Quarterly Journal of Experimental Physiology (1976) 61, 309-319
THE EFFECT OF INTRAVENOUS INFUSION OF HYPEROSMOTIC SODIUM AND POTASSIUM CHLORIDE SOLUTIONS ON CEPHALIC BLOOD FLOW IN CONSCIOUS SHEEP. By A. M. BEAL. From The Agricultural Research Council Institute of Animal Physiology, Babraham, Cambridge CB2 4AT, England. (Receivedfor publication 5th March 1976) The rate of flow of plasma and blood through the head of conscious sheep was measured before, during and after the intravenous infusion of I mol. 1-1 NaCl and 1 mol. 1-1 KCl at 0-8-1 0 ml. min-' for 2 hours. The plasma flow was estimated by indicator-dilution technique using sodium para-aminohippurate which was shown to be a satisfactory indicator substance. Short periods of rumination were found to cause marked increases in cephalic blood flow. The infusion of hyperosmotic sodium chloride caused no consistent changes in the rates of cephalic plasma flow and blood flow. During potassium infusion plasma and blood flows increased as the plasma potassium concentration increased up to approximately 6 mmol.1-1. Further increases in plasma potassium concentration were associated with a progressive return of these flow rates to or below the preinfusion levels. This pattern of change in the rate of plasma flow through the head of the sheep was very similar to that previously reported for renal plasma flow during hyperkalaemia in conscious sheep. At its maximum the cephalic plasma flow was 1-163 ±0-029 (S.E. of mean) times the pre-infusion flow rate. Cephalic blood flow tended to reach maximum rates at slightly higher plasma potassium concentrations and thereafter, to fall more slowly than the plasma flow due to concomitant increases in haematocrit. Maximum cephalic blood flow was 1-176 + 0-032 times the preinfusion flow rate. The lowest rates of cephalic plasma and blood flow occurred during the first 30 minutes following cessation of potassium infusion.
The intravenous infusion of potassium chloride into conscious sheep causes a fall in the rate of secretion of parotid saliva. Since equimolar infusions of sodium chloride did not produce a similar depression of salivary flow, the effect was attributed to the hyperkalaemia caused by the potassium infusion [Beal, Budtz-Olsen and Clark, 1975; Beal, Clark and Budtz-Olsen, 1975]. As the rate of production of saliva has been shown to vary with the rate of blood flow through the parotid gland during occlusion of the carotid artery [Coats, Denton, Goding and Wright, 1956; Denton, 1957] the depression of salivary flow during potassium infusion may also result from a reduction in blood supply if hyperkalaemia causes haemodynamic changes which result in lowered carotid arterial blood flow. Hyperkalaemia reduces or increases vascular resistance in the anaesthetized dog, depending on the rate of infusion and the plasma potassium concentration [Emanuel, Scott and Haddy, 1959; Scott, Emanuel and Haddy, 1959]. High plasma potassium concentration is also associated with depressed cardiac contractility [Garb, 1951; Sarnoff, Gilmore, McDonald, Daggett, Weisfeldt and Mansfield, 1966; Logic, Krotkiewski, Koppius and Surawicz, 1968]. Because the vascular innervation of the conscious animal is fully competent, changes in blood flow during hyperkalaemia cannot be assumed to be the same in both the conscious and anaesthetized states. Direct measurements of blood flow rates in the carotid arteries by indicatordilution technique were found to be unsatisfactory due to inadequate mixing 309
310 Beal of the indicator. However as the common carotid arteries provide almost the whole of the blood supply to the head of the sheep [Baldwin and Bell, 1963] measurement of total cephalic blood flow should closely approximate the rate of flow in both carotid arteries. The object of the experiments described in this paper was to investigate and compare the effects of intravenous infusions of hyperosmotic sodium chloride and potassium chloride solutions on the flow of blood through the head of the conscious sheep. METHODS General methods of management were the same as reported previously [Beal, 1976].
Experimental Procedures Nine conscious sheep (6 Clun Forest ewes weighing 32-057-5 kg and 3 Merino x Clun Forest x Welsh Mountain ewes weighing 34 5-39 0 kg) were used. Each Clun Forest sheep had one carotid artery exteriorized in a skin loop and a permanent vinyl cannula in the other carotid artery (0 9 mm i.d., 1P5 mm o.d.; Portex Limited). The Merino cross-bred sheep had both carotid arteries exteriorized in skin loops. During the afternoon of the day before experiment a vinyl sampling cannula (1-4 mm i.d., 2-0 mm o.d.; Portex Limited) was inserted into each external jugular vein using the technique of Seldinger [1953]. The tips of these cannulae were placed caudal to and within a few centimeters of the junction of the external and internal maxillary veins. A third venous cannula of similar size to the sampling cannulae was inserted into one jugular vein caudal to the previous jugular cannulation to allow the hyperosmotic solutions to be infused into the general circulation. The exteriorized carotid artery was cannulated with 2 disposable plastic cannulae, one directed caudally and the other directed cranially (Braunula, size 0 5; Armour Pharmaceutical Company Ltd). In the sheep with 2 carotid loops a third cannula (Braunula) of the same size was inserted in a cranial direction into the other carotid artery. On the morning of experiment the sheep were transferred to and restrained on canvas stretchers in a normal upright position with their feet just off the floor. The bladder was catheterized with a self-retaining catheter (Casper pattern; Rusch) and vinyl extensions were connected to the vascular cannulae so that all manipulations during infusion and sampling occurred in the region of the animal's shoulder. After a short period in which the sheep were allowed to become accustomed to the experimental situation a blood sample was taken from the caudally-directed carotid cannula. A priming injection (40 ml) of an isotonic solution containing 03 g sodium para-aminohippurate (PAH)/100 ml of NaCl/KCl solution (135: 3-5 mmol.h1') was given, followed by a constant intracarotid infusion of the same solution by peristaltic pump (1-1 ml.min-1 simultaneously into both cranially-directed cannulae). After 2 h of infusion (stabilization period) the peristaltic pump was replaced with a 2 channel syringe pump driving 20 ml glass syringes at a rate sufficient to continue the PAH infusion at 1-1 ml.min'1 into each carotid artery. After 10min infusion by syringe pump, blood samples were collected from the caudally-directed carotid cannula and the 2 jugular sampling cannulae during the next 5 min. The syringe pump was then re-filled and the cycle repeated so that blood samples were taken every 15 min. During the period of syringe filling the continuous infusion was maintained with the peristaltic pump. Several sham blood collections were made to accustom the animals to this routine and then the sheep were subjected to one of the following treatments: (1) Isotonic control treatment (10 replications). The sheep received only the isotonic PAH infusion for the entire period of experiment (4 5 h). (2) Hyper. osmotic NaCl treatment (10 replications). After an initial 45-60 min (3 or 4 blood samples) during which the sheep received only the intracarotid PAH infusion, an additional infusion of 1 mol.l-'NaCl was given intravenously at 0-8-1-0 ml.min-1 into the caudally-placed jugular cannula for a period of 2 h. The PAH infusion and blood sampling were continued for 2 h after hyperosmotic NaCl administration had been terminated. (3) Hyperosmotic KCI
Head blood flow during hypertonic infusion
311
treatment (10 replications). The format of this experiment was identical to the hyperosmotic NaCl treatment with the exception that 1 mol.lI- KCl was infused instead of 1 mol.I-1 NaCl. Blood samples were taken into 10 ml plastic syringes heparinized with 1 drop of heparin (5000 i.u./ml). Approximately 5 ml arterial blood was withdrawn from the carotid artery below the site of PAH infusion followed by 10 ml venous blood from each jugular sampling cannula and a further 5 ml arterial blood from the carotid artery. To minimize errors resulting from non-uniform blood flow the withdrawal of blood was made slowly over a period of several minutes and the volumes taken were well in excess of the amount necessary for analysis. Depending on the haematocrit of the animals, 4-5 ml samples of arterial blood and 2-3 ml samples of venous blood were kept to provide sufficient plasma for analysis and the excess was returned to the animals immediately.
