- Email: [email protected]
Dependence of csf on plasma bicarbonate during hypocapnia and hypoxemic hypocapnia
Respiration Physiology (1976) 26,11-26;
North-Holland Publishing Company, Amsterdam
DEPENDENCE OF CSF ON PLASMA BICARBONATE DURING HYPOCAPNIA AND HYPOXEMIC HYPOCAPNIA “’
DALE A. PELLIGRINO
and JEROME A. DEMPSEY3
Pulmonary Physiology Laboratory, Department of Preventive Medicine, University qf Wisconsin, Madison, Wis. 53706, U.S.A.
Abstract. We have previously shown pH compensation
to be similar in CSF and arterial blood during chronic hypoxemic hypocapnia in man and pony, and postulated that the compensatory reduction in CSF [HCO;] was dependent upon corresponding changes in [HCO;]a. We tested this hypothesis in anesthetized, paralyzed dogs by determining the effects of 7 or 14 hours of hypocapnia (Pa,,, 20 and 30 mm Hg), hypoxemia (Pa,, 30,38 and 48 mm Hg) and hypocapnic hypoxemia on CSF acid-base status. [HCO;]a was either permitted to fall normally or was held near control levels by NaHCO, infusion. In hypocapnia and hypoxemic hypocapnia, the decrease in [HCO;] and % pH compensation in CSF were 5 that in arterial blood. Most (SlG39%) of the compensatory decrease in CSF [HCO;] was prevented by preventing the corresponding reduction in [HCO;]a. This dependence of changes in CSF on plasma [HCO;] required a concurrent decrease in CSF PcoZ, but was largely independent of variations in plasma pH. A minor but significant portion of the decrease in CSF [HCO;] was achieved independently of corresponding changes in [HCO;]a. The contribution of this local mechanism to CSF [HCO;] regulation increased with increasing severity of hypocapnia or hypoxemia and was usually associated with a selective increase in CSF lactate. It was concluded that [HCO;] regulation in the CSF during hypoxemic hypocapnia was primarily dependent upon, and therefore limited by, the concomitant decrease in plasma [HCO;]. Cerebrospinal fluid Hypocapnia Hypoxia
Acid-base status Bicarbonate Blood
In chronic disorders of systemic acid-base status it is well-recognized that the protection or compensation of pH in cerebral spinal fluid (CSF) is determined primarily through the regulation qf CSF [HCO;] (Leusen, 1972; SiesjG, 1972). The problem of exactly which mechanisms are responsible for CSF [HCO;] regulation remains controversial. Two broad categories of regulatory mechanisms may be identified (Siesjli, 1972) : a “non-specific” regulation of CSF [HCO;] influenced by correAcceptedfor
publication 19 September 1975.
’ Preliminary findings have been reported in abstract form (Fed. Proc. 33(3): 1369, 1974). ’ This study was supported in part by grants from NIH (1-ROl-HL-15469-02) and the U.S. Army Research and Development Command (No. 17-74-C-4020). ’ D. A. Pelligrino is an NIH pre-doctoral trainee (HL05626) and J. A. Dempsey is the recipient of an NIH Carreer Development Award (HLOO149). 11
12
D.
and J.
A. PELLIGRINO
A. DEMPSEY
sponding changes in plasma [HCO;] and a “CSF-specific” or local regulation via mechanisms which are largely independent of changes in plasma [HCO;]. It has been commonly held that CSF-specific mechanisms dominate CSF [HCO;] regulation because of the relatively close protection afforded CSF pH in systemic acidosis and alkalosis (Leusen, 1972); Mitchell et al., 1965b; Severinghaus et al., 1963). However, recent reviews (Loeschcke, 1972; Siesjii, 1972) have emphasized that a near-complete protection of CSF pH is consistently achieved only in primary metabolic acid-base disorders where changes in CSF [HCO;] are substantially less than those in plasma. In the chronic steady-state of primary respiratory acidosis, compensatory increases in CSF [HCO;] are similar to those in arterial plasma and pH compensation is incomplete in both fluid compartments. This apparent influence of a changing plasma [HCO;] on CSF [HCO;] regulation is also strongly indicated in several studies of short- and long-term respiratory alkalosis. These data provided the rationale for the present study and are summarized in fig. 1. The relevant finding here is the parallel reductions in CSF and plasma [HCO;] during: (a) long-term hypoxemic hypocapnia in awake man and pony (fig. 1 A); and (b) short- and long-term moderate normoxic hypocapnia in awake man and anesthetized dogs (fig. 1 B). -8L"
" ' " ,*'I A[HCOS] CSF(meq/l) ,,j"__
-6- @
+/,/'i
"
0’ ,’ 8
"
/'
,,/#7i
-- @
-
,‘-IDENTITY
-4,$L-
/‘,
,
,
-2 Fig. 1, Reductions
,
in CSF and arterial
during
nonnoxic
A: Dam obtained
in awake
(0,
,
,
-4
,
-6
(
,
-8
blood bicarbonate
and hypoxic
hypocapnia.
man and pony during
ranged
between
,
,
,
,
-4
,
-6
below normocapnic, Values
26 hours
w), 3400 m (A) and 4300 m (A, [7). (Pa,? - 3855
for % pH compensation
,
-2
are mean
normoxic
changes
to 5 weeks sojourn
,
-8
at altitudes
mm Hg; Pacol - 25-32 mm Hg).
50 and 75 % ; and were similar
control
levels,
(+ SEM). of 3100 m
Averagevalues
in CSF and plasma
(Dempsey
et
al., 1974, 1975; Forster
et al., 1975; Orr et al., 1975). B : Data obtained in normoxic hypocapnia in awake man (26 hours - 30 mm Hg Pacol) (Dempsey er al., 1975), and anesthetized dogs (48 hours - 20 mmg Hg Pa,,*) (K azemi et al., 1967 and 1973 ; Plum and Posner, 1967; Van Varenberg et al., 1965; Wischer and Kazemi, 1975).
This close correspondence shown between CSF and plasma suggests that the compensatory reduction in CSF [HCO;] may be truly dependent upon, and therefore limited by, changes in plasma [HCO;]. To test this hypothesis in anesthetized dogs we determined the effects of varying degrees of arterial hypocapnia and/or hypoxemia on [HCO;] regulation and pH compensation in CSF, under conditions
CSF [H+]
COMPENSATION
IN HYPOCAPNIA
13
where plasma [HCO;] was either permitted to fall normally or was held constant near normoxic, normocapnic control levels. Methods
Mongrel dogs (average weight 20 kg) were anesthetized with intravenous pentobarbital (15-20 mg/kg and 46 mg/kg every 2 hours thereafter) and paralyzed with intravenous gallamine (100 mg initially and 60 mg every 2 hours thereafter). Ventilation was maintained through a tracheostomy tube by a constant volume respirator and rectal temperature was maintained at 3838.5 “C with a water blanket. Endtidal Pco2 and P,, were continuously monitored with rapidly responding analyzers (Godart Capnograph and Beckman OM-11). Cathethers were inserted into a femoral vein for injections and infusions and into a femoral artery for continuous monitoring of blood pressure and heart rate and for anaerobic sampling (3 ml per sample). A 20-gauge spinal needle was placed suboccipitally into the cisterna magna for anaerobic sampling of CSF (3-4 ml per sample). The study was aimed at quantifying the effect of plasma [HCO;] on the reduction in CSF [HCO;] during hypocapnia and/or hypoxia. A total of 12 groups, each with 4 to 5 dogs, were studied prior to sacrifice (total N = 54). Each of 6 of these groups was subjected to a specific level of normoxic hypocapnia (Pa,,, - 20 or 30 mm Hg), normocapnic hypoxemia (Pa,, - 30, 38 or 48 mm Hg) or hypocapnic hypoxemia (Pa,, - 48 + Paco2 - 25 mm Hg) for 7 or 14 hours, during which [HCO;la was permitted to fall in a normal fashion. Six additional groups of dogs were also each subjected to 7 or 14 hours at one of these specific levels of hypocapnia and/or hypoxia; but in these animals [HCO;]a was prevented from falling below control levels by continuous intravenous infusion of NaHCO,. The key comparison in this study, then, focused on the differences in changes in CSF [HCO;] between groups which were at similar levels of hypocapnia and/or hypoxemia; but at different, i.e., control us reduced, levels of plasma [HCO;]. Three additional dogs were studied for 7 hours in normoxic normocapnia with NaHC03 infusion to determine the effects of NaHCO, infusion alone on CSF [HCO;]. The protocol for a single experiment was as follows. Immediately following anesthesia, tracheal intubation and paralysis, the FIEFand ventilator were set to provide a Paco2 - 40 mm Hg and Pa,, - 100 mm Hg. The cisternal needle was then inserted. A 2-l/2 hour ‘control’ period of normoxic normocapnia followed, during which arterial blood was sampled every 20-30 minutes and CSF sampled twice, one hour apart. Then (in about one-half the animals) hypoxia and/or hypocapnia were gradually induced over a 15-25 minute period by mechanical hyperventilation and/or manipulation of the inspired 0, and N2 concentrations. The desired levels of PaoZ and PaYoZ were then maintained over the subsequent 7 or 14 hour period. For studies which required an unchanging [HCO;]a, a low rate of NaHCO, infusion was begun 10 minutes prior to the experimental period and the infusion continued while hypoxia and/or hypocapnia was gradually induced and then maintained, as outlined above. The infusion rate was frequently adjusted (020.8 meq/min NaHC03) to maintain
14
D. A. PELLIGRINO and J. A. DEMPSEY
[HCO;]a at or within 1.5 meq/l above control levels throughout the experimental period. Arterial blood was sampled and analyzed every l&15 minutes during the first hour and at 30-45 minute intervals thereafter. Constant Pao2 and Paco2 (and [HCO;]a where appropriate) were maintained throughout the experimental period by continuous monitoring of end-tidal gases and frequent analyses of arterial blood samples. Cistemal CSF samples were obtained at the 3rd, 5th and 7th (or 14th) hours of the experimental period. CSF samples contaminated with air bubbles or blood were discarded. Arterial samples were analyzed for Po,, Pco2 and pH and CSF samples for Pco2 and pH with electrodes (Radiometer, Copenhagen) at 38 “C, with corrections for small deviations in the animal’s temperature. [HCO;] was calculated using the appropriate pk’ and S values (Mitchell et aE., 1965a). Lactic acid concentration was analyzed in plasma and CSF by a modified calorimetric method (Barker and Summerson, 1941). The technique, reproducibility and validity of these measurements have been previously described (Dempsey et al., 1974, 1975). The percentage of pH compensation in CSF and arterial blood during hypocapnia was calculated after Siesjo (1971).
TABLE Arterial blood and CSF acid-base status in control Arterial blood +7or
Control NaHCO, infusion
l4hr
[HCO;]
Lact
P07
Pco*
PH
Moderate normoxic hypocapnia (7 hr) No 90 41.5 7.345 +0.006 +2.6 kO.5 Yes 96 40.7 7.349 +0.4 *0.012 k5.6
22.0 kO.5 21.8 +0.5
2.72 +0.51 2.36 *0.53
96.0 * 1.4 85 +4.9
29.3 kO.7 29.8 +I.0
7.417 f 0.009 7.484 kO.026
Severe normoxic hypocapnia (7 hr) No 103 40.8 7.358 +0.7 rf:0.007 f5.3 Yes II4 41 .o 7.350 +0.7 kO.008 k4.5
22.4 kO.4 22.1 kO.3
2.13 kO.27 2.03 kO.15
95 +4.7 100 k6.3
19.0 kO.4 20.7 +0.7
7.525 kO.016 7.643 *0.010
Severe normoxic hypocapnia (I4 hr) No 122 40.2 7.374 +0.012 + 1.0 k7.0 Yes II8 41.1 7.353 +3.8 +0.6 + 0.002
22.9 kO.3 22.2 +0.4
I .90 rfr0.27 2.19 kO.25
96 f 1.8 87 + 1.2
22.3 +2.6 21.5 * 1.2
1.475 f0.051 7.644 *0.018
Pol
Pco2
PH
PO1and Pcol in mm Hg; HCO; and lactate in meq/l. All values are mean *SEM with 4 or 5 dogs per condition.
