Clinical Science (1970)38, 503-518.

THE HYPOXIC DRIVE TO BREATHING I N CHRONIC BRONCHITIS A N D EMPHYSEMA D. C. FLENLEY, D. H. F R A N K L I N

AND

J . S . MILLAR

Department of Medicine, University of Edinburgh (Received 3 October 1969)

SUMMARY

1. We measured the ventilatory response to CO, at two levels of arterial Po, in twelve patients who suffered from chronic obstructive bronchitis. We also determined the lumbar cerebrospinal fluid bicarbonate in ten of these patients. 2. The CO, response was depressed in nine patients who suffered from hypoxia and CO, retention when breathing air. The hypoxic drive to breathing was normal in six cases, increased in one and absent in two cases who had severe chronic hypoxia and secondary polycythaemia. 3. The slope of the acute on chronic whole body CO, titration line expressed in terms of arterial H+ and arterial Pco, was the same in the hypercapnic patients as in normal men. This relationship allows predictions of the duration of hypercapnia in clinical practice. 4. Increased buffering in cerebrospinal fluid does not account for the depressed ventilatory response to CO, in these patients. 5. By calculating the probable rise in jugular venous Pco, which will follow correction of chronic hypoxia in these patients, we conclude that the administration of oxygen will remove a peripheral chemoreceptor stimulus to breathing but increase the central stimulus by a rise in cerebrospinal fluid acidity. Patients suffering from chronic obstructive bronchitis (Medical Research Council, 1965) complicated by arterial hypoxaemia and carbon dioxide retention often develop a further rise in arterial carbon dioxide tension (Pco,) when their hypoxaemia is corrected (Donald, 1949; Comroe, Bahnson & Coates, 1950; Westlake, Simpson & Kaye, 1955). This phenonienon is usually attributed to the removal of a major ventilatory stimulus from hypoxia, whereas a rise in arterial Pco, has long been recognized as a relatively inefficient stimulus to ventilation in these patients (Scott, 1920). We have examined the effects of high and low levels of arterial oxygen tension (Po,) on the steady state ventilatory response to induced increases in arterial Correspondence: Dr D. C. Flenley, Department of Medicine, University of Edinburgh, Royal Inlirmary, Edinburgh 3, Scotland.

503

504

D. C. FIenley, D. H. Franklin and J. S. Millar

Pco, in twelve bronchitic patients, nine of whom had both hypoxaemia and CO, retention when breathing air. The chemical control of ventilation can be studied by relating the minute ventilation to the arterial Pco,, during CO, inhalation, whilst the arterial Po, is held constant, this linear relationship being termed the isoxic C02 response (Lloyd, Jukes & Cunningham, 1958). In most normal subjects hypoxia increases the slope of the isoxic CO, response line, so that the hypoxic drive can be expressed as the potentiation of the ventilatory response to CO,. Following acclimatization to the chronic hypoxia of altitude, normal lowlanders show an increase in hypoxic drive (Milledge & Lahiri, 1967), whereas it is considerably diminished in residents at high altitude and particularly in patients at altitude who suffer from chronic mountain sickness (Severinghaus, Bainton & Carcelen, 1966; Milledge & Lahiri, 1967; Lefrancois, Hautier & Pasquis, 1968). Permanent loss of the hypoxic ventilatory drive has now been attributed to chronic hypoxia within the first 2 years of life (Sorensen & Severinghaus, 1968). Do similar variations in hypoxic drive exist in persistent hypoxia resulting from chronic bronchitis ? The depressed CO, response of these bronchitic patients cannot be entirely explained by the increased buffering power of their blood (Flenley & Millar, 1967), nor by their increased rate of inspiratory work (Flenley & Millar, 1968). The central stimulation of ventilation by CO, is probably mediated by changes in the acidity of the cisternal CSF (Leusen, 1954; Mitchell et al., 1963), which will in turn depend upon both the bicarbonate concentration (HCO;) and the Pco, of the cisternal CSF. From our measurements of the HCO; in lumbar CSF we have estimated the probable changes in cisternal CSF hydrogen ion activity (H') which would have occurred during our CO, response studies. The Pco, of reduced whole blood rises when the blood is oxygenated, if the content of CO, in the blood is unaltered (Christiansen, Douglas & Haldane, 1914). We examined the role of this CDH effect on the relationship between Pco, and the C02 content of whole arterial blood in v i v a Part of the rise in Pco, following the administration of oxygen to these hypoxic patients may result from this CDH effect. By use of reported values for cerebral blood flow in similar patients (Patterson, Heyman & Duke, 1952), we have predicted the changes in Pco, and H f of cisternal CSF which may follow from the correction of their arterial hypoxaemia.

METHODS Patients The twelve patients complained of chronic productive cough and dyspnoea on effort from 5 to 20 years. They consented to the procedures after the nature and purpose had been fully explained, but two patients declined lumbar puncture following the CO, response study. Chronic obstructive bronchitis and emphysema were diagnosed from the clinical history, the physical and radiological examination and also from the reduced vital capacity, the increased residual volume and the persistent reduction in forced expiratory volume in 1 sec (FEV, .,). They had various degrees of hypoxia, hypercapnia, respiratory acidosis and secondary polycythaemia (Table 1). At the time of study cor pulmonale was controlled with maintenance doses of digoxin and frusemide in cases 4-12, and ventilatory failure (Campbell, 1965) was diagnosed in these cases from the blood gas tensions when they were breathing air (Table 1). The clinical condition and FEV, .o were unchanged for at least 1 week before the study in all

80.0 59.1 61.0 55.3 59-5 69.0 55.0

52.5 72.0 55-3 52.2 71.0

165

154

168

166

177

164

168

171

164

166

162

171

61

68

61

57

54

70

71

62

69

44

65

46

1

2

3

4

5

6

7

8

9

10

11

12

2.32 (3.64) 1.58 (2.89) 1.98 (3.80) 1.72 (3.75) 2.26 (4.39) 1.94 (3.39) 1.33 (3.58) 1-70 (3.93) 1.37 (3.41) 1.35 (4.73) 1.00 (3.39) 1.60 (4-28)

(1.)

