Clinical Science and Molecular Medicine (1977) 53, 387-395.
Effect of 30% oxygen on local matching of perfusion
and ventilation in chronic airways obstruction
NOEMI M. EISER, HAZEL A. JONES
AND
J. M. B. HUGHES
Department of Medicine, Royal Postgraduate Medical School,
Hammersmith Hospital, London
(Received 10 February 1977; accepted 15 June 1977)
s-m
1. Sixteen patients with chronic bronchitis
and airways obstruction were given radioactive
nitrogen (13N) by intravenous injection and
by inhalation, while breathing air and after
10-20 min breathing 30% oxygen. The clearance of 13Nfrom four zones of each patient's
whole lung field was monitored.
2. The I3N clearance of each region in these
patients with chronic bronchitis was much
slower than in normal subjects. Oxygen
breathing produced a significant delay in the
clearance of intravenously administered 3N in
23 zones in 10 patients but no systematic change
in clearance after inhaled 13N.
3. With inhalation of 30% oxygen there was
no significant change in the mean minute
ventilation, tidal volume or arterial Pcoz.
4. The results suggest that local hypoxic
vasoconstriction is present in some patients on
breathing air and that this is relieved by 30%
oxygen, resulting in a diversion of local blood
flow from well-ventilated to more poorly
ventilated areas. The fall. in PA/^ on 30%
oxygen is insufficient to increase arterial Pcoz.
Key words : gas exchange, hypoxic vasoconstriction, nitrogen ( l 3N),lung clearance curves,
regional ventilation and perfusion.
Abbreviations: FEVl.o, forced expiratory volume in 1 s; FRC, functional residual capacity;
0, pulmonary blood flow; Tg0,time for radioCorrespondence: Dr J. M. B. Hughes, Department
of Medicine, Royal Postgraduate Medical School,
Hammersmith Hospital, Ducane Road, London WIZ
OHS.
activity to fall by 90% from peak value; TLC,
total lung capacity; PA,alveolar ventilation; PE
minute ventilation; Veso,total volume expired
in T g 0 ;VT,tidal volume.
Introduction
From changes of minute ventilation and arterial
blood gas tensions, Pain, Read & Read (1965)
argued that breathing 100% oxygen must have
increased the maldistribution of ventilation
and perfusion in patients with chronic bronchitis and emphysema. Since breathing oxygen
is known to reduce pulmonary vascular resistance in these patients (Holt & Branscomb,
1965), they suggested that oxygen inhibits local
hypoxic vasoconstriction so that a greater
proportion of the cardiac output perfuses the
poorly ventilated parts of the lung that were
formerly hypoxic. There is, as yet, no direct
evidence to support this hypothesis.
An increase in oxygen concentration could
affect either the distribution of ventilation or of
perfusion, or of both. Inspiring 30% oxygen
decreased airway resistance in patients with
chronic bronchitis (Astin, 1970), but breathing
oxygen produced no consistent changes in
inert gas wash-out in such patients (Cumming
& Jones, 1967). In fact, there are stronger
arguments in favour of an oxygen-dependent
redistribution of local blood flow (von Euler
& Liljestrand, 1946; Barer, Howard & Shaw,
1970; Grant, Davis, Jones & Hughes, 1976).
We have devised a method, using the clearance of radioactive nitrogen (13N), to distinguish between a redistribution of local ventilation and of local blood flow caused by oxygen
387
N. M. Eiser, H.A . Jones and J. M. B. Hughes
388
I
I3N
I
Air breothing
Intravenous I3N
I
1
Oxygen breathing
FIG. 1 . Diagram of radioactivity counting fields (interrupted circles) containing two ventilatory units and
their blood supply. On the left (air breathing) one unit
is poorly ventilated because of airways obstruction,
and poorly perfused because of hypoxic vasoconstriction. During oxygen breathing (on the right) perfusion
LO the poorly ventilated unit increases when hypoxic
vasoconstriction is relieved. The stippling shows the
distribution of blood flow after intravenous injection
of ISN. The subsequent clearance of I3N from the
radioactivity counting field by ventilation will be less
efficient during O2 breathing.
breathing. By comparing the regional clearance
of inhaled 13N gas with the clearance of 13N
dissolved in sodium chloride solution and
injected intravenously, an index of the matching
of local ventilation to perfusion within each
radioactivity counting zone was obtained.