CSF [H’]
~MPENSATION
15
IN HYPOCAPNIA
Results PRE-EXPERIMENTAL CONTROLS
The acid-base status during normo~pnic, normoxic control conditions are shown for each group in tables 1 and 2. The control data obtained in the 12 different groups of animals were consistent. For example, the range of values for rnekAl [HCO;] was 2.5 meq/l in arterial blood (20.4 to 22.9) and 1.7 meq/l in CSF (21.7 to 23.4). Mean values in CSF, compared to arterial blood, showed a lower pH (N 0.02 units) and higher P,,, (N 6 mm Hg) and [HCO;] (w 1 meq/l). The control arterial [HCO;] in these anesthetized, paralyzed animals were in the mid-range of mean values reported for unanesthetized, ‘non-panting’ dogs (19.8 to 24.5 meq/l) (Wise, 1973; Sylvester et al., 1973; Gillespie and Hyatt, 1974; Feigl and D’Alecy, 1972; Jennings and Sparling, 1974). Data obtained periodically during the 2-l/2 hour pre-experimental control periods (not shown), showed no systematic changes in arterial acid-base status and close agreement between two CSF samples. EFFECTS OF
INFUSION ALONE
NaHCO,
Normocapnic-normoxic
conditions
were maintained
for 7 hours and NaHCO,
t md at the end of each period of normoxic hypocapnia ~~~
_.._Cisternal CSF Control
HCO;]
Lact
Pco1
8.5
1.31
47.6
to.3
& 0.05
+1.3
‘2.1
1.76
47.5
k1.0
iO.55
kO.5
-.
+7 or 14 hr PH
[HCO;]
Lact
Pco2
PH
[HCO,]
Lact
7.325 +0.011
23.3 +0.4
2.61 +0.34
38.1
7.367
20.7
2.31
+ 0.009
io.7
10.14
7.311 *0.005
22.5
1.94
10.6 39.1
7.386
22.2
I .96
+0.2
+0.09
to.5
*0.01
+0.6
t-o.12
5.6
1.78
45.7
2.09 +0.12
19.5
3.02
kO.6
22.6 +0.4
7.436
50.27
7.329 F0.005
30.4
to.4
kO.8
!2.5
2.34
46.9
7.318
22.6
2.10
30.5
* 0.009 7.472
50.4 21.4
4.20
+o.s
kO.39
+0.6
,0.007
+0.3
+0.15
+1.1
~0.008
+0.5
kO.49
6.2
1.40 +0.13
45.1
7.338
22.8
2.05
30.4
7.410
18.1
2.73
& 0.004
kO.4
kO.11
+ 2.6
7.303
21.7
2.40
28.2
+ 0.036 7.464
k0.I 19.4
4.67
,0.007
+0.4
kO.16
20.3
*0.004
+0.3
20.69
to.3 !3.3
2.20
kO.5 46.7
* 1.2
&0.60
kO.8
+0.93
+0.35
16
D. A. PELLIORINO
and
J. A. DEMPSEY
1 Arterial
Arterial
blood and CSF acid-base
status
in control
am
blood
Control
f7
hr
NaHCO, infusion Hypoxia No Yes
Po
2
+ hypocapnia
Lact
PO2
Pco,
pH
7.417
(7 hr) 7.335
20.4
2.52
45.0
26.3
k4.6
+0.6
+0.010
kO.7
+0.36
k3.1
kO.9
kO.024
108
41.4
7.337
21.7
2.05
48.3
27.8
7.525
+5.5
+0.9
kO.012
+0.3
+0.11
kO.8
kO.7
k 0.008
7.337
21.2
1.98
46.1
39.1
7.330
+0.005
kO.5
+0.32
* 1.3
+ 1.9
* 0.020
7.338
normocapnic 105 +4.9
Severe normocapnic
hypoxia
(7 hr)
40.7 kO.9 hypoxia
(7 hr)
120
39.9
7.339
20.9
1.60
37.9
36.9
+4.5
kO.08
* 0.007
+0.5
+0.20
+ 1.6
* 1.3
*0.010
118
41.0
7.362
22.7
1.96
38.3
36.0
7.430
k5.3
* 1.5
kO.016
+0.7
kO.36
+2.4
f 1.8
+0.006
Very severe normocapnic Yes
[HCO;]
39.3
No
Yes
PH
102
Moderate
No
P co2
hypoxia
(7 hr)
119
40.6
7.355
22.1
2.31
30.8
38.4
7.391
k4.6
kO.6
* 0.007
+0.4
kO.14
+0.1
+0.7
+0.013
PO2 andPco,in mm Hg ; HCO;
and lactate in meq/l. All values are mean + SEM with 4 or 5 dogs per condition.
infused at a rate which produced a 3.0-3.5 meq/l increase in [HCO;]a above control (table 3). No measureable changes in CSF [HCO;] accompanied these increased levels of [HCO;]a. The increases in [HCO;]a above control levels in these studies
TABLE Effects Arterial
of NaHCO,
infusion
3
during
normoxic
normocapnia
CSF
blood
PH
P co2
WC%1
PH
P co2
t-HCO ;1
Control
7.33
42
22.0
7.32
47
22.6
+3hr +5hr
7.38 7.38
42 43
24.8 25.0
7.31 7.32
48 47
22.8 22.6
+7hr
7.39
42
25.2
7.30
48
22.4
P co* in mm Hg; [HCO;]
in meqil. Mean values (N = 3). Arterial
Po, was - 100 mmHg
throughout
CSF [H +]
17
COMPENSATION IN HYF’OCAPNIA
id of each period of hypoxemia and hypocapnic hypoxemia _ _
Cisternal CSF +7 hr
Control [CO;]
Lad
Pco1
PH
[HCO,]
Lact
P,,,
pH
[HCO,]
Lact
,.6
45.2 kO.8 47.6 +0.5
7.346 +0.004 7.314 kO.004
23.3 +os 22.7 +0.2
2.67
.8 0.5
1.42 kO.18 1.67 kO.18
+0.11 2.29 * 0.09
3.54 to.6 36.7 +1.1
7.386 kO.015 7.400 &0.009
19.6 kO.8 21.3 +0.5
3.79 rto.52 3.55 kO.18
I.1 1.0
I .90 kO.48
47.4 +0.7
7.326 kO.004
23.3 to.5
2.49 +0.10
47.5 & 1.4
7.304 kO.015
22.1 kO.3
2.49 +0.86
.3 0.9 .5 1.3
1.53 io.34 0.96 +0.21
44.0 k1.0 46.2 +1.0
7.346 &0.004 7.339 +0.007
22.7 +0.5 23.4 +0.3
2.30 +0.20 2.34 +0.16
42.2 &I.6 42.2 +0.8
7.339 kO.013 7.346 20.014
21.4 kl.1 22.0 +0.4
2.53 7to.37 2.82 kO.48
.8 0.5
2.53 kO.46
45.5 +1.9
7.337 +0.013
23.0 kO.4
2.39 +0.13
40.0 +1.2
7.335 +0.006
20.1 +0.8
3.77 &OS1
0.6
in normoxic normocapnia were 2-3 times greater than the 0.3-1.2 meq/l increases produced during the experimental periods of hypocapnia and/or hypoxia (fig. 3). We interpret these data to mean that the NaHCO, infusion alone had no significant effect on CSF [HCO;] during the experimental periods. These limited data obtained in not-moxie normocapnia agree with a number of reported studies(Leusen, 1972). ACIItBASE
CHANGES IN HYPOCAPNIA
AND/OR HYPOXEMIA
The changes in the arterial and CSF acid-base status over the duration of each experiment are shown in fig. 2 - for studies without NaHCO, infusion, and in fig. 3 for studies in which NaHCO, infusion accompanied the hypocapnia and/or hypoxemia. The mean data obtained in the control and at the end of each experimental period are summarized in table 1 (hypocapnic experiments) and table 2 (hypoxic hypocapnia and hypoxic experiments).
After 7 hours of moderate hypocapnia (PacoZ N 29 mm Hg) (table 1 - no NaHCO,
18
and J. .4. DEMPSEY
D. A. PELLIGRINO
ARTERIAL
BLOOD
CISTERNAL
CSF
A PC02 (mm Hg )
5 0
20
f
Fig. 2. Effects of hypocapnia are mean changes of normoxic
hypocapnia
the upper left-hand
and/or
from normocapnic, (0,
0).
panel are shown
hypoxemia normoxic normocapnic the various
on acid-base control
status
hypoxemia
blood
(A, A) and hypocapnic
levels of arterial
Paol (in parenthesis),
in arterial
and CSF. Values
levels, Symbols denote the experimental hypocapnia
conditions
hypoxemia
(ordinate)
( x ). In
and the average
for each condition.
infusion), Pcol fell 2.7 mm Hg more in arterial blood (- 12.2 mm Hg) than in CSF (-9.5 mm Hg) and the reduction in [HCO;] was 0.9 meq/l greater in blood (- 3.5 meq/l) than in CSF (- 2.5 meq/l). After 7 hours of the more severe hypocapnia (Pace, - 19 mmg H) (table 1 - no NaHCO, infusion), Pco2 fell 6.5 mm Hg more in blood (-21.8 mm Hg) than in CSF (- 15.3 mm Hg). The fall in [HCO;] was 3.7 meq/l greater in blood (-6.8 me& than in the CSF (- 3.1 meq/l). These changes after 7 hours at both levels of hypocapnia resulted in a significant rise in pH in both blood and CSF, with pH compensation - 12-l 5 % more complete in blood than in CSF. After a longer period of 14 hours at the more severe level of hypocapnia (table 1 - no NaHCO, infusion), changes in blood (- 17.9 mm Hg APco, and -6.7 meq/l A[HCO;]) more closely paralleled those in CSF (- 14.7 mm Hg) and -4.7 meq/l), and pH compensation was similar in blood and CSF (- 57 %). When hypocapnia was induced and plasma [HCO;] was prevented from falling below control levels (table 1 - NaHC03 infusion). CSF [HCO,] remained un-
CSF [H’]
COMPENSATION
ARTERIAL
BLOOD
CISTERNAL
A Pcop (mn $11
CSF
Hg) 15
l-
A
19
IN HYPOCAPNIA
IHCO
0.4Or
I
I I
I 2
I 3
1 4
I
5
I 6
IdIll 7’10’14
I
I I
I 2
I 3
I 4
I 5
I 6
IllI 7’10’14e
Fig. 3. Effects of hypocapnia and/or hypoxemia plus NaHCO; infusion on acid-base status in arterial blood and CSF. Values are mean changes from normocapnic, normoxic control levels. As in fig. 2 symbols denote the experimental conditions of normoxic hypocapnia (0, a), normocapnic hypoxemia (A., 0) and hypocapnic hypoxemia (x). In the upper left-hand panel are shown the various levels of arterial hypocapnia (ordinate) and the average Paol, for each condition.
changed at 30 mm Hg Pacoz and fell 1.2 meq/l after 7 hours and 2.3 meq/l after 14 hours at 20 mm Hg Paco2. Lactate rose only in the CSF (+ 2 to 2.5 meq/l) and only during the more severe level of hypo~pnia. These average reductions in CSF [HCO;] were all significantly less (P < 0.01) than those obtained when plasma [HCO;] was permitted to fall in a normal fashion. Hypocapnia plus hypoxemia
With the combined hypocapnic hypoxemia (table 2 - no NaHC03 infusion) Pc,, fell 2.3 mm Hg more in arterial blood (- 13.0 mm Hg) than in CSF (- 10.7 mm Hg). [HCO;] fell 3.8 meq/l in both compartments and pH compensation was l&12% more complete in the CSF than in the blood. Lactate rose only in the CSF (+ 1.2 meqil). When plasma [HCO;] was prevented from falling during hypocapnic hypoxemia, CSF [HCO;] fell 1.4 meq/l - a decrease which was signi~cantiy less than that seen when [HCO; ]a was permitted to fall (P < 0.01). Lactate rose only in the CSF (+ 1.3 meq/l).