Analytical Procedures were as described previously [Beal, 1976]. PAH was estimated in duplicate by the method of Bratton and Marshall [1939] as adapted for the Technicon autoanalyser by Harvey and Brothers [1962]. Mathematical and Statistical Procedures Unilateral cephalic plasma and blood flows were calculated as follows: Plasma flow (ml.minm
PAH infusion rate venous PAH -arterial PAH
Blood flow (ml.min') = Plasma flow x 100 100-Ht Total cephalic flows were calculated as the sum of these values for the left and right sides of the head. Analysis of covariance was applied to the data for each corresponding sample during the period of hyperosmotic infusion and for two samplings after the hyperosmotic infusion terminated. The two covariates used in the analysis of each variable were the values of that variable for the ante-penultimate blood sample and for the last blood sample of the stabilization period preceding the hyperosmotic infusions (i.e. the first and third samples on the graphs in Fig. 1). When the analyses of covariance produced significant variance ratios the differences between individual treatments were found using Tukey's w-procedure (i.e. honestly significant difference (hsd) procedure) as described by Steel and Torrie (1960]. The results of the hsd analyses are presented as levels of significance in Table I.
RESULTS Experiments validating method of bloodflow measurement Samples of carotid arterial blood and jugular venous blood were collected from 5 sheep during the third hour of bilateral intracarotid infusions of PAH (0 3 0 PAH at 1 1 ml. min 1). The concentration of PAH was estimated in plasma and in whole blood after lysis of the red cells by repeated freezing and thawing. The plasma PAH concentration was adjusted to whole blood values by use of the haematocrit and compared with the measured whole blood concentration. The mean ratio for 30 comparisons of calculated whole blood PAH concentration to measured whole blood concentration was 0997+00011 (S.E. of mean). The results showed no measurable uptake of PAH by red blood cells in arterial blood or in the venous blood leaving the head. Five blood samples were taken simultaneously from the carotid arteries and jugular veins of 6 sheep during the third hour of an intravenous infusion of
Beal 312 PAH (0 5 % at 2-2 ml .min '). The concentration of PAH in the plasma of the venous blood was expressed as a ratio of the concentration in the corresponding sample of arterial plasma (30 comparisons). The mean value for venous PAH/ arterial PAH was 1 003 ±0 002 (S.E. of mean). The mean increase in haematocrit from arterial to venous blood was 0 18 + 0049 (S.E. of mean) %. During the third hour of bilateral intracarotid infusion of PAH (03 at 1-1 ml. min-') a sample of cerebrospinal fluid (CSF) was taken by lumbar puncture from each of 6 sheep. The PAH concentration of the CSF was extremely low and was calculated to be 0-35 ± 0 007 (S.E. of mean) % of the jugular plasma concentration (-= cephalic plasma concentration). Parotid saliva was collected under the same infusion conditions from 6 sheep and the concentration of PAH in the saliva was always less than 0 20 % of the PAH concentration of cephalic plasma.
900
_~~~~
CEPHALIC BLOOD FLOW 600
-
(ml./min)
300
-
0
2
1
TIME
(hours)
FIG. 1. Mean values for blood flow through the head of sheep before, during and after a short episode of rumination (10 observations). The solid bar indicates the timing and the maximum duration of the period of rumination (means ± S.E. of mean).
In preliminary experiments, when the sheep were fasted for a shorter period before experiment, sporadic rumination occurred during the period of flow measurement which was associated with such large increases in blood flow that the experiments had to be discarded. The mean blood flow rates before, during and after 10 such incidents of rumination in which the sheep never chewed for longer than 15 min are shown in Fig. 1 and demonstrate the necessity to prevent rumination if reasonably stable cephalic flow rates are to be obtained.
Head blood flow during hypertonic infusion
313
Isotonic control treatment The plasma sodium concentration showed a small increase during the first 60-90 min of blood flow measurement and thereafter remained reasonably stable for the remainder of the experiment. At the same time the plasma potassium concentration fell slightly and the osmolality of the plasma showed little change. The haematocrit fell throughout the entire experiment with the greatest rate of fall occurring in the initial stabilization before blood flow measurement began and with a slower rate of decline (averaging about 2%) during the 4 5 h period of flow measurement. In individual experiments the cephalic plasma flow tended to oscillate slowly during the period of observation. The mean difference in flow rate between maximum and minimum flows was 80 ± 10 (S.E. of mean) ml. min- and the time interval between the occurrence of these maxima and minima was 135 +23 (S.E. of mean) min. Since these changes in plasma flow did not occur at the same time in each experiment the mean plasma flow through the head showed no significant changes during the entire 4-5 h of measurement which could be attributed to the isotonic infusion. The mean oscillation in cephalic blood flow was 106 + 13 ml. min-' and the periods of maximum and minimum flow coincided with those of plasma flow since alterations in haematocrit during the period of flow measurement were not large enough to prevent this synchronization (Fig. 2).
Hyperosmotic NaCI treatment Ilp to the beginning of hyperosmotic infusion the treatment of all sheep was identical and thus the results obtained to that point in the 1 mol.1-1 NaCl and 1 mol.l1 KCI experiments were very similar to those of the control experiment. When hyperosmotic NaCl was infused the concentration of sodium in plasma and the osmolality of the plasma rose to levels significantly greater than those in the control experiment, whereas the plasma potassium concentration declined slowly as in the control treatment (Fig. 2, Table I). The haematocrit fell throughout the hyperosmotic NaCl experiment and during the period of 1 mol. 11 NaCl infusion appeared to fall more rapidly than during the corresponding time in the control experiment (Fig. 2) but this difference between the two experiments was not statistically significant. Cephalic plasma flow was not significantly influenced by the infusion of hyperosmotic NaCl and the pattern of changes in flow throughout the entire period of measurement was similar to that observed in the control treatment. The mean difference in flow rate between the maximum and minimum flows was 65 + 10 (S.E. of mean) ml. min- and the time interval between these maxima and minima was 153 + 25 min. Cephalic blood flow showed oscillations of 88 + 11 ml. min -' coincident with those for plasma flow and as these changes in plasma and blood flow did not occur at the same time in each experiment the mean plasma and blood flows through the head of the sheep were reasonably stable over the entire period of flow measurement (Fig. 2). D
_o\.ou.' / fi.(DU\
314
Beal
(]
,
! NaCI C
Control
155
Plasma
KCl
_
145Lff4{{f 9
(mmol/l)IJa: Plasma
_*@@O@S@@@O@@@@@@@.@
Osmolaility
*
(mosmollkg)290E Haematocrit
a' a'
L
*
*S@eeeeeee,.,...,.,
+
.
L*1J++*+ *4*YTT
i
40_
_
t
i
4"0
25L 450 _
Plasma Flow (ml /min)
Cepholic
.
.
TT?a
,
t
Blood Flow
(ml/min)
,
_ 200 600 _
-t .1
_
300_*
**
2
4
_
IIIf
I
0
1
0
Time
2
4
0
2
4
(hours)
FIo. 2. The plasma sodium and potassium concentration, plasma osmolality, haematocrit, cephalic plasma flow and cephalic blood flow before, during and after the intravenous infusion of 1 mol. 1- I NaCl and 1 mol. 1- I KCI at 0 8-1 0 ml. min-1 for 2 h. The control treatment received isosmotic infusion only. The filled squares give the values for parameters estimated on the initial blood sample taken before any other experimental procedure was commenced (N = 10; means ±S.E. of mean).