Patients with chronic bronchitis and emphysema
were studied breathing air and breathing
oxygen. The 30% oxygen was chosen because
it is the highest inspired oxygen concentration
generally used in the routine management of
patients with hypercapnic respiratory failure.
Theory and model
Fig. 1 is a schematic diagram of two ventilatory
units within a radioactivity counting zone.
When a solution of a very insoluble gas such as
I3N is injected intravenously, 99% is rapidly
evolved into the alveolar gas phase as it passes
through the lung. The regional distribution of
radioactivity will thus be proportional to the
local blood flow. During tidal breathing the
rate of clearance of the isotope from alveolar
gas reflects mainly the ventilation of the betterperfused alveoli. If bronchial or bronchiolar
narrowing hinders ventilation of some units
(Fig. l), local alveolar oxygen tension will fall
and local blood flow will be reduced because
of hypoxic vasoconstriction. As a result, a
greater proportion of the local perfusion will
be received by the better-ventilated unit in the
radioactivity counting field (Fig. I), and the
efficiencyof clearance of 13Nwill increase. Thus
hypoxic vasoconstriction can be regarded as
a ‘compensatory mechanism’, which improves
local gas exchange (Hughes, 1975). Breathing
an oxygen-enriched gas mixture may raise the
oxygen tension in the poorly ventilated unit
above the threshold for vasoconstriction and
increase local blood flow (Fig. 1). When inhomogeneity of ventilation within the counting
zone exists on breathing air, the redistribution
of blood flow when breathing oxygen will
deliver a larger proportion of the injected 13N
to more poorly ventilated units, so that the
overall clearance of the isotope from the zone
will be less efficient than when breathing air.
A second measurement must be made to
assess any effect of oxygen on the distribution
of ventilation itself. If I3N is inhaled continuously until equilibrium is achieved, all
alveoli will be equally labelled. The subsequent
clearance reflects the distribution and efficiency
of local ventilation only, and is independent of
blood flow. By relating the local clearance of
intravenous I3N (ventilation- and perfusiondependent) to inhaled 3N (ventilation-dependent) the efficiency with which local blood flow is
matched to local ventilation can be calculated.
Materials and methods
Patients and procedure
Details of the 16 patients studied have been
deposited as Clinical Science and Molecular
Medicine Table no. 77/12 with the Librarian,
the Royal Society of Medicine (1 Wimpole
Street, London W 1M SAE), from whom copies
may be obtained on request. All had chronic
bronchitis and airways obstruction unaffected
by bronchodilators. All subjects were outpatients and none was acutely ill at the time of
study. Some (nos. 1-7) had a high total lung
capacity (TLC) and low single-breath carbon
monoxide transfer factor (DLco) suggesting
emphysema (type A); others (nos. 8-11) had a
normal TLC and DLco but a raised mixed
venous Pcoz (type B). The remaining patients
(nos. 12-16) had emphysema and a raised
mixed venous Pco2 (type M). Nine patients
(no. 1, no. 2, nos. 9-15) had had episodes of
peripheral oedema.
During the studies the patients sat in a
specially constructed chair with their backs
Local hypoxic vasoconstriction in bronchitis
389
I
I
I
I
I
Tso air
I
Tso oxygen
I
I
2
4
8
6
I0
12
Time Grin)
FIG.2. Clearance of radioactivity on a log scale plotted against time for the left upper zone of a patient
(no. 8) with irreversible airways obstruction, after intravenous injection of I3N. The clearance of
radioactivity is delayed when breathing 30% oxygen ( 0 )as compared with breathing air (a),both in
the time for activity to fall to 10% of the peak (T90) and also in relation to the volume expired by the
) . upper cross-hatched column shows
whole lung when this radioactivity value is reached ( V E ~ ~The
V,,, (air) (= 66.6 litres) and the lower column shows V E 9 o (oxygen) (= 98.5 litres).