20
D.
A. PELLIGRINO
and J.
A. DEMPSEY
Hypoxemia alone (46 or 38 mm Hg Pao,,) produced no systematic changes in arterial blood [HCO;], lactate or pH; but CSF [HCO;] fell 1.2-1.5 meq/l below control levels (P < 0.01) with no change in CSF lactate (table 2). When NaHCO, infusion was superimposed at 2 levels of hypoxemia, i.e., Pao2 38 and 30 mm Hg, CSF [HCO;] was reduced 1.4 and 3.0 meq/l, respectively (P < 0.01). CSF lactate rose 1.4 meq/L at 30 mm Hg Pao2. DEPENDENCE OF
CSF
ON PLASMA
[HCO;]
Table 4 compares pertinent findings obtained with and without NaHCO, infusion; and shows the calculated effects of preventing the fall in plasma [HCO;] during hypocapnia and/or hypoxemia on [HCO;] regulation and pH compensation in the CSF. In normoxic hypocapnia and hypoxic hypocapnia more than half of the reduction in CSF [HCO;] was dependent upon a concomitant reduction in arterial [HCO;] . Accordingly, the completeness of pH compensation in the CSF was reduced substantially by preventing the fall in arterial [HCO;] (.z.e., when compensation of arterial blood pH was completely prevented). This dependence of CSF [HCO;] regulation on a reduction in plasma [HCO;] was greatest in moderate hypocapnia (89%) and least in more severe, prolonged hypocapnia (51%). At all levels of normocapnic hypoxemia, CSF [HCO;] was reduced independently of corresponding changes in plasma [HCO;]. This reduction in CSF [HCO;] was similar at Pao,, 46 and 38 mm Hg (- 1.3 meq/l) and greatest at 30 mm Hg Pa,, ( - 3 .Omeq/l). Diiussion
Present findings in anesthetized dogs have shown that bicarbonate regulation and pH compensation in CSF during short-term hypocapnia or hypoxemic hypocapnia was primarily dependent upon corresponding changes in plasma bicarbonate. In addition, we quantified the contribution of a component to CSF [HCO;] regulation during hypocapnia which was independent of changes in the plasma [HCO;]. The contribution of this local or ‘CSF-specific’ mechanism to CSF [HCO;] regulation during hypocapnia was always less than that attributable to a reduction in plasma [HCO;] ; but was shown to increase with increasing severity of hypocapnia, was significant at even moderate levels of hypoxemia and was usually associated with a corresponding rise in CSF lactic acid concentration. CSF [HCO,]
REGULATION IN RESPIRATORY ALKALOSIS
Present findings are consistent with the view that quite different mechanisms might be responsible for brain ECF [HCO;] regulation during the different types of systemic acid-base disturbances (Siesjii, 1972). Furthermore, a consideration of past and present findings suggests that the regulation of CSF [HCO;] during respiratory alkalosis might be unique, because of the relatively large extent to which it is dependent upon concomitant changes in plasma [HCO;]. For example, in primary
CSF [H+]
COMPENSATION
IN HYPOCAPNIA
21
metabolic alkalosis or acidosis, CSF [HCOY] and therefore CSF pH change little with respect to plasma (Leusen, 1972; Loeschcke 1972; Siesjii, 1972). Most evidence points to appropriate changes in the CSF/plasma electrical potential difference sensitive to a changing pHa - as a primary protector of CSF [HCO;] and [H+] in metabolic acid-base disturbances (Siesj8, 1969). In primary respiratory acidosis, CSF pH appears to be less well-protected, but - at least over 6 or 7 hours of hypercapnia (see below) - the rise in CSF [HCO;] clearly exceeds that in plasma (Leusen, 1972; Pannier et al., 1971; Siesjii and Ponten, 1966). This compensatory rise in CSF [HCO;] has been shown to be due in small part to the corresponding rise in plasma [HCO;], but is determined primarily by local mechanisms; namely, CO2 hydration at the site(s) of CSF formation and increased brain ammonia production (Kazemi et al., 1973 ; Kelley and Kazemi, 1974; Vogl and Maren, 1975; Wischer and Kazemi, 1975). Present findings in primary respiratory alkalosis have shown that the compensatory reduction in CSF [HCO;] is even more dependent upon the concomitant fall in plasma [HCO;] and is influenced locally only to a limited extent and possibly through only the single mechanism of an increased brain lactic acid production. Our findings and interpretations are supported in part by the work of Kazemi and colleagues (Kazemi et al., 1973 ; Wischer and Kazemi, 1975), which showed that the reduction in CSF [HCO;] during 4 to 6 hours of hypocapnia in dogs was unaffected by brain carbonic anhydrase inhibition and was not accompanied by significant changes in brain ammonia or glutamic acid concentrations. It is not clear exactly which mechanism(s) are responsible for the observed dependence of changes in CSF [HCO;] on plasma bicarbonate during respiratory alkalosis. The most straightforward explanation would be that bicarbonate loss from CSF simply followed the decline in brain capillary plasma bicarbonate, per se, via passive diffusion. However, in view of the relative stability of CSF [HCO;] consistently shown during studies of systemic non-respiratory acidosis (see above), this process must account for only a very minor portion of the observed dependence of CSF upon plasma [HCO;] during hypocapnia. This conclusion requires the assumption that the permeability of the blood-CSF ‘barrier’ is similar in metabolic acidosis and respiratory alkalosis. Information concerning the permeabilities of [HCO;] and [H+] and their variability, is notably sparse (Woodbury, 1971). A second explanation would link the observed dependence of CSF on plasma [HCO;] to local regulatory mechanisms. One such mechanism might be a changing CSF-plasma potential difference (E) which is known to become more positive in systemic acidosis and more negative in alkalosis, thereby returning the relevant electro-chemical potential differences toward normal (Hornbein and Pavlin, 1975; Siesjii, 1969 and 1972). However, this explanation does not fit present observations, if the common view is taken that changes in E are solely the result of changes in plasma pH (Held et al., 1964; Hornbein and Pavlin, 1975). For example, by comparing conditions of equal hypocapnia at widely different pHa (i.e. with us without maintained [HCO;]a), we consistently observed that the greatest reductions in CSF [HCO;] occurred when the rise in pHa was least. These data do not rule out
22
D. A. PELLIGRINO
and
J. A. DEMPSEY
changes in E as a key mediator of CSF [HCO;] regulation during respiratory alkalosis, if one accepts the minority view that such changes in the potential are regulated at least in part by factors other than plasma pH : for example, by changes in plasma [HCO;] and/or [Cl-] (Loeschcke, 1971) or by unknown factors associated with acute changes in cerebral blood flow (Besson et al., 1971). A second means of local control over the ionic composition of newly formed CSF during respiratory alkalosis might be found in some mechanism which is sensitive to a low intra-cranial Pco2. Such a ‘CO,-sensitive’ regulator has been suggested from studies in respiratory acidosis (see above) ; and limited data in chronic hypocapnia indicate that some of the inter-subject variability in changes in CSF [HCO;] may be related to variations in CSF Pco2 rather than in plasma Pco2 or [HCO;] (Dempsey et al., 1975; Forster et al., 1975). However, present data does not support a significant role for reduced CSF P co2, per se, as an independent regulator of CSF [HCO;]. For example, comparing conditions of maintained us reduced plasma [HCO;] during ‘moderate’ arterial hypocapnia, the decreases in CSF Pco2 were equal in the paired conditions, but the compensatory reduction in CSF [HCO;] was only observed when [HCO;]a was permitted to fall. Hence, available data indicate that most of the compensatory movement of bicarbonate out of the CSF during hypocapnia is unrelated to variations in plasma pH and is not attributable to the independent effects of a reduced plasma [HCO;] or CSF Pco2 alone. We can only conclude, then, that [HCO;] transfer out of the CSF requires a concomitant reduction in both plasma bicarbonate and CSF Pco2. APPLICATION
OF FINDINGS
TO CHRONIC
HYPOXEMIC
HYPOCAPNIA
These studies were designed to interpret the close correspondence observed in pH compensation between CSF and blood during long-term, steady-state hypoxemic hypocapnia in man. In general, present findings permit the conclusion that this close correspondence between fluid compartments was attributable primarily to a true dependence of bicarbonate loss from CSF on a reduction in plasma bicarbonate. They imply, therefore, that renal compensation (Gledhill et al., 1975; Genari et aE., 1972) provides an important limitation to the rate and extent of CSF bicarbonate regulation during long-term hypocapnia, with the result that pH compensation will be similar and incomplete in both blood and CSF (Forster et al., 1975; Siesjo, 1972). Present findings also indicated a role for some factor which affected a variable, yet significant reduction in CSF [HCO;] - independently of a constant [HCO;]a during all those conditions of hypoxemia and/or hypocapnia chosen to simulate arterial blood gases in healthy man during sojourn to 3100-4300 m altitude. Even the relatively ‘moderate’ levels of hypoxemia encountered during man’s first 24 hours (- 3840 mm Hg Pa,,) or after his first week (- 45-50 Pa,,) at 4300 m altitude (Forster et al., 1975) were associated with local regulatory influences which were sufficient to account for - one-third of the total bicarbonate reduction or pH compensation in CSF in the hypocapnic, hypoxemic dog (table 4). This data agrees
CSF [H’]
COMPENSATION
TABLE 4 regulation and pH com~nsation in the CSF on plasma and/or hypoxemia ..-.I ~__..^ -.._.. ___~.. ____ ._-.-..
Dependence of [HCO;] ____...--.
Change in CSF [HCO;] Without PaCoL Pa,, (mm Hg) (mm Hg) HCO; infusion .____ 30 20 20* 26 39 37 38
.._> 90 > 90 > 90 45 46 38 30
* Paco2 N 20 mm ** [HCO;]a did *** E.g., during [HCO;] without
(mesit)
(meq/l) ----2.6 -3.1 -4.7 -3.7 - I .2** - 1.3 -
from control ~_
With HCO; infusion
23
IN HYPOCAPNIA
[HCO;] ~.
during hypox~mia
--.-
_.-_-.
- ..-
% CSF pH compensation .----I___ ._--.
‘Dependence’ on plasma[HCO;] (% of change without infusion)***
Without HCO; infusion
With HCO; infusion
89% 61% 51% 65% (0 %) (0%) -
40% 35% 55% 60%
5O’ /0 15% 20% 20 9’ 10 -
.l_-__--0.3 -1.2 -2.3 - t.3 - 1.4 - 3.0
Hg for 14 hours. All other conditions were maintained for 7 hours. nat change significantly from normocapnic, normoxic control levels during these studies. moderate normoxic hypocapnia (Pam2 N 30 mm Hg) the compensatory reduction in CSF NaHCO; infusion = 2.6 meq/l; and the reduction in CSF [HCO;] when [HCO;]a was not
permitted to fall = 0.3 meq/l. Thus, the fall in CSF [HCO;] which was ‘dependent’ upon a falling [HCO;]a = 2.3 meq/I; or z&i 100 = 89% of the total change observed without NaHCO; infusion.
