Importance of changes in plasma HCO3− on regulation of CSF HCO3− in respiratory alkalosis

RespirationPhysiology(1976) 26, 265-278; North-Hoknd PublishingCompany, Amsterdam

IMPORTANCE OF CHANGES IN PLASMA HCO, ON REGULATION OF CSF HCO; IN RESPIRATORY ALKALOSIS’

L. CHOMA’

and H. KAZEMI

Medical Services (Pulmonary Unit), MassachusettsGeneral Hospitaland Departmentof Medicine, Harvard Medical School, Boston, Mass. 02114, U.S.A.

Ah&r&: In respiratory alkalosis the fall in CSF bicarbonate is in part due to increased CSF lactate. The rest of CSF HCO; fall may be actively regulated or as more recent evidence suggests is dependent on plasma HCO; fall. Therefore, the relationship between plasma and CSF HCO; changes was studied during 4 hours of respiratory alkalosis (Pso-,, = 20 mm Hg) in anesthetized dogs when plasma HCO; : (1) fell normally, (2) kept ‘normal’ by NaHCO, infusion, (3) increased by infusing more NaHCO,, and (4) reduced by infusing HCl. In respiratory alkalosis plasma and CSF HCO; fell 4.6 and 3.8 mEq/L, respectively. In hypocapnia and ‘normal’ plasma HCO; CSF HCO; fell 2 mEq/L and lactate increased 1.33 ml&/L. In hypocapnia and metabolic alkalosis plasma HCO; increased 6.5 mEq/L and CSF HCO; remained unchanged and lactate increased 2.12 mEq/L. In combined hypocapnia and metabolic acidosis plasma HCO; fall 10.5 mEq/L but CSF HCO; fell 3.1 mEq/L and CSF pH returned to normal at 4 hours. Therefore CSF HCO; fall in hypocapnia is primarily and critically dependent on the simultaneous fall in plasma HCO; content, with a minimal contribution from CNS lactate increase. When CSF pH has returned to normal, however, CSF HCO; fall is stopped despite further falls in plasma HCO; . CSF pH H+ homeostasis HCO; movement

Lactate Regulation of respiration

The ionic composition of the cerebrospinal fluid (CSF), particularly its [H”], is important in the central chemical control of ventilation (Pappe~e~~r, 1967; Leusen, 1972), as well as in the regulation of cerebral blood flow (Fencl et al., 1969), and in consciousness (Posner and Plm, 1967). In respiratory acid-base disturbances there is better CSF pH compensation than in blood where the CSF [I-I’] is regulated primarily by appropriate changes in the CSF bicarbonate content (see Leusen, 1972). In instances of respiratory alkalosis there is a greater fall in CSF HCO; than that in blood (Van Vaerenbergh et al., 1965b; Plum and Posner, 1967; Kazemi et al., Acceptedfor publication15 January 1976

* Supported by grants HL-16925 and HL-06664 from National Institutes of Health. * Work done during a fellowship in part supported by: Conselbo de Ensino e P&quisa da Universidade Federal do Parana, Brasil. 265

266

L. CHOMA AND H. KAZEMI

1967, 1969). M~hanisms responsible for CSF HCO; regulation in respiratory alkalosis have not ‘been clearly identified, although active regulation of CSF bicarbonate through local mechanisms in the CNS and at the brain-blood barrier has been proposed (Severinghaus et al., 1963), and more recently Hombein and Pavlin (1975) have presented evidence to suggest that distribution of H+ and HCO; in the CSF in respiratory alkalosis is a function of the d-c potential difference between blood and CSF. Brain and CSF lactate concentrations are selectively increased in respiratory alkalosis (Leusen and Demeester, 1966; Kazemi et al., 1969) and the increase in lactic acid can reduce the bicarbonate concentration at both sites (valenca et al., 1971). Since aside from the lactate increase in the CNS no other specific m~hanisms have been demonstrated for regulation of CSF HCO; in respiratory alkalosis, possible dependence of CSF HCO; fall on plasma HCO; reduction has been emphasized {Dempsey et al., 1974, 1975). Earlier studies from this laboratory have shown that less than one third of the fall in CSF HCO; in respiratory alkalosis can be attributed to the simultaneous increase in CSF lactate (Kazemi et al., 1969) and that brain organic buffers and ammonia which contribute to H+ buffering in respiratory acidosis play no role in H+ homeostatis in the CSF during hypocapnia (Kazemi et al., 1973). More recent studies have shown that in both respiratory acidosis and alkalosis CSF HCO; changes are influenced by the concurrent changes in plasma bicarbonate concentration (Wichser and Kazemi, 1975). In respiratory acidosis the increase in plasma HCO; is reflected equally in the CSF, but in addition there is local formation of HCO; in the CNS (a reaction catalyzed by the carbonic anhydrase present in the choroid plexus and glial cells) allowing for a greater increase in CSF HCO; than in blood and thus better CSF pH compensation. In respiratory alkalosis the concomitant increase in CSF lactate allows for a greater fall in CSF HCO,, but that once the CSF HCO; decrease due to the simultaneous CSF lactate increase is accounted for the fall in CSF HCO; becomes identical to the fall in arterial plasma HCO; . The present investigation was then undertaken to quantitate the relationship between plasma and CSF HCO; falls in respiratory alkalosis when at a Pace, of 20 mm Hg: (a) plasma HCO; fall was prevented by infusion of NaHCO,, (2) plasma HCO; was increased by infusing larger quantities of NaHCO,, and (3) plasma HCO; was markedly reduced by infusion of hydr~hlori~ acid. The results indicate significant dependence of CSF HCO; alterations on s~ul~neous changes in plasma HCO; in respiratory alkalosis within broad limits.