Hyperosmotic KCl treatment The infusion of 1 mol . 1- KCI increased the plasma potassium concentration to levels well in excess of those occurring in the previous two treatments (Fig. 2; Table I). The sodium concentration of the plasma fell slightly during this infusion but the difference in concentration between this treatment and the control treatment was not statistically significant (Table I). The rise in plasma
Head blood flow during hypertonic infusion
315
TABLE I. Summary ofthefinal hsd analyses of cephalic plasma and bloodflows, haematocritandplasma electrolyte concentrations for the 8 blood samples during and the first 2 blood samples after the 2 hour intravenous infusion of 1 mol. 1- I NaCl and 1 mol. 1- 1 KCl and the corresponding periods of the control experiment. Every variable on the ordinate of Fig. 2 was tested and omission of a comparison from this table means that no significant differences were found. Differences are indicated as levels of significance (* = P < 0-05; ** = P < 0-01; *** = P < 0001; - = not significant). P is the probability of erroneous rejection of the null hypothesis (i.e. type 1 error). Variable Comparison Blood Sample degrees of NaCl KCI Test Infusion Recovery 7 1 2 6 8 freedom = 25 versus versus 2 3 4 1 5
Plasma Na concn. Plasma K concn. Plasma Osmolality
Haematocrit Plasma Flow
Blood Flow
Control
* * * * *** *** NaCl - *** *** *** Control *** * * * * * * *** *** *** *** *** *** *** *** *** NaCl * * - ** * *** *** * Control Control ** * * ** * *** *** ** ** ** * *** *** *** Control * * * ** *** *** *** *** *** *** *** NaCl * * * Control * NaCl * * * * * ** ** Control - ** ** NaCl
osmolality during potassium infusion was of similar magnitude to that occurring during the hyperosmotic NaCl infusion. In contrast to the previous treatments the haematocrit rose during the infusion of KCl, at a slow rate initially and at a much more rapid rate during the second part of the infusion (Fig. 2). The point of inflection in the curve for mean haematocrit coincided with the peak values for mean plasma and blood flows, but this did not occur in individual experimnents. During the initial part of all potassium infusions the plasma flow through the head was always elevated above the pre-infusion (stabilization) values reaching a maximum in most experiments during the first hour. Although the size of this increment in cephalic plasma flow varied considerably between experiments (range 6-31 %) the rates of flow after 45 and 60 min of potassium infusion were always greater than those for the corresponding times in the control and hyperosmotic NaCl treatments (Fig. 2, Table I). The mean value for the maximum increase in cephalic plasma flow was 1-163 +±0029 (S.E. of mean) times the mean plasma flow before potassium was infused and the mean plasma potassium concentration at the peak of the increase in plasma flow was 5-92 + 0O20 mmol. I . As the potassium infusion continued the rate of flow declined so that in all but 3 experiments the cephalic plasma flow rate at peak hyperkalaemia was less than pre-infusion levels. When the potassium infusion was stopped plasma flow did not recover immediately and frequently continued to fall rapidly for a further 15-30 min so that mean plasma flow rates had not returned to pre-infusion values until 45 min after the potassium infusion had
316 Beal ceased (Fig. 2, Table I). Cephalic blood flow also increased during the initial period of potassium infusion and declined during the latter part of the infusion but as a result of the concomitant increase in haematocrit the increment in blood flow at the beginning of potassium infusion was slightly larger than that for plasma flow. The mean value for the maximum increase in cephalic blood flow was 1-176 +0-032 (S.E. of mean) times the mean blood flow before potassium infusion and the mean plasma potassium concentration at maximum blood flow was 6 03 + 0-22 mmol . 1-1. Although the rate of flow of blood through the head declined during the second half of the potassium infusion it did not fall as rapidly as plasma flow due to the increasing haematocrit. At the end of potassium infusion blood flow rates were less than the pre-infusion values in only 4 of 10 experiments. Blood flow continued to fall during the first part of the recovery period following the infusion of potassium but, unlike plasma flow, the fall in cephalic blood flow was not statistically significant (Fig. 2, Table I). DIsCUSSION In the sheep the common carotid arteries provide the total blood supply of the extracerebral region and almost all the cerebral region except the caudal medulla oblongata [Baldwin and Bell, 1963]. Although each side of the head behaves as a discrete circulation they are interconnected and partial or complete occlusion of a carotid artery results in an immediate redistribution of supply via the remaining patent vessels. Therefore, the repeated measurement of cephalic blood flow in the conscious animal at short intervals over a number of hours necessitates the use of a technique which will allow simultaneous measurement of flow through both sides of the head and does not itself cause marked changes in flow. The indicator-dilution technique used in the present experiments fulfils the requirement satisfactorily. The use of PAH as the indicator substance has the advantage that rapid extraction of the compound from the plasma occurs in the kidney and keeps the level of recirculation low and reasonably stable. Thus the arterio-venous difference in concentration becomes relatively large which increases the accuracy of flow measurement. The major source of error in the use of an indicator-dilution technique for the measurement of a regional circulation is the possibility that the indicator may escape from the circulation and be stored, excreted or metabolized within the region under study. The 2 h period of stabilization before blood sampling should have been sufficient for equilibration to occur between the blood and any extravascular space in the head into which PAH could diffuse rapidly. Diffusion gradients taking longer than 2 h to equilibrate would have no significant effect on the accuracy of flow measurement. The close agreement between the plasma PAH concentrations of jugular and carotid blood during intravenous PAH infusion supports the above thesis and also indicates that no significant metabolism or excretion of PAH occurs in the head. No attempt was made to measure PAH concentrations in cephalic lymph but estimations of PAH in cerebrospinal fluid and parotid saliva showed that neither of these secretions contained sufficient PAH to materially affect the accuracy of flow measurement.