VC

0.1 5.77 (582) 5.09 (4.86) 5.69 (6.08) 4.86 (5.90) 8.20 (6.81) 4.76 (5.73) 4.06 (6.08) 6.95 (6.34) 4.84 (5.46) 5.54 (5.90)

(5.56) 5.92 (6-34)

(1.1 3.45 (2.05) 3.51 (1.87) 3.71 (2.13) 3.14 (2.02) 5.94 (2.24) 2.82 (2-17) 2.73 (2.30) 5.25 (2.22) 3.47 (2.15) 4.19 (1.78)

(2.03) 4.32 (1.95)

4.06 (3.08) 3.64 (1.39) 4.40 (4.01) 3.50 (4.07) 6.54 (4.47) 3.87 (3.60) 3.15 (4.41) 5.91 (453) 3.57 (3.46) 4-59 (3.85) (4.07) 4.42 (3.54)

-

TLC

RV

FRC (1.)

(2.21) 0.60 (2.92) 0.65 (2.93) 0.60 (3.43) 0.55 (2.22) 0.70 (2.63) 0.49 (3.00) 0.65 (2.55) 0.21 (3-32) 0.60 (2.58) 0.71 (3.45)

0-85

1.20 (3.83)

FEVl.o (1.)

(87) 44 (93) 38 (88) 40 (93)

44

(91) 61 (87) 50 (87) 62 (89)

55

67 (89) 60 (87) 67 (89) 53 (90)

Po2 (mHg)

66

58

7.32

7.39

33.5

34.0

34.0

26.8

22.1

22.4

22.6 38.0 7-39 64

7-34

19.9 35.0 7.41 52

66

21.0 36.6 7.43

21.3 56

34.0

19.1 31.1 7.40

52

7.41

21.0 34.1 7.42

53

55

19.9

20.4

26-2

31.0

20.1

7.39

7.45

46

26.0

HCOj Cap@ (mEq/l) (ml/lOO ml)

44

7.36

pH

40

(mmHg)

Pco2

Arterial blood when breathing air

Abbreviations: VC, vital capacity; FRC, functional residual capacity; RV, residual volume; TLC, total lung capacity; FEVl.o, forced expiratory volume in 1 sec; Po2, oxygen tension; Pco2, carbon dioxide tension; HCO,, bicarbonate concentration; Cap 02, oxygen capacity. The predicted normal values of lung volumes (Cotes, 1965) and arterial Po2 (Conway, Payne & Tomkin, 1965) are in parentheses beneath the measured values.

Age Height Weight Orr) (cm) (kg)

Case

TABLE 1. Physical characteristics and results of pulmonary function tests in the twelve patients

z

u l

s-

506

D. C.Flenley, D. H. Franklin and J. S. Millar

cases. Subdivisions of lung volume were measured in the sitting position by helium dilution (Meneely 8z Kaltreider, 1941), and the FEV,., with a McDermott dry spirometer (Collins, McDermott & McDermott, 1964).

Procedure The humidified, warmed, inspired gas mixture was delivered to the inspiratory valve from a rotameter mixing device which allowed the concentrations of oxygen and carbon dioxide in the mixture to be changed rapidly. Expired gas was collected in a Tissot spirometer, the expiratory resistance being 0.5 cm water at 1 l/sec. Arterial blood was sampled from a small bore nylon catheter inserted into the brachial artery by a modified Seldinger technique (Berneus et al., 1954). An experienced nurse made sure that the seated patient was comfortable during the study, but did not allow him to fall asleep.

"

Ventilation

Arterial sample

I 0

ffl

w

ffl

El

ffl

ffl

w

I

2

3

4

5

6

7

I

I

I

I

I

I

I

15

30

45

60

75

90

105

0

I IX,

Time (rnin)

FIG.1. The inspired oxygen and COz tensions during a typical study (case 7) showing the time of sampling of arterial blood, and of ventilatory measurements.

Two isoxic CO, response lines were obtained in each subject, the arterial Po, onzthe low isoxic line being between 40 and 50 mmHg in three subjects, and between 50 and 60 mmHg in the rest. The arterial Po, on the high isoxic line was above 96 mmHg, in all subjects, but was usually between 150 and 200 mmHg (Table 2). In order to maintain a constant arterial Poz, the oxygen concentration in the inspired gas was lowered empirically as ventilation increased, for the changes in alveolar to arterial oxygen tension gradients during hyperventilation precluded accurate predictions in these patients. The inspired gas mixture was held constant for 15 min periods, the arterial Po, being measured after 5 min to ensure that isoxic conditions had been achieved. The expired gas was collected over the last 2 min of each period, at the same time as 15 ml of arterial blood were withdrawn (Fig. 1). In ten cases CSF was

Hypoxic drive in chronic bronchitis

507

sampled by lumbar puncture 4-5 hr after completing the CO, response study, anaerobic CSF and arterial blood samples being drawn simultaneously.