against a gamma camera (Nuclear Enterprises
Mk. I11 A) linked to a digital computer (Hewlett Packard 2100). Position was adjusted by
means of radioactive markers. Throughout
the study the patients breathed quietly through a
mouthpiece and two-way valve. From a gas
meter on the inspiratory line, minute ventilation
( pE), tidal volume (VT) and respiratory frequency (f) were measured. I3N (1-5 mCi
dissolved in sodium chloride solution) was
injected rapidly into a central venous catheter
inserted via an antecubital vein. Clearance of
the isotope from the lungs was followed down
to 5% of the peak radioactivity. Afterwards
10-15 mCi of I3N gas in a 50 ml syringe was
added to the inspiratory line from a pump at
10 ml/min for 5 min until the radioactivity over
the lung fields was steady. Then clearance was
followed as before. The measurements were
repeated after the patients had breathed 30%
oxygen for 20 min, except in eight cases where
the intravenous "N measurement was made
earlier (after 5-1 5 rnin). Oxygen breathing
continued during the clearance measurements.
The radiation dose, which is virtually confined
to the lungs, was 500 mrad for each patient.
Two samples of arterialized capillary blood
were taken from the ear lobe before and after
20 min oxygen breathing. On a separate occasion, functional residual capacity was measured
in a body plethysmograph before and after 20
min of breathing 30% oxygen.
Calculations
O n the gamma camera display the lung fields
were divided into four zones, upper and lower
for each lung. Radioactivity accumulated over
5 s periods and clearance curves were constructed for each zone and for the total field.
After correction for radioactive decay and
background radioactivity, activity as % of
maximum counts was plotted on a logarithmic
scale against time (Fig. 2). The time taken for
radioactivity to fall by 90% from its peak ( 7 ' g 0 )
was found for each region and, from the tracing
of overall ventilation, the total volume in
litres expired to this time ( v E 9 0 ) was calculated
for each T 9 0 . The larger the VEg0in litres for a
region the slower the clearance.
For the whole lung, we calculated in addition
a ventilatory efficiency (Prowse & Cumming,
390
N.M . Eiser, H. A . Jones and J. M. B. Hughes
Oxygen
FIG.3. VEso(litres) on air compared with breathing 30% oxygen after (a) intravenous 13N and (b) inhaled 13N. The scatter of points around the line of identity is small after inhaled "N, indicating little
alteration of clearance of inhaled 13N by 30% oxygen. The larger scatter to the right of the line after
intravenous injection shows the slower clearance of some regions when breathing oxygen. Each point
(0)represents one zone. Cross-hatched area represents the repeatability of a single measurement when
breathing air.
1973) by relating the VEg0for the whole radioactivity counting field to functional residual
capacity (FRC). Efficiency was obtained by
comparing the turnover number (TO) for 90%
elimination (VEso/FRC) to that for an ideal
lung, after a small correction for respiratory
frequency (Cumming & Jones, 1966). Thus
TOso = VEso(litres)/FRC (litres), where TOs0
is the turnover number for activity to fall to
10% of its initial value. For example, ideally a
chamber of 1 litre capacity filled with 100% N1
must be flushed with 2-3 litres (1 litre+log, 10)
of oxygen for the nitrogen concentration
to fall to 10%. Therefore efficiency,, = TOpo
(ideal)/T09, (patient).
Efficiency sets a standard, independent of the
bulk flow of ventilation (except for that spent
in the anatomical dead space) and lung volume,
enabling comparisons to be made under
different conditions. It was only possible to
calculate efficiency for the whole lung, since the
and FRC was not
regional share of overall
known. The V,,O in litres was chosen as the
variable to compare changes in I3N wash-out
in individual zones since it corrected for any
changes in minute ventilation associated with
oxygen breathing.
Since oxygen breathing may affect the distribution of ventilation as well as the distribution of local blood flow,only changes in intravenous I3N clearance which exceed the inhaled
vE
13Nclearance are relevant to the notion of local
hypoxic vasoconstriction.The regional clearance
of nitrogen (or VEso)after inhalation to equilibrium is proportional to the local ventilation
per unit ventilated volume ( ~/VA~,,,,.)
and the
clearance after 13Ninjection is proportional to
local ventilation per unit perfused volume (v/
V A ~ ~ ~The
~ . )ratio
.