Methods Mongrel dogs (average wt. 17.5 kg) were anesthetized with intravenous sodium pentobarbital 30 mg/kg initially and 4.5 mg/kg every 2 hours throughout the experiments. They were paralyzed by subcutaneous injections of 40 mg succinylchohne initially and 20 mg every 2 hours thereafter. They were intubated with an endo-

CSF HCO;

REGULATION IN RBSPIRATORY ALKALGSIS

267

tracheal tube and ventilation was maintained on room air by a constant volume respirator. Tidal volume and frequency of the respirator were adjusted to keep the initial Pko, at about 38 f 2 mm Hg. Atelectasis was prevented by an intermittent hyperinflation valve (Harvard Apparatus Co., Millis, Mass.) which produced hyperinflation of the lungs every 10 minutes by introduction of 30 cm H,O pressure at the expiratory outlet of the respirator. A polyethylene catheter was inserted into a femoral artery and the arterial blood pressure was monitored osciilographically through a Statham transducer. A femoral vein was also catheterized for adniinistration of fluids. Another polye~ylene catheter was inserted into the saggital sinus and the tip of the catheter was directed posteriorly toward the occiput. An # 18 gauge spinal needle with a three way stopcock attached was inserted into the cisterna magna.

MEAsuREMEm

At the begixming of the experiments and then periodically 7 ml samples of femoral and saggital sinus blood were collected anaerobically in heparinized syringes. Cisternal CSF samples (3 ml) were drawn an~robi~lly in previously dry syringes whose dead space and stopcock were flushed with CSF just prior to sampling. CSF samples contaminated with air bubbles or blood were discarded. Duplicate determinations of blood Po,, Pco, and pH as well cistemal CSF Pco, and pH were made at 37 “C in appropriate electrodes (Models 123 and 125, Instrumentation Laboratory, Inc., Boston, Mass). The bicarbonate concentrations were calculated from the Henderson-Hasselbalch equation, using pK’ values of 6.15 for CSF and 6.10 for blood and CO, solubility factors of 0.0324 and0.0301 respectively (Alexander et al., 1961). Cerebral oxygenation was assessed by determination of 0, content in arterial and saggital sinus blood by means of an 0, analyzer (Lex-O&on, Lexington Instruments Corp, Waltham, Mass). Samples of blood and CSF were i~~ia~ly transferred to test tubes containing 16 % per&lo& acid, centrifuged and filtered for analysis of lactate content by the enzymatic method (Lactate TC-B, No 15972, C. F. Boehringer and Son, Mannheim, Germany).

DESIGN OF EXPERIMENTS

In order to study the effect of variations in plasma bicarbonate level during respiratory alkalosis on the CSF bicarbonate content four separate groups of experiments were conducted where the Pace, was maintained at about 20 mm Hg and the plasma HCO; content altered by infusion of varying quantities of NaHCO or HCl. In these experiments blood acid-base measurements were made at 0, 1,2,3 and 4 hours of hypocapnia and CSF measurements at 0, 2 and 4 hours. All animals were ven-

268

L. CHOMA AND H. KAZFiMI

tilated with room air and maintains 0 time measur~ents.