Analytical methods Arterial blood was sampled into siliconed syringes, with the dead space filled with 1000 units/ml heparin solution. The Po,, Pco, and pH of these samples were immediately measured in duplicate by each of two separate sets of electrodes (Radiometer Company). Both sets of blood gas electrodes were calibrated with the patient's blood, which was tonometered during the study with two gas mixtures of high and low oxygen and CO, concentrations (Flenley, 1967). Effluent gas from the tonometer was analysed by the Lloyd Haldane apparatus, duplicate analyses agreeing to 0.03%. The pH electrodes were calibrated with precision buffers between each blood sample. The total carbon dioxide content of anaerobically separated arterial plasma (C art. plasma co,) and of anaerobic CSF was determined on each sample by the manometric method (Van Slyke & Neill, 1924), duplicate analyses agreeing 0.2 mEq/l. These Pco, electrodes, when calibrated with tonometered blood, can measure blood Pco, with 95% confidence limits of k2.6 mmHg in this range (Flenley, Millar & Rees, 1967) reducing to k 1-3mmHg with two independent electrodes. The arterial Pco, was also calculated from the mean of the duplicate readings of pH by the two glass electrodes and the C art. plasma co,, by the Henderson-Hasselbalch equation, with a pK' of 6.10 and a solubility coefficient of 0.0306 (Severinghaus, 1966). The arterial Pco, reported is the mean of these three independent measurements. The arterial Po, reported is the mean of the two sets of duplicate measurements on the separate electrodes, and therefore has 95% confidence limits of k3.5 mmHg (Flenley et al., 1967). Blood oxygen capacity was measured spectrophotometrically (King & Wootton, 1956)with previous calibrations against the manometric method. RESULTS

Isoxic CO, response The minute ventilation (VE I/niin) and arterial Pco, (mmHg) or hydrogen ion activity (H', n-mole/l), for the isoxic lines at the two levels of Po, in each patient were fitted by least squares regression to equations:

where SCO,and SHare the slopes of the CO, response lines, and BCO,and BH the intercepts of the lines on the Pco, axis and H+ axis respectively (Table 2, Fig. 2). The low isoxic lines lie to the left of the high isoxic line (Fig. 2) in all cases but 10 and 12, so that a hypoxic ventilatory drive was demonstrated in all except these two cases. The slope of the low isoxic lines averaged 1.93 (SEM 0.53) 1 min-l mmHg-l, excluding case 12, who showed no significant correlation between either arterial Pco, or H', and VE.For the high isoxic lines, again excluding case 12, Sco, averaged 0.88 (SEM 0.17) 1 min-l mmI3g-l. In thirteen normal men Sco, was 1.74 (SEM 0.14) 1 min-l mmHg-', when the arterial Po, was over 100 mmHg (Cunningham et al., 1961; Flenley & Millar, 1967; Falchuk, Lamb & Tenney, 1966). In our cases the intercept on the Pco, axis (Bco,) averaged 45 (SEM 2.4) H

D. C. Flenley, D. H. Franklin and J. S. Millar

508

TABLE 2. Constants of the equations: h = Sco, (Pco, -Bco,) and VE = SH(H+ -BH), for arterial blood values Case

Arterial

Carbon dioxide

Hydrogen ion

Po, (mmHg)

sco, (1 min-I mmHg-')

1 2 3 4 5 6 7 8 9

10 11 12

49, 50, 55, 49 176, 189, 171, 112 48, 55, 56, 55 235,224, 212, 220 58, 56, 52 141, 142, 149, 133 42, 42, 42 142, 176, 161, 155 52, 52, 58, 55 127, 148, 152, 110 54, 55, 54, 58 157,201, 162, 139 52, 56, 57 140, 125, 106, 190 43, 40, 45, 47 197,203, 194, 222 54, 56, 60, 60 171, 172, 159 52, 51, 48 98, 157, 155,96 36, 36, 36, 45 187, 146,77 52, 57, 56, 48 151, 131,230

3.16 1.67 4.30 2.18 5.56 0.86 2.93 0.87 1.20 0.84 0.47 0.65 0.65 0.46 1.02 0.39 0.81 0.86 0.80

0.81 0.39 0.20 0.06 0.18

Bco, (mmHg) 34 30 40 41 43 46 51 47 47 50 30 54

44 54 46 46 56 62 58 59 42 40 51 32

r

SH (1' min-' n-mole-')

0.96 1-00 0.92 0.99 0.97 0.93 0.95 0.92 0.99 0.99 0.93 0.99 0.92 0*49* 0.99 0.96 0.98 0.97 0.96 0.92 0.93 0.81 0.70* 0*55*

4.36 2.37 4.17 2.34 8.10 1.30 3.58 1.43 1.90 1.42 0.98 1.32 1.50 2.24 1.15 0.51 1.38 2.10 1.96 1.27 0.70 0.26 0.09 0.74

BH

r

(n-mole I-') 32 29 29 29 41 36 39 32 36 37 27 39 31 39 30 32 36 42

44 43 30 22 25 40

0.97 0.95 0-86 0.89 0.70* 0.97 0*67* 0*74* 0.9 1 0.96 0.93 0.98 0.96 0.92 0*62* 0.66* 0.94 0-99 1s o 0 0.92 0.94 0*78* 0.75* 0.78*

The arterial Po, are the values for each point on the isoxic ventilation/Pco, line. r is the correlation coefficient. *r value not signilicant (Pz0.01).

mmHg for the low isoxic line and 47 (SEM 3.0) mmHg for the high isoxic line, excluding case 12. The relationship between VE and arterial H + was significant in nine cases (Table 2), at either one or both levels of Po,. The slope of these lines, SH of equation 2, averaged 2.12 (SEM 0.48) l2 min-' n-mole-' for the low isoxic line, and 1-79 (SEM 0.17) l2 min-' n-mole-' for the high lines. The intercept values, BH, averaged 33 (SEM 1.8) n-mole/l for the low lines, and 37 (SEM 1.8) n-mole/l for the high lines. Cerebrospinalfluid acid base balance Values of Pco,, pH and HCO; of lumbar CSF and arterial blood are shown in Table 3. When these samples were taken, between 4 and 5 hr after the C 0 2 response study, the patients

Hypoxic drive in chronic bronchitis

509

were breathing air. The Pco, and pH of this blood did not differ significantly from that obtained at the start of the C02 response study (Table 1) but mean plasma HCO; rose significantly by 0.6 mEq/l in this period. The effects of these changes on the composition of CSF during the CO, response study are discussed later. In cases 2 , 4 , 5, 6, 9, 10, 11 and 12, who 0

20

15

5t-

Arterial Po, 40-60

iPf

" 40 50 60 70 80 40 50 60 70 80 40 50 60 70 80 Arterial Pco, (mmHg1

FIG2. The ventilatory response to C02 in the twelve patients. Regressionlines at low arterial Po2 (0) and high arterial Po2 ( 0 ) are shown.