V,,, (inhalationlinjection)
is proportional to V A " ~ , V
, ~A
. / , ~ ~This
~ . ratio
tends towards unity if perfusion is distributed
uniformly in relation to local alveolar volume,
or if ventilation is itself uniform. Thus, if there
is hypoxic vasoconstriction, the ratio should
be greater than one, becoming less with oxygen
breathing. Blood flow per unit alveolar volume
for each zone was calculated by relating the
peak radioactivity counts after 3N injection
to the plateau value achieved at the end of 13N
inhalation. The ratio for each zone was expressed as a percentage of that for the whole
field.
Results
Abnormal clearance curves (Fig. 2), compared
with those of normal subjects, were obtained
in all zones studied, indicating marked inhomogeneities of ventilation, with well and
poorly ventilated populations of alveoli within
each zone. Fig. 2 also shows the slowing of
Local hypoxic vasoconstriction in bronchitis
391
TABLE
1. Mean clearance of 13Nin allpatients
V E ~ intravenous
,
"N: expired volume (litres) for lung radioactivity to fall to 10% of
its peak value after injection of 13N. VEgo inhaled 13N: expired volume (litres) for
lung radioactivity to fall to 10% of its plateau value after I3N equilibration. Mean
values+ 1 SD are shown.
Change
Air
30% Oxygen oxygen-air
P
(%)
V E intravenous
~ ~
13N (1)
All regions
Upper zones
Lower zones
inhaled 3N (1)
All regions
Upper zones
Lower zones
VE90 inhaled 13N
VEgo intravenous 13N
All regions
Upper zones
Lower zones
Mean of four regions
in each patient
f 38
f21
+32
< 0.006
< 0.1
c005
58.4f29.2
490f 14-4
68.1 3 4 6
f 0.4
-5
N.S.
N.S.
N.S.
1*1+0-4
1.lk0-4
1.1 0.4
- 15
1.3 & 0.4
1.2k0-4
+
-16
- 12
1.2k0.4
1.1k0.4
- 14
50.7f28.5
4 3 - 2 2 15.4
58.2f 30-4
70.02 67.0
52.3k28.3
77.1 f 55.3
VEgo
58-12 23.7
51-82 12.9
6 4 - 4 226-2
1-3+04
clearance of intravenous 13N on 30% oxygen
which occurred in some zones.
A comparison of the changes seen in each
zone in all patients when breathing air and 30%
oxygen is shown in Fig. 3, after intravenous
13N and after inhaled 13N. Points below the
line of identity indicate zones in which the
clearance was slower when breathing oxygen
than when breathing air. The reproducibility
of I3N clearance was measured in a separate
study by comparing regional clearance curves
on two successive occasions in 20 zones in
five other patients, similar to those in this series.
There was SD of 12%for each zone. The hatched
area in Fig. 3 indicates 2 SD and so only zones
outside this area reflect a significant change
when oxygen is breathed.
The mean Veso values for all the patients
when breathing air and oxygen are shown in
Table 1, expressed either as a mean of all
zones, or as a mean of the two upper or two
lower zones of each patient. Since nearly all
individual zones had VEgovalues when breathing air well in excess of the normal range of
11-16 litres after intravenous 13N and 14-17
litres after inhaled 13N (Ewan, Eiser, Jones,
Obdrzalek, Rhodes & Hughes, 1976), the mean
clearances were also very slow. The slower
clearance in the lower compared with the upper
4-6
zones is an inversion of the pattern seen in
normal subjects. The clearance of 13N was
faster after intravenous injection than after
inhalation, implying that only the betterventilated alveoli were being perfused. When
30% oxygen was breathed, the clearance of
the intravenous 13N slowed, becoming even
slower than the clearance of the inhaled 13N.
Fig. 4 shows the change in V E g O on oxygen
expressed as a percentage of VEg0on air:
V,,,[(oxygen - air)/air]%; the more positive
the value of this index, the slower the clearance
of 13N when breathing oxygen. A high value
for the intravenous 13N ratio with a low or
negative value for inhaled 3NcIearance implies
a change in the local distribution of blood flow.
