Oxygen Supplementation during Exercise in Cystic Flbrosls'"
PATRICIA A. NIXON, DAVID M. ORENSTEIN, SCOTT E. CURTIS, and ELIZABETH A. ROSS
Introduction
Exercise intolerance and exertional dys­
pnea are common problems for patients
with cystic fibrosis (CF). As lung func­
tion deteriorates, the CF patient's abili­
ty to exercise invariably diminishes. In
patients with severe lung disease, oxy­
hemoglobin desaturation during exercise
may be severe (1). We have previously
shown that patients with a forced expi­
ratory volume in 1 s (FEV 1)less than 50070
of forced vital capacity (FVC) are more
likely to desaturate during exercise than
are patients with an FEV 1 > 50070 of FVC
(1). Supplemental oxygen is often pre­
scribed in order to minimize desatura­
tion during everyday activities and im­
prove mobility, but the research examin­
ing its effects on oxygenation and exercise
tolerance in CF patients is scant (2, 3).
Coates and coworkers (2) found that
the physical work capacity (PWC, i.e.,
maximal work load accomplished on a
progressive exercise test) of 20 children
with CF did not improve significantly
with hyperoxic air. However, peak min­
ute ventilation and heart rate were sig­
nificantly lower during exercise with ox­
ygen supplementation, suggesting that
both pulmonary and cardiac work were
lessened. Oxyhemoglobin saturation was
not measured during exercise, making
it impossible to determine if the benefi­
cial effects of oxygen supplementation
were related to improved oxyhemoglobin
saturation.
Cropp and colleagues (3), in a prelim­
inary presentation, likewise reported re­
duced minute ventilation in CF patients
during exercise with hyperoxic air with
no differences in the peak work attained.
Hyperoxic air appeared to enable patients
to ride longer during an endurance cy­
cle ergometer test (80070 PWC), with the
greatest improvement occurring in pa­
tients who desaturated the most with nor­
moxie air. Other studies have shown that
oxygen supplementation enables adults
with chronic obstructive pulmonary dis­
ease (COPD) to prolong their submax­
imal exercise endurance time without
improving maximal physical work capa­
city (4-6). Furthermore, subjects with nor­
SUMMARY Fourteen female and 22 male patients with cystic fibrosis (CF), 8 to 29 yr of age, per­
formed two progressive exercise tests to exhaustion on a cycle ergometer, breathing nonnoxic 8jt
(21oliJ Oz) for one test, and hyperoxic air (30% Oz) for the other test. The order of gas administration
was randomized. Minute ventilation eVE), oxygen uptake (VOz) , end-tidal CO z tension (PETC02 ) , W()(1(
rate, oxyhemoglobin saturation (Sa02), and heart rate (HR) were measured throughout the tests.
The Saoz of 11 patients at peak exercise was 90% or less ("Low Sat" group). The Sa02 of 23 patients
remained above 90% throughout the exercise ("High Sat" group). Hyperoxic air minimized desatura­
tion during exercise In the Low Sat group to 2 ± 2% compared to a decrease of 10 ± 5% with
normoxic air. The decrease in saturation was not significant for the High Sat group (1 ± 1% tor
both 21% and 30% O 2). Peak work rate and V0 2 did not differ significantly between normoxic and
hyperoxlc conditions. However, VE and HR at peak exercise tended to be lower, and PETC02 was
higher during peak exercise with 30% O 2than 21% O 2for both groups. During submaximal exercise.,
Oz desaturation was diminished and HR was significantly lower with supplemental O 2, specfficaly
in the Low Sat group. VE was significantly lower for both groups during submaximal exercise with
hyperoxic air. The results suggest that O 2supplementation minimizes Oz desaturation and enab~
patients with CF to exercise with reduced ventilatory and cardiovascular work.
AM REV RESPIR DIS 1990; 142:807-811
mal lungs mayor may not increase their
maximum exercise performance breath­
ing hyperoxic air (7).