had a raised Pco, when breathing air, the CSF HCO; was above normal. The Pco, of both arterial blood (average 55, SEM 3.0 mmHg) and CSF (average 61, SEM 3.1 mmHg) showed a similar variance between the cases, but the pH of the CSF (average 7-32, SEM 0.001) varied less than that of the arterial blood (average 7.40, SEM 0.030). DISCUSSION Technical problems. Reliable studies of the isoxic steady state ventilatory response to CO, depend upon the achievement of a steady state of ventilation; the maintenance of a constant

D. C.Flenley, D. H. Franklin and J. S. Millar

510

arterial Po,, particularly on the low Po, isoxic line (Lloyd et al., 1958; Cunningham et al., 1961); and accurate measurements of arterial Pco,. In our study the expired volume collected from the 7th to 9th min did not differ significantly (P>O.Ol) from that collected over the 13th to the 15th min in any of the 15 min 'steady state' periods. Thus a constant ventilation was obtained (cf. Bernards,Dejours &Lacaisse, 1966).Themeanvariation inarterialPo, was f 3 mmHg TABLE 3. CSF and arterial blood acid-base values at the time of lumbar puncture Arterial blood

Lumbar cerebrospinal fluid

Case

Pcoz

pH

(mmHg) 1 2 3 4 5 6 9 10 11 12 Normal SEM

38 46 43 60 52 52 65 61 57 68 38.6 k0.5

7.46 7.45 7.40 7.38 7.40 7.43 7.40 7.36 7.39 7-32 7.39 kO.01

HCO, (mEq/l) 26.0 31.2 26.1 35.0 31.0 35.0 39.0 34.0 34.0 34.4 22.9 k2.3

Pco, (mmHg) 44 63 50 64 57 56 70 72 59 74 47.9 k5.7

pH 7.36 7.32 7.32 7.30 7.33 7.38 7.29 7-28 7.36 7.29 7.31 k0.03

HCOi (mEq/l) 25.0 29.4 25.8 30.6 28.4 30.3 31-9 31.7 30.6 33.6 23.4 +2,4

SH

;(I, min-' n-mole-') 2.0 2.5 1.o 1*4 1.1 1*o 1.3 1.0 0.3 0.3 1.7-2'0

-

on the low isoxic line, which is acceptable by the criteria of Falchuk, Lamb & Tenney (1966), and indeed is barely outside the probable error of measurement of the arterial Po, (Flenley et al., 1967). In normal subjects, with an arterial Po, between 40 and 50 mmHg, Sco, could vary by 237% to -22% if the arterial Pco, measurement varied by f 1 mmHg. By the use of two independent CO, electrodes, both calibrated with tonometered blood, and also independent pH and HCO; measurements, we believe that our Pco, values are accurate to within f 1 mmHg. The hypoxic ventilatory drive. All the nine cases of ventilatory failure show a depressed ventilatory response to a rise in arterial Pco, without hypoxia, but in ten of the twelve cases ventilation at any Pco, was greater when the arterial Po, was between 40 and 60 mmHg than when it was over 100 mmHg. In cases 1,2,4,5,8 and 11 the isoxic lines tend to converge on a common point on the Pco, axis, so that the mean difference between the BCO,for the high and low isoxic lines is not significant in these cases. Thus the ratio of the slopes of the two lines is an adequate description of the hypoxic stimulus. In cases 1, 2 and 5 this ratio of the slope of the low isoxic line to that of the high line averages 1.76 (SEM 0.17), which does not differ significantly from the ratio of slopes of lines at the same level of Po, in three normal subjects (Falchuk et al., 1966). In cases 4, 8 and 11 where the arterial Po, on the low line was between )~ again does not differ 40 and 50 mmHg, the ratio of slopes averaged 2-50 (SEM 0 ~ 2 9 which significantly from that in five normal subjects for this level of Po, (Cunningham et al., 1961;

Hypoxic drive in chronic bronchitis

51 1

Falchuk et al., 1966). In these six cases, therefore, we can show no quantitative difference in the hypoxic drive to ventilation from that in these normal subjects. In cases 6, 7 and 9, however, hypoxia appears mainly as an additive stimulus, increasing ventilation by a constant amount at any level of Pco,. This 'shift to the left' in the CO, response line was not due to a metabolic acidosis in hypoxia in these patients, for the arterial blood was less acidic during hypoxia (Table 2). The ratio of slopes is 6.46 in case 3, so that his hypoxic drive is greater than normal (Falchuk et al., 1966). He appeared on clinical grounds to suffer from chronic infective asthma (Bates & Christie, 1964), without having sustained any prolonged period of severe hypoxia. Case 10 showed a depressed ventilatory response to CO, with an absent hypoxic drive. This patient was the youngest studied (44 years), and when breathing air was known to have an arterial Po, averaging 56 mmHg and Pco, 53 mmHg over the previous 4 years. Case 12, in whom we also failed to demonstrate a hypoxic drive, had in addition no ventilatory response to CO,. Polycythaemia was present in him 10 years before this study, and his hypoxaemia and hypercapnia (Table 1) had persisted for at least the preceding 4 years. These two patients resemble cases of chronic mountain polycythaemia (Severinghaus et al., 1966), in that they suffered from chronic hypoxaemia with polycythaemia, and impairment of the hypoxic drive to ventilation. It is possible, but unproven, that in these two cases the peripheral arterial chemoreceptors had been deficient for many years, or even from birth, and that a combination of this deficiency with subsequent chronic obstructive bronchitis had resulted in an unusual degree of hypoxaemia at an early age. The whole body COZ titration line. Prolonged elevation of arterial Pco, leads to increased reabsorption of bicarbonate in the renal tubules (Relman, Etsten & Schwartz, 1953), so that the arterial H + is only slightly raised. This mechanism operates maximally after 5-7 days of hypercapnia so that the initial values of arterial H + and Pco, in our patients would reflect this renal compensation, unlike the arterial H + during the 15 min periods of acute CO, inhalation (Schwartz, Brackett & Cohen, 1965). The equation H + = 30.6+0.19 Pco, describes the linear regression of arterial H + on the arterial Pco, for the samples taken at the start of the study when our patients were breathing air. We believe their chronic hypercapnia was then fully compensated. This equation is similar to that derived by Engel et al. (1968) to describe their results in twenty-eight steady state situations in thirteen patients who were not receiving diuretics. The acute on chronic whole body CO, titration line of each patient is shown in Fig. 3 in terms of arterial H + and Pco,. The slope of these lines was not significantly different between the cases, and averaged 0.58 (SEM 0.04) n-molel-' mmHg-l. This value differs significantly from the slope of the chronic whole body C02 titration line (Fig. 3), but this average slope is the same as that of the acute whole body CO, titration curve in normal young men (Brackett, Cohen & Schwartz, 1966). The similarity of slope of these acute on chronic whole body CO, titration lines permits one measurement of arterial H + and Pco, to describe the nature of hypercapnia. Measurements are plotted on an H+/Pco, diagram (Fig. 4) which also includes the acute (Brackett et al., 1965) and chronic (Engel et al., 1968) whole body CO, titration lines and their significance bands (Cohen & Schwartz, 1966). Values at A indicate acute hypercapnia, superimposed upon a previous chronic hypercapnia (B). An acute rise or fall in Pco, in such a case will lie along the line AB, which is parallel to the acute CO, titration line. However, if this patient