A wide range of response to oxygen is seen, and
several patterns emerge. Seven patients showed
little change of either inhaled or intravenous
13N wash-out on oxygen. However, in six
patients perfusion was redistributed, as shown
by a slowing of intravenous 13N clearance of
36-130% on oxygen, and much smaller associated change in inhaled 3N clearance from 15 %
slower to 10% faster. Inhaled 13N clearance
was affected by oxygen in three patients; in
one, a 36% deterioration in Vagofor inhaled
13N was accompanied by a 49% fall in VE90
for intravenous 3N, indicating increasing mal-
392
N. M. Eiser, H . A . Jones and J . M. B. Hughes
inholed "N
( intr~venour'~N1
Each zone
T
Mean of four zones
in each potient
Fm. 5 . Changes in V,,, ratio (inhaled '3N/intravenous
13N) on 30% oxygen compared with air breathing, for
individual zones (left) and for the mean of four zones
in each patient (right). This ratio will alter when there
is more change in the wash-out of intravenous 13N
than of inhaled 13N, i.e. when 310 matching is influenced by a redistribution of perfusion. Negative
values imply that a greater proportion of local blood
flow perfuses better-ventilated units on air compared
with oxygen.
this ratio means that 30% inspired oxygen
slowed
the clearance of intravenous 13N more
FIG. 4. Change in clearance of 13N with 30% oxygen
breathing for the mean of four zones in each patient.
than the inhaled 13N clearance, suggesting
For each patient the percentage change in V,,, on
relaxation of hypoxic vasoconstriction.
oxygen [V,,, (oxygen-airlair) %] for intravenous "N
Values of this ratio in the individual zones
and inhaled 13N is joined by a line. There was no
significant change on comparison of intravenous with
are plotted on the left in Fig. 5 and the mean
inhaled 13N wash-out in seven patients (0)hut in six
values of four zones in each patient on the right.
( 0 ) intravenous 13N wash-out was delayed much more
A change of 0-2 or more occurred in 23 of the
than inhaled 13N wash-out.
64 zones and in six of the 16 patients. For the
distributionof both ventilationand perfusion. In
group as a whole a mean change of 0.2 occurred.
By the paired t-test, this was highly significant
afurther two patientstherewas a faster clearance
of both the inhaled 13N (18 and 19%) and the
for the individual zones (P<O.O002), but not
so significant for the mean of four zones in
intravenous 3N(40and 23 %), compatible with
an improvement in ventilatory efficiency and
each patient (P = 0.08). This result suggests
better matching of perfusion to ventilation.
that in our patients the increased ventilationIf all ventilated units were equally perfused,
perfusion mismatching when breathing oxygen
~ ~ ~ was the result of a redistribution of local perthe VEg0ratio (inhaled 1 3 N / i n t r a ~ e n13N)
would be 1.0. However, in 44/64 zones (69%)
fusion rather than of local ventilation.
the VEg0ratio (inhaled 13N/intravenous 13N)
The mean clearance of injected 13N from
was greater than unity, suggesting that, in these
the whole lung expressed as a ventilatory
zones, only the better-ventilated units were
efficiency (Table 2) was 29% ( -t 15% SD) combeing perfused. Fig. 5 shows the change in
pared with a value of more than 50% in normal
this ratio when breathing oxygen compared
subjects (Ewan et a[., 1976). When breathing
with breathing air. A large negative change in
oxygen this fell to 24rtlOX. This is just
Local hypoxic vasoconstriction in bronchitis
393
TABLE
2. Change of overall lung function on 30% oxygen
Normal ranges quoted are results from our laboratory. Mean valuesfl
Mean
30% Oxygen difference
Air
Overall efficiency 90%
(a) Intravenous I3N
(b) Inhaled "N
Minute ventilation (1)
Tidal volume (1)
Functional residual capacity (1)
Arterialized capillary blood
Pas02 (kPa)
Pa,oz (kPa)
29t.15
24+ 8
9*0+2.3
054t.016
5.4+ 1-3
6*3+1.2
8-8+ 1.7
SD
24k 10
24+ 8
8.3f1.4
051+0.14
5.1f1.1
6.4k1.7
14.0f 3.4
-5
0
are shown.