Head blood flow during hypertonic infusion 317 The venous sampling cannulae were placed a few centimeters below the junction of the external and internal maxillary veins to ensure that blood samples were a reasonable mixture of blood from these vessels. This positioning of the cannulae results in blood flow from the thyroid, larynx and pharynx being included in the total cephalic flow measurements. The necessity to fast the animals before experiment reduces the water intake of the sheep which in turn, leads to a reduction in the total volume of body water and presumably a reduction in cardiac output. Therefore the values obtained for cephalic plasma and blood flow may represent the lower end of the range of flow values found in conscious sheep. However shorter periods of fasting are associated with much greater likelihood of rumination which results in marked increases in cephalic blood flow (Fig. 1). The high initial haematocrit values obtained in these experiments have been shown under similar experimental conditions to be due to splenic contraction. Likewise the large increments in haematocrit which occurred during potassium infusion, although partly due to redistribution of fluid between body compartments, were also mainly the result of release of red cells from the spleen [Beal, 1976]. The intravenous infusion of hyperosmotic potassium chloride solution into the sheep invariably resulted in an increase in cephalic plasma and blood flows as the plasma potassium concentration rose to about 6 mmol . 1-1 followed by a progressive decline in the flow rates as the level of hyperkalaemia increased further. These increased flow rates could be caused by a decrease in peripheral resistance or an increase in arterial blood pressure (B.P.). In anaesthetized dogs and cats raised plasma osmolality is associated with a reduction in vascular peripheral resistance [Overbeck, Molnar and Haddy, 1961; Mellander, Johansson, Gray, Jonsson, Lundvall and Ljung, 1967; Jelks and Emerson, 1974]. The infusions of 1 mol .l-1 NaCl and 1 mol .l-1 KCI into the sheep resulted in approximately equal elevation of plasma osmolality and presumably also of plasma chloride concentration but only the potassium infusion was associated with plasma and blood flow changes which would suggest an effect on peripheral resistance. As the arterial BP of sheep remains constant during hyperosmotic sodium chloride infusion [Beal, 1976] any tendency for increased plasma osmolality to cause a fall in vascular resistance must be offset by the ability of the conscious animal to prevent vasodilation reflexly. In the conscious intact sheep, arterial BP was not depressed until the plasma potassium concentration exceeded 6 mmol .l-1 and frequently BP rose slightly at the onset of potassium infusion [Beal, 1976]. The increases in plasma flow and blood flow through the head of the sheep during the early part of the potassium infusions were therefore the result of decreased peripheral resistance which in some cases may have been coupled with a slight increase in arterial BP. This vasodilatory action is consistent with the reported effect of hyperkalaemia on various vascular beds in the anaesthetized dog [Scott, Daugherty, Overbeck and Haddy, 1968]. At higher plasma potassium concentrations the plasma flow through the head of the sheep declined and frequently fell to flow rates below the levels which existed before potassium infusion. These changes in cephalic
Beal 318 plasma flow match well with the changes in renal plasma flow observed during acute hyperkalaemia in the conscious sheep [Beal, Budtz-Olsen, Clark, Cross and French, 1975]. The tendency for cephalic plasma and blood flows to fall at high plasma potassium concentrations suggests that arterial BP was falling and this would seem to be confirmed by the BP data for conscious intact sheep receiving potassium infusion [Beal, 1976]. However, the fall in arterial BP accompanying the fall in cephalic plasma flow cannot be as substantial as would be expected from the changes in the rate of plasma flow since cephalic blood flow reaches maximum values at higher plasma potassium concentrations than plasma flow and falls more slowly due to the concurrent increase in haematocrit. When the potassium infusion was terminated, the rates of cephalic plasma and blood flow fell precipitously to the lowest values observed in the experiments and in general, had not returned to normal values until 45 min of the recovery period had elapsed. As no fall in arterial BP was observed immediately following the cessation of potassium chloride infusion [Beal, 1976] these low flow rates were apparently the result of vasoconstriction. The observation that the haematocrit of conscious sheep increases with the severity of hyperkalaemia provides a simple explanation for a previous finding on renal haemodynamics. Beal et al. [1975] observed that the glomerular filtration rate of conscious sheep undergoing a potassium infusion rises and often reaches maximum values at higher plasma potassium concentrations than does the renal plasma flow and that the subsequent fall in glomerular filtration rate at even higher plasma potassium concentrations was relatively slower than the corresponding fall in renal plasma flow. Obviously renal blood flow at high plasma potassium concentrations would not fall in direct proportion with the fall in renal plasma flow due to the increasing volume of red cells in the blood and this would result in better maintenance of arterial BP, filtration pressure and glomerular filtration rate than would be expected from the changes in renal plasma flow. ACKNOWLEDGMENT I am indebted to Mr P. V. Burrow for skilful technical assistance.
REFERENCES BALDWIN, B. A. and BELL, F. R. (1963). The anatomy of the cerebral circulation of the sheep and ox. The dynamic distribution of the blood supplied by the carotid and vertebral arteries to cranial regions. Journal of Anatomy, 97, 203-215. BEAL, A. M. (1976). Changes in arterial blood pressure, heart rate and haematocrit during acute hyperkalaemia in conscious sheep. Quarterly Journal of Experimental Physiology. 61, 297-308. BEAL, A. M., BUDTZ-OLSEN, 0. E. and CLARK, R. C. (1975). The effect of potassium chloride infusion on parotid salivary flow and composition in conscious sheep. Quarterly Journal of Experimental Physiology, 60, 161-169. BEAL, A. M., CLARK, R. C. and BUDTZ-OLSEN, 0. E. (1975). The composition and flow of parotid saliva during acute hyperkalaemia in sodium-deficient sheep. Quarterly Journal of Experimental Physiology, 60, 315-323.
Head blood flow during hypertonic infusion
319
BEAL, A. M., BUDTZ-OLSEN, 0. E., CLARK, R. C., CROSS, R. B. and FRENCH, T. J. (1975). Changes in renal haemodynamics and electrolyte excretion during acute hyperkalaemia in conscious adrenalectomized sheep. Quarterly Journal of Experimental Physiology, 60, 207-221. BRATTON, A. C. and MARSHALL, E. K. (1939). A new coupling component for sulphanilamide determination. Journal of Biological Chemistry, 128, 537-550. COATS, D. A., DENTON, D. A., GODING, J. R. and WRIGHT, R. D. (1956). Secretion by the parotid gland of the sheep. Journal of Physiology, 131, 13-3 1. DENTON, D. A. (1957). The study of sheep with permanent unilateral parotid fistulae. Quarterly Journal of Experimental Physiology, 42, 72-95. EMANUEL, D. A., ScoTr, J. B. and HADDY, F. J. (1959). Effect of potassium upon small and large blood vessels of the dog forelimb. American Journal of Physiology, 196, 637-742. GARB, S. (1951). The effects of potassium, ammonium, chloride, calcium, strontium and magnesium on the electrogram and myogram of mammalian heart muscle. Journal of Pharmacology and Experimental Therapeutics, 101, 317-326. HARVEY, R. B. and BROTHERS, A. J. (1962). Renal extraction of para-aminohippurate and creatinine measured by continuous in vivo sampling of arterial and renal vein blood. Annals of the New York Academy of Sciences, 102, 46-54. JELKS, G. W. and EMERSON, T. E., JR. (1974). Effects of plasma electrolyte abnormalities on total peripheral resistance and other hemodynamic parameters in dogs. Proceedings of the Society for Experimental Biology and Medicine, 146, 59-65. LoGic, J. R., KROTKIEWSKI, A., KoPPius, A. and SURAWICZ, B. (1968). Negative inotropic effect of K+: its modification by Ca"+ and acetyl strophanthidin in dogs. American Journal of Physiology, 215, 14-22. MELLANDER, S., JOHANSSON, B., GRAY, S., JONSSON, O., LUNDVALL, J. and LJUNG, B. (1967). The effects of hyperosmolarity on intact and isolated vascular smooth muscle. Possible role in exercise hyperemia. Angiologica, 4, 310-322. OVERBECK, H., MOLNAR, J. I. and HADDY, F. J. (1961). Resistance to blood flow through the vascular bed of the dog forelimb. American Journal of Cardiology, 8, 533-541. SARNOFF, S. J., GILMORE, J. P., MCDONALD, R. H., DAGGETr, W. M., WEISFELDT, M. L. and MANSFIELD, P. B. (1966). Relationship between myocardial K+ balance, 02 consumption and contractility. American Journal of Physiology, 211, 361-375. SCOTT, J. B., EMANUEL, D. A. and HADDY, F. J. (1959). The effect of potassium on renal vascular resistance and urine flow rate. American Journal of Physiology, 197, 305-308. ScoTr, J. B., DAUGHERTY, R. M., JR., OVERBECK, H. W. and HADDY, F. J. (1968). Vascular effects of ions. Federation Proceedings, 27, 1403-1407. SELDINGER, S. I. (1953). Catheter replacement of the needle in percutaneous arteriography. Acta Radiologica, 39, 368-376. STEEL, R. G. D. and TORRIE, J. H. (1960). Principles andProcedures in Statistics, pp. 109-1 10. McGraw-Hill, New York.