512

D. C. Flenley, D . H . Franklin and J. S. Millar

has sustained hypercapnia for 5-1 days, with normal renal function the values will move to point C, on the chronic CO, titration line. Point D indicates acute hypercapnia alone, which, if maintained for 5-7 days, will produce values as at F. Within the overlap of the two significance bands prediction is not possible.

1

1

l

l

l

1

1

1

1

1

1

l

The whole body CO, titration curve of whole blood. Reduced haemoglobin combines reversibly with CO,, forming carbamino compounds (Roughton, 1964), this reaction being the basis of the Christiansen-Douglas-Haldane effect (Christiansen et al., 1914), whereby the PCO, of a reduced whole blood sample rises on oxygenation. The extent of this rise depends upon the change in oxygen saturation (So,), on pH, and on the blood oxygen capacity, as shown by the equation: C blood co2 C plasma co,

= I -

0.0215 (oxygen capacity) (2.244 -0.00422 SO,) (8.74 -PH)

where C plasma co2 is the plasma CO, content (ml/lOO ml), the oxygen capacity is in ml/ 100 ml, and So, is a percentage (Van Slyke & Sendroy, 1928; Visser, 1960). We calculated C blood CO, from measurements of C plasma co,, and related this arterial C blood co2 to the measured Pco, at the two levels of So, studied in each patient (Fig. 5). These regression lines are linesr estimates of segments of the whole body CO, titration curves but as only three or four points were available for each line, correlation was only significant at the 5% level in eight c.ases when hypoxic, and in five cases with So2 over 98%. The slopes of the lines for

Hypoxic drive in chronic bronchitis

513

partially reduced blood (So, 70.2-90.1 %) averaged 0.37 (SEM 0.32) ml 100 ml-l mmHg-l, which is not significantly different from the average slope when the blood was oxygenated (0.35 (SEM 0-06)ml100 ml-I mmHg-'). Ifthe CO, content of whole blood is unchanged, the rise in Pco, following oxygenation of partially reduced whole blood will be greater if the slope of the whole blood CO, content/Pco, line is reduced (Fig. 5). In our bronchitic patients these acute on chronic CO, titration lines are substantially flatter than the chronic whole body CO, titration lines (Fig. 3, Fig. 5). Indeed, as shown above, the patients' titration lines are parallel to the acute titration lines of normal men. The implications of the CDH effect in ventilatory control of the chronic bronchitic are discussed later in terms of the changes in CSF acidity following the relief of hypoxia in these patients.

30

I

I

I

I

I

40

50

60

70

'

I

1

I

80

90

100

Arteriol Pco, (mmHg)

FIG.4. The nature of hypercapnia. The acute whole body CO,titration line with 95% significance bands (stippled) (Brackett et al., 1965; Cohen & Schwartz, 1966) and the chronic whole body CO, titration line with significance bands (cross-hatched) (Engel et al., 1968). Points A to F are discussed in the text.

The buffering power of cerebrospinalfluid. If the central stimulation of ventilation by Cot arises from changes in the H + of the cisternal CSF (Leusen, 1954; Mitchell et al., 1963), this will, in turn, depend upon the Pco, and HCO; of the CSF. We propose that the HCO; of lumbar CSF taken 4 hr after the CO, response study is similar to that of cisternal CSF during that study. It is not feasible to prove this by direct sampling of cisternal CSF in unanaesthetized man during CO, inhalation. Therefore we base our proposal upon the following evidence: (1) cisternal and lumbar CSF have the same HCO; in man (Van Heijst, Maes & Visser, 1966); (2) the arterial Pco, in our patients before the CO, response study was the same

D. C. Flenley, D. H. Franklin and J. S. Millar

514

as that at the time of lumbar puncture (Table 1 and Table 3); (3) the small but significant rise in average arterial HCO; in these samples was only 0.6 mEq/l, which would be expected to increase CSF HCO; by 0.3 mEq/l (Alroy & Flenley, 1967), after 50 days (Bradley, Spencer & Semple, 1965), so that important changes within 8 hr appear unlikely; (4) although there is thought to be active transport of bicarbonate across the blood-brain barrier so as to stabilize CSF H + in face of changes in arterial Pco, (Mitchell et al., 1965), such transport is slow. In the dog cisternal CSF HCO; rose after 30 min of 7% CO, inhalation, but not after 12 min (Michel, 1964). If CSF HCO; had risen significantly during our 2 hr long studies, the increase 70 -

