ClfnicalScience(1979) 51,225-23 I
Sodium excretion in man, and adaptation to a low-sodium diet: effect of intravenous sodium chloride
D. GORDON A N D W . S. PEART Medical Unit, St Mary's Hospital Medical School, London
(Received 22 December 1978; accepted 18 April 1979) dose of sodium chloride is excreted more rapidly than an equal intravenous load. This suggests that there may be a gastrointestinal or portal monitor of sodium intake. Similar results have been found in the rabbit (Lennane, Peart, Carey & Shaw, 1975b) and there is evidence for osmoreceptors in the portal circulation in the rat (Haberich, Aziz & Nowacki, 1965; Adachi, Niijima & Jacobs, 1976) and dog (Daly, Roe & Horrocks, 1967), although Kapteina, Motz, Schwartz-Porsche & Gauer (1978) found no evidence for such portal osmoreceptors in the dog. There is evidence for receptors responding to hypertonic sodium chloride, but not to hypertonic sucrose, in the cat (Passo, Thornborough & Rothballer, 1973). If the gastrointestinal or portal monitor proposed by Lennane et al. (l975a) is an essential part of the control mechanism for the rapid reduction in the renal excretion of sodium that follows a change to a low-sodium diet, then the natriuresis that normally follows infusion of sodium chloride might be diminished or absent if the infusion is given at the time of a reduction in oral sodium intake, because of the influence of the gastrointestinal sodium monitor. It is important to establish how much of the same intravenous load is retained if there is no reduction in oral sodium intake. These experiments were therefore designed to test the renal response to intravenous sodium at the time of change to a lowsodium diet, or when the intravenous sodium was given in addition to a diet providing a constant intake of sodium similar to that seen on a normal free diet. The changes during intravenous sodium chloride infusion in a number of non-hormonal blood factors possibly related to sodium excretion were also studied.
Summary 1. The aim of this study was to test whether a postulated gastrointestinal or portal monitor of sodium intake plays any part in adjusting renal sodium excretion when dietary sodium is reduced. 2. Normal male subjects were given 50 mmol of sodium chloride intravenously three times daily for 3 days to replace or to supplement a constant oral intake of sodium chloride. 3. When oral sodium chloride was replaced with intravenous sodium chloride, renal sodium excretion remained constant. 4. When oral sodium chloride was kept constant, sodium administered as intravenous sodium chloride was promptly excreted in three out of four subjects. There was a delay in the increase in sodium excretion in the fourth subject. 5. Infusions containing 50 mmol of sodium chloride in 50 ml given intravenously over 22 min produced a rise in plasma sodium concentration and a fall in concentration of total plasma solids. 6. These results provide no evidence for a gastrointestinal or portal monitor of sodium intake, but do not disprove the existence of such a monitor.
Key words: intravenous infusions, salt-free diet, sodium chloride. Introduction
The responses to oral and intravenous sodium chloride in sodiumdeplete man can differ (Lennane, Carey, Goodwin & Peart, 1975a). An oral Correspondence: Dr D. Gordon, Medical Unit, St Mary's Hospital Medical School, London W2 IPG. 15
225
226
D. Gordon and W. S . Peart
Methods
In Experiment A, four normal male subjects aged 23 or 24 years were studied in pairs. There were two study periods of 6 days each. At least 1 week elapsed between the study periods. The subjects were asked to avoid dietary or alcoholic excess on the day before, and during the 6 days of the experiment they abstained from alcohol, drugs and excessive physical work. They continued their normal daily activities. During each study period they ate a diet of weighed quantities of food calculated to provide 5 mmol of sodium, 71 mmol of potassium and 7500 kJ each day (McCance & Widdowson, 1960), taken in three meals at 08.30, 12.30 and 18.30 hours. The fluid intake (1.7-2.6 I/day) was kept constant for each subject. With each meal on days 1-3 they took two gelatin capsules each containing 1.5 g of sodium chloride, providing 154 mmol/day of sodium and chloride. With each meal on days 4-6 they took two capsules, identical in appearance, but containing Dglucose. The subjects were unable to detect the change in the contents of the capsules. After breakfast every morning the subjects were weighed on a scale accurate to 50 g, blood pressure was measured while sitting by using a random-zero sphygmomanometer (Gelman Hawksley Ltd) and, on day 3, blood was taken by venepuncture. Immediately after each meal on days 4-6 the subject sat in a chair and 50 ml of solution was infused at a constant rate over 22 min by using a syringe pump through a 21 G needle in a forearm vein. The solution contained either 50 mmol (‘strong’) or 1.5 mmol (‘weak’) of sodium chloride. One subject of each pair