THE EFFECT OF INTRAVENOUS INFUSION OF HYPEROSMOTIC SODIUM AND POTASSIUM CHLORIDE SOLUTIONS ON CEPHALIC BLOOD FLOW IN CONSCIOUS SHEEP. By A. M. BEAL. From The Agricultural Research Council Institute of Animal Physiology, Babraham, Cambridge CB2 4AT, England. (Receivedfor publication 5th March 1976) The rate of flow of plasma and blood through the head of conscious sheep was measured before, during and after the intravenous infusion of I mol. 1-1 NaCl and 1 mol. 1-1 KCl at 0-8-1 0 ml. min-' for 2 hours. The plasma flow was estimated by indicator-dilution technique using sodium para-aminohippurate which was shown to be a satisfactory indicator substance. Short periods of rumination were found to cause marked increases in cephalic blood flow. The infusion of hyperosmotic sodium chloride caused no consistent changes in the rates of cephalic plasma flow and blood flow. During potassium infusion plasma and blood flows increased as the plasma potassium concentration increased up to approximately 6 mmol.1-1. Further increases in plasma potassium concentration were associated with a progressive return of these flow rates to or below the preinfusion levels. This pattern of change in the rate of plasma flow through the head of the sheep was very similar to that previously reported for renal plasma flow during hyperkalaemia in conscious sheep. At its maximum the cephalic plasma flow was 1-163 ±0-029 (S.E. of mean) times the pre-infusion flow rate. Cephalic blood flow tended to reach maximum rates at slightly higher plasma potassium concentrations and thereafter, to fall more slowly than the plasma flow due to concomitant increases in haematocrit. Maximum cephalic blood flow was 1-176 + 0-032 times the preinfusion flow rate. The lowest rates of cephalic plasma and blood flow occurred during the first 30 minutes following cessation of potassium infusion.
The intravenous infusion of potassium chloride into conscious sheep causes a fall in the rate of secretion of parotid saliva. Since equimolar infusions of sodium chloride did not produce a similar depression of salivary flow, the effect was attributed to the hyperkalaemia caused by the potassium infusion [Beal, Budtz-Olsen and Clark, 1975; Beal, Clark and Budtz-Olsen, 1975]. As the rate of production of saliva has been shown to vary with the rate of blood flow through the parotid gland during occlusion of the carotid artery [Coats, Denton, Goding and Wright, 1956; Denton, 1957] the depression of salivary flow during potassium infusion may also result from a reduction in blood supply if hyperkalaemia causes haemodynamic changes which result in lowered carotid arterial blood flow. Hyperkalaemia reduces or increases vascular resistance in the anaesthetized dog, depending on the rate of infusion and the plasma potassium concentration [Emanuel, Scott and Haddy, 1959; Scott, Emanuel and Haddy, 1959]. High plasma potassium concentration is also associated with depressed cardiac contractility [Garb, 1951; Sarnoff, Gilmore, McDonald, Daggett, Weisfeldt and Mansfield, 1966; Logic, Krotkiewski, Koppius and Surawicz, 1968]. Because the vascular innervation of the conscious animal is fully competent, changes in blood flow during hyperkalaemia cannot be assumed to be the same in both the conscious and anaesthetized states. Direct measurements of blood flow rates in the carotid arteries by indicatordilution technique were found to be unsatisfactory due to inadequate mixing 309
310 Beal of the indicator. However as the common carotid arteries provide almost the whole of the blood supply to the head of the sheep [Baldwin and Bell, 1963] measurement of total cephalic blood flow should closely approximate the rate of flow in both carotid arteries. The object of the experiments described in this paper was to investigate and compare the effects of intravenous infusions of hyperosmotic sodium chloride and potassium chloride solutions on the flow of blood through the head of the conscious sheep. METHODS General methods of management were the same as reported previously [Beal, 1976].
Experimental Procedures Nine conscious sheep (6 Clun Forest ewes weighing 32-057-5 kg and 3 Merino x Clun Forest x Welsh Mountain ewes weighing 34 5-39 0 kg) were used. Each Clun Forest sheep had one carotid artery exteriorized in a skin loop and a permanent vinyl cannula in the other carotid artery (0 9 mm i.d., 1P5 mm o.d.; Portex Limited). The Merino cross-bred sheep had both carotid arteries exteriorized in skin loops. During the afternoon of the day before experiment a vinyl sampling cannula (1-4 mm i.d., 2-0 mm o.d.; Portex Limited) was inserted into each external jugular vein using the technique of Seldinger [1953]. The tips of these cannulae were placed caudal to and within a few centimeters of the junction of the external and internal maxillary veins. A third venous cannula of similar size to the sampling cannulae was inserted into one jugular vein caudal to the previous jugular cannulation to allow the hyperosmotic solutions to be infused into the general circulation. The exteriorized carotid artery was cannulated with 2 disposable plastic cannulae, one directed caudally and the other directed cranially (Braunula, size 0 5; Armour Pharmaceutical Company Ltd). In the sheep with 2 carotid loops a third cannula (Braunula) of the same size was inserted in a cranial direction into the other carotid artery. On the morning of experiment the sheep were transferred to and restrained on canvas stretchers in a normal upright position with their feet just off the floor. The bladder was catheterized with a self-retaining catheter (Casper pattern; Rusch) and vinyl extensions were connected to the vascular cannulae so that all manipulations during infusion and sampling occurred in the region of the animal's shoulder. After a short period in which the sheep were allowed to become accustomed to the experimental situation a blood sample was taken from the caudally-directed carotid cannula. A priming injection (40 ml) of an isotonic solution containing 03 g sodium para-aminohippurate (PAH)/100 ml of NaCl/KCl solution (135: 3-5 mmol.h1') was given, followed by a constant intracarotid infusion of the same solution by peristaltic pump (1-1 ml.min-1 simultaneously into both cranially-directed cannulae). After 2 h of infusion (stabilization period) the peristaltic pump was replaced with a 2 channel syringe pump driving 20 ml glass syringes at a rate sufficient to continue the PAH infusion at 1-1 ml.min'1 into each carotid artery. After 10min infusion by syringe pump, blood samples were collected from the caudally-directed carotid cannula and the 2 jugular sampling cannulae during the next 5 min. The syringe pump was then re-filled and the cycle repeated so that blood samples were taken every 15 min. During the period of syringe filling the continuous infusion was maintained with the peristaltic pump. Several sham blood collections were made to accustom the animals to this routine and then the sheep were subjected to one of the following treatments: (1) Isotonic control treatment (10 replications). The sheep received only the isotonic PAH infusion for the entire period of experiment (4 5 h). (2) Hyper. osmotic NaCl treatment (10 replications). After an initial 45-60 min (3 or 4 blood samples) during which the sheep received only the intracarotid PAH infusion, an additional infusion of 1 mol.l-'NaCl was given intravenously at 0-8-1-0 ml.min-1 into the caudally-placed jugular cannula for a period of 2 h. The PAH infusion and blood sampling were continued for 2 h after hyperosmotic NaCl administration had been terminated. (3) Hyperosmotic KCI
Head blood flow during hypertonic infusion
311
treatment (10 replications). The format of this experiment was identical to the hyperosmotic NaCl treatment with the exception that 1 mol.lI- KCl was infused instead of 1 mol.I-1 NaCl. Blood samples were taken into 10 ml plastic syringes heparinized with 1 drop of heparin (5000 i.u./ml). Approximately 5 ml arterial blood was withdrawn from the carotid artery below the site of PAH infusion followed by 10 ml venous blood from each jugular sampling cannula and a further 5 ml arterial blood from the carotid artery. To minimize errors resulting from non-uniform blood flow the withdrawal of blood was made slowly over a period of several minutes and the volumes taken were well in excess of the amount necessary for analysis. Depending on the haematocrit of the animals, 4-5 ml samples of arterial blood and 2-3 ml samples of venous blood were kept to provide sufficient plasma for analysis and the excess was returned to the animals immediately.