65

-E 2

-

0

0

>E

N

8

60-

-

-0

0

a

i

55-

0

c,

50 -

45'

x,

I

I

40

I

I

50

I

I

60

I

I

70

t

I 80

Arterial Pco, (mrnHg)

FIG.5. The COz titration lines, in terms of Pcoz and COz content of whole blood (C art. blood coz) at an oxygen saturation of 86% (solid lines) and 100% (dotted lines). The acute in vivo lines are calculated from the data of Brackett et ul. (1965), the in vitro lines from data assembled by Root (1958), and the chronic in vivo lines from the data of Engel ef al. (1968). The heavy lines are calculated from our data in cases 1, 4,5 and 12.

in CSF buffering would result in a progressive decline in acidic ventilatory drive, as the highest arterial Pco, levels occurred after 1 hr (Fig. 1). Thus the ventilatory response would tend to show a convexity towards the Pco, axis at high Pco, levels. No such effect was seen in our results (Fig. 2) in our previous studies of bronchitic patients (Flenley & Millar, 1967), or in normal subjects (Falchuk et al., 1966). We conclude that the CSF HCO; did not rise significantly during our ventilatory response studies. By equating cisternal and lumbar CSF HCO; we have calculated the probable changes of cisternal H+ which occurred in our patients during CO, inhalation under normoxic conditions. The change in CSF Pco, is assumed to be nine-tenths of the change in arterial Pco, (Bradley

Hypoxic drive in chronic bronchitis

515

et al., 1965), and the change in cisternal H' is then obtained from the Henderson-Hasselbalch equation for CSF (Severinghaus, 1966; Mitchell et al., 1965). The values of SH,CSF in the regression equations

VE = SH,CSF (H',

CSF -BH,CSF)

which are so calculated are shown in Table 3 for the high isoxic line only. The values of SH,CSF given for the normal subjects are based on an Sco, for the arterial blood (Table 2) of 1-5-2-0 1 min-' mmHg-l for this high isoxic line (Cunningham et al., 1961; Flenley & Miller, 1967; Falchuk et al., 1966), and also on a CSF HCO-j of 23.4 mEq/l (Posner, Swanson & Plum, 1964). SH,CSF is normal in cases 1 and 2, who did not have CO, retention. In the other cases, SH,CSF is below normal, so that their depressed ventilatory response to CO, is not due to increased buffering in the CSF. Oxygenation and cerebrospinalfluid acidity. Jugular venous blood has the same Pco, as the cisternal CSF in man (Bradley & Semple, 1962). What will be the effect on the jugular venous Pco,, and thence the cisternal Pco, and H' which would be predicted as a result of correcting arterial hypoxaemia in our patients? This can be calculate'd from the carbon dioxide content of arterial blood, at the two levels of oxygenation, if we know the differences in oxygen and CO, content across the cerebral circulation. We did not measure cerebral blood flow in our patients, but they appear to be clinically similar to the twelve patients studied by Patterson et al. (1952), who had an average arterial Po, of 39 mmHg, Pco, of 58 mmHg, and an arteriovenous oxygen difference of 4.6 m1/100 ml across the cerebral circulation when breathing air. On breathing oxygen the average arterial Po, rose to 80 mmHg, the Pco, to 70 mmHg, and the cerebral arterio-venous oxygen difference fell to 4.4 m l / l O O ml. From these figures, and our calculations of the CDH effect in vivo, by assuming a brain respiratory quotient of 1 we have calculated the jugular venous CO, content, and thence the jugular venous Pco, and pH (Table 4) from the Sigaard-Anderson curve nomogram, on the basis that the CO, dissociation curve across the brain has the same slope as the in vitro CO, dissociation curve (Michel, 1968). From the calculated values of jugular venous Pco, at the two levels of oxygenation, we have calculated the change in cisternal CSF H', from the lumbar CSF HCO; (Table 3) (Severinghaus, 1966). If such a change in CSF H' had occurred under normoxic conditions, ventilation would have increased by the amount predicted in Table 3, which is based upon the values of SH,CSF for normoxic conditions. However, these predictions exceed the observed changes by over 700% in cases 5, 6 , 9 and 10, so that in these cases at least the hypoxic drive is considerably more potent than one might predict from the isoxic CO, responses based upon arterial blood values. Relief of hypoxia reduces the peripheral chemoreceptor drive but increases the central drive by increasing CSF acidity. The fact that ventilation is often unchanged or even falls when hypoxia is corrected indicates the great power of the hypoxic drive. These calculations cannot substitute for direct measurements of the changes in cisternal CSF acidity which actually occur on relief of hypoxia, but as such measurements cannot be made in unanaesthetized patients, it appears reasonable to employ this arithmetically assisted speculation in order to assess the probable importance of these factors. A similar rise in the Pco, of jugular venous blood samples was noted when normal subjects were given 100% oxygen to breathe (Lambertsen et al., 1953). The effect is increased when the changes in saturation following oxygen administration are greater, as in chronic hypoxia.