was given ‘strong’ infusions, and the other ‘weak’ infusions, during the first study period, and the reverse during the second study period. Therefore, each subject had infusions on days 4-6 of one study period which were equivalent to the sodium chloride taken in capsules on days 1-3; in the other study period he had infusions of equal volume but providing little sodium chloride. Blood was taken without stasis through an indwelling 21 G needle from a vein in the other arm, before, 10-15 min (mean 12.3) after the start of, and at the end of, the first infusion of the day. Not more than 50 ml of blood was taken in one day. Urine was collected every 4 h for the 6 days starting at 08.00 hours on day 1. Subjects woke themselves to micturate at 04.00 hours by using an alarm clock. The subjects in Experiment B were again normal males, and two of the four had taken part in
Experiment A. The age range was 19-24 years. The protocol was identical with that of Experiment A apart from two modifications. The first was that the capsules on days 4-6 each contained 1.5 g of sodium chloride, so that the intake of sodium by mouth was 159 mmol/day for all 6 days. Therefore, the total sodium intake on days 4-6 was 309 mmol/day (three ‘strong’ infusions daily) or 164 mmol/day (‘weak’ infusions). The second modification was that additional bood samples were taken 2 - 4 min (mean 2.6) and 5-8 min (mean 7.1) after the start of the first infusion of the day on days 4-6. The subjects, who were paid, were medical students. They gave consent to the experimental procedures after full explanation, and were aware that the capsules and infusions would contain various amounts of sodium chloride. They were told that exact knowledge of the amounts of sodium chloride used might frustrate the object of the experiment. Urine and plasma sodium and potassium were measured by flame photometry (Corning EEL). Plasma chloride was measured by a potentiometric method (Corning EEL). Total plasma solids were measured by drying a weighed plasma sample to constant weight. Urine creatinine was measured in the AutoAnalyzer (Technicon) by colorimetry after reaction with alkaline picrate. Plasma renin activity was measured by radioimmunoassay of angiotensin I (Boyd, Adamson, Fitz & Peart, 1969) generated from endogenous substrate on incubation of plasma. Plasma (2 ml) was taken, and 0.2 ml of EDTA (0.2 mol/l) and 0.1 ml of a solution containing 2,3-dimercaptopropanol (0.17 mol/l) with 8-hydroxyquinoline sulphate (0.05 mol/l) were added. The pH was brought to 6.0 with 0.1 ml of HCI (! mol/l). The mixture was incubated in closed tubes at 37OC for 2 h, and then placed in boiling water for 90 s to coagulate protein and denature renin. The protein coagulum was stirred, and the sample centrifuged until a clear supernatant was obtained. A portion (0.2 ml) of the supernatant was taken for immunoassay. The results from day 1 were disregarded during the analysis, as effects due to adaptations from a free diet might have been particularly prominent. The renal excretion of sodium was drawn as a cusum (Woodward & Goldsmith, 1964; Wohl, 1964), taking days 2 and 3 as the control period, to illustrate the time at which sodium excretion changed. The significance of any changes in sodium or potassium excretion from values in days 2 and 3 was estimated with a sequential t-test (Armitage,
Sodium excretion and intravenous sodium 1954), comparing the excretion in each 4 h on days 4-6 with the mean for the corresponding 4 h on days 2 and 3. Changes were described as significant when P 0.05.Differences between means were compared by using Student’s t-test, paired or unpaired as appropriate, and means are quoted
100 to 146 mmHg and diastolic pressures from 46 to 86 mmHg. Mean creatinine excretion was 14.62 f 0.14 mmol/day. No single 24 h creatinine excretion varied by more than 2 SD from the mean for the particular subject, suggesting that urine collections were not grossly incomplete. Plasma sodium and potassium showed no consistent trend from day to day in either experiment. Plasma renin activity showed a significant increase in the control period of Experiment A, i.e. while on a low sodium diet with ‘weak’ infusions (day 3 d a y 6,2050 ? 240 to 5480 f 960 pmol h-’ l-l, t = 4.63,O.Ol < 2P < 0.05) but no consistent changes from day to day at other times. Renal sodium excretion during Experiment A is shown in Fig. 1. With ‘weak’ infusions there was a fall in sodium excretion, which became significant between 24 h (subject D) and 48 h (subject S) after the change in sodium intake. When oral sodium chloride was almost exactly replaced with ‘strong’ infusions, sodium excretion did not change significantly except in subject G. The fall in sodium excretion in subject G became significant by 36 h after the first infusion, and may be because he habitually takes a diet rich in sodium (about 300 mmol/day) and therefore took longer than the others to come into equilibrium with the fixed intake of 159 mmol of sodium/day. Renal