Normal
range
49-67
44-56
-0.7
-0.03
-0.3
-0.1
5.3
4.7-6.0
11*0-13*5
(on air)
significant ( P = 0.05). The efficiency of clearance of inhaled 13N was also 24%, compared
with more than 45% in normal subjects, but
this did not change when breathing oxygen.
In normal subjects breathing 30% oxygen does
not affect either efficiencies. There were no
significant changes in either VT or FRC when
breathing 30% oxygen but there was a significant decrease in the mean VEof the group of
0.71 litre (Pi0.05), mainly due to a 15-33%
fall in & in four patients. Arterial Po, increased
as expected but the mean changes in arterial
PCOZfor the group were insignificant. Perfusion per unit alveolar volume was not
affected by 30% oxygen on an interzonal basis.
Table 3 summarizes the changes associated
with oxygen breathing, for patients with A,
chronic bronchitis and emphysema, B, chronic
bronchitis with a raised mixed venous Pcoz
but normal gas transfer and M, chronic
bronchitis and emphysema with a raised mixed
venous Pco2. The patients (no. 1, no. 2 and
nos. 9-15) with past or present oedema are
also indicated. The most striking feature is
that the patients who had most increase in
ventilation-perfusion mismatching, as shown
by a large change in VEg0for 13N given intravenously, had past or present oedema. However,
the oxygen-induced changes in VA0 (intravenous 13N) and the VBgoratio (inhaled 13N/
intravenous I 3 N ) (Fig. 5) were not correlated
with FEVl.o, TLC, DLco,initial arterial Pco,
or changes in arterial Pcoz when breathing
oxygen. The fall in Vnwas not correlated with
any rise in arterial Pcoz (Table 3). Only three
patients increased their arterial Pcoz when
breathing 30% oxygen by more than 1.0 kPa
(range 1-1 - 1.6 kPa). The largest rise of arterial
Pcoz (patient no. 15) was associated with
a 15% fall in minute ventilation and a small
deterioration (8 %) in ventilatory efficiency.
In patient no. 14, a 1.2 kPa rise in arterial Pcoz
could be attributed to an increase in ventilationperfusion mismatching, mainly due to a worsening in the distribution of ventilation alone
[36% change in V,,, (inhaled 13N)].Surprisingly, there was no significant change in minute
ventilation or in 13N clearance in patient no.
16, whose arterial PCOZrose by 1.1 kPa. Thus
in general there was no significant change in
arterial Pcoz on 30% oxygen in these patients,
and in the three patients whose arterial Pcoz
did rise this could not be attributed to a redistribution of perfusion, nor, indeed, to any
single cause.
Discussion
It is unlikely that the arterial Po, reached a
plateau in those patients who breathed 30%
oxygen for only 5-10 min before intravenous
3N was given. Consequently any redistribution
of local perfusion may have been underestimated. However, there was no correlation
between the length of the oxygen breathing
period and the changes in YE,,.
The inhalation of 13N was restricted to 5
min to avoid undue radiation exposure, and to
prevent altering the 'tail' of the clearance
curves by absorption of 13N into the tissues
and blood (Matthews & Dollery, 1965). Of
course, this absorption would have been rela-
N . M . Eiser, H . A . Jones nnd J. M . B. Hughes
394
TABLE
3 . Effects of breathing 30% oxygen in each patient
Type of patient (see the Materials and methods section): A, chronic bronchitis and
emphysema; B, chronic bronchitis, raised Pv,coz and relatively normal gas transfer; M,
chronic bronchitis and mixed picture.
= past oedema;
= present oedema.
AVE(%) (inhaled 13N) = change (%) in minute ventilation on oxygen during inhaled
clearance. 13N i.v. AVEQo(%) = change PA) in V,,, on oxygen after intravenous "N.
13N inhaled A V,Q,(%) = change (%) in VE9, on oxygen after inhaled 13N.
++
+
A 3,
Patient
no.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Type
APa,coZ(kPa)
(inhaled I3N)
l3N i.v.