The purpose of the present investiga­
tion was to explore further the effects of
oxygen supplementation on exercise
tolerance and oxyhemoglobin saturation
in patients with CF. Although we were
interested in its effects on peak perfor­
mance, daily activity does not often in­
clude maximal exercise.Consequently, we
also examined responses to a sub maximal
amount of work that may require less
motivation by the patient and may more
closely mimic real-life situations for
which oxygen is likely to be prescribed
and used. We also chose to study a larger
sample of patients so that comparisons
could be made between patients who
desaturated during exercise and those
who did not.
Methods
Subjects
Fourteen female and 22 male patients between
8 and 29 yr of age volunteered. All patients
were recruited from the Cystic Fibrosis Cen­
ter at Children's Hospital of Pittsburgh and
had a diagnosis of CF confirmed by a pos­
tive sweat test and typical respiratory/diges­
tive symptoms and/or family history (8). To
include patients who were likely to desaturate
and those who were unlikely to desaturate dur­
ing exercise, we recruited roughly equal num­
bers of patients with FEV 1 < 500/0 FVC and
FEV! > 500/0 FVC. Patients who used sup­
plemental oxygen continuously were exclud­
ed. Patients were subsequently categorized in­
to two groups based on whether or no! they
desaturated during exercise with normoxc air.
Patients were labeled "High Sat" if their peak
Sao2 stayed above 900/0 during exercise. anc
"Low Sat" if their Sao2 was 90070 or lower a:
peak exercise. This division was based 0:1 the
most commonly used cutoff for clinica, pre­
scription of supplemental oxygen (9). Writ­
ten informed consent was obtained from the
patient (and one guardian if the patien: Wa5
17 yr of age or younger).
Pulmonary Function Testing
Prior to each exercise test, standard spireme­
try was performed, with values recorded 1'0;:­
FVC, FEV 17 and maximal voluntary vectila­
(Received in original form July 5, /989 csd
revised form January 17, 1990)
:~7
1 From the Pulmonology and Cystic Fibrous 01.­
vision, Department of Pediatrics, Univerury cc
Pittsburgh School of Medicine; the Children's Hos­
pital of Pittsburgh, Pittsburgh, Pennsylvania; ant:
the Department of Pediatrics, University o: Ala
barna, Birmingham, Alabama.
2 Supported in part by the Cystic Fibrosis ?ou:-:··
dation and National Institutes of Health Crar L:
S07 RR0550724, MOl RROOO84, and R _- '.
HL35334-01.
3 Correspondence and
req uests for re:r:r:cJ
should be addressed to Patricia A. Nixon, ?t. D..
Cystic Fibrosis Center, Department of Pednrrics.
Children's Hospital of Pittsburgh, 3705 5n Ave,
and DeSoto St., Pittsburgh, PA 15213.
808
tion (MVV). Tests were repeated until repro­
ducible results were obtained (at least three
times).
Exercise Testing
Progressive exercise tests. Each patient per­
formed two progressiveexercisetests separated
by a 90-min rest period. One test was per­
formed breathing hyperoxic air (30070 O 2), and
the other breathing normoxic air (21070 O 2 ) .
The order in which the gas mixtures were ad­
ministered was randomized, with the patient
and the supervising physician unaware of the
mixture.
The gas mixture for each test was humidi­
fied and contained in a Douglas bag connect­
ed to the inspiratory port of the breathing cir­
cuit. The patient breathed the gas mixture
while he or she was sitting on the cycle er­
gometer for 12min immediately preceding the
exercise. This period of time was sufficient
for resting values of oxyhemoglobin satura­
tion to become stable. The patient then be­
gan the exercise test following the Godfrey
protocol (10). The work rate was set at zero W
(or no resistance) for the first minute, and in­
creased by 10, 15, or 20 W (depending on the
height of the patient) each minute thereafter
until the patient was unable to continue. The
patient was encouraged to give a maximal ef­
fort. Minute ventilation (VE), oxygen uptake
(V02), carbon dioxide production (Ve02), and
end-tidal CO 2 tension (PETe02) were deter­
mined for each minute of rest and exercise
using a Medical Graphics 2001 metabolic cart
(breath-by-breath analysis; Medical Graph­
ics Corp., St. Paul, MN). (Calibration of the
pneumotachometer using the humidified
hyperoxic gas mixture did not yield different
results than calibration using room air.) The
patient's electrocardiogram was monitored
continuously, and heart rate was determined
for each minute. Oxyhemoglobin saturation
was measured continuously by a Hewlett­
Packard ear oximeter. Physical work capaci­
ty (PWC) was defined as the highest work rate
(watts) maintained for 1 min.