= 2.6-0.3
with that previously shown during progressive, acute hypoxemia in the anesthetized rat and dog (Kogure et al., 1970; MacMillan and Siesjii, 197i), and may be explained at least in part by a selective cerebral tissue lactacidosis (Mines and S#rensen, 1971; Plum and Posner, 1967). A strict application of these findings to chronic hypoxemic hypocapnia in man suggests that CSF [HCO;] was significantly influenced by locaf regulatory mechanisms, even though the reduction in bicarbonate and completeness of pH compensation in the CSF was paralleled or even exceeded by that in plasma (fig. 1). Although the available data appear to justify this postulate, serious limitations exist in our attempt to apply findings obtained in semi-acute conditions to the truly chronic steady-state of hypoxemic hypocapnia. Some findings are available which, while incomplete, strongly suggest that the participation of local mechanisms in the regulation of CSF [HCO;] in hypoxia and respiratory acid-base disturbances might be time dependent. For example, in the awake pony the reduction in CSF [HCO;] (and rise in CSF lactate concentration) was shown to exceed that in plasma during the initial six hours of hypoxic exposure but not after two or more days of continued exposure (Orr et al., 1975); lactic acid concentrations in CSF remain fairly constant near sea-level control levels in healthy man after one to three weeks of high altitude sojourn (Dempsey et al., 1974; Forster et al., 1975); studies in the human native of high altitudes have failed to detect an increased level of brain anaerobic glycolysis
24
0. A. PELLIGRINO
and J. A. DEMPSEY
(Blayo et al., 1973; Sorensen et al., 1974); and in normoxic hypocapnia, cerebral blood flow undergoes an acute reduction followed by a gradual return toward normal, thereby presenting at least a potential time-dependent effect on the local regulation of bicarbonate movement out of the CSF (Lassen, 1968). Similarly, in experimental respiratory acidosis, studies of 4-24 hours duration show a highly significant regulation of the increased CSF [HCO;] via local mechanisms (see above); whereas, data obtained in the truly chronic steady-state of hypercapnia shows that CSF and plasma have undergone comparable increases in [HCO;] and that pH is incompletely compensated in both fluid compa~ments (Bisgard et al., 1975; Bleich et al., 1964; Messeter and Siesjo, 1971; Siesjo, 1972). Hence, available findings point to the probability of a changing contribution of ‘specific’ and ‘non-specific’ mechanisms to CSF bicarbonate and pH regulation during respiratory acid-base disorders, which may be time dependent. However, to quantify the relative contributions of these factors in the chronic steady state requires more definitive studies in which, for example, the effects of a changing plasma bicarbonate are controlled throughout the entire time course of the derangement. Whatever the magnitude of this contribution from local mechanisms to CSF [HCO;] regulation might be during the chronic steady state of hypocapnic hypoxemia, it must be substantially iess than that offered by a falling plasma bicarbonate and is insufficient to achieve a seiective [H’] homeostasis in the CSF.
Acknowledgements We acknowledge the expert technical assistance provided by Ms. Monique Wanner and Ms. Jean Vaughn and the participation of our colleagues, Charles Irvin, Martin Mastenbrook and Timothy Musch. We thank Drs. Peter Lipton, Warren Dennis and Thomas Maren for their helpful criticism of the manuscript. References Barker, S. B. and W. H. Summerson (1941). The calorimetric determination of lactic acid in biological material. J. Biol. Chem. 138: 535-544. Besson, J. M., C. D. Woody and W. H. Marshall (1971). Influences of respiratory acidosis and of cerebral blood flow variations on the DC potential. In : Ion Homeostasis of the Brain, edited by B. K. Siesjo and S. C. Sdrensen. New York, Academic Press, pp. 97-123. Bisgard, G., H. V. Forster, C. A. Rawlings, J. A. Orr, B. Rassmussen and D. D. Buss (1975). CSF acidbase balance in chronic hypercapnia following carotid body denervation. J. Appf. Physiol. (In press). Blayo, M. C., J. P. Marc-Vergnes and J. J. Pocidalo (1973). pH, Pco, and PoI of cisternal cerebrospinal fluid in high altitude natives. Respir. Physiol. 19: 298-311. Bleich, H. L., P. M. Be&man and W. B. Schwartz (1964). The response of cerebrospinal fluid compensation to sustained hypercapnia. J. Clin. Invest. 43: 1 l-16. Dempsey, J. A., H. V. Forsterand G. A. doPico (1974). Ventilatory acciimatization to moderate hypoxemia in man: the role of spinal fluid [FL’]. J. C&z. Invest. 53: 1091-1100. Dempsey, J. A., H. V. Forster, N. Gledhill and G. A. dePico (1975). Effects of moderate hypoxemia and hypocapnia on CSF [H ‘1 and ventilation in man. J. Appl. Physiol. 38: 665-674.
CSF
Feigl, E. 0. and L. G. D’Alecy H. V., J. A. Dempsey
during Genari,
(1972). Normal
acclimatization
and L. W. Chosy
Gillespie,
blood pH, oxygen
and carbon
dioxide
tensions
in
(1975). Incomplete
compensation
of CSF [H+]
in man
(43OOm). J. Appl. Physiol. 38: 1067-1072.
to high altitude
F. J.. M. B. Goldstein
hypocapnia.
arterial
25
IN HYPOCAPNIA
dogs. J. Appl. Physiol. 32: 152-153.
unanesthetized Forster,
[H +] COMPENSATION
and W. B. Schwartz
(1972). The nature
of renal adaptation
to chronic
J. Clin. Invest. 51: 1722-1730.
D. J. and R. E. Hyatt
(1974). Respiratory
mechanics
dog. J. Appl. Physiol.
in the unanesthetized
36(l): 98-102. Gledhill,
N., G. J. Beirne and J. A. Dempsey
term hypocapnia
(1975). Adaptation
phase of renal compensation
to short-
in man. Kidney Inr. (In press).
Held, D., V. Fencl and J. L. Taylor
(1964). Electrical
potential
fluid. J. Neurophysiol. 27:
of cerebrospinal
942-959. Hornbein,
T. F. and E. G. Pavlin (1975). Distribution
respiratory Jennings,
alkalosis
D. B. and J. Sparling
Kazemi,
and alkalosis.
H., N. S. Shore,
acidosis
between
CSF and blood during
and alkalosis.
(1974). Effects of low 0, and high CO, on cardiorespiratory
Am. J. Physiol. 226: 431438. and E. Caravallo-Gil (1967). Brain CO,
conscious resting dogs. Kazemi, H., D. C. Shannon acidosis
of H+ and HCO;
in dogs. .4/n. J. Ph~~iol. 22X: I 149~ I 154.
buffering
function
in
capacity
in respiratory
buffers
in respiratory
J. Appl. Physiol. 22: 241-246. V. E. Shih and D. C. Shannon
(1973). Brain
organic
J. Appl. Physiol. 34: 478482.
Kelley, M. A. and H. Kazemi
(1974). Role of ammonia
as a buffer in the central
system. Respir.
nervous
Physiol. 22: 345-359. Kogure,
K., P. Scheinberg,
vasodilation Lassen,
in hypoxia.
0. M. Reinmuth,
N. A. (1968). Brain extracellular
J. C&z. Lab. Invest. 22: 247-25 Leusen, I. (1972). Regulation 52: l-56. Loeschcke,
the main factor
ventilation.
cerebral
of cerebral flow. Stand.
blood
with reference
New York, Academic des Liquor
Physiol. Rev.
to breathing.
between CSF and blood. In: Ion Homeostasis
of the Brain, edited
Press, pp. 77796. cerebrospinalis
und seine Regulation
durch
Klin. Wschr. 50: 581-593.
V. and B. K. Siesjii (1971). The effect of arterial
arterial
controlling
fluid composition
H. H. (1972). Der Slure-Basenstatus
MacMillan, Messeter,
pH:
of cerebrospinal
H. H. (1971). DC potentials
die Hungen
and R. Busto (1970). Mechanisms
1.
by B. J. Siesjii and S. C. Sorensen. Loeschcke,
J. Fujishina
J. Appl. Physiol. 29 : 223-228.
blood and cisternal
cerebrospinal
K. and B. K. Siesjo (1971). Regulation
hypoxemia
upon acid-base
parameters
in
fluid of the rat. Acfa. Physiol. &and. 83 : 454459. ofthe CSF pH in acute and sustained
respiratory
acidosis.
Acta. Physiol. &and. 83: 21-30. Mines, A. H. and S. C. Sorensen
(1971). Changes
in the electrochemical
between blood and cerebrospinal fluid and in cerebrospinal hypoxia. Acta. Physiol. Stand. 81 : 225-230.
potential
difference
fluid lactate concentration
Mitchell, R. A., D. A. Herbert and C. T. Carman (1965a). Acid-base for cerebrospinal fluid. J. Appl. Physiol. 20: 27-30.
constants
for HCO;
during isocarbic
and temperature
coefficients
Mitchell, R. A., C. T. Carman, J. W. Severinghaus, B. W. Richardson, M. M. Singer and S. Shnider (1965b). Stability of cerebrospinal fluid pH in chronic acid-base disturbances in blood, J. A&. Physiol. 20: 4431152. Orr, J. A., G. E. Bisgard, H. V. Forster, D. D. Buss, J. A. Dempsey fluid alkalosis during high altitude sojourn in unanesthetized
and J. A. Will (1975). Cerebrospinal ponies. Respir. Physiol. 25: 23-37.
Pannier, J. L., J. Weyne and I. Leusen (1971). The CSF/blood potential and the regulation bonate concentration of CSF during acidosis in the cat. Life Sci. 10: 287-300. Plum, F. and J. B. Posner (1967). Blood J. Phvsiol. 212: 864870.
and cerebrospinal
fluid lactate
during
of the bicar-
hyperventilation.
Am.
26
and
D. A. PELLIGRINO
Severinghaus,
J. W., R. A. Mitchell,
at high altitude
suggesting
B. W. Richardson
active transport
Siesjii, B. K. and U. Ponten (1966). Factors
J. J. A. DEMPSEY
and M. M. Singer (1963). Respiratory
affecting
control
of CSF pH. J. Appl. Physiol. 18: 115551162.
regulation
the cerebrospinal
fluid (CSF) bicarbonate
concentra-
tion. Experientia 22: 61 l-614. Siesjii, B. K. (1969). A new theory
for the regulation
of the extra-cellular
of pH regulation
in hypercapnia
pH in the brain.
&and. J. Clin.
Lab. Invest. 24: l-9. SiesjB, B. K. (1971). Quantification Invest. 28
Siesjo, B. K. (1972). The regulation Sorensen,
S. C., N. A. Lassen,
metabolism Sylvester,
and cerebral
Van Varenbergh,
Wischer,
J. Coudert
blood flow in high altitude
and M. P. Zamora
residents.
(1973). Ventilatory
(1974). Cerebral
glucose
J. Appl. Physiol. 37: 305-310. control
during
brief infusions
of CO,-
dog. J. Appl. Physiol. 35: 178-186.
P., G. Demeester
and I. Leusen
(1965). Lactate
in cerebrospinal
fluid during
hyper-
Arch. Int. Physiol. Biochim. 73: 738-747.
Vogl, B. P. and T. H. Maren brospinal
fluid pH. Kidney Int. 1: 360-374.
of cerebrospinal
J. W. Severinghaus,
J. T., B. J. Whipp and K. Wasserman
laden blood in the awake ventilation
Stand. J. Clin. Lab.
and hypocapnia.
: 113-I 16 (Editorial).
(1975). Sodium,
chloride
and bicarbonate
movement
from plasma
to cere-
fuid in cats. Am. J. Physiol. 228: 673-683.
J. and H. Kazemi
(1975). CSF bicarbonate
regulation
in respiratory
acidosis
and alkalosis.
J. Appl. Physiol. 38 : 504-511. Wise, W. C. (1973). Normal
arterial
blood
gases and chemical
components
in the unanesthetized
dog.
J. Appl. Physiol. 35 : 427-429. Woodbury,
J. W. (1971). Alfred pp. 275-289.
Benson
Symposium.
