The first clearance study served as a timecontrol experiment. In the other two clearance studies either 75mmol of NaCl or 75mmol of KCl (random order) were given orally after the first 2 collection hours. On the day of the clearance experiments the subjects did not have breakfast, and consumed their lunch and dinner after the clearance studies.
Study II: KCI versus NaCl intravenously
A second group of seven subjects (three males, four females, age range 19-24 years) also underwent one time-control clearance experiment and two test experiments. These two test experiments concerned a NaCl load versus a KCl load intravenously to exclude the possibility that any differences in response to either load in study I were due to different rates of intestinal reabsorption of the loads. The results of study I (see below) did not show the presumed 'symmetrical' effect of NaCl and KCl on electrolyte excretion alluded to above, but instead showed significant effects after the KC1 load only. We therefore decided to abandon the 100mmol Na+/100mmol K + diet, and instead to prescribe a 150mmol Na+/100mmol K + diet, which would more closely resemble the normal sodium intake of the subjects. This diet was taken throughout the study, and the first clearance experiment (timecontrol study) was performed after at least 4 days on this diet. The two test experiments (random order) were performed after an interval of at least 2 days. Clearance studies were performed as described above, except that the hourly urine collections were made for 6h, from 09.00 until 15.00 hours (this adaptation compared with the protocol of study I was made since the changes in electrolyte excretion aimed at had occurred within this time period). In the test experiments either NaCl C0.05 mol/l NaCl in 1% (w/v) glucose] or KCl c0.05 mol/l KCl in 1% (w/v) glucose] was infused over 2 h, from 10.00 until 12.00 hours. The infusion rate was adjusted individually ( 15m12 h - kg - ) in order to administer 0.75mmol of Na+ or K + 2h-' kg-'. By following this design of a 2 h infusion, we tried to generate roughly similar changes in body NaCl and KCl load, and in plasma potassium concentration, as obtained after oral loading. In the time-control experiment 5% (w/v) glucose was infused at a rate of 3 m12 h-' kg-', and additional water was given to match the fluid load of the NaCl and KCl infusions. The KCl infusions were given directly into the subclavian vein by a catheter introduced through a cubital vein; the other infusions were performed via a cubital vein. On the study days the meals were taken as in study I.
'
'
Study 111: NaCl versus KHC03 orally
A third group of seven subjects (four males, three females, age range 21-26 years) underwent one timecontrol and two test clearance experiments. The two test experiments used an oral KCl load and a KHCOJ load to establish the specificity of potassium instead of the accompanying anion in the natriuretic and anti-aciduretic effects of a potassium load. The subjects consumed a diet containing 150mmol of Naf and 100mmol of K+/day, and underwent three clearance experiments of 6 h each.
K+ load and electrolyte excretion
The first clearance study served as a time-control study. In the two test experiments either KCl or KHC03 (1mmol/kg, random order) was taken orally after the first 60min collection period. In addition to the determination of sodium, potassium and chloride, these samples were also assessed for bicarbonate, ammonium, titratable acid and pH. These measurements were made immediately after urine collection. Laboratory methods
Plasma and urine samples were analysed for sodium and potassium by standard flame photometry. Chloride was measured by an autoanalyser. Plasma bicarbonate concentration and venous plasma pH were determined by a ABL-3 radiometer immediately after the blood samples were taken. Urinary bicarbonate, ammonia and titratable acid concentrations were assessed by potentiometric titration, and urinary pH was determined with a Radiometer PHM 60 pH-meter. Plasma aldosterone concentration, plasma renin activity and plasma atrial natriuretic peptide concentration were determined by standard radioimmunoassays.
569
Table I.Steady-state data on the days before the clearance studies. Values are means &$EM. Abbreviations: UN,+v, 24h urinary sodium excretion; UK+V, 24h urinary potassium excretion. UN,+V (mmol/24 h)
uK+v (mmoll24h)
Body wt. (kg)
Study I Control Oral KCI Oral NaCl
107.2 f9.3 II I.3 II.l 99.4 k 12.8
*
87.4k6.9 79.4k 12.0 83. I k0.9
65.7k3.1 65.6k2.9 65.9 f 2.6
Study II Control i.v. KCI i.v. NaCl
133.6+ 12.2 121.7+19.6 136.4+ 19.0
73.4k6.2 68.2+5.1 82.7 f 3.4
67.7k3.9 67.6k4.0 61.1 f4.0
Study 111 Control Oral KCI Oral KHCO,
134.8+ 11.9 129.4+ 10.2 122.6k 15.0
75.0k9.3 12.9 +7.8 73.4k5.0
72.3k2.6 71.9 k7.0 72.1 kl.7
Ti
I
Statistical analysis and calculations
Values are given as means fSEM. Plasma aldosterone concentrations and plasma renin activity were averaged after logarithmic transformation. Statistical comparisons were only made within each study, not between the three studies. Differences in baseline data and in cumulative electrolyte excretion between control and KC1-load and NaC1- or KHC03-load experiments were analysed by oneway analysis of variance. Statistical analysis of the clearance data was performed by two-way analysis of variance for repeated measures. If the variance ratio reached statistical significance, the differences alysed by the method of the least e, using the analysis of varia square error to calculate the least significant difference. RESULTS
Appropriate balance before the clearance studies was confirmed by the 24h excretion of sodium and potassium (Table 1). Study I: KCI versus NaCl orally
The main variables measured in this study are given as hourly data in Fig. 1. Baseline sodium and potassium excretions were comparable. Immediate increments in electrolyte excretion were seen only after the KC1 load, with peak excretion rates in the second hour after the load. Therefore, detailed data for the baseline collection hour and the second hour after the load are given in Table 2 and were used for statistical analysis. Compared with the time-
t
o'-'i
Time (h)
o
I
i i i ii i Time (h)
Fig. 1. Plasma potassium and aldosterone (ALDO) concentrations and urinary potassium (UK+V) and sodium (U,,+ V) excretion during study I (KCI versus NaCl orally). Plasma concentrations: each point represents the mean value of the blood samples taken at the midpoint llection period. Urine excretion: each point represents the rate over the previous 60min. 0-0, Timecontrol study; 0-0, KCI study; A-A, NaCl study. The ingestion of the test meal is denoted by the arrows. For statistical analysis, see the text and Table 2.
control experiment, the KCl load induced significant increments in plasma potassium and aldosterone concentrations and urinary potassium excretion (P