Submaximal exercise tests. Twenty-four pa­
tients returned on a second day (1 to 2 wk
after initial testing) to undergo submaximal
exercise testing breathing 21070 O 2 for one test
and 30070 O 2 for another test. The order of
the tests was randomized, with the patient and
supervising physician unaware of the gas mix­
ture. Before the exercise, the patient breathed
the gas mixture while seated on the cycle er­
gometer. After 12 min of resting measure­
ments, the patient pedaled for 2 min at a work
load equivalent to 25070 of the peak watts at­
tained on the normoxic maximal exercisetest.
The work rate then increased to 50070 of
PWC 21'l.02 and remained at this rate for 10
min (or less if the patient was unable to pedal
for 10 min). Minute ventilation, V02, Ve0 2,
PETe02, heart rate, and Sao, were measured
throughout rest and exercise as for the maxi­
mal tests.
After a 90-min rest period, the patient
repeated the test breathing the other gas
mixture.
NIXON, ORENSTEIN, CURTIS, AND ROSS
Data Analysis
Independent t tests were used to make com­
parisons between the High and Low Sat
groups for anthropometric and resting phys­
iologic variables. Two-factor analysis of vari­
ance was used to make comparisons between
Sat groups (High versus Low) for exercisetests
results with the two gas mixtures (21070 versus
30070 O 2). For the submaximal exercise tests,
the data for each minute of exercise at 50070
of PWC were averaged for the 10-min exer­
cise period to derive a mean submaximal ex­
ercise response. The mean responses weresub­
jected to a two-factor analysis of variance (Sat
group x gas mixture). A Tukey test for post
hoc comparisons was used to probe for sig­
nificant interactions (Sat group x gas mixture).
Results
High Sat versus Low Sat
Group Comparisons
The data of two younger subjects who
were unable to cooperate sufficiently are
excluded from the results. Twenty-three
patients maintained Sao, levels above
90070 throughout progressive exercise with
normoxic air, constituting the High Sat
group. The Low Sat group consisted of
11 patients whose Sao, levels were 90070
or below at peak exercise. Anthropomet­
ric and resting physiologic data of sub­
jects in the two Sat groups are presented
in Table 1. Subjects in the High Sat group
were significantly younger and shorter
and tended to weigh less than subjects
in the Low Sat group. The High Sat sub­
jects also had significantly better pulmo­
nary function than Low Sat subjects as
indicated by a higher mean FEV 1 070 pre­
dicted. The resting Sao, (breathing 21070
O 2) was significantly lower in the Low
Sat subjects. Resting end-tidal CO 2 ten­
sion did not differ significantly between
the Low and High Sat groups. No patient
had a PETeo2 greater than 43 mm Hg.
As presented in table 2, statistical analy-
sis confirmed that the Low Sat group had
a significantly lower mean Sao, than the
High Sat group at peak exercise breath­
ing 21070 O 2. For the total group of pa­
tients, Sao, (with 21070 O 2 ) attained at
peak exercise was significantly related to
resting Sao, (with 21070 O 2 ) (r = 0.823,
p < 0.0001).
The Low Sat group had significantly
worse exercise tolerance than the High
Sat group under both normoxic and hy­
peroxic conditions. Regardless of gas
mixture, subjects in the High Sat group
were able to attain a significantly higher
peak work rate (110 ± 36 versus 83 ±
44 W) and oxygen consumption (31.7 ±
7.0 versus 18.6 ± 5.9 ml/kg/min) than
subjects in the Low Sat group. Minute
ventilation (L/min) at peak exercise (peak
VEl was also significantly higher in the
High Sat group compared to the Low Sat
group (63.9 ± 15.9 versus 45.7 ± 14.3
L/min). However, when peak work rate
was considered, peak Vn/watts did not
differ between the two Sat groups (0.60
± 0.13 versus 0.61 ± 0.22 L/min/W).