(%)
A VEQO (%)
-2
A+
A+
A
A
+-0.2
-24
A
A
A
-0 4
0
-24
1-0.1
- 12
B
+0.1
07
$7
B+
B+
B+
M+
M+
M++
M++
M-
I
-
-
+0.3
-0.3
- 0.9
-
+ 1.2
+ 1.6
+ 1.1
tively trivial compared with that of the more
soluble Is3Xe (Ronchetti, Ewan, Jones &
Hughes, 1975). Since complete equilibration of
inhaled "N would have taken 15-20 min, the
most poorly ventilated units could not have
been adequately labelled. Thus the rate of
clearance of inhaled 13N was probably overestimated when breathing either air or oxygen.
Nevertheless, since the radioisotope was inhaled over the same period when breathing
either air or oxygen, and its subsequent clearance was unchanged by oxygen, it is unlikely
that the oxygen significantly affected the distribution of ventilation irrmost zones.
The inhalation and intravenous injection of
13Nmust have labelled different populations of
alveoli; some alveolar units, ventilated but
completely unperfused, were seen only with the
inhaled 13N, and others, perfused but very
poorly ventilated, only with the injected 13N.
A change in the patients' position between
studies might affect the comparison between
results with air and oxygen but the radioactivity
counting fields were relatively large and the
patients were anchored by a fixed headrest and
mouthpiece. No attempt was made to random-
+3
- 13
+I
- 15
+1
0
-11
$4
+4
- 15
-4
+ 102
"N inhaled
A VEQO
+6
+ 36
-7
+5
- 23
-40
- 14
- 19
- 18
0
-9
$9
I5
+I
+21
43
+
+ 10
+ 58
+ 130
+ 16
+49
+I5
-6
0
+
0
+6
- 10
-5
+36
+8
-5
ize the order of wash-outs; wash-outs on air
always preceded those on oxygen. Had the
order been reversed the studies would have been
lengthened by at least 20 min to allow the
arterial blood gas tensions to revert to their
normal values when breathing air.
Pain et al. (1965), investigating the effect of
100% oxygen on ventilation-perfusion matching in chronic bronchitis, found increases in
arterial Pcoz out of proportion to the changes
seen in minute ventilation. They attributed
this to a redistribution of pulmonary blood
flow rather than a change in the distribution of
ventilation, but no measurements of regional
distributions were made. Our observation that
the clearance of 13Nafter intravenous injection
was faster than that after inhalation when
breathing air implies a preferential perfusion of
better-ventilated alveoli. The subsequent slowing of clearance of the intravenous 13N, but
not the inhaled 13N, when breathing 30%
oxygen, which occurred in six of 16 patients,
strongly suggested a local redistribution of
perfusion, probably through relief of hypoxic
vasoconstriction. Perfusion per unit volume
(estimated by dividing the peak radioactivity
Local hypoxic vasoconstriction in bronchitis
counts after intravenous I3N by the equilibration plateau for the whole zone) was unaffected
on an interzonal basis. Nevertheless, within
zones, a local redistribution of perfusion was
found in 23/64 zones (36%) in ten of our patients.
There is strong experimental evidence linking
local pulmonary vasoconstriction with alveolar
hypoxia (Dirken & Heemstra, 1948; Rahn &
Bahnson, 1953; Arborelius, 1966; Grant et al.,
1976) but the mechanisms remain obscure.
Pulmonary vasodilatation in patients with
chronic airways obstruction when 100% oxygen
is breathed was shown by Holt & Branscomb
(1969, who found pulmonary arterial pressure
and pulmonary vascular resistance to fall,
without a change in heart rate, pulmonary
wedge pressure or systemic arterial pressure.
Oxygen in concentrations as low as 30%
is responsible for an increase in mismatching of
local perfusion to ventilation in some patients
with chronic airways obstruction due to a
redistribution of local pulmonary perfusion.