D. C. Flenley, D. H . Franklin and J. S. Millar

516

TABLE 4. Predicted changes in cisternal CSF H + and ventilation on correction of hypoxia Arterial blood

Jugular venous blood (calculated)

Case

Po2 (-g) 1 2 3 4 5 6 9 10 11 12

PCOZ

pH

(mmHg) 112 49 216 48 142 58 142 42 108 52 139 54 159 55 94 43 177 36 131 46

37 38

44 42 47

44 56 55 60 54 67 49 70 65 69 66 68 58 65 64

Pco2

pH

Cisternal CSF AH+ (n-mole/l)

(-g)

7.46 7-45 7.48 7.50 7.36 7.40 7.42 7.43 7.36 7.39 7.35 7.43 7.35 7.39 7-31 7.33 7.35 7.41 7.34 7-32

44 45 53 51 52 51 53 53 74 64 74 61 90 84 86 79

7,43 7.42 7.45 7.45 7.35 7.38 7-45 7.45 7.30 7.35 7.33 7.37 7.28 7.32 7.26 7.29

78 70 80 77

7.32 7.37 7-29 7.29

Ventilation A VE Predicted Observed (I/min) (l/min)

- 1

-

2.0

-0.4

+ 1

+ 2.5

-0.8

+ l

+

1.0

+5.1

0

-7.3

0

+ 10 + 9 + 4 + 5 + 5

+ 3

+ll.O

+ 9.0 + 5.2 + 5.0 + 1.5 + 1-5

+ 1.2

+ 1-3 +0*5

+0.2 - 1.4 - 1.8

Cisternal CSF AH+ is the calculated increase in cisternal CSF H+ when the arterial Po2 rises from the low to the high value in each case. AVE is the predicted or observed increase in minute volume for this rise in Poz.

In conclusion, it would seem possible that the importance of the hypoxic ventilatory stimulus in chronic hypoxia is underestimated by consideration of the ventilatory response to arterial Pco, alone, for a rise in arterial Po, will also probably induce an increase in the acidity of the cisternal CSF. The size of these changes will depend upon the validity of the assumptions presented above, and in particular upon any concomitant changes in their cerebral blood flow. ACKNOWLEDGMENTS

We gratefully acknowledge the assistance of Dr G. G. Alroy. This work was supported by the Medical Research Council, the British Heart Foundation, and the John Risk Bequest. REFERENCES

ALROY, G.G. & FLENLEY, D.C. (1967) The acidity of the cerebrospinal fluid in man with particular reference to chronic ventilatory failure. Clinical Science, 33, 335-343. BATES,D.V. & CHRISTIE,R.V. (1964) Respiratory Function in Disease. Saunders, Philadelphia.

Hypoxic drive in chronic bronchitis

517

BERNARDS, J.A., DEJOURS,P. & LACAISSE, A. (1966) Ventilatory effects in man of breathing successively CO, free, COz rich and COz free gas mixtures with low, normal or high oxygen concentration. Respiratory Physiology, 1, 390-397. BERNEUS, B., CAWTEN,A., HOLMGREN, A. & SELDINGER, S.I. (1954) Percutaneous catheterisation of peripheral arteries as a method for blood sampling. Scandinavian Journal of Clinical and Laboratory Investigation, 6, 217-221. W.B. (1965) Carbon dioxide titration curve of normal man. BRACKETT, N.A., JR, COHEN, J.J. & SCHWARTZ, New England Journal of Medicine, 212, 6. BRADLEY, R.D. & SEMPLE,S.J.G. (1962) A comparison of certain acid base characteristics of arterial blood, jugular venous blood and cerebrospinal fluid in man, and the effect on them of some acute and chronic acid-base disturbances. Journal of Physiology, 160,381-391. BRADLEY, R.D., SPENCER, G.T. & SEMPLE, S.J.G. (1965) Rate of change of CSF Pco2 and (HCO;) in acidbase disturbances in man. In: CSF and Regulation of Respiration (Ed. by Brooks, Kao and Lloyd). Blackwell Scientific Publications, Oxford. CAMPBELL, E.J.M. (1965) Respiratory failure. British Medical Journal, i, 1451-1460. CHRISTIANSEN, J., DOUGLAS, C.C. & HALDANE, J.S. (1914) The absorption and dissociation of carbon dioxide by human blood. Journal of Physiology, 48,244-271. W.B. (1966) Evaluation of acid-base equilibrium in pulmonary insufficiency. COHEN,J.J. & SCHWARTZ, American Journal of Medicine, 41, 163-167. COLLINS, M.M., MCDERMOIT, M. & MCDERMOTT, T.J. (1964) Bellows spirometer and transistor timer for the measurement of forced expiratory volume and vital capacity. Journal of Physiology, 172,39~-41~. COMROE, J.H., BAHNSON, E.R. & ~OATES, E.O. (1950) Mental changes occurring in chronically anoxaemic patients during oxygen therapy. Journal of the American Medical Association, 143, 1044-1048. P.J. (1965) Arterial oxygen tensions of patients waiting surgery. CONWAY, C.M., PAYNE,J.M. & TOMKIN, British Journal of Anaesthesia, 37, 405408. COTES, J.E. (1965) Lung Function: Assessment and Application in Medicine. Blackwell Scientific Publications, Oxford. CUNNINGHAM, D.C.J., SHAW, D.G., LAHIRI,S . & LLOYD,B.B. (1961) The effect of maintained ammonium chloride acidosis on the relation between pulmonary ventilation and alveolar oxygen and carbon dioxide in man. Quarterly Journal of Experimental Physiology, 46,323-334. DONALD, K.W. (1949) Neurological effects of oxygen. Lancet, ii, 1056. ENGEL, K., DELL,R.B., RAHILL, W.J., DENNING, C.R. &WINTERS, R.W. (1968) Quantitative displacement of acid-base equilibrium in chronic respiratory acidosis. Journal of Applied Physiology, 24,288-295. S.M. (1966) Ventilatory response to hypoxia and COz following FALCHUK, K.H., LAMB,T.W. & TENNEY, COz exposure and NaHCO; ingestion. Journal of Applied Physiology, 21,393-398. FLENLEY, D.C. (1967) A rotating disc tonometer. Respiratory Physiology, 3,256-260. FLENLEY, D.C. & MILLAR, J.S. (1967) Ventilatory response to oxygen and carbon dioxide in chronic respiratory failure. Clinical Science, 33, 319-334. FLENLEY, D.C. & MILLAR,J.S. (1968) The effects of carbon dioxide inhalation on the inspiratory work of breathing in chronic ventilatory failure. Clinical Science, 34, 385-395. FLENLEY, D.C., MILLAR,J.S. & REES,H.A. (1967) The accuracy of oxygen and carbon dioxide electrodes. British Medical Journal, ii, 349-352. KING,E.J. & WOOTTON, I.D.P. (1956) Micro-analysis in Medical Biochemistry, 3rd edn. Churchill, London. LAMBERTSEN, C.J, KOUGH,R.H., COOPER,D.Y., EMMEL,G.L., LOESCHCICE,H.H. & SCHMIDT,C.F. (1953) Oxygen toxicity: Effects in man of oxygen inhalation at 1 and 3.5 atmospheres upon blood gas transport, cerebral circulation and cerebral metabolism. Journal of Applied Physiology, 5, 471486. LEFRANCOIS, R., HAUTIER,H. & PASQUIS, P. (1968) Ventilatory oxygen drive in acute and chronic hypoxia. Respiratory Physiology, 4, 217-228. I.R.(1954) Chemosensitivity of the respiratory centre. Influence of changes in the H+ and total buffer LEUSEN, concentration in the cerebral ventricles on respiration. American Journal of Physiology, 176, 45-51. LLOYD, B.B., JUICES, M.G.M. & CUNNINGHAM, D.J.C. (1958) The relation between oxygen pressure and the respiratory response to carbon dioxide in man. Quarterly Journal of Experimental Physiology, 43,214-227. COUNCIL(1965) Delinition and classification of chronic bronchitis, for clinical and MEDICALRESEARCH epidemiologicalpurposes. Lancet, i, 775-779.