The ventilatory equivalent for oxygen
(VE/V02 ) (53.5 ± 11.7 versus 58.0 ± 17.9)
and the ratio of VE to resting MVV
(VE/MVV) (0.86 ± 0.21 versus 0.81 ±
0.22) at peak exercise also did not differ
significantly between the High and Low
Sat groups, respectively.
The peak heart rate of the High Sat
group was significantly higher than that
of the Low Sat group (177 ± 14 versus
164 ± 14 bpm). The Veo 2 at peak exer­
cise was also significantly higher in the
High Sat than Low Sat group (1.41 ±
0.52 versus 0.98 ± 0.43 L/min). Further­
more, end-tidal CO 2tension (pETeD:!.) in­
creased significantly more from rest to
peak exercise in the Low Sat group (9 ±
7 mm Hg) than in the High Sat group
(2 ± 4 mm Hg), reaching mean peak lev­
els of 44 and 34 mm Hg, respectively,
TABLE 1
SUBJECT CHARACTERISTICS·
High Sat Group
Parameter
Age, yr
Sex, M/F
Weight, kg
Height, cm
FEV 1 , % predicted
Sao2, 0/0
PETc0 2 , mm Hg
(n = 23)
13.9 ± 4.1
13/6
39.0 ± 9.9
151 ± 12
64 ± 16
96 ± 2
34 ± 3
Low Sat Group
(n
=
11)
20.8 ± 4.5
9/6
45.6 ± 10.6
164 ± 12
28 ± 8
92 ± 3
35 ± 4
p Value
<
0.0001
NS
< 0.01
< 0.0001
< 0.(Xl1
NS
Definition of abbreviations: High Sat group = patients with oxyhemoglobin saturation (SS02) > 90%
at peak exercise breathing 21 % O 2; Low Sat group = patients with SS02 " 90% at peak exercise
breathing 21 % O 2; SS02 = resting oxyhemoglobin saturation breathing 21 % O 2; PETC0 2 = resting end­
tidal CO 2 tension breathing 21% O 2,
• Values are expressed as mean ± 1 SO.
o,
809
SUPPLEMENTATION DURING EXERCISE IN CF
TABLE 2
PEAK EXERCISE RESULTS·
High Sat Group
Low Sat Group
Parameter
21 % O2
30%
Sa02, %
95 ± 2
97 ± 1
O2
21 % O2
30%
83 ± 7
95 ± 3
O2
Significant Differences
Low Sat, 21% O2 < all other combinations;
Low Sat, 30% O2 < High Sat 21 % and 30%
p Value
< 0000;
O2
Work rate, W
110 ± 34
110 ± 39
79 ± 42
87 ± 47
High Sat> Low Sat
< 0 lJ7
Work rate, W/kg
2.8 ± 0.6
2.8 ± 0.7
1.6 ± 0.7
1.8 ± 0.7
High Sat> Low Sat
<
V0 2, mllkg/min
31.7 ± 6.6
31.6 ± 7.6
18.5 ± 6.5
18.6 ± 5.5
High Sat> Low Sat
< 0001
VE, L/min
65.0 ± 15.4
62.8 ± 16.7
46.8 ± 15.0
44.6 ± 14.3
High Sat> Low Sat
< 0004
VE, L/minlW
0.60 ± 0.11
0.60 ± 0.16
0.64 ± 0.23
0.68 ± 0.22
21 % O2 > 30%
< 006
VE, L/min/cm
0.43 ± 0.08
0.41 ± 0.09
0.28 ± 0.09
0.27 ± 0.08
High Sat> Low Sat;
21 % O2 > 30% O2
VE/MVV
0.88 ± 0.21
0.85 ± 0.21
0.81 ± 0.26
0.80 ± 0.19
NS
Ve.N0 2
55.5 ± 9.5
51.5 ± 13.5
60.0 ± 18.8
56.1 ± 17.6
21 % O2 > 30%
Heart rate, bpm
179 ± 13
175 ± 15
165 ± 16
163 ± 11
PETc0 2, mm Hg
34 ± 3
36 ± 5
43 ± 8
46 ± 10
0.97 ± 0.45
0.98 ± 0.43
VC02, L/min
1.43 ± 0.51
1.38 ± 0.55
O2
0001
< 0.001
< 006
~~S
O2
< 0.04
High Sat> Low Sat
< 002
High Sat < Low Sat;
21 % O2 < 30% O2
< 0001
< 0005
High Sat> Low Sat
< 003
Definition of abbreviations: Work rate = peak work rate; V0 2 = peak oxygen uptake; VE = peak minute ventilation; VElMVV = peak minute ventilation/resting maximal voluntary vennanon:
VE1V02 = peak minute ventilation/peak oxygen uptake; heart rate = peak heart rate; Sao 2 = oxyhemoglobin saturation at peak exercise; PETC0 2 = end-tidal CO 2 tension at peak exercise vee:
= peak volume of CO2 produced.