at a normal Pao,z for one hour prior to the

Group I: Respiratory alkaiosis

Six dogs were mechanically hyperventilated on room air. The Pa,, was reduced to 20 mm Hg by increasing the frequency and stroke volume of the respirator and maintained there for 4 hours and measurements obtained as described above. These animals formed the control studies for all the subsequent experiments. Group II: Respiratory alkalosis and ‘normal’ plasma HCO, Twelve animals were m~hanically hy~~entilat~ as in group I, but with the onset of hypocapnia an intravenous infusion of 1 molar NaHCO, solution was also started by a continuous infusion pump at the rate of approximately 0.5 ml/min. in order to keep plasma bicarbonate between 22 to 24 mEq/L. Measurements were obtained as in group I. At the end of 4 hours the mean quantity of NaHCO, infused in each animal was 108 mEq. In 6 of the 12 animals both the hypocapnia and NaHCO, infusion were continued for another 4 hours (total of 8 hours) in order to study the effect of time on CSF HCO; alterations. In these six animals the mean total quantity of NaHCO; infused was 162 mEq over the 8 hours. Group 111: Respiratory alkalosis and met~olic

alkalosis

Six dogs were m~hani~lly hy~~entilated for 4 hours as in previous experiments, but with the onset of hypocapnia intravenous infusion of 1 molar NaHCO, was started by a continuous infusion pump at the rate of 1 ml/min in order increase plasma HCO; to about 29 mEq/L and maintain it at that level for 4 hours. The mean total quantity of NaHCO, infused in each animal was 217 mEq/L. An additional control group of experiments were conducted in this series to evaluate the effect of raising only the plasma bicarbonate on CSF bicarbonate. In six additional dogs the Pqo, was kept at 40 f 2 mm Hg for 4 hours while the animals received a continuous infusion of 1 molar NaHCO,. The mean total quantity of NaHCO, infused during the 4 hours was 150 mEq. Group IV:

Respiratory alkalosis and the metabolic acidosis

Six animals were mechanically hyperventilated for 4 hours as in previous groups and simultaneously received a continuous intravenous infusion of 0.2 N HCl in order to lower and keep plasma bicarbonate at about 12 mEq/L during the 4 hours of hypocapnia. The mean total quantity of HCl infused in each animal was 46 mEq. STATISTICAL

ANALYSIS

Results of all the experiments t-test (Hill, 1971).

were subjected to statistical analysis by Student’s

CSF HCO,

CRAYON

IN MPIRATORY

ALKALOSIS

269

RdtS RESPIRATORY ALKALOSIS (fig.

1)

The fall in the mean Pace, from 38 to 20 mm Hg for 4 hours was associated with a fall in the mean CSF Pco, in the six animals from 46 to 30 mm Hg (P < 0.01). The mean arterial pH increased from 7.384 to 7.554 at 4 hours (P < 0.01) and the CSF 60

?.BOf

r

OL

7.20L I

26 2’

I

I

I

I

I

I

I

I

O----O

ARTERIAL

B

CISTERNAL

I

CSF

Fig. 1. Arterial and cisterual CSF acid-base parameters and lactate during 4 hours of respiratory alkalosis produced by mechanical hyperventilation on room air (group I). Mean values in 6 dogs. Vertical bars are SEM.

pH from 7.323 to 7.419 (P < 0.01). The mean arterial HCO; decreased from 21.9 to 17.3 mEq/L at 4 hours and the mean CSF HCO; from 21.8 to 18.1 mEq/L (P < 0.05 for both). Arterial lactate increased from 0.62 to 1.04 mEq/L, and CSF lactate from 0.72 to 1.33 mEq/L (P < 0.01) after 4 hours of hypocapnia. Arterial oxygenation was adequate with mean Pao, above 87 mm Hg during the experiments, and the arterial-saggital sinus 0, content difference increased from 4.0 ~01% to 6.4 ~01% after 4 hours of hypocapnia, indicating some fall in cerebral blood flow (table 1).

RESPIRATORY ALKALOSIS AND ‘NORMAL’ PLASMA

HCO;

(fig. 2)

During 4 hours of respiratory alkalosis and intravenous NaHCO, infusion the mean Phz in 12 dogs was maintained at 20 mm Hg and the mean CSF Pea, fell from 43 mm Hg to 28 (P < 0.01). The mean arterial pH became alkaline at 7.711 (P < 0.001 compared to control), and mean CSF pH increased from 7.316 to 7.464 (P < 0.01). Infusion of bicarbonate caused a slight increase in plasma HCO;,

270

L. CHOMA AND H. KAZEMI

OL

7.2oL I

I

I

I

L

I

26 \’ 5

22

Ic?