Ion Homeostasis
of the Brain,
3rd. Copenhagen,
North-Holland Publishing Company, Amsterdam
DEPENDENCE OF CSF ON PLASMA BICARBONATE DURING HYPOCAPNIA AND HYPOXEMIC HYPOCAPNIA “’
DALE A. PELLIGRINO
and JEROME A. DEMPSEY3
Pulmonary Physiology Laboratory, Department of Preventive Medicine, University qf Wisconsin, Madison, Wis. 53706, U.S.A.
Abstract. We have previously shown pH compensation
to be similar in CSF and arterial blood during chronic hypoxemic hypocapnia in man and pony, and postulated that the compensatory reduction in CSF [HCO;] was dependent upon corresponding changes in [HCO;]a. We tested this hypothesis in anesthetized, paralyzed dogs by determining the effects of 7 or 14 hours of hypocapnia (Pa,,, 20 and 30 mm Hg), hypoxemia (Pa,, 30,38 and 48 mm Hg) and hypocapnic hypoxemia on CSF acid-base status. [HCO;]a was either permitted to fall normally or was held near control levels by NaHCO, infusion. In hypocapnia and hypoxemic hypocapnia, the decrease in [HCO;] and % pH compensation in CSF were 5 that in arterial blood. Most (SlG39%) of the compensatory decrease in CSF [HCO;] was prevented by preventing the corresponding reduction in [HCO;]a. This dependence of changes in CSF on plasma [HCO;] required a concurrent decrease in CSF PcoZ, but was largely independent of variations in plasma pH. A minor but significant portion of the decrease in CSF [HCO;] was achieved independently of corresponding changes in [HCO;]a. The contribution of this local mechanism to CSF [HCO;] regulation increased with increasing severity of hypocapnia or hypoxemia and was usually associated with a selective increase in CSF lactate. It was concluded that [HCO;] regulation in the CSF during hypoxemic hypocapnia was primarily dependent upon, and therefore limited by, the concomitant decrease in plasma [HCO;]. Cerebrospinal fluid Hypocapnia Hypoxia
Acid-base status Bicarbonate Blood
In chronic disorders of systemic acid-base status it is well-recognized that the protection or compensation of pH in cerebral spinal fluid (CSF) is determined primarily through the regulation qf CSF [HCO;] (Leusen, 1972; SiesjG, 1972). The problem of exactly which mechanisms are responsible for CSF [HCO;] regulation remains controversial. Two broad categories of regulatory mechanisms may be identified (Siesjli, 1972) : a “non-specific” regulation of CSF [HCO;] influenced by correAcceptedfor
publication 19 September 1975.
’ Preliminary findings have been reported in abstract form (Fed. Proc. 33(3): 1369, 1974). ’ This study was supported in part by grants from NIH (1-ROl-HL-15469-02) and the U.S. Army Research and Development Command (No. 17-74-C-4020). ’ D. A. Pelligrino is an NIH pre-doctoral trainee (HL05626) and J. A. Dempsey is the recipient of an NIH Carreer Development Award (HLOO149). 11
12
D.
and J.
A. PELLIGRINO
A. DEMPSEY
sponding changes in plasma [HCO;] and a “CSF-specific” or local regulation via mechanisms which are largely independent of changes in plasma [HCO;]. It has been commonly held that CSF-specific mechanisms dominate CSF [HCO;] regulation because of the relatively close protection afforded CSF pH in systemic acidosis and alkalosis (Leusen, 1972); Mitchell et al., 1965b; Severinghaus et al., 1963). However, recent reviews (Loeschcke, 1972; Siesjii, 1972) have emphasized that a near-complete protection of CSF pH is consistently achieved only in primary metabolic acid-base disorders where changes in CSF [HCO;] are substantially less than those in plasma. In the chronic steady-state of primary respiratory acidosis, compensatory increases in CSF [HCO;] are similar to those in arterial plasma and pH compensation is incomplete in both fluid compartments. This apparent influence of a changing plasma [HCO;] on CSF [HCO;] regulation is also strongly indicated in several studies of short- and long-term respiratory alkalosis. These data provided the rationale for the present study and are summarized in fig. 1. The relevant finding here is the parallel reductions in CSF and plasma [HCO;] during: (a) long-term hypoxemic hypocapnia in awake man and pony (fig. 1 A); and (b) short- and long-term moderate normoxic hypocapnia in awake man and anesthetized dogs (fig. 1 B). -8L"
" ' " ,*'I A[HCOS] CSF(meq/l) ,,j"__
-6- @
+/,/'i
"
0’ ,’ 8
"
/'
,,/#7i
-- @
-
,‘-IDENTITY
-4,$L-
/‘,
,
,
-2 Fig. 1, Reductions
,
in CSF and arterial
during
nonnoxic
A: Dam obtained
in awake
(0,
,
,
-4
,
-6
(
,
-8
blood bicarbonate
and hypoxic
hypocapnia.
man and pony during
ranged
between
,
,
,
,
-4
,
-6
below normocapnic, Values
26 hours
w), 3400 m (A) and 4300 m (A, [7). (Pa,? - 3855
for % pH compensation
,
-2
are mean
normoxic
changes
to 5 weeks sojourn
,
-8
at altitudes
mm Hg; Pacol - 25-32 mm Hg).
50 and 75 % ; and were similar
control
levels,
(+ SEM). of 3100 m
Averagevalues
in CSF and plasma
(Dempsey
et
al., 1974, 1975; Forster
et al., 1975; Orr et al., 1975). B : Data obtained in normoxic hypocapnia in awake man (26 hours - 30 mm Hg Pacol) (Dempsey er al., 1975), and anesthetized dogs (48 hours - 20 mmg Hg Pa,,*) (K azemi et al., 1967 and 1973 ; Plum and Posner, 1967; Van Varenberg et al., 1965; Wischer and Kazemi, 1975).
This close correspondence shown between CSF and plasma suggests that the compensatory reduction in CSF [HCO;] may be truly dependent upon, and therefore limited by, changes in plasma [HCO;]. To test this hypothesis in anesthetized dogs we determined the effects of varying degrees of arterial hypocapnia and/or hypoxemia on [HCO;] regulation and pH compensation in CSF, under conditions
CSF [H+]
COMPENSATION
IN HYPOCAPNIA
13
where plasma [HCO;] was either permitted to fall normally or was held constant near normoxic, normocapnic control levels. Methods
Mongrel dogs (average weight 20 kg) were anesthetized with intravenous pentobarbital (15-20 mg/kg and 46 mg/kg every 2 hours thereafter) and paralyzed with intravenous gallamine (100 mg initially and 60 mg every 2 hours thereafter). Ventilation was maintained through a tracheostomy tube by a constant volume respirator and rectal temperature was maintained at 3838.5 “C with a water blanket. Endtidal Pco2 and P,, were continuously monitored with rapidly responding analyzers (Godart Capnograph and Beckman OM-11). Cathethers were inserted into a femoral vein for injections and infusions and into a femoral artery for continuous monitoring of blood pressure and heart rate and for anaerobic sampling (3 ml per sample). A 20-gauge spinal needle was placed suboccipitally into the cisterna magna for anaerobic sampling of CSF (3-4 ml per sample). The study was aimed at quantifying the effect of plasma [HCO;] on the reduction in CSF [HCO;] during hypocapnia and/or hypoxia. A total of 12 groups, each with 4 to 5 dogs, were studied prior to sacrifice (total N = 54). Each of 6 of these groups was subjected to a specific level of normoxic hypocapnia (Pa,,, - 20 or 30 mm Hg), normocapnic hypoxemia (Pa,, - 30, 38 or 48 mm Hg) or hypocapnic hypoxemia (Pa,, - 48 + Paco2 - 25 mm Hg) for 7 or 14 hours, during which [HCO;la was permitted to fall in a normal fashion. Six additional groups of dogs were also each subjected to 7 or 14 hours at one of these specific levels of hypocapnia and/or hypoxia; but in these animals [HCO;]a was prevented from falling below control levels by continuous intravenous infusion of NaHCO,. The key comparison in this study, then, focused on the differences in changes in CSF [HCO;] between groups which were at similar levels of hypocapnia and/or hypoxemia; but at different, i.e., control us reduced, levels of plasma [HCO;]. Three additional dogs were studied for 7 hours in normoxic normocapnia with NaHC03 infusion to determine the effects of NaHCO, infusion alone on CSF [HCO;]. The protocol for a single experiment was as follows. Immediately following anesthesia, tracheal intubation and paralysis, the FIEFand ventilator were set to provide a Paco2 - 40 mm Hg and Pa,, - 100 mm Hg. The cisternal needle was then inserted. A 2-l/2 hour ‘control’ period of normoxic normocapnia followed, during which arterial blood was sampled every 20-30 minutes and CSF sampled twice, one hour apart. Then (in about one-half the animals) hypoxia and/or hypocapnia were gradually induced over a 15-25 minute period by mechanical hyperventilation and/or manipulation of the inspired 0, and N2 concentrations. The desired levels of PaoZ and PaYoZ were then maintained over the subsequent 7 or 14 hour period. For studies which required an unchanging [HCO;]a, a low rate of NaHCO, infusion was begun 10 minutes prior to the experimental period and the infusion continued while hypoxia and/or hypocapnia was gradually induced and then maintained, as outlined above. The infusion rate was frequently adjusted (020.8 meq/min NaHC03) to maintain
14
D. A. PELLIGRINO and J. A. DEMPSEY
[HCO;]a at or within 1.5 meq/l above control levels throughout the experimental period. Arterial blood was sampled and analyzed every l&15 minutes during the first hour and at 30-45 minute intervals thereafter. Constant Pao2 and Paco2 (and [HCO;]a where appropriate) were maintained throughout the experimental period by continuous monitoring of end-tidal gases and frequent analyses of arterial blood samples. Cistemal CSF samples were obtained at the 3rd, 5th and 7th (or 14th) hours of the experimental period. CSF samples contaminated with air bubbles or blood were discarded. Arterial samples were analyzed for Po,, Pco2 and pH and CSF samples for Pco2 and pH with electrodes (Radiometer, Copenhagen) at 38 “C, with corrections for small deviations in the animal’s temperature. [HCO;] was calculated using the appropriate pk’ and S values (Mitchell et aE., 1965a). Lactic acid concentration was analyzed in plasma and CSF by a modified calorimetric method (Barker and Summerson, 1941). The technique, reproducibility and validity of these measurements have been previously described (Dempsey et al., 1974, 1975). The percentage of pH compensation in CSF and arterial blood during hypocapnia was calculated after Siesjo (1971).
TABLE Arterial blood and CSF acid-base status in control Arterial blood +7or
Control NaHCO, infusion
l4hr
[HCO;]
Lact
P07
Pco*
PH
Moderate normoxic hypocapnia (7 hr) No 90 41.5 7.345 +0.006 +2.6 kO.5 Yes 96 40.7 7.349 +0.4 *0.012 k5.6
22.0 kO.5 21.8 +0.5
2.72 +0.51 2.36 *0.53
96.0 * 1.4 85 +4.9
29.3 kO.7 29.8 +I.0
7.417 f 0.009 7.484 kO.026
Severe normoxic hypocapnia (7 hr) No 103 40.8 7.358 +0.7 rf:0.007 f5.3 Yes II4 41 .o 7.350 +0.7 kO.008 k4.5
22.4 kO.4 22.1 kO.3
2.13 kO.27 2.03 kO.15
95 +4.7 100 k6.3
19.0 kO.4 20.7 +0.7
7.525 kO.016 7.643 *0.010
Severe normoxic hypocapnia (I4 hr) No 122 40.2 7.374 +0.012 + 1.0 k7.0 Yes II8 41.1 7.353 +3.8 +0.6 + 0.002
22.9 kO.3 22.2 +0.4
I .90 rfr0.27 2.19 kO.25
96 f 1.8 87 + 1.2
22.3 +2.6 21.5 * 1.2
1.475 f0.051 7.644 *0.018
Pol
Pco2
PH
PO1and Pcol in mm Hg; HCO; and lactate in meq/l. All values are mean *SEM with 4 or 5 dogs per condition.