Analytical Procedures were as described previously [Beal, 1976]. PAH was estimated in duplicate by the method of Bratton and Marshall [1939] as adapted for the Technicon autoanalyser by Harvey and Brothers [1962]. Mathematical and Statistical Procedures Unilateral cephalic plasma and blood flows were calculated as follows: Plasma flow (ml.minm
PAH infusion rate venous PAH -arterial PAH
Blood flow (ml.min') = Plasma flow x 100 100-Ht Total cephalic flows were calculated as the sum of these values for the left and right sides of the head. Analysis of covariance was applied to the data for each corresponding sample during the period of hyperosmotic infusion and for two samplings after the hyperosmotic infusion terminated. The two covariates used in the analysis of each variable were the values of that variable for the ante-penultimate blood sample and for the last blood sample of the stabilization period preceding the hyperosmotic infusions (i.e. the first and third samples on the graphs in Fig. 1). When the analyses of covariance produced significant variance ratios the differences between individual treatments were found using Tukey's w-procedure (i.e. honestly significant difference (hsd) procedure) as described by Steel and Torrie (1960]. The results of the hsd analyses are presented as levels of significance in Table I.
RESULTS Experiments validating method of bloodflow measurement Samples of carotid arterial blood and jugular venous blood were collected from 5 sheep during the third hour of bilateral intracarotid infusions of PAH (0 3 0 PAH at 1 1 ml. min 1). The concentration of PAH was estimated in plasma and in whole blood after lysis of the red cells by repeated freezing and thawing. The plasma PAH concentration was adjusted to whole blood values by use of the haematocrit and compared with the measured whole blood concentration. The mean ratio for 30 comparisons of calculated whole blood PAH concentration to measured whole blood concentration was 0997+00011 (S.E. of mean). The results showed no measurable uptake of PAH by red blood cells in arterial blood or in the venous blood leaving the head. Five blood samples were taken simultaneously from the carotid arteries and jugular veins of 6 sheep during the third hour of an intravenous infusion of
Beal 312 PAH (0 5 % at 2-2 ml .min '). The concentration of PAH in the plasma of the venous blood was expressed as a ratio of the concentration in the corresponding sample of arterial plasma (30 comparisons). The mean value for venous PAH/ arterial PAH was 1 003 ±0 002 (S.E. of mean). The mean increase in haematocrit from arterial to venous blood was 0 18 + 0049 (S.E. of mean) %. During the third hour of bilateral intracarotid infusion of PAH (03 at 1-1 ml. min-') a sample of cerebrospinal fluid (CSF) was taken by lumbar puncture from each of 6 sheep. The PAH concentration of the CSF was extremely low and was calculated to be 0-35 ± 0 007 (S.E. of mean) % of the jugular plasma concentration (-= cephalic plasma concentration). Parotid saliva was collected under the same infusion conditions from 6 sheep and the concentration of PAH in the saliva was always less than 0 20 % of the PAH concentration of cephalic plasma.
900
_~~~~
CEPHALIC BLOOD FLOW 600
-
(ml./min)
300
-
0
2
1
TIME
(hours)
FIG. 1. Mean values for blood flow through the head of sheep before, during and after a short episode of rumination (10 observations). The solid bar indicates the timing and the maximum duration of the period of rumination (means ± S.E. of mean).
In preliminary experiments, when the sheep were fasted for a shorter period before experiment, sporadic rumination occurred during the period of flow measurement which was associated with such large increases in blood flow that the experiments had to be discarded. The mean blood flow rates before, during and after 10 such incidents of rumination in which the sheep never chewed for longer than 15 min are shown in Fig. 1 and demonstrate the necessity to prevent rumination if reasonably stable cephalic flow rates are to be obtained.
Head blood flow during hypertonic infusion
313
Isotonic control treatment The plasma sodium concentration showed a small increase during the first 60-90 min of blood flow measurement and thereafter remained reasonably stable for the remainder of the experiment. At the same time the plasma potassium concentration fell slightly and the osmolality of the plasma showed little change. The haematocrit fell throughout the entire experiment with the greatest rate of fall occurring in the initial stabilization before blood flow measurement began and with a slower rate of decline (averaging about 2%) during the 4 5 h period of flow measurement. In individual experiments the cephalic plasma flow tended to oscillate slowly during the period of observation. The mean difference in flow rate between maximum and minimum flows was 80 ± 10 (S.E. of mean) ml. min- and the time interval between the occurrence of these maxima and minima was 135 +23 (S.E. of mean) min. Since these changes in plasma flow did not occur at the same time in each experiment the mean plasma flow through the head showed no significant changes during the entire 4-5 h of measurement which could be attributed to the isotonic infusion. The mean oscillation in cephalic blood flow was 106 + 13 ml. min-' and the periods of maximum and minimum flow coincided with those of plasma flow since alterations in haematocrit during the period of flow measurement were not large enough to prevent this synchronization (Fig. 2).
Hyperosmotic NaCI treatment Ilp to the beginning of hyperosmotic infusion the treatment of all sheep was identical and thus the results obtained to that point in the 1 mol.1-1 NaCl and 1 mol.l1 KCI experiments were very similar to those of the control experiment. When hyperosmotic NaCl was infused the concentration of sodium in plasma and the osmolality of the plasma rose to levels significantly greater than those in the control experiment, whereas the plasma potassium concentration declined slowly as in the control treatment (Fig. 2, Table I). The haematocrit fell throughout the hyperosmotic NaCl experiment and during the period of 1 mol. 11 NaCl infusion appeared to fall more rapidly than during the corresponding time in the control experiment (Fig. 2) but this difference between the two experiments was not statistically significant. Cephalic plasma flow was not significantly influenced by the infusion of hyperosmotic NaCl and the pattern of changes in flow throughout the entire period of measurement was similar to that observed in the control treatment. The mean difference in flow rate between the maximum and minimum flows was 65 + 10 (S.E. of mean) ml. min- and the time interval between these maxima and minima was 153 + 25 min. Cephalic blood flow showed oscillations of 88 + 11 ml. min -' coincident with those for plasma flow and as these changes in plasma and blood flow did not occur at the same time in each experiment the mean plasma and blood flows through the head of the sheep were reasonably stable over the entire period of flow measurement (Fig. 2). D
_o\.ou.' / fi.(DU\
314
Beal
(]
,
! NaCI C
Control
155
Plasma
KCl
_
145Lff4{{f 9
(mmol/l)IJa: Plasma
_*@@O@S@@@O@@@@@@@.@
Osmolaility
*
(mosmollkg)290E Haematocrit
a' a'
L
*
*S@eeeeeee,.,...,.,
+
.
L*1J++*+ *4*YTT
i
40_
_
t
i
4"0
25L 450 _
Plasma Flow (ml /min)
Cepholic
.
.
TT?a
,
t
Blood Flow
(ml/min)
,
_ 200 600 _
-t .1
_
300_*
**
2
4
_
IIIf
I
0
1
0
Time
2
4
0
2
4
(hours)
FIo. 2. The plasma sodium and potassium concentration, plasma osmolality, haematocrit, cephalic plasma flow and cephalic blood flow before, during and after the intravenous infusion of 1 mol. 1- I NaCl and 1 mol. 1- I KCI at 0 8-1 0 ml. min-1 for 2 h. The control treatment received isosmotic infusion only. The filled squares give the values for parameters estimated on the initial blood sample taken before any other experimental procedure was commenced (N = 10; means ±S.E. of mean).