Sodium excretion in man, and adaptation to a low-sodium diet: effect of intravenous sodium chloride
D. GORDON A N D W . S. PEART Medical Unit, St Mary's Hospital Medical School, London
(Received 22 December 1978; accepted 18 April 1979) dose of sodium chloride is excreted more rapidly than an equal intravenous load. This suggests that there may be a gastrointestinal or portal monitor of sodium intake. Similar results have been found in the rabbit (Lennane, Peart, Carey & Shaw, 1975b) and there is evidence for osmoreceptors in the portal circulation in the rat (Haberich, Aziz & Nowacki, 1965; Adachi, Niijima & Jacobs, 1976) and dog (Daly, Roe & Horrocks, 1967), although Kapteina, Motz, Schwartz-Porsche & Gauer (1978) found no evidence for such portal osmoreceptors in the dog. There is evidence for receptors responding to hypertonic sodium chloride, but not to hypertonic sucrose, in the cat (Passo, Thornborough & Rothballer, 1973). If the gastrointestinal or portal monitor proposed by Lennane et al. (l975a) is an essential part of the control mechanism for the rapid reduction in the renal excretion of sodium that follows a change to a low-sodium diet, then the natriuresis that normally follows infusion of sodium chloride might be diminished or absent if the infusion is given at the time of a reduction in oral sodium intake, because of the influence of the gastrointestinal sodium monitor. It is important to establish how much of the same intravenous load is retained if there is no reduction in oral sodium intake. These experiments were therefore designed to test the renal response to intravenous sodium at the time of change to a lowsodium diet, or when the intravenous sodium was given in addition to a diet providing a constant intake of sodium similar to that seen on a normal free diet. The changes during intravenous sodium chloride infusion in a number of non-hormonal blood factors possibly related to sodium excretion were also studied.
Summary 1. The aim of this study was to test whether a postulated gastrointestinal or portal monitor of sodium intake plays any part in adjusting renal sodium excretion when dietary sodium is reduced. 2. Normal male subjects were given 50 mmol of sodium chloride intravenously three times daily for 3 days to replace or to supplement a constant oral intake of sodium chloride. 3. When oral sodium chloride was replaced with intravenous sodium chloride, renal sodium excretion remained constant. 4. When oral sodium chloride was kept constant, sodium administered as intravenous sodium chloride was promptly excreted in three out of four subjects. There was a delay in the increase in sodium excretion in the fourth subject. 5. Infusions containing 50 mmol of sodium chloride in 50 ml given intravenously over 22 min produced a rise in plasma sodium concentration and a fall in concentration of total plasma solids. 6. These results provide no evidence for a gastrointestinal or portal monitor of sodium intake, but do not disprove the existence of such a monitor.
Key words: intravenous infusions, salt-free diet, sodium chloride. Introduction
The responses to oral and intravenous sodium chloride in sodiumdeplete man can differ (Lennane, Carey, Goodwin & Peart, 1975a). An oral Correspondence: Dr D. Gordon, Medical Unit, St Mary's Hospital Medical School, London W2 IPG. 15
225
226
D. Gordon and W. S . Peart
Methods
In Experiment A, four normal male subjects aged 23 or 24 years were studied in pairs. There were two study periods of 6 days each. At least 1 week elapsed between the study periods. The subjects were asked to avoid dietary or alcoholic excess on the day before, and during the 6 days of the experiment they abstained from alcohol, drugs and excessive physical work. They continued their normal daily activities. During each study period they ate a diet of weighed quantities of food calculated to provide 5 mmol of sodium, 71 mmol of potassium and 7500 kJ each day (McCance & Widdowson, 1960), taken in three meals at 08.30, 12.30 and 18.30 hours. The fluid intake (1.7-2.6 I/day) was kept constant for each subject. With each meal on days 1-3 they took two gelatin capsules each containing 1.5 g of sodium chloride, providing 154 mmol/day of sodium and chloride. With each meal on days 4-6 they took two capsules, identical in appearance, but containing Dglucose. The subjects were unable to detect the change in the contents of the capsules. After breakfast every morning the subjects were weighed on a scale accurate to 50 g, blood pressure was measured while sitting by using a random-zero sphygmomanometer (Gelman Hawksley Ltd) and, on day 3, blood was taken by venepuncture. Immediately after each meal on days 4-6 the subject sat in a chair and 50 ml of solution was infused at a constant rate over 22 min by using a syringe pump through a 21 G needle in a forearm vein. The solution contained either 50 mmol (‘strong’) or 1.5 mmol (‘weak’) of sodium chloride. One subject of each pair was given ‘strong’ infusions, and the other ‘weak’ infusions, during the first study period, and the reverse during the second study period. Therefore, each subject had infusions on days 4-6 of one study period which were equivalent to the sodium chloride taken in capsules on days 1-3; in the other study period he had infusions of equal volume but providing little sodium chloride. Blood was taken without stasis through an indwelling 21 G needle from a vein in the other arm, before, 10-15 min (mean 12.3) after the start of, and at the end of, the first infusion of the day. Not more than 50 ml of blood was taken in one day. Urine was collected every 4 h for the 6 days starting at 08.00 hours on day 1. Subjects woke themselves to micturate at 04.00 hours by using an alarm clock. The subjects in Experiment B were again normal males, and two of the four had taken part in
Experiment A. The age range was 19-24 years. The protocol was identical with that of Experiment A apart from two modifications. The first was that the capsules on days 4-6 each contained 1.5 g of sodium chloride, so that the intake of sodium by mouth was 159 mmol/day for all 6 days. Therefore, the total sodium intake on days 4-6 was 309 mmol/day (three ‘strong’ infusions daily) or 164 mmol/day (‘weak’ infusions). The second modification was that additional bood samples were taken 2 - 4 min (mean 2.6) and 5-8 min (mean 7.1) after the start of the first infusion of the day on days 4-6. The subjects, who were paid, were medical students. They gave consent to the experimental procedures after full explanation, and were aware that the capsules and infusions would contain various amounts of sodium chloride. They were told that exact knowledge of the amounts of sodium chloride used might frustrate the object of the experiment. Urine and plasma sodium and potassium were measured by flame photometry (Corning EEL). Plasma chloride was measured by a potentiometric method (Corning EEL). Total plasma solids were measured by drying a weighed plasma sample to constant weight. Urine creatinine was measured in the AutoAnalyzer (Technicon) by colorimetry after reaction with alkaline picrate. Plasma renin activity was measured by radioimmunoassay of angiotensin I (Boyd, Adamson, Fitz & Peart, 1969) generated from endogenous substrate on incubation of plasma. Plasma (2 ml) was taken, and 0.2 ml of EDTA (0.2 mol/l) and 0.1 ml of a solution containing 2,3-dimercaptopropanol (0.17 mol/l) with 8-hydroxyquinoline sulphate (0.05 mol/l) were added. The pH was brought to 6.0 with 0.1 ml of HCI (! mol/l). The mixture was incubated in closed tubes at 37OC for 2 h, and then placed in boiling water for 90 s to coagulate protein and denature renin. The protein coagulum was stirred, and the sample centrifuged until a clear supernatant was obtained. A portion (0.2 ml) of the supernatant was taken for immunoassay. The results from day 1 were disregarded during the analysis, as effects due to adaptations from a free diet might have been particularly prominent. The renal excretion of sodium was drawn as a cusum (Woodward & Goldsmith, 1964; Wohl, 1964), taking days 2 and 3 as the control period, to illustrate the time at which sodium excretion changed. The significance of any changes in sodium or potassium excretion from values in days 2 and 3 was estimated with a sequential t-test (Armitage,
Sodium excretion and intravenous sodium 1954), comparing the excretion in each 4 h on days 4-6 with the mean for the corresponding 4 h on days 2 and 3. Changes were described as significant when P 0.05.Differences between means were compared by using Student’s t-test, paired or unpaired as appropriate, and means are quoted
100 to 146 mmHg and diastolic pressures from 46 to 86 mmHg. Mean creatinine excretion was 14.62 f 0.14 mmol/day. No single 24 h creatinine excretion varied by more than 2 SD from the mean for the particular subject, suggesting that urine collections were not grossly incomplete. Plasma sodium and potassium showed no consistent trend from day to day in either experiment. Plasma renin activity showed a significant increase in the control period of Experiment A, i.e. while on a low sodium diet with ‘weak’ infusions (day 3 d a y 6,2050 ? 240 to 5480 f 960 pmol h-’ l-l, t = 4.63,O.Ol < 2P < 0.05) but no consistent changes from day to day at other times. Renal sodium excretion during Experiment A is shown in Fig. 1. With ‘weak’ infusions there was a fall in sodium excretion, which became significant between 24 h (subject D) and 48 h (subject S) after the change in sodium intake. When oral sodium chloride was almost exactly replaced with ‘strong’ infusions, sodium excretion did not change significantly except in subject G. The fall in sodium excretion in subject G became significant by 36 h after the first infusion, and may be because he habitually takes a diet rich in sodium (about 300 mmol/day) and therefore took longer than the others to come into equilibrium with the fixed intake of 159 mmol of sodium/day. Renal