CSF [H’]
~MPENSATION
15
IN HYPOCAPNIA
Results PRE-EXPERIMENTAL CONTROLS
The acid-base status during normo~pnic, normoxic control conditions are shown for each group in tables 1 and 2. The control data obtained in the 12 different groups of animals were consistent. For example, the range of values for rnekAl [HCO;] was 2.5 meq/l in arterial blood (20.4 to 22.9) and 1.7 meq/l in CSF (21.7 to 23.4). Mean values in CSF, compared to arterial blood, showed a lower pH (N 0.02 units) and higher P,,, (N 6 mm Hg) and [HCO;] (w 1 meq/l). The control arterial [HCO;] in these anesthetized, paralyzed animals were in the mid-range of mean values reported for unanesthetized, ‘non-panting’ dogs (19.8 to 24.5 meq/l) (Wise, 1973; Sylvester et al., 1973; Gillespie and Hyatt, 1974; Feigl and D’Alecy, 1972; Jennings and Sparling, 1974). Data obtained periodically during the 2-l/2 hour pre-experimental control periods (not shown), showed no systematic changes in arterial acid-base status and close agreement between two CSF samples. EFFECTS OF
INFUSION ALONE
NaHCO,
Normocapnic-normoxic
conditions
were maintained
for 7 hours and NaHCO,
t md at the end of each period of normoxic hypocapnia ~~~
_.._Cisternal CSF Control
HCO;]
Lact
Pco1
8.5
1.31
47.6
to.3
& 0.05
+1.3
‘2.1
1.76
47.5
k1.0
iO.55
kO.5
-.
+7 or 14 hr PH
[HCO;]
Lact
Pco2
PH
[HCO,]
Lact
7.325 +0.011
23.3 +0.4
2.61 +0.34
38.1
7.367
20.7
2.31
+ 0.009
io.7
10.14
7.311 *0.005
22.5
1.94
10.6 39.1
7.386
22.2
I .96
+0.2
+0.09
to.5
*0.01
+0.6
t-o.12
5.6
1.78
45.7
2.09 +0.12
19.5
3.02
kO.6
22.6 +0.4
7.436
50.27
7.329 F0.005
30.4
to.4
kO.8
!2.5
2.34
46.9
7.318
22.6
2.10
30.5
* 0.009 7.472
50.4 21.4
4.20
+o.s
kO.39
+0.6
,0.007
+0.3
+0.15
+1.1
~0.008
+0.5
kO.49
6.2
1.40 +0.13
45.1
7.338
22.8
2.05
30.4
7.410
18.1
2.73
& 0.004
kO.4
kO.11
+ 2.6
7.303
21.7
2.40
28.2
+ 0.036 7.464
k0.I 19.4
4.67
,0.007
+0.4
kO.16
20.3
*0.004
+0.3
20.69
to.3 !3.3
2.20
kO.5 46.7
* 1.2
&0.60
kO.8
+0.93
+0.35
16
D. A. PELLIORINO
and
J. A. DEMPSEY
1 Arterial
Arterial
blood and CSF acid-base
status
in control
am
blood
Control
f7
hr
NaHCO, infusion Hypoxia No Yes
Po
2
+ hypocapnia
Lact
PO2
Pco,
pH
7.417
(7 hr) 7.335
20.4
2.52
45.0
26.3
k4.6
+0.6
+0.010
kO.7
+0.36
k3.1
kO.9
kO.024
108
41.4
7.337
21.7
2.05
48.3
27.8
7.525
+5.5
+0.9
kO.012
+0.3
+0.11
kO.8
kO.7
k 0.008
7.337
21.2
1.98
46.1
39.1
7.330
+0.005
kO.5
+0.32
* 1.3
+ 1.9
* 0.020
7.338
normocapnic 105 +4.9
Severe normocapnic
hypoxia
(7 hr)
40.7 kO.9 hypoxia
(7 hr)
120
39.9
7.339
20.9
1.60
37.9
36.9
+4.5
kO.08
* 0.007
+0.5
+0.20
+ 1.6
* 1.3
*0.010
118
41.0
7.362
22.7
1.96
38.3
36.0
7.430
k5.3
* 1.5
kO.016
+0.7
kO.36
+2.4
f 1.8
+0.006
Very severe normocapnic Yes
[HCO;]
39.3
No
Yes
PH
102
Moderate
No
P co2
hypoxia
(7 hr)
119
40.6
7.355
22.1
2.31
30.8
38.4
7.391
k4.6
kO.6
* 0.007
+0.4
kO.14
+0.1
+0.7
+0.013
PO2 andPco,in mm Hg ; HCO;
and lactate in meq/l. All values are mean + SEM with 4 or 5 dogs per condition.
infused at a rate which produced a 3.0-3.5 meq/l increase in [HCO;]a above control (table 3). No measureable changes in CSF [HCO;] accompanied these increased levels of [HCO;]a. The increases in [HCO;]a above control levels in these studies
TABLE Effects Arterial
of NaHCO,
infusion
3
during
normoxic
normocapnia
CSF
blood
PH
P co2
WC%1
PH
P co2
t-HCO ;1
Control
7.33
42
22.0
7.32
47
22.6
+3hr +5hr
7.38 7.38
42 43
24.8 25.0
7.31 7.32
48 47
22.8 22.6
+7hr
7.39
42
25.2
7.30
48
22.4
P co* in mm Hg; [HCO;]
in meqil. Mean values (N = 3). Arterial
Po, was - 100 mmHg
throughout
CSF [H +]
17
COMPENSATION IN HYF’OCAPNIA
id of each period of hypoxemia and hypocapnic hypoxemia _ _
Cisternal CSF +7 hr
Control [CO;]
Lad
Pco1
PH
[HCO,]
Lact
P,,,
pH
[HCO,]
Lact
,.6
45.2 kO.8 47.6 +0.5
7.346 +0.004 7.314 kO.004
23.3 +os 22.7 +0.2
2.67
.8 0.5
1.42 kO.18 1.67 kO.18
+0.11 2.29 * 0.09
3.54 to.6 36.7 +1.1
7.386 kO.015 7.400 &0.009
19.6 kO.8 21.3 +0.5
3.79 rto.52 3.55 kO.18
I.1 1.0
I .90 kO.48
47.4 +0.7
7.326 kO.004
23.3 to.5
2.49 +0.10
47.5 & 1.4
7.304 kO.015
22.1 kO.3
2.49 +0.86
.3 0.9 .5 1.3
1.53 io.34 0.96 +0.21
44.0 k1.0 46.2 +1.0
7.346 &0.004 7.339 +0.007
22.7 +0.5 23.4 +0.3
2.30 +0.20 2.34 +0.16
42.2 &I.6 42.2 +0.8
7.339 kO.013 7.346 20.014
21.4 kl.1 22.0 +0.4
2.53 7to.37 2.82 kO.48
.8 0.5
2.53 kO.46
45.5 +1.9
7.337 +0.013
23.0 kO.4
2.39 +0.13
40.0 +1.2
7.335 +0.006
20.1 +0.8
3.77 &OS1
0.6
in normoxic normocapnia were 2-3 times greater than the 0.3-1.2 meq/l increases produced during the experimental periods of hypocapnia and/or hypoxia (fig. 3). We interpret these data to mean that the NaHCO, infusion alone had no significant effect on CSF [HCO;] during the experimental periods. These limited data obtained in not-moxie normocapnia agree with a number of reported studies(Leusen, 1972). ACIItBASE
CHANGES IN HYPOCAPNIA
AND/OR HYPOXEMIA
The changes in the arterial and CSF acid-base status over the duration of each experiment are shown in fig. 2 - for studies without NaHCO, infusion, and in fig. 3 for studies in which NaHCO, infusion accompanied the hypocapnia and/or hypoxemia. The mean data obtained in the control and at the end of each experimental period are summarized in table 1 (hypocapnic experiments) and table 2 (hypoxic hypocapnia and hypoxic experiments).
After 7 hours of moderate hypocapnia (PacoZ N 29 mm Hg) (table 1 - no NaHCO,
18
and J. .4. DEMPSEY
D. A. PELLIGRINO
ARTERIAL
BLOOD
CISTERNAL
CSF
A PC02 (mm Hg )
5 0
20
f
Fig. 2. Effects of hypocapnia are mean changes of normoxic
hypocapnia
the upper left-hand
and/or
from normocapnic, (0,
0).
panel are shown
hypoxemia normoxic normocapnic the various
on acid-base control
status
hypoxemia
blood
(A, A) and hypocapnic
levels of arterial
Paol (in parenthesis),
in arterial
and CSF. Values
levels, Symbols denote the experimental hypocapnia
conditions
hypoxemia
(ordinate)
( x ). In
and the average
for each condition.
infusion), Pcol fell 2.7 mm Hg more in arterial blood (- 12.2 mm Hg) than in CSF (-9.5 mm Hg) and the reduction in [HCO;] was 0.9 meq/l greater in blood (- 3.5 meq/l) than in CSF (- 2.5 meq/l). After 7 hours of the more severe hypocapnia (Pace, - 19 mmg H) (table 1 - no NaHCO, infusion), Pco2 fell 6.5 mm Hg more in blood (-21.8 mm Hg) than in CSF (- 15.3 mm Hg). The fall in [HCO;] was 3.7 meq/l greater in blood (-6.8 me& than in the CSF (- 3.1 meq/l). These changes after 7 hours at both levels of hypocapnia resulted in a significant rise in pH in both blood and CSF, with pH compensation - 12-l 5 % more complete in blood than in CSF. After a longer period of 14 hours at the more severe level of hypocapnia (table 1 - no NaHCO, infusion), changes in blood (- 17.9 mm Hg APco, and -6.7 meq/l A[HCO;]) more closely paralleled those in CSF (- 14.7 mm Hg) and -4.7 meq/l), and pH compensation was similar in blood and CSF (- 57 %). When hypocapnia was induced and plasma [HCO;] was prevented from falling below control levels (table 1 - NaHC03 infusion). CSF [HCO,] remained un-
CSF [H’]
COMPENSATION
ARTERIAL
BLOOD
CISTERNAL
A Pcop (mn $11
CSF
Hg) 15
l-
A
19
IN HYPOCAPNIA
IHCO
0.4Or
I
I I
I 2
I 3
1 4
I
5
I 6
IdIll 7’10’14
I
I I
I 2
I 3
I 4
I 5
I 6
IllI 7’10’14e
Fig. 3. Effects of hypocapnia and/or hypoxemia plus NaHCO; infusion on acid-base status in arterial blood and CSF. Values are mean changes from normocapnic, normoxic control levels. As in fig. 2 symbols denote the experimental conditions of normoxic hypocapnia (0, a), normocapnic hypoxemia (A., 0) and hypocapnic hypoxemia (x). In the upper left-hand panel are shown the various levels of arterial hypocapnia (ordinate) and the average Paol, for each condition.
changed at 30 mm Hg Pacoz and fell 1.2 meq/l after 7 hours and 2.3 meq/l after 14 hours at 20 mm Hg Paco2. Lactate rose only in the CSF (+ 2 to 2.5 meq/l) and only during the more severe level of hypo~pnia. These average reductions in CSF [HCO;] were all significantly less (P < 0.01) than those obtained when plasma [HCO;] was permitted to fall in a normal fashion. Hypocapnia plus hypoxemia
With the combined hypocapnic hypoxemia (table 2 - no NaHC03 infusion) Pc,, fell 2.3 mm Hg more in arterial blood (- 13.0 mm Hg) than in CSF (- 10.7 mm Hg). [HCO;] fell 3.8 meq/l in both compartments and pH compensation was l&12% more complete in the CSF than in the blood. Lactate rose only in the CSF (+ 1.2 meqil). When plasma [HCO;] was prevented from falling during hypocapnic hypoxemia, CSF [HCO;] fell 1.4 meq/l - a decrease which was signi~cantiy less than that seen when [HCO; ]a was permitted to fall (P < 0.01). Lactate rose only in the CSF (+ 1.3 meq/l).