• Values are mean :t:: 1 SO.
Normoxic versus
Hyperoxic Conditions
Progressive exercise. Statistical analy­
ses revealed that a significant interaction
between the Sat group and gas mixture
existed for change in Sa0 2 from rest to
peak exercise (figure 1). That is, the sig­
nificantly greater drop in Sa02 seen with
normoxic air could be accounted for al­
most entirely by the Low Sat group. Sub­
jects in the Low Sat group demonstrated
significant oxyhemoglobin desaturation
with normoxic air, reaching an Sa0 2 of
83010 at peak exercise. However, in the
same group with hyperoxic air, Sa02 fell
by only 2010, reaching a minimum level
of 95010. The change in Sa02 during exer­
cise did not differ significantly between
normoxic and hyperoxic conditions for
the High Sat group.
Other differences observed between
normoxic and hyperoxic gas conditions
were independent of the Sat group. The
peak work rate (100 ± 39 W for 21010 O 2
versus 103 ± 42 W for 30010 O 2) and V02
(27.2 ± 9.1 versus 27.1 ± 9.3 ml/kg/
min for 21010 and 30010 O 2) attained did
not differ significantly between normoxic
and hyperoxic conditions. However, VE
(56.5 ± 18.0 versus 58.7 ± 17.4 L/min)
and heart rate (171 ± 15 versus 174 ±
15 bpm) at peak exercises were signifi­
cantly lower when subjects breathed 30010
O 2 than with 21010 O 2, respectively. The
lower VEfor a similar V02was associated
with a significantly lower VE/V0 2 for the
hyperoxic condition (53.1 ± 14.9 versus
57.0 ± 13.3), suggesting that less venti­
lation was needed to attain the same
amount of oxygen uptake. Test order did
not affect these results.
Hyperoxic air was associated with a
significantly greater increase in PETC02
(5.9 ± 6.8 mm Hg) from rest to peak ex­
ercise than the increase (3.2 ± 5.4 mm
Hg) observed with normoxic air. The ra­
tio of VE to MVV at peak exercise did
not differ significantly between normoxic
and hyperoxic conditions.
100
95
:::::-:---
.
--------~.
85
80
.....a-
Rest
.......
_
Peak
Exercise
Fig. 1. Oxyhemoglobin saturation (Sao 2) at rest and
peak exercise in High Sat and Low Sat groups breath­
ing 30% and 21% 02. The Low Sat group exhibited sig­
nificantly less oxyhemoglobin desaturation with exer­
cise breathing 30% O2 than 21% 02' The change in
Sao2during exercise did not differ significantly between
normoxic and hyperoxic conditions for the High Sat
High Sat, 30% 02; (.
.)
group. ( .