16

s

14 0+

I

I

I

o----o

ARTERIAL

-

CISTERNAL

I

CSF

o-4

4

HOURS

HOURS

Fig. 2. Arterial and cistemal CSF acid-base parameters and lactate during 4 hours of respiratory alkalosis produced by mechanical hyperventilation on room air while plasma HCO; was kept relatively ‘normal by a continuous intravenous infusion of NaHCO; (group II). Mean values in 12 dogs. Vertical bars are SEM.

reaching a value of 24.3 mEq/L at 3 and 4 hours. The mean CSF HCO, fell from 20.7 mEq/L to 18.7 (P < 0.05) at 4 hours. There was an increase in the mean CSF lactate from 0.92 mEq/L to 2.25 at 4 hours (P < 0.05). The rise in arterial lactate was less going from 0.65 mEq/L to 1.53 at 4 hours (P < 0.05). TABLE 1 Arterial PO1and arterial-sag&al sinus 0, content differences

%-cvo2 w %I Group I control at 4 hours

87.2k6.1 111.8*3.5*

4.0*0.3 6.4f0.6*

Group II control at 4 hours

83.4k2.4 105.8f3.1*

6.0 kO.5 8.7 fO.S*

Group III control at 4 hours

88.3 f 3.7 118.3f4.8*

4.6 kO.7 9.1 f0.8*

Group IV control at 4 hours

94.3 f 1.5 121.2+5.7*

5.OkO.6 16.0 f 1.O*

Mean values in at least 6 dogs + SE. l P value < 0.025 as compared to control values.

7.537 +0.03*

50.7kl.7

39.5+1.1

4 hours

7.683 &0.04

7.382 kO.08

30.4*0.7

19.7+0.2

8 hours

7.699kO.01

A

38.8 + 1.7 48.1 f 1.9

28.lkO.3

19.4kO.4

4 hours

-

PH

0 time

C

Hg)

A

:z

* Statistically significant at Pvalue < 0.025 as compared to 0 time.

Normocapnia + NaHCO, infusion

Hypocapnia + ‘normal plasma HCO;

time

32.6fl.3*

22.3f0.08

7.324kO.01 7.336 f0.01

22.7fO.l

24.4kO.4

A

7.41 +0.01

7.457kO.01

C

(m&/L)

HCO;

25.2f0.8*

23.Ok1.14

17.9f0.4

18.7 &OS

C

1.37f0.13*

0.88+0.09

2.37hO.73

1.27f0.32

0.74+0.05

2.25k0.19

1.53kO.28

0.84+0.15

C

A

(m&/L)

Lact.

TABLE 2 Acid-base changes in arterial blood (A) and CSF (C) from 4 to 8 hours of hypocapnia and ‘normal’ plasma HCO; , and during 4 hours of normocapnic metabolic alkalosis (mean values in 6 dogsIf:SEM)

272

L. CHOMA AND H. KAZEMI

The mean Pao2 was above 83 mm Hg in these experiments, and the arterial-sagittal sinus 0, content difference increased from 6.9 ~01% to 8.7 % at 4 hours (table 1) again indicating a fall in cerebral blood flow. In six animals in this group hypocapnia and NaHCO, infusion were carried out for 8 hours to see if further changes in CSF HCO; level occurred as a function of time. There were no significant changes in arterial or cistemal CSF pH, Pco2, HCO; or lactate between 4 and 8 hours (table 2) and therefore the changes seen at 4 hours represent reasonable steady state values under these experimental conditions.

RESPIRATORY ALKALOSIS AND METABOLIC ALKALOSIS (fig.

3)

In six animals when a larger quantity of NaHCO, was infused in association with hypocapnia for 4 hours the mean Pko, was reduced from 40 to 20 mm Hg and the mean CSF Pco, fell from 44 to 29 mm Hg. The mean arterial pH rose from 7.391 to 7.791 (P < O.OOl),while the mean CSF pH increased from 7.317 to 7.506 at 4 hours (P < 0.01). The mean arterial HCO; increased to 29.5 mEq/L at 2 hours and was the same at 4 hours. The mean CSF HCO; was 20.8 mEq/L at 0 time and remained essentially the same (2 1.4) at 4 hours. The mean arterial lactate increased from 0.66 mEq/L to 2.10 at 4 hours, and the mean CSF lactate increased from 0.91 mEq to 3.04 (P < 0.002) at 4 hours. The mean Pao2 was above 88 mm Hg in these animals, and the cerebral a-v 0, content difference increased from 4.6 ~01% to 9.1 vol y0 (table 1) indicating approximately a 50 % reduction in cerebral blood.

O----O

ARTERIAL

_

CISTERNAL

CSF

Fig. 3. Arterial and cistemal CSF acid-base parameters and fact&e during 4 hours of respiratory aIkalosis produced by mechanical hyperventifation on room air with simultaneous metabolic alkalosis induced by intravenous infusion of NaHCO, (group III). Mean values in 12 dogs. Vertical bars are SEM.