Hyperosmotic KCl treatment The infusion of 1 mol . 1- KCI increased the plasma potassium concentration to levels well in excess of those occurring in the previous two treatments (Fig. 2; Table I). The sodium concentration of the plasma fell slightly during this infusion but the difference in concentration between this treatment and the control treatment was not statistically significant (Table I). The rise in plasma
Head blood flow during hypertonic infusion
315
TABLE I. Summary ofthefinal hsd analyses of cephalic plasma and bloodflows, haematocritandplasma electrolyte concentrations for the 8 blood samples during and the first 2 blood samples after the 2 hour intravenous infusion of 1 mol. 1- I NaCl and 1 mol. 1- 1 KCl and the corresponding periods of the control experiment. Every variable on the ordinate of Fig. 2 was tested and omission of a comparison from this table means that no significant differences were found. Differences are indicated as levels of significance (* = P < 0-05; ** = P < 0-01; *** = P < 0001; - = not significant). P is the probability of erroneous rejection of the null hypothesis (i.e. type 1 error). Variable Comparison Blood Sample degrees of NaCl KCI Test Infusion Recovery 7 1 2 6 8 freedom = 25 versus versus 2 3 4 1 5
Plasma Na concn. Plasma K concn. Plasma Osmolality
Haematocrit Plasma Flow
Blood Flow
Control
* * * * *** *** NaCl - *** *** *** Control *** * * * * * * *** *** *** *** *** *** *** *** *** NaCl * * - ** * *** *** * Control Control ** * * ** * *** *** ** ** ** * *** *** *** Control * * * ** *** *** *** *** *** *** *** NaCl * * * Control * NaCl * * * * * ** ** Control - ** ** NaCl
osmolality during potassium infusion was of similar magnitude to that occurring during the hyperosmotic NaCl infusion. In contrast to the previous treatments the haematocrit rose during the infusion of KCl, at a slow rate initially and at a much more rapid rate during the second part of the infusion (Fig. 2). The point of inflection in the curve for mean haematocrit coincided with the peak values for mean plasma and blood flows, but this did not occur in individual experimnents. During the initial part of all potassium infusions the plasma flow through the head was always elevated above the pre-infusion (stabilization) values reaching a maximum in most experiments during the first hour. Although the size of this increment in cephalic plasma flow varied considerably between experiments (range 6-31 %) the rates of flow after 45 and 60 min of potassium infusion were always greater than those for the corresponding times in the control and hyperosmotic NaCl treatments (Fig. 2, Table I). The mean value for the maximum increase in cephalic plasma flow was 1-163 +±0029 (S.E. of mean) times the mean plasma flow before potassium was infused and the mean plasma potassium concentration at the peak of the increase in plasma flow was 5-92 + 0O20 mmol. I . As the potassium infusion continued the rate of flow declined so that in all but 3 experiments the cephalic plasma flow rate at peak hyperkalaemia was less than pre-infusion levels. When the potassium infusion was stopped plasma flow did not recover immediately and frequently continued to fall rapidly for a further 15-30 min so that mean plasma flow rates had not returned to pre-infusion values until 45 min after the potassium infusion had
316 Beal ceased (Fig. 2, Table I). Cephalic blood flow also increased during the initial period of potassium infusion and declined during the latter part of the infusion but as a result of the concomitant increase in haematocrit the increment in blood flow at the beginning of potassium infusion was slightly larger than that for plasma flow. The mean value for the maximum increase in cephalic blood flow was 1-176 +0-032 (S.E. of mean) times the mean blood flow before potassium infusion and the mean plasma potassium concentration at maximum blood flow was 6 03 + 0-22 mmol . 1-1. Although the rate of flow of blood through the head declined during the second half of the potassium infusion it did not fall as rapidly as plasma flow due to the increasing haematocrit. At the end of potassium infusion blood flow rates were less than the pre-infusion values in only 4 of 10 experiments. Blood flow continued to fall during the first part of the recovery period following the infusion of potassium but, unlike plasma flow, the fall in cephalic blood flow was not statistically significant (Fig. 2, Table I). DIsCUSSION In the sheep the common carotid arteries provide the total blood supply of the extracerebral region and almost all the cerebral region except the caudal medulla oblongata [Baldwin and Bell, 1963]. Although each side of the head behaves as a discrete circulation they are interconnected and partial or complete occlusion of a carotid artery results in an immediate redistribution of supply via the remaining patent vessels. Therefore, the repeated measurement of cephalic blood flow in the conscious animal at short intervals over a number of hours necessitates the use of a technique which will allow simultaneous measurement of flow through both sides of the head and does not itself cause marked changes in flow. The indicator-dilution technique used in the present experiments fulfils the requirement satisfactorily. The use of PAH as the indicator substance has the advantage that rapid extraction of the compound from the plasma occurs in the kidney and keeps the level of recirculation low and reasonably stable. Thus the arterio-venous difference in concentration becomes relatively large which increases the accuracy of flow measurement. The major source of error in the use of an indicator-dilution technique for the measurement of a regional circulation is the possibility that the indicator may escape from the circulation and be stored, excreted or metabolized within the region under study. The 2 h period of stabilization before blood sampling should have been sufficient for equilibration to occur between the blood and any extravascular space in the head into which PAH could diffuse rapidly. Diffusion gradients taking longer than 2 h to equilibrate would have no significant effect on the accuracy of flow measurement. The close agreement between the plasma PAH concentrations of jugular and carotid blood during intravenous PAH infusion supports the above thesis and also indicates that no significant metabolism or excretion of PAH occurs in the head. No attempt was made to measure PAH concentrations in cephalic lymph but estimations of PAH in cerebrospinal fluid and parotid saliva showed that neither of these secretions contained sufficient PAH to materially affect the accuracy of flow measurement.