Low Sat, 30% 02: (.- - -.) High Sat, 21% 02:
(. - - - .) Low Sat, 21% 02'
.>
Submaximal exercise. The mean values
for submaximal exercise under normox­
ic and hyperoxic conditions for High and
Low Sat groups are shown in Table 3. As
with peak exercise, a significant interac­
tion between Sat group and gas mixture
was found for change in Sao,during sub­
maximal exercise (figure 2). Similarly, the
significantly greater drop in Sao, seen
with normoxic air could be accounted for
almost entirely by the Low Sat group.
Oxyhemoglobin saturation decreased sig­
nificantly (-7 ± 7010) during exercise
with normoxic air in the Low Sat group.
With hyperoxic air, the mean decrease
inSa02(-1 ± 2%)oftheLowSatgroup
was not significant and did not differ sig­
nificantly from the mean change in Sao,
of the High Sat group during submax­
imal exercise breathing either 30018 O 2 or
21010 O 2.
The mean heart rate response was also
affected by gas mixture and Sat group
interactively. As presented in figure 3, the
mean heart rate was significantly lower
during exercise with hyperoxic air than
normoxic air, with the lowest heart rate
occurring in the Low Sat group during
exercise with hyperoxic air.
Differences between normoxic and
hyperoxicconditions that were evident:
for the remaining submaximal test results
occurred regardless of the Sat group. For
the same work rate (50010 of the normox­
ic PWC), mean V02 did not differ signif­
icantly between normoxic and hyperox­
810
NIXON , ORENSTEIN , CURTIS , AND ROSS
TABLE 3
SUBMAXIMAL EXERCISE TEST RESULTS '
High Sat Group
Parameter"
21% 0 ,
95
21.4
38.7
0.75
0.25
47.1
147
36
0.84
Sao " %
vo, mllkg/min
VE , Llmin
VE , Llm inlW
VE , Llmin/cm
VENO,
Heart rate, bpm
PET eo" mm Hg
vco, Llmin
±
±
±
±
±
±
±
±
±
• Values are expr essed as mean
30% 0 ,
1
3.0
4.4
0.22
0.02
8.4
13
4
0.20
:t
Low Sat Group
97
20.9
37.8
0.73
0.25
47.3
143
36
0.82
21% 0 ,
± 1
± 2.8
± 4.3
± 0.21
± 0.02
± 10.0
± 11
87
15.7
39.9
0.96
0.24
54.9
148
39
0.81
± 4
± 0.20
30% 0 ,
± 8
±
±
±
±
±
±
±
2.7
9.1
0.40
0.06
19.1
17
7
± 0.22
96
15.1
36.7
0.89
0.22
52.8
135
39
0.76
±
±
±
±
±
±
±
±
±
2
2.7
5.0
0.37
0.03
15.9
20
6
0.22
Significant DiHerences
Low Sat, 21% 0 , < all other combinations
High Sat > Low Sat
21% 0 , > 30% 0,
21% 0 , > 30% 0,
21% 0 , > 30% 0 ,
NS
Low Sat, 30% 0 , < all other combinations
NS
21% 0 , > 30% 0 ,
< O C<JO€
< 0 C·jO'
< O C6
< O' j
< O C6
t\S
< O C3
1\"
< O C~
1 SO.
t Value for each parameter is the average of the 10-min exercise period.
ic conditions (19.3 ± 3.9 versus 18.8 ±
3.9 mllkg/min, respectively). However,
mean VE was significantly lower during
exercise with hyperoxic air (37.4 ± 4.4
L/min) than with normoxic air (39.1 ±
6.4 L/min). In spite of the lower VE for
the similar V02, VE/V0 2 was not signifi ­
cantly lower for the hyperoxic test (49 .3
± 12.4) than for the normoxic test (50.0
± 13.4).
During submaximal exercise,the mean
Ve0 2 was significantly lower with hyper­
oxic air (0.80 ± 0.21 L/min) than with
normoxic air (0 .83 ± 0.20 L/min). De­
spite this difference in Ve0 2, the mean
increase in PETe02 did not differ between
normoxic (2 ± 4 mm Hg) and hyperoxic
(3 ± 4 mm Hg) conditions. Similarly, the
absolute levelsof PETe02attained did not
differ significantly between normoxic (37
± 5 mm Hg) and hyperoxic (37 ± 5 mm
Hg) conditions.