CSF HCO;

REGULATION IN RESPIRATGRYALKALGSIS

273

In order to assess the effect of high plasma HCO; on CSF HCO; in the absence of hypocapnia, NaHCO, was infused in 6 dogs while their Pa,,, was kept in the normal range. In these animals the mean plasma HCO; was increased from 22.3 mEq/L to 32.6 at 4 hours and there was a slight but significant increase in the mean CSF HCO; from 23.0 to 25.2 mEq/L (table 2). There was also a significant increase in the mean CSF lactate from 0.7 to 1.37 mEq/L (P < 0.05), but there was no change in the arterial lactate (table 2). SPERRY

ALKALOSIS AND METABOLIC ACDOSIS (fig.

4)

In six animals infusion of 0.2 N hydrochloric acid was added to 4 hours of hypocapnia and again the mean Paco2 was maintained at 20 mm Hg as in the previous three groups of experiments. However, the mean CSF PC-,, was somewhat higher at 33.3 mm Hg after 4 hours of hypocapnia than in the other experiments. The mean arterial pH had returned to its control value of 7.380 at 4 hours. The mean CSF pH was 7.371 at 4 hours, a value slightly but insignificantly (P = 0.2) above the control value of 7.341. The mean arterial HCO; fell from 22.3 mEq/L to 12.2 at 2 hours (P < 0.01) and did not change significantly thereafter. The mean CSF HCO; fell from 21.0 mEq/L to 17.9 (P < 0.05) at 4 hours. There was almost an identical increase in arterial and CSF lactate ~on~ntrations at 4 hours - the arterial going from 0.99 mEq/L to 1.60 and the CSF from 0.99 to 1.88. The mean P%, was above 94 mm Hg in these animals, but the cerebral a-v 0, content difference increased from 5.0 to 16.0 ~01% (table 1) indicating a marked reduction in cerebral blood flow.

fj”r

24 \”

o----O

ARTERIAL

m

CISTERNAL

CSF

20

8 0LL-S-U HOURS

4

0L-L2u-J4 MWt?S

Fig. 4. Arterial and cisternal CSF acid-base parameters and factate during 4 hours of respiratory akalo$s produced by me&an&I hy~~~tiIatiou on room air with simultaneous metabolic acidosis induced by intiavenous infusion of 0.2 N HCl (group IV). Mean values in 6 dogs. Vertical bars are SEM.

274

L. CHOMA AND H. KAZEMI

Stability of CSF pH in chronic states of acid-base imbalance has been stressed (~i~hell et al., 1965) and in a review of the subject Fencl(l971) pointed out CSF pH appears to be better protected when acid-base derangements are of ‘metabolic’ origin than when they are of ‘respiratory’ origin. Regardless of the cause of acid-base derangement CSF pH is better compensated than blood and in respiratory acidosis there is a greater increase in CSF HCO; than in plasma HCO; and in respiratory alkalosis a greater decrease in CSF HCO; than in plasma HCO; (Leusen, 1972). Our earlier studies suggested that the greater fall in CSF HCO; than plasma HCO; in respiratory alkalosis was essentially all due to the selective rise in brain and subsequently the CSF lactate content (Kazemi et al., 1969; Wichser and Kazemi, 1975). The increase in the CNS lactate content is due to increased brain lactic acid content and not to any increase in blood lactate, since there are blood-brain (Klein and Olsen, 1947) and blood-CSF (Van Vaerenbergh et al., 1965a) barriers to lactate diffusion, The lactate increase in the brain in hypocapnia is probably related to several factors which include increased anaerobic glycolysis because of reduced cerebral blood flow due to hypocapnia (Kety and Schmidt, 1948; Harper and Glass, 1965) accentuation of the Bohr effect because of the alkaline pH and thus less 0, delivery to brain tissue (Leusen and Demeester, 1966) and as proposed by Laborit A HC03mEq/L

At LoCiOtt

-6 m---o

RESfWAIY

D-9

RL +‘Nmd

o-4

RI

+

ALAAURAS

YETABOLZ

IRA1

\

10

‘r

PLLSMA HC03TALMLOSls

--aO.6

0w Wl?S

Fig. 5. Changes in bicarbonate content (dHC0;) in cisternal CSF and arterial plasma during 4 hours of respiratory alkalosis in the 4 experimental groups. The change in HCO; during the experiments is compared to each groups’ contro1 value at 0 time. Numbers of the right hand side of each panel indicate the rise in lactate (Id lactate) after 4 hours of hypocapnia in comparison with 0 time. For discussion see

CSF HCO;