Head blood flow during hypertonic infusion 317 The venous sampling cannulae were placed a few centimeters below the junction of the external and internal maxillary veins to ensure that blood samples were a reasonable mixture of blood from these vessels. This positioning of the cannulae results in blood flow from the thyroid, larynx and pharynx being included in the total cephalic flow measurements. The necessity to fast the animals before experiment reduces the water intake of the sheep which in turn, leads to a reduction in the total volume of body water and presumably a reduction in cardiac output. Therefore the values obtained for cephalic plasma and blood flow may represent the lower end of the range of flow values found in conscious sheep. However shorter periods of fasting are associated with much greater likelihood of rumination which results in marked increases in cephalic blood flow (Fig. 1). The high initial haematocrit values obtained in these experiments have been shown under similar experimental conditions to be due to splenic contraction. Likewise the large increments in haematocrit which occurred during potassium infusion, although partly due to redistribution of fluid between body compartments, were also mainly the result of release of red cells from the spleen [Beal, 1976]. The intravenous infusion of hyperosmotic potassium chloride solution into the sheep invariably resulted in an increase in cephalic plasma and blood flows as the plasma potassium concentration rose to about 6 mmol . 1-1 followed by a progressive decline in the flow rates as the level of hyperkalaemia increased further. These increased flow rates could be caused by a decrease in peripheral resistance or an increase in arterial blood pressure (B.P.). In anaesthetized dogs and cats raised plasma osmolality is associated with a reduction in vascular peripheral resistance [Overbeck, Molnar and Haddy, 1961; Mellander, Johansson, Gray, Jonsson, Lundvall and Ljung, 1967; Jelks and Emerson, 1974]. The infusions of 1 mol .l-1 NaCl and 1 mol .l-1 KCI into the sheep resulted in approximately equal elevation of plasma osmolality and presumably also of plasma chloride concentration but only the potassium infusion was associated with plasma and blood flow changes which would suggest an effect on peripheral resistance. As the arterial BP of sheep remains constant during hyperosmotic sodium chloride infusion [Beal, 1976] any tendency for increased plasma osmolality to cause a fall in vascular resistance must be offset by the ability of the conscious animal to prevent vasodilation reflexly. In the conscious intact sheep, arterial BP was not depressed until the plasma potassium concentration exceeded 6 mmol .l-1 and frequently BP rose slightly at the onset of potassium infusion [Beal, 1976]. The increases in plasma flow and blood flow through the head of the sheep during the early part of the potassium infusions were therefore the result of decreased peripheral resistance which in some cases may have been coupled with a slight increase in arterial BP. This vasodilatory action is consistent with the reported effect of hyperkalaemia on various vascular beds in the anaesthetized dog [Scott, Daugherty, Overbeck and Haddy, 1968]. At higher plasma potassium concentrations the plasma flow through the head of the sheep declined and frequently fell to flow rates below the levels which existed before potassium infusion. These changes in cephalic
Beal 318 plasma flow match well with the changes in renal plasma flow observed during acute hyperkalaemia in the conscious sheep [Beal, Budtz-Olsen, Clark, Cross and French, 1975]. The tendency for cephalic plasma and blood flows to fall at high plasma potassium concentrations suggests that arterial BP was falling and this would seem to be confirmed by the BP data for conscious intact sheep receiving potassium infusion [Beal, 1976]. However, the fall in arterial BP accompanying the fall in cephalic plasma flow cannot be as substantial as would be expected from the changes in the rate of plasma flow since cephalic blood flow reaches maximum values at higher plasma potassium concentrations than plasma flow and falls more slowly due to the concurrent increase in haematocrit. When the potassium infusion was terminated, the rates of cephalic plasma and blood flow fell precipitously to the lowest values observed in the experiments and in general, had not returned to normal values until 45 min of the recovery period had elapsed. As no fall in arterial BP was observed immediately following the cessation of potassium chloride infusion [Beal, 1976] these low flow rates were apparently the result of vasoconstriction. The observation that the haematocrit of conscious sheep increases with the severity of hyperkalaemia provides a simple explanation for a previous finding on renal haemodynamics. Beal et al. [1975] observed that the glomerular filtration rate of conscious sheep undergoing a potassium infusion rises and often reaches maximum values at higher plasma potassium concentrations than does the renal plasma flow and that the subsequent fall in glomerular filtration rate at even higher plasma potassium concentrations was relatively slower than the corresponding fall in renal plasma flow. Obviously renal blood flow at high plasma potassium concentrations would not fall in direct proportion with the fall in renal plasma flow due to the increasing volume of red cells in the blood and this would result in better maintenance of arterial BP, filtration pressure and glomerular filtration rate than would be expected from the changes in renal plasma flow. ACKNOWLEDGMENT I am indebted to Mr P. V. Burrow for skilful technical assistance.
REFERENCES BALDWIN, B. A. and BELL, F. R. (1963). The anatomy of the cerebral circulation of the sheep and ox. The dynamic distribution of the blood supplied by the carotid and vertebral arteries to cranial regions. Journal of Anatomy, 97, 203-215. BEAL, A. M. (1976). Changes in arterial blood pressure, heart rate and haematocrit during acute hyperkalaemia in conscious sheep. Quarterly Journal of Experimental Physiology. 61, 297-308. BEAL, A. M., BUDTZ-OLSEN, 0. E. and CLARK, R. C. (1975). The effect of potassium chloride infusion on parotid salivary flow and composition in conscious sheep. Quarterly Journal of Experimental Physiology, 60, 161-169. BEAL, A. M., CLARK, R. C. and BUDTZ-OLSEN, 0. E. (1975). The composition and flow of parotid saliva during acute hyperkalaemia in sodium-deficient sheep. Quarterly Journal of Experimental Physiology, 60, 315-323.
Head blood flow during hypertonic infusion
319
BEAL, A. M., BUDTZ-OLSEN, 0. E., CLARK, R. C., CROSS, R. B. and FRENCH, T. J. (1975). Changes in renal haemodynamics and electrolyte excretion during acute hyperkalaemia in conscious adrenalectomized sheep. Quarterly Journal of Experimental Physiology, 60, 207-221. BRATTON, A. C. and MARSHALL, E. K. (1939). A new coupling component for sulphanilamide determination. Journal of Biological Chemistry, 128, 537-550. COATS, D. A., DENTON, D. A., GODING, J. R. and WRIGHT, R. D. (1956). Secretion by the parotid gland of the sheep. Journal of Physiology, 131, 13-3 1. DENTON, D. A. (1957). The study of sheep with permanent unilateral parotid fistulae. Quarterly Journal of Experimental Physiology, 42, 72-95. EMANUEL, D. A., ScoTr, J. B. and HADDY, F. J. (1959). Effect of potassium upon small and large blood vessels of the dog forelimb. American Journal of Physiology, 196, 637-742. GARB, S. (1951). The effects of potassium, ammonium, chloride, calcium, strontium and magnesium on the electrogram and myogram of mammalian heart muscle. Journal of Pharmacology and Experimental Therapeutics, 101, 317-326. HARVEY, R. B. and BROTHERS, A. J. (1962). Renal extraction of para-aminohippurate and creatinine measured by continuous in vivo sampling of arterial and renal vein blood. Annals of the New York Academy of Sciences, 102, 46-54. JELKS, G. W. and EMERSON, T. E., JR. (1974). Effects of plasma electrolyte abnormalities on total peripheral resistance and other hemodynamic parameters in dogs. Proceedings of the Society for Experimental Biology and Medicine, 146, 59-65. LoGic, J. R., KROTKIEWSKI, A., KoPPius, A. and SURAWICZ, B. (1968). Negative inotropic effect of K+: its modification by Ca"+ and acetyl strophanthidin in dogs. American Journal of Physiology, 215, 14-22. MELLANDER, S., JOHANSSON, B., GRAY, S., JONSSON, O., LUNDVALL, J. and LJUNG, B. (1967). The effects of hyperosmolarity on intact and isolated vascular smooth muscle. Possible role in exercise hyperemia. Angiologica, 4, 310-322. OVERBECK, H., MOLNAR, J. I. and HADDY, F. J. (1961). Resistance to blood flow through the vascular bed of the dog forelimb. American Journal of Cardiology, 8, 533-541. SARNOFF, S. J., GILMORE, J. P., MCDONALD, R. H., DAGGETr, W. M., WEISFELDT, M. L. and MANSFIELD, P. B. (1966). Relationship between myocardial K+ balance, 02 consumption and contractility. American Journal of Physiology, 211, 361-375. SCOTT, J. B., EMANUEL, D. A. and HADDY, F. J. (1959). The effect of potassium on renal vascular resistance and urine flow rate. American Journal of Physiology, 197, 305-308. ScoTr, J. B., DAUGHERTY, R. M., JR., OVERBECK, H. W. and HADDY, F. J. (1968). Vascular effects of ions. Federation Proceedings, 27, 1403-1407. SELDINGER, S. I. (1953). Catheter replacement of the needle in percutaneous arteriography. Acta Radiologica, 39, 368-376. STEEL, R. G. D. and TORRIE, J. H. (1960). Principles andProcedures in Statistics, pp. 109-1 10. McGraw-Hill, New York.