From the clinical standpoint, the increase in
V A / mismatching
~
was not severe enough to
raise the arterial Pco2 significantly. There was
no systematicchange in the distribution of local
ventilation nor in the pattern of breathing on
the 30% oxygen mixture. Minute ventilation
fell in only four of our patients and on its
own this fall was insufficient to influence the
arterial Pcoz. Apart from some association
with a history of salt and water retention, it
was not possible to predict either from clinical
or laboratory data which patients would show
the greatest changes in v A / e matching when
breathing 30% oxygen.
Acknowledgments
We thank Mr J. C . Clark, Mr P. D. Buckingham,
Mr P. L. Horlock and Mr R. D. Williams of
the MRC Cyclotron Unit for preparation of
3N,and Mr A. Herring and Miss Sue Moss of
the Department of Medical Physics for data
processing.
395
References
ARBORELIUS,
M., JR (1966) "'Kr in the study of pulmonary circulation and ventilation during unilateral
hypoxia. Scandinavian Journal of RespiratoryDiseases,
62,105-108.
ASTIN,T.W. (1970) The relationship between arterial
blood oxygen saturation, carbon dioxide tension
and pH and airways resistance during 30% oxygen
breathing in patients with chronic bronchitis with
airways obstruction. American Review of Respiratory
Diseases, 102, 382-387.
BARER,G.R., HOWARD,P. & SHAW, J.W. (1970)
Stimulus-response curves for the pulmonary vascular
bed to hypoxia and hypercapnia. JournalofPhysiology
(London), 211, 139-155.
CUMMING,
G. & JONES,J.G. (1966) The construction
and repeatability of lung nitrogen clearance curves.
Respiration Physiology, 1, 238-248.
CUMMING,
G. & JONES, J.G. (1967) Effect of oxygen in
high concentration on lung NZ clearance curves in
patients with chronic bronchitis. Journal of Applied
Physiology, 22, 277-281.
DIRKEN,
M.N.J. & HEEMSTRA,
H. (1948) Alveolar
oxygen tension and lung circulation. QuarterIy
Journal of Experimental Physiology, 34, 193-21 1.
EWAN,P.W., EISER,N.M.,JONES,H.A., OBDRZALEK,
J., RHODES,
C.G. & HUGHES,J.M.B. (1976) Intraregional inhomogeneity of ventilation and perfusion.
Bulletin EuropPen Physio-pathologie Respiratoire. 12,
195~.
GRANT,B.J.B., DAVIES,E.E., JONES,H.A. & HUGHES,
J.M.B. (1976) Local regulation of pulmonary blood
flow and ventilation-perfusion ratios in the coatimundi. Journal of Applied Physiology, 40,216-228.
HOLT,J.H. & BRANSCOMB,
B.V. (1965) Haemodynamic
responses to controlled 100% oxygen breathing in
emphysema. Journal of Applied Physiology, 20,
215-220.
HUGHES,
J.M.B. (1975) Efficiency of gas exchange in
the lung. Bullerin de Physio-pathologie Respiratoire,
11,921-935.
MATTHEWS,
C.M.E. & DOLLERY,C.T. (1965) Interpretations of 133Xelung washin and washout curves
using an analogue computer. Clinical Science, 28,
573-590.
PAIN,M.C.F., READ,D.J.C. &READ,J. (1965) Changes
in arterial carbon dioxide tensions in patients with
chronic lung disease breathing oxygen. Australasian
Annals of Medicine, 14, 195-204.
G. (1973) Effect of lung
PROWSE,K. & CUMMING,
volume and disease on the lung nitrogen decay
curve. Journal of Applied Physiology, 34,23-33.
RAHN,H. & BAHNSON,
H.T. (1953) Effect of unilateral
hypoxia on gas exchange and calculated pulmonary
blood flow in each lung, Journal of AppliedPhysiology,
6,105-112.
RONCHETTI,R., EWAN, P.W., JONES,T. & HUGHES,
J.M.B. (1975) Use of 13N for regional clearance
curves compared with 133Xe. Bulletin de Physiopathologie Respiratoire. 11, 1 2 4 ~ - 1 2 5 ~ .
VON EULER,U.S. & LILIESTRAND, G. (1946) Observations on the pulmonary arterial blood pressure in the
cat. Aeta Physiologica Scandinavica, 12,301-320.