RJJGUL.ATION IN RIBPIRATORY

ALKALOSIS

275

(1966) possibly a direct effect of reduced Pco, on delivery of fuel into the Krebs cycle with carboxylation of pyruvic acid to mahc acid resulting in increased lactic acid which had been in equilibrium with pyruvic acid. The increased brain lactate is then reflected in CSF, but the lactate increase in the CSF is insufficient stoichiometrically to aFount for the entire HCO; fall in the CSF. It has been suggested that active local m~hanisms in the CNS allowed for further reduction of CSF HCO; (Severinghaus et al., 1963) or that the CSF HCO; reduction was influenced by the concurrent fall in plasma HCO; (Wichser and Kazemi, 1975; Dempsey et al., 1975). The results of the present experiments suggest that the changes in plasma HCO; in respiratory alkalosis have a dominant and significant effect on the changes in CSF HCO; con~tration and, therefore, on hydrogen ion homeostasis in the CSF. The changes in CSF and arterial plasma bicarbonate content in the four groups of animals with hypocapnia are summarized in fig. 5, and changes in lactate concentration of both sites at the end of 4 hours of hypocapnia are also given on the right hand side of each panel. In pure respiratory alkalosis CSF bicarbonate decreased by 3.8 r&q/L (P < 0.01) while the arterial plasma HCO; decreased 4.6 mEq/L at 4 hours (P < 0.01). Simultaneously the CSF lactate increased by 0.6 mEq/L and arterial lactate by 0.4 mEq/L. The relative fall in CSF HCO; in these experiments was somewhat less than reported by us earlier (Kazemi et al., 1969; Wichser and Kazemi, 1975) as well as by others (Van Vaerenbergh et al., 1965b; Plum and Posner, 1967). This may in part be due to the very modest increase in CSF lactate of 0.6 mEq/L in these experiments, as opposed to increases of 2-3 mEq/L reported earlier. There is no ready explanation for the limited lactate increase in the CSF in the present series, but it could be due to a smaller reduction in cerebral blood in these experiments, as shown by the cerebral a-v 0, content difference (table 1), than reported earlier. However, with the limited increase in CSF lactate in these experiments, the fall in CSF HCO, was almost of the same magnitude as the fall in plasma HCO; content after 4 hours of hypocapnia. In group II experiments when the fall in plasma HCO; was prevented by infusion of NaHCO, and thus plasma HCO; kept relatively ‘normal’ by keeping it between 22 and 24 mEq/L during 4 hours of hypocapnia, the CSF HCO; fell 2.0 mEq/L (P < O.Ol), a value significantly less than the fall in CSF HCO; in pure respiratory alkalosis (P = 0.01). This fall of 2 mEqfL in CSF HCO; was matched almost equally by an increase in CSF lactate of 1.33 mEq/L (P < 0.01). Therefore when plasma HCO; fall was prevented in respiratory alkalosis the CSF HCO; fell only by an amount almost equal to the rise in CSF lactate, indicating the significant dependence of CSF HCO; reduction in hypocapnia on the simultaneous fall in plasma HCO;. Prolonging these experiments to 8 hours had no further effect on CSF HCO; content (table 2). Therefore without decreasing plasma HCO; the CSF HCO; does not fall in hypocapnia as a function of time, again emphasizing the dependence of CSF HCO; fall on the simultaneous plasma HCO; reduction. In group III experiments where a significant metabolic alkalosis was superimposed

276

L. CHOMA AND H. KAZJ3hB

on respiratory alkalosis, plasma HCO; increased by 6.5 mEq/L at the end of 4 hours (P < O.OOl),but there was no significant change in CSF HCO; . In these experiments CSF lactate increased by 2.12 mEq/L (P < O.OOl), and therefore it is possible CSF HCO; would have increased by about 2 mEq/L if there had been no increase in CSF lactate. This would suggest that some bicarbonate may have entered the CSF from plasma; where the HCO; level was markedly increased. For this reason another control series of experiments were performed, when during 4 hours of normocapnia arterial plasma HCO; was maintained about 10 mEq/L above control. In these experiments CSF HCO; increased by 2 mEq/L at the end of 4 hours (table 2), suggesting some entry of HCO, from plasma into CSF where plasma HCO; was sustained at a high level for 4 hours. In group IV experiments when metabolic acidosis was su~~pos~ on 4 hours of respiratory alkalosis, the plasma HCO; fell by 10.5 mEq/L (P < O.Ol), a value twice as great as that in pure respiratory alkalosis (group I) but CSF HCC); fell by 3.1 mEq/L at 4 hours (P < 0.01) a value almost identical to that in pure respiratory alkalosis. Therefore in hypocapnia when plasma HCO; fall is accentuated, CSF HCO; reduction does not follow the plasma HCO; blindly, but reaches a lower limit, which in these experiments was similar to that reached in respiratory alkalosis. CSF lactate increased by 0.9 mEq/L (P < 0.05) in these experiment at 4 hours, despite a marked reduction in cerebral blood flow as shown by the large cerebral a-v 0, content difference (table 1). The fact that CSF lactate did not increase further may be related to the relatively higher CSF Po,, of 33 mm Hg (fig. 4) in these experiments which at the cellular level could counteract the effect of reduced cerebral blood flow in causing increased lactic acid formation. It is noteworthy that in these experiments CSF pH had returned to normal at 4 hours (fig. 4) and thus purely from the stand point of Hf homeostasis in the CSF no further falls in CSF HCO; were necessary. Taken altogether these data show that the fall in plasma HCO; in respiratory alkalosis is the predominant factor in CSF HCO; reduction and that if the fall in plasma HCO; is prevented, then the reduction in CSF HCO; is limited and is primarily due to the increase in CSF lactate. Therefore, the reduction in CSF HCO; content in respiratory alkalosis is due to two factors : (1) the simultaneous reduction in plasma HCO; content and (2) increased lactate in the CNS. Since the rise in lactate in the CNS is relatively modest in hypocapnia and not particularly significant until Pace, is less than 20 mm Hg (Weyne et al., 1970), the plasma HCO; reduction becomes the dominant event in CSF HCO; regulation and thus in H+ homeostasis in the CSF in respiratory alkalosis. The adjustments in CSF HCO; in hypocapnia are dependent on the reduction in plasma HCO;, however, the exact mode of transfer of HCO; out of the CSF is not clear. Our data would suggest that in pure respiratory alkalosis there is an equal reduction in CSF HCO; content for each milliequivalent fall in plasma HCO,, and probably no active transport mechanism for transfer of HCO; out of the CSF needs to be invoked. The question of changes in d-c potential being responsible for