20
D.
A. PELLIGRINO
and J.
A. DEMPSEY
Hypoxemia alone (46 or 38 mm Hg Pao,,) produced no systematic changes in arterial blood [HCO;], lactate or pH; but CSF [HCO;] fell 1.2-1.5 meq/l below control levels (P < 0.01) with no change in CSF lactate (table 2). When NaHCO, infusion was superimposed at 2 levels of hypoxemia, i.e., Pao2 38 and 30 mm Hg, CSF [HCO;] was reduced 1.4 and 3.0 meq/l, respectively (P < 0.01). CSF lactate rose 1.4 meq/L at 30 mm Hg Pao2. DEPENDENCE OF
CSF
ON PLASMA
[HCO;]
Table 4 compares pertinent findings obtained with and without NaHCO, infusion; and shows the calculated effects of preventing the fall in plasma [HCO;] during hypocapnia and/or hypoxemia on [HCO;] regulation and pH compensation in the CSF. In normoxic hypocapnia and hypoxic hypocapnia more than half of the reduction in CSF [HCO;] was dependent upon a concomitant reduction in arterial [HCO;] . Accordingly, the completeness of pH compensation in the CSF was reduced substantially by preventing the fall in arterial [HCO;] (.z.e., when compensation of arterial blood pH was completely prevented). This dependence of CSF [HCO;] regulation on a reduction in plasma [HCO;] was greatest in moderate hypocapnia (89%) and least in more severe, prolonged hypocapnia (51%). At all levels of normocapnic hypoxemia, CSF [HCO;] was reduced independently of corresponding changes in plasma [HCO;]. This reduction in CSF [HCO;] was similar at Pao,, 46 and 38 mm Hg (- 1.3 meq/l) and greatest at 30 mm Hg Pa,, ( - 3 .Omeq/l). Diiussion
Present findings in anesthetized dogs have shown that bicarbonate regulation and pH compensation in CSF during short-term hypocapnia or hypoxemic hypocapnia was primarily dependent upon corresponding changes in plasma bicarbonate. In addition, we quantified the contribution of a component to CSF [HCO;] regulation during hypocapnia which was independent of changes in the plasma [HCO;]. The contribution of this local or ‘CSF-specific’ mechanism to CSF [HCO;] regulation during hypocapnia was always less than that attributable to a reduction in plasma [HCO;] ; but was shown to increase with increasing severity of hypocapnia, was significant at even moderate levels of hypoxemia and was usually associated with a corresponding rise in CSF lactic acid concentration. CSF [HCO,]
REGULATION IN RESPIRATORY ALKALOSIS
Present findings are consistent with the view that quite different mechanisms might be responsible for brain ECF [HCO;] regulation during the different types of systemic acid-base disturbances (Siesjii, 1972). Furthermore, a consideration of past and present findings suggests that the regulation of CSF [HCO;] during respiratory alkalosis might be unique, because of the relatively large extent to which it is dependent upon concomitant changes in plasma [HCO;]. For example, in primary
CSF [H+]
COMPENSATION
IN HYPOCAPNIA
21
metabolic alkalosis or acidosis, CSF [HCOY] and therefore CSF pH change little with respect to plasma (Leusen, 1972; Loeschcke 1972; Siesjii, 1972). Most evidence points to appropriate changes in the CSF/plasma electrical potential difference sensitive to a changing pHa - as a primary protector of CSF [HCO;] and [H+] in metabolic acid-base disturbances (Siesj8, 1969). In primary respiratory acidosis, CSF pH appears to be less well-protected, but - at least over 6 or 7 hours of hypercapnia (see below) - the rise in CSF [HCO;] clearly exceeds that in plasma (Leusen, 1972; Pannier et al., 1971; Siesjii and Ponten, 1966). This compensatory rise in CSF [HCO;] has been shown to be due in small part to the corresponding rise in plasma [HCO;], but is determined primarily by local mechanisms; namely, CO2 hydration at the site(s) of CSF formation and increased brain ammonia production (Kazemi et al., 1973 ; Kelley and Kazemi, 1974; Vogl and Maren, 1975; Wischer and Kazemi, 1975). Present findings in primary respiratory alkalosis have shown that the compensatory reduction in CSF [HCO;] is even more dependent upon the concomitant fall in plasma [HCO;] and is influenced locally only to a limited extent and possibly through only the single mechanism of an increased brain lactic acid production. Our findings and interpretations are supported in part by the work of Kazemi and colleagues (Kazemi et al., 1973 ; Wischer and Kazemi, 1975), which showed that the reduction in CSF [HCO;] during 4 to 6 hours of hypocapnia in dogs was unaffected by brain carbonic anhydrase inhibition and was not accompanied by significant changes in brain ammonia or glutamic acid concentrations. It is not clear exactly which mechanism(s) are responsible for the observed dependence of changes in CSF [HCO;] on plasma bicarbonate during respiratory alkalosis. The most straightforward explanation would be that bicarbonate loss from CSF simply followed the decline in brain capillary plasma bicarbonate, per se, via passive diffusion. However, in view of the relative stability of CSF [HCO;] consistently shown during studies of systemic non-respiratory acidosis (see above), this process must account for only a very minor portion of the observed dependence of CSF upon plasma [HCO;] during hypocapnia. This conclusion requires the assumption that the permeability of the blood-CSF ‘barrier’ is similar in metabolic acidosis and respiratory alkalosis. Information concerning the permeabilities of [HCO;] and [H+] and their variability, is notably sparse (Woodbury, 1971). A second explanation would link the observed dependence of CSF on plasma [HCO;] to local regulatory mechanisms. One such mechanism might be a changing CSF-plasma potential difference (E) which is known to become more positive in systemic acidosis and more negative in alkalosis, thereby returning the relevant electro-chemical potential differences toward normal (Hornbein and Pavlin, 1975; Siesjii, 1969 and 1972). However, this explanation does not fit present observations, if the common view is taken that changes in E are solely the result of changes in plasma pH (Held et al., 1964; Hornbein and Pavlin, 1975). For example, by comparing conditions of equal hypocapnia at widely different pHa (i.e. with us without maintained [HCO;]a), we consistently observed that the greatest reductions in CSF [HCO;] occurred when the rise in pHa was least. These data do not rule out
22
D. A. PELLIGRINO
and
J. A. DEMPSEY
changes in E as a key mediator of CSF [HCO;] regulation during respiratory alkalosis, if one accepts the minority view that such changes in the potential are regulated at least in part by factors other than plasma pH : for example, by changes in plasma [HCO;] and/or [Cl-] (Loeschcke, 1971) or by unknown factors associated with acute changes in cerebral blood flow (Besson et al., 1971). A second means of local control over the ionic composition of newly formed CSF during respiratory alkalosis might be found in some mechanism which is sensitive to a low intra-cranial Pco2. Such a ‘CO,-sensitive’ regulator has been suggested from studies in respiratory acidosis (see above) ; and limited data in chronic hypocapnia indicate that some of the inter-subject variability in changes in CSF [HCO;] may be related to variations in CSF Pco2 rather than in plasma Pco2 or [HCO;] (Dempsey et al., 1975; Forster et al., 1975). However, present data does not support a significant role for reduced CSF P co2, per se, as an independent regulator of CSF [HCO;]. For example, comparing conditions of maintained us reduced plasma [HCO;] during ‘moderate’ arterial hypocapnia, the decreases in CSF Pco2 were equal in the paired conditions, but the compensatory reduction in CSF [HCO;] was only observed when [HCO;]a was permitted to fall. Hence, available data indicate that most of the compensatory movement of bicarbonate out of the CSF during hypocapnia is unrelated to variations in plasma pH and is not attributable to the independent effects of a reduced plasma [HCO;] or CSF Pco2 alone. We can only conclude, then, that [HCO;] transfer out of the CSF requires a concomitant reduction in both plasma bicarbonate and CSF Pco2. APPLICATION
OF FINDINGS
TO CHRONIC
HYPOXEMIC
HYPOCAPNIA
These studies were designed to interpret the close correspondence observed in pH compensation between CSF and blood during long-term, steady-state hypoxemic hypocapnia in man. In general, present findings permit the conclusion that this close correspondence between fluid compartments was attributable primarily to a true dependence of bicarbonate loss from CSF on a reduction in plasma bicarbonate. They imply, therefore, that renal compensation (Gledhill et al., 1975; Genari et aE., 1972) provides an important limitation to the rate and extent of CSF bicarbonate regulation during long-term hypocapnia, with the result that pH compensation will be similar and incomplete in both blood and CSF (Forster et al., 1975; Siesjo, 1972). Present findings also indicated a role for some factor which affected a variable, yet significant reduction in CSF [HCO;] - independently of a constant [HCO;]a during all those conditions of hypoxemia and/or hypocapnia chosen to simulate arterial blood gases in healthy man during sojourn to 3100-4300 m altitude. Even the relatively ‘moderate’ levels of hypoxemia encountered during man’s first 24 hours (- 3840 mm Hg Pa,,) or after his first week (- 45-50 Pa,,) at 4300 m altitude (Forster et al., 1975) were associated with local regulatory influences which were sufficient to account for - one-third of the total bicarbonate reduction or pH compensation in CSF in the hypocapnic, hypoxemic dog (table 4). This data agrees
CSF [H’]
COMPENSATION
TABLE 4 regulation and pH com~nsation in the CSF on plasma and/or hypoxemia ..-.I ~__..^ -.._.. ___~.. ____ ._-.-..
Dependence of [HCO;] ____...--.
Change in CSF [HCO;] Without PaCoL Pa,, (mm Hg) (mm Hg) HCO; infusion .____ 30 20 20* 26 39 37 38
.._> 90 > 90 > 90 45 46 38 30
* Paco2 N 20 mm ** [HCO;]a did *** E.g., during [HCO;] without
(mesit)
(meq/l) ----2.6 -3.1 -4.7 -3.7 - I .2** - 1.3 -
from control ~_
With HCO; infusion
23
IN HYPOCAPNIA
[HCO;] ~.
during hypox~mia
--.-
_.-_-.
- ..-
% CSF pH compensation .----I___ ._--.
‘Dependence’ on plasma[HCO;] (% of change without infusion)***
Without HCO; infusion
With HCO; infusion
89% 61% 51% 65% (0 %) (0%) -
40% 35% 55% 60%
5O’ /0 15% 20% 20 9’ 10 -
.l_-__--0.3 -1.2 -2.3 - t.3 - 1.4 - 3.0
Hg for 14 hours. All other conditions were maintained for 7 hours. nat change significantly from normocapnic, normoxic control levels during these studies. moderate normoxic hypocapnia (Pam2 N 30 mm Hg) the compensatory reduction in CSF NaHCO; infusion = 2.6 meq/l; and the reduction in CSF [HCO;] when [HCO;]a was not
permitted to fall = 0.3 meq/l. Thus, the fall in CSF [HCO;] which was ‘dependent’ upon a falling [HCO;]a = 2.3 meq/I; or z&i 100 = 89% of the total change observed without NaHCO; infusion.