100
. ~:
95
.-- -------- -- - -- -~
ti-
N
.
o
Ul
90
Discussion
The results of this study indicate that sup­
plemental O 2 (F102 = 30070) minimizes
oxyhemoglobin desaturation in patients
with CF who desaturate during exercise
with normoxic air. However, in spite of
improved oxygenation, peak work capac­
ity and oxygen uptake are not improved
with supplemental O 2, The results sup­
port previous reports (2, 3) that oxygen
supplementation enables patients with
CF to perform the same work load with
lower minute ventilation and heart rate
than they employ breathing normoxic air.
Similar results have been demonstrated
in normal subjects (7) and in adults with
COPD (4, 5). Since exercise limitation
in some patients with chronic lung dis­
ease may be caused by fatigue of the ven­
tilatory muscles, decreasing ventilatory
muscle work may in itself increase a pa ­
tient's ability to tolerate exertion (11). (It
is po ssible that even higher inspired oxy­
gen concentrations might improve work
capacity since some studies in normal
subjects have demonstrated improved
work capacity only when F102 exceeded
70% [7]. However, we chose not to study
FI0 2concentrations that were exceedingly
high and impractical for real-life use in
ambulatory patients.)
We also found that the improved oxy­
genation and lower minute ventilation at
peak exercise with hyperoxic air were ac­
companied by a statistically signifi cant
increase in end -tidal CO 2 tension, as is
also seen in patients with COPD (4). .-\1­
though statistically significant, the differ­
ence between end-tidal CO 2 tension at
peak exercise with 21% O 2 (37 mm Hgl
and 30% O 2 (40 mm Hg) in the tota l
group probably ha s little clinical signifi­
cance. In five patients in the Low Sat
group, notable CO 2 retention occurred.
exceeding 48 mm Hg at peak exercise with
30070 O 2, Although PETe02cannot be tak­
en as a direct index of Pae02 during ex­
ercise,the fact that minute ventilation was
less with hyperoxia while Ve02 was not
different, suggests that the higher PETco :
is, in fact, an indication of hypercapnea,
Ventilation during exercise in CF pat ients
is commonly thought to be limited by me­
chanical factors, with minute ventilatio n
(VE) often said to approach or exceed
resting MVV (12-14). Yet these Low Sac
patients with an elevated PETe02 all had
a VE/MVV < 83%, suggesting that ven­
tilation had not reached an absolute me­
chanicallimit. This suggests further tha:
the lower VE, even with elevated PETeo :.
was determined by ventilatory drive and
was not the dire consequence of a veri­
tilatory apparatus pushed to its rr.axi­
mum capacity. Coates and coworkers (I ~ I
have also shown that a group of CF pa­
1 70
8 5 ' - - - -........-
Re st
- - -- -.......- - ­
Exercise
Fig . 2. Oxyhemoglobin saturation (Sao,) at rest and
during submaximal exercise in High Sat and Low Sat
groups breathing 30% and 21% 0,. Sao, decreased
significa nlly during exercise breath ing normoxic air in
the Low Sat group . During exercise breathing hyps rox­
ic air, the decrease in Sao, of the Low Sat group was
not significant and did not differ significanll y from the
. )
change in Sao, of the High Sat group . ( .
High Sat, 30% 0 , ; ( .
. ) Low Sat, 30% 0,;
(. - - -. ) High Sat, 21% 0 , ; ( . - - - .) Low Sat,
21%0,.
~
16 0
8
Po
15 0
s
cr::
......
14 0
,0
'-'
Fig . 3. Heart rate during suorn eurnar
exercise in the High Sat (open OO"ll anc
Low Sat (hatched bars) groups tJ-earr:·
ing 21% and 30% 0 , . Heart rat: was
significantly lower in the Low Sal ;rouc
during exercise breathing hyperoc c ai r.
ell
ell
13 0
120
0
::t
11 0
10 0
21 % 02
30% 02
811