CSF HCO;

RBGUL.ATION IN RESPIRATORY AIKALOSIS

277

the alterations in CSF H+ and HCO; in respiratory alkalosis (Hombein and Pavlin, 1975) cannot be answered from our data since no d-c potential measurements were made. However, changes in d-c potential seem unlikely to explain all our findings since changes in CSF/plasma d-c potential difference, E, are dependent on plasma pH (Held et al., 1964) and despite significantly different arterial pH’s in pure respiratory alkalosis (group I, fig. 1) and combined respiratory alkalosis and metabolic acidosis (group IV, fig. 4) the same CSF HCO; value was reached after 4 hours of hypocapnia. The present experiments support the concept of diffusion of HCO; from CSF into plasma in instances of hypocapnia. However, when in respiratory alkalosis plasma HCO; was markedly reduced by infusion of HCl (group IV, fig. 4) the fall in CSF HCO; was significantly less than that in plasma (fig. 5) and was of such a degree that resulted in a normal CSF pH after 4 hours of hypocapnia. Thus there appears to be a lower limit below which CSF HCO; will not fall in hypocapnia regardless of further reductions in plasma HCO; . This level may be determined by the CSF pH, so that HCO; movement from the CSF is stopped when CSF pH is fully compensated and a normal [H’] is reestablished in the CSF. This theory is in accord with our observation on the regulation of CSF HCO; in respiratory acidosis, where both the local and systemic factors responsible for CSF HCO, regulation are coordinated by a governing mechanism sensitive to H+ in the extracellular fluid of the brain (Kazemi and Hasan, 1975; Hasan and Kazemi, 1976). To fully extend this H+ theory to instances of respiratory alkalosis experiments of longer duration with variously lowered plasma HCO, contents need to be performed. In summury, then, these studies indicate the dependence of CSF HCO; on plasma HCO; content in respiratory alkalosis and show that changes in plasma HCO; are of critical importance to CSF HCO; regulation in hypocapnia and that the increase in brain and CSF lactath (the local mechanisms) are of limited value in CSF HCO; regulation in the absence of a fall in plasma HCO;. They also show that falls in plasma HCO; are not followed indefinitely by the CSF, and that CSF HCO; will not fall below a lower limit despite greater falls in plasma HCO; . This lower limit may well be regulated by [H’] in the CSF and that once a normal CSF pH is achieved no further falls in CSF HCO; occur. References Alexander, S. R., R. Gelfand and C. J. Lambertsen (1961). The pK’ of carbonic acid in cerebrospinal fluid. J. Biol. Chem. 236: 592-596. Dempsey, J. A., H. V. Forster and G. A. doPico (1974). Ventilatory acclimatization tomoderate hypoxemia in man. The role of spinal fluid H+. J. Clin. Inoest. 53: 1091-1100. Dempsey, J. A., H. V. Forster, N. Gledhill and G. A. doPico (1975). Effects of moderate hypoxemia and hypocapnia on CSF H+ and ventilation in man. J. Appl. Physiol. 38 : 665-674. Fencl, V., J. R. Vale and J. A. Broth (1969). Respiration and cerebral blood flow in metabolic acidosis and alkalosis in humans. J. Appl. Physiol. 27 : 67-76.