= 2.6-0.3
with that previously shown during progressive, acute hypoxemia in the anesthetized rat and dog (Kogure et al., 1970; MacMillan and Siesjii, 197i), and may be explained at least in part by a selective cerebral tissue lactacidosis (Mines and S#rensen, 1971; Plum and Posner, 1967). A strict application of these findings to chronic hypoxemic hypocapnia in man suggests that CSF [HCO;] was significantly influenced by locaf regulatory mechanisms, even though the reduction in bicarbonate and completeness of pH compensation in the CSF was paralleled or even exceeded by that in plasma (fig. 1). Although the available data appear to justify this postulate, serious limitations exist in our attempt to apply findings obtained in semi-acute conditions to the truly chronic steady-state of hypoxemic hypocapnia. Some findings are available which, while incomplete, strongly suggest that the participation of local mechanisms in the regulation of CSF [HCO;] in hypoxia and respiratory acid-base disturbances might be time dependent. For example, in the awake pony the reduction in CSF [HCO;] (and rise in CSF lactate concentration) was shown to exceed that in plasma during the initial six hours of hypoxic exposure but not after two or more days of continued exposure (Orr et al., 1975); lactic acid concentrations in CSF remain fairly constant near sea-level control levels in healthy man after one to three weeks of high altitude sojourn (Dempsey et al., 1974; Forster et al., 1975); studies in the human native of high altitudes have failed to detect an increased level of brain anaerobic glycolysis
24
0. A. PELLIGRINO
and J. A. DEMPSEY
(Blayo et al., 1973; Sorensen et al., 1974); and in normoxic hypocapnia, cerebral blood flow undergoes an acute reduction followed by a gradual return toward normal, thereby presenting at least a potential time-dependent effect on the local regulation of bicarbonate movement out of the CSF (Lassen, 1968). Similarly, in experimental respiratory acidosis, studies of 4-24 hours duration show a highly significant regulation of the increased CSF [HCO;] via local mechanisms (see above); whereas, data obtained in the truly chronic steady-state of hypercapnia shows that CSF and plasma have undergone comparable increases in [HCO;] and that pH is incompletely compensated in both fluid compa~ments (Bisgard et al., 1975; Bleich et al., 1964; Messeter and Siesjo, 1971; Siesjo, 1972). Hence, available findings point to the probability of a changing contribution of ‘specific’ and ‘non-specific’ mechanisms to CSF bicarbonate and pH regulation during respiratory acid-base disorders, which may be time dependent. However, to quantify the relative contributions of these factors in the chronic steady state requires more definitive studies in which, for example, the effects of a changing plasma bicarbonate are controlled throughout the entire time course of the derangement. Whatever the magnitude of this contribution from local mechanisms to CSF [HCO;] regulation might be during the chronic steady state of hypocapnic hypoxemia, it must be substantially iess than that offered by a falling plasma bicarbonate and is insufficient to achieve a seiective [H’] homeostasis in the CSF.
Acknowledgements We acknowledge the expert technical assistance provided by Ms. Monique Wanner and Ms. Jean Vaughn and the participation of our colleagues, Charles Irvin, Martin Mastenbrook and Timothy Musch. We thank Drs. Peter Lipton, Warren Dennis and Thomas Maren for their helpful criticism of the manuscript. References Barker, S. B. and W. H. Summerson (1941). The calorimetric determination of lactic acid in biological material. J. Biol. Chem. 138: 535-544. Besson, J. M., C. D. Woody and W. H. Marshall (1971). Influences of respiratory acidosis and of cerebral blood flow variations on the DC potential. In : Ion Homeostasis of the Brain, edited by B. K. Siesjo and S. C. Sdrensen. New York, Academic Press, pp. 97-123. Bisgard, G., H. V. Forster, C. A. Rawlings, J. A. Orr, B. Rassmussen and D. D. Buss (1975). CSF acidbase balance in chronic hypercapnia following carotid body denervation. J. Appf. Physiol. (In press). Blayo, M. C., J. P. Marc-Vergnes and J. J. Pocidalo (1973). pH, Pco, and PoI of cisternal cerebrospinal fluid in high altitude natives. Respir. Physiol. 19: 298-311. Bleich, H. L., P. M. Be&man and W. B. Schwartz (1964). The response of cerebrospinal fluid compensation to sustained hypercapnia. J. Clin. Invest. 43: 1 l-16. Dempsey, J. A., H. V. Forsterand G. A. doPico (1974). Ventilatory acciimatization to moderate hypoxemia in man: the role of spinal fluid [FL’]. J. C&z. Invest. 53: 1091-1100. Dempsey, J. A., H. V. Forster, N. Gledhill and G. A. dePico (1975). Effects of moderate hypoxemia and hypocapnia on CSF [H ‘1 and ventilation in man. J. Appl. Physiol. 38: 665-674.
CSF
Feigl, E. 0. and L. G. D’Alecy H. V., J. A. Dempsey
during Genari,
(1972). Normal
acclimatization
and L. W. Chosy
Gillespie,
blood pH, oxygen
and carbon
dioxide
tensions
in
(1975). Incomplete
compensation
of CSF [H+]
in man
(43OOm). J. Appl. Physiol. 38: 1067-1072.
to high altitude
F. J.. M. B. Goldstein
hypocapnia.
arterial
25
IN HYPOCAPNIA
dogs. J. Appl. Physiol. 32: 152-153.
unanesthetized Forster,
[H +] COMPENSATION
and W. B. Schwartz
(1972). The nature
of renal adaptation
to chronic
J. Clin. Invest. 51: 1722-1730.
D. J. and R. E. Hyatt
(1974). Respiratory
mechanics
dog. J. Appl. Physiol.
in the unanesthetized
36(l): 98-102. Gledhill,
N., G. J. Beirne and J. A. Dempsey
term hypocapnia
(1975). Adaptation
phase of renal compensation
to short-
in man. Kidney Inr. (In press).
Held, D., V. Fencl and J. L. Taylor
(1964). Electrical
potential
fluid. J. Neurophysiol. 27:
of cerebrospinal
942-959. Hornbein,
T. F. and E. G. Pavlin (1975). Distribution
respiratory Jennings,
alkalosis
D. B. and J. Sparling
Kazemi,
and alkalosis.
H., N. S. Shore,
acidosis
between
CSF and blood during
and alkalosis.
(1974). Effects of low 0, and high CO, on cardiorespiratory
Am. J. Physiol. 226: 431438. and E. Caravallo-Gil (1967). Brain CO,
conscious resting dogs. Kazemi, H., D. C. Shannon acidosis
of H+ and HCO;
in dogs. .4/n. J. Ph~~iol. 22X: I 149~ I 154.
buffering
function
in
capacity
in respiratory
buffers
in respiratory
J. Appl. Physiol. 22: 241-246. V. E. Shih and D. C. Shannon
(1973). Brain
organic
J. Appl. Physiol. 34: 478482.
Kelley, M. A. and H. Kazemi
(1974). Role of ammonia
as a buffer in the central
system. Respir.
nervous
Physiol. 22: 345-359. Kogure,
K., P. Scheinberg,
vasodilation Lassen,
in hypoxia.
0. M. Reinmuth,
N. A. (1968). Brain extracellular
J. C&z. Lab. Invest. 22: 247-25 Leusen, I. (1972). Regulation 52: l-56. Loeschcke,
the main factor
ventilation.
cerebral
of cerebral flow. Stand.
blood
with reference
New York, Academic des Liquor
Physiol. Rev.
to breathing.
between CSF and blood. In: Ion Homeostasis
of the Brain, edited
Press, pp. 77796. cerebrospinalis
und seine Regulation
durch
Klin. Wschr. 50: 581-593.
V. and B. K. Siesjii (1971). The effect of arterial
arterial
controlling
fluid composition
H. H. (1972). Der Slure-Basenstatus
MacMillan, Messeter,
pH:
of cerebrospinal
H. H. (1971). DC potentials
die Hungen
and R. Busto (1970). Mechanisms
1.
by B. J. Siesjii and S. C. Sorensen. Loeschcke,
J. Fujishina
J. Appl. Physiol. 29 : 223-228.
blood and cisternal
cerebrospinal
K. and B. K. Siesjo (1971). Regulation
hypoxemia
upon acid-base
parameters
in
fluid of the rat. Acfa. Physiol. &and. 83 : 454459. ofthe CSF pH in acute and sustained
respiratory
acidosis.
Acta. Physiol. &and. 83: 21-30. Mines, A. H. and S. C. Sorensen
(1971). Changes
in the electrochemical
between blood and cerebrospinal fluid and in cerebrospinal hypoxia. Acta. Physiol. Stand. 81 : 225-230.
potential
difference
fluid lactate concentration
Mitchell, R. A., D. A. Herbert and C. T. Carman (1965a). Acid-base for cerebrospinal fluid. J. Appl. Physiol. 20: 27-30.
constants
for HCO;
during isocarbic
and temperature
coefficients
Mitchell, R. A., C. T. Carman, J. W. Severinghaus, B. W. Richardson, M. M. Singer and S. Shnider (1965b). Stability of cerebrospinal fluid pH in chronic acid-base disturbances in blood, J. A&. Physiol. 20: 4431152. Orr, J. A., G. E. Bisgard, H. V. Forster, D. D. Buss, J. A. Dempsey fluid alkalosis during high altitude sojourn in unanesthetized
and J. A. Will (1975). Cerebrospinal ponies. Respir. Physiol. 25: 23-37.
Pannier, J. L., J. Weyne and I. Leusen (1971). The CSF/blood potential and the regulation bonate concentration of CSF during acidosis in the cat. Life Sci. 10: 287-300. Plum, F. and J. B. Posner (1967). Blood J. Phvsiol. 212: 864870.
and cerebrospinal
fluid lactate
during
of the bicar-
hyperventilation.
Am.
26
and
D. A. PELLIGRINO
Severinghaus,
J. W., R. A. Mitchell,
at high altitude
suggesting
B. W. Richardson
active transport
Siesjii, B. K. and U. Ponten (1966). Factors
J. J. A. DEMPSEY
and M. M. Singer (1963). Respiratory
affecting
control
of CSF pH. J. Appl. Physiol. 18: 115551162.
regulation
the cerebrospinal
fluid (CSF) bicarbonate
concentra-
tion. Experientia 22: 61 l-614. Siesjii, B. K. (1969). A new theory
for the regulation
of the extra-cellular
of pH regulation
in hypercapnia
pH in the brain.
&and. J. Clin.
Lab. Invest. 24: l-9. SiesjB, B. K. (1971). Quantification Invest. 28
Siesjo, B. K. (1972). The regulation Sorensen,
S. C., N. A. Lassen,
metabolism Sylvester,
and cerebral
Van Varenbergh,
Wischer,
J. Coudert
blood flow in high altitude
and M. P. Zamora
residents.
(1973). Ventilatory
(1974). Cerebral
glucose
J. Appl. Physiol. 37: 305-310. control
during
brief infusions
of CO,-
dog. J. Appl. Physiol. 35: 178-186.
P., G. Demeester
and I. Leusen
(1965). Lactate
in cerebrospinal
fluid during
hyper-
Arch. Int. Physiol. Biochim. 73: 738-747.
Vogl, B. P. and T. H. Maren brospinal
fluid pH. Kidney Int. 1: 360-374.
of cerebrospinal
J. W. Severinghaus,
J. T., B. J. Whipp and K. Wasserman
laden blood in the awake ventilation
Stand. J. Clin. Lab.
and hypocapnia.
: 113-I 16 (Editorial).
(1975). Sodium,
chloride
and bicarbonate
movement
from plasma
to cere-
fuid in cats. Am. J. Physiol. 228: 673-683.
J. and H. Kazemi
(1975). CSF bicarbonate
regulation
in respiratory
acidosis
and alkalosis.
J. Appl. Physiol. 38 : 504-511. Wise, W. C. (1973). Normal
arterial
blood
gases and chemical
components
in the unanesthetized
dog.
J. Appl. Physiol. 35 : 427-429. Woodbury,
J. W. (1971). Alfred pp. 275-289.
Benson
Symposium.
Ion Homeostasis
of the Brain,
3rd. Copenhagen,