518

D. C. Flenley, D. H. Franklin and J. S. Millar

MENEELY, G.R. & KALTREIDER, N.L. (1941) Use of helium for determination of pulmonary capacity. Proceedings of the Society for Experimental Biology and Medicine, 46,266-269. MICHEL,C.C. (1964) CSF (HCO;) during respiratory acid-base disturbance. Journal of Physiology, 170, 6 6 ~ 67~. MICHEL,C.C. (1968) The buffering behaviour of blood during hypoxaemia and respiratory exchange: Theory. Respiratory Physiology, 4, 283-291. MILLEDGE, J.S. & LAHIRI,S. (1967) Respiratory control in lowlanders and Sherpa highlanders at altitude. Respiratory Physiology, 2, 310-322. MITCHELL,R.A., CARMAN, C.T., SEVERINGHAUS, J.W., RICHARDSON, B.W., SINGER,M.M. & SCHNIDER, S. (1965) Stability of cerebrospinal fluid pH in chronic acid-base disturbances in blood. Journal of Applied Physiology, 20, 443-452. MITCHELL, R.A., HERBERT, D.A. & CARMAN,C.T. (1965) Acid-base constants and temperature coefficients for cerebrospinal fluid. Journal of Applied Physiology, 20,27-30. MITCHELL, R.A., LOESCHCKE, H.H., MASSION,W.H. & SEVERINGHAUS, J.W. (1963) Respiratory responses mediated through superficial chemosensitive areas on the medulla. Journal of Applied Physiology, 18,525533. PATTERSON, J.L., HEYMAN, A. &. DUKE, T.W. (1952) Cerebral circulation and metabolism in chronic pulmonary emphysema. American Journal of Medicine, 12, 331-348. POSNER,J.B., SWANSON, A.G. & PLUM,F. (1965) Acid-base balance in cerebrospinal fluid. Archives of Neurology, 12,479-496. RELMAN, A.S., ETSTEN,B. & SCHWARTZ, W.B. (1953) The regulation of ienal bicarbonate reabsorption by plasma carbon dioxide tension. Journal of Clinical Investigation, 32, 972-978. ROOT,R.W. (1958) Blood CO, absorption as a function of CO, pressure: Man. In: Handbook of Respiration (Ed. by D. S. Dittmer and R. M. Grebe). Saunders, Philadelphia. ROUGHTON, F.J.W. (1964) Transport of oxygen and carbon dioxide. Handbook of Physiology, Section 3. Respiration, Volume 1 (Ed. by W. 0. Fenn and H. Rahn), chap. 31, pp. 767-826. American Physiological Society, Washington. J.J. (1965) The response of extracellular hydrogen ion concentraSCHWARTZ, W.B., BRACKETT, N.A. & COHEN, tion to graded degrees of chronic hypercapnia. Journal of Clinical Investigation, 44,291-301. SCOTT,R.W. (1920) Observations on the pathologic physiology of chronic pulmonary emphysema. Archives of Internal Medicine, 26, 544-560. SEVERINGHAUS, J.W. (1966) Blood gas calculator. Journal of Applied Physiology, 21, 1108-1 116. SEVERINGHAUS, J.W., BANON, C.R. & CARCELEN, A. (1966) Respiratory insensitivity to hypoxia in chronically hypoxic man. Respiratory Physiology, 1, 308-334. SORENSEN, S.C. & SEVERINGHAUS, J.W. (1968) Respiratory insensitivity to acute hypoxia persisting after correction of tetralogy of Fallot. Journal of Applied Physiology, 25, 221-223. VANHEIJST,A.N.P., MAES,A.H.J. & VISSER,B.F. (1966) Comparison of the acid-base balance in cisternal and lumbar cerebrospinal fluid. Pfliigers Archiv fur die gesampte Physiologie, 287, 242-246. VANSLYKE,D.D. & NEILL,J.M. (1924) The determination of gases in blood and other solutions by vacuum extraction and manometric measurement. Journal of Biological Chemistry, 61, 523-573. VANSLYKE,D.D. & SENDROY, J. (1928) Studies of gas and electrolyte equilibria in blood. XV. Line charts for graphic calculations by the Henderson-Hasselbalch equations and for calculating plasma carbon dioxide content from whole blood content. Journal of Biological Chemistry, 79, 781-798. VISSER,B.F. (1960) Pulmonary diffusion of carbon dioxide. Physics in Medicine and Biology, 5,155-166. WESTLAKE, E.K., SIMPSON,T. & KAYE, M. (1955) Carbon dioxide narcosis in emphysema. Quarterly Journal of Medicine, 24, 155-173.