278

L. CHOMA AND H. ICAZEMI

Fen&, V. (1971). Distribution of H* and HCO; in cerebral fluids. In: Ion Homeostasis of the Brain, edited by B. K. Siesjo and S. C. !I$rensen. New York, Academic Press, pp. 174-185. Harper, A. M. and H. I. Glass (1965). Effect of alterations in arterial carbon dioxide tension on the blood flow through the cerebral cortex at normal and low arterial blood pressures J. Neural. Neurosurg. Psych&. 28 : 449452. Hasan, F. M. and H. Kazemi (1976). The dual contribution theory of regulation of CSF HCO; in respiratory acidosis. J. Appl. Physiol. (in press). Held, D., V. Fencl and J. R. Pappenbeimer (1964). Electrical potential of cerebrospinal fluid. J. Neurophysioi. 21: 942-959. Hill, B. A. (1971). Principles of Medical Statistics, 9th ed. London, The Lancet Limited, p. 146. Hornbein, T. F. and E. G. Pavlin (1975). Distribution of H+ and HCO; between CSF and blood during respiratory alkalosis. Am. J. Physiol. 228: 1149-l 154. Kaxemi, H., D. C. Shannon and E. Carvallo-Gil (1967). Brain CO, buffering capacity in respiratory acidosis and alkaIosis. J. Appt. Physiol. 22: 241-246. Kaxemi, H., L. M. Valenca and D. C. Shannon (1969). Brain and cerebrospinal fluid lactate concentration in respiratory acidosis and alkalosis. Respir. Physiol. 6: 178-186. Kazemi, H., N, S. Shore, V. E. Shih and D. C. Shannon (1973). Brain organic buffers in respiratory acidosis and alkalosis. J. Appt. Physiol. 34: 478-482. Kaxemi, H. and F. M. Hasan (1975). Regulation of CSF bicarbonate in respiratory acidosis. Fed, Proc. 34: 357. Kety, S. S. and C. F. Schmidt (1948). The effects of altered arterial tensions of carbon dioxide and oxygen in cerebral blood flow and cerebral oxygen consumption of normal young men. J. Clin. Invest. 27: 484-492. Klein, J. R. and N. S. Olsen (1947). Distribution of intravenously injected glutumate, lactate, pyruvate and succinate between blood and brain. J. Biol. Chem. 167: l-5. Laborit, H. (1966). De la prophylaxie a la therapeutique du choc dit ‘irrhersibie’. Presse Mkd. 74: 2079-2083. Leusen, I. and G. Demeester (1966). Lactate and pyruvate in brain of rats during hyperventilation. Arch. Int. Physiol. Biochim. 74: 25-34. Heusen, I. (1972). Regulation of cerebrospinal fluid composition with reference to breathing. Physiof. Reu. 52: l-56. Mitchell, R. A., C. T. Carman, J. W. Severinghaus, B. W. Richardson, M. M. Singer and S. Shnider (1965). Stability of cerebrospinal fluid pH in chronic acid-base disturbances in blood. J. Appl. Physiol. 20. 443-452. Pappenheimer, J. R. (1%7). The ionic composition of cerebral extracelhtlar fluid and its relation to control of breathing. Haruey Lect. Ser. 61: 71-94. Plum, F. and J. B. Posner (1967). Blood and cerebrospinal fluid lactate during hyperventilation. Am. J. Physiol. 212: 864870. Posner, J. B. and F. Pllun (1967). Spinal fluid pH and neurologic symptoms in systemic acidosis. N. Engl. J. Med. 277: 605613. Severinghaus, J. W., R. A. Mitchell, B. W. Richardson and M. M. Singer (1963). Respiratory control at high altitude suggesting active transport regulation of CSF pH. J. Appf. Physiol. 18: 1149-l 159. Valenca, L. M., D. C. Shannon and H. &emi (1971). Clearance of lactate from the cerebrospinal fluid. Neurology 21: 615-620. Van Vaerenbergh, P. J., E. Lacroix, G. Demeester and I. Leusen (1965a). Lactate in cerebrospinal fluid during muscular exercise. Arch. Znt. Physiol. Biochim. 13 : 729-737. Van Vaerenbergh, P. J. J., G. Demeester and I. Leusen (1965b). Lactate in cerebrospinal fluid during hy~~en~ation. Arch. Int. Physiol. Biochim. 73: 738-747. Weyne, J., G. Demeester and I. Leusen (1970). Effects of carbon dioxide, bicarbonate and pH on lactate and pyruvate in the brain of rats. Pji’iQersArch. 314: 292-311. Wichser, J. and H. Kaxemi (1975). CSF bicarbonate regulation in respiratory acidosis and alkalosis. J. Appl. Physiol. 38: 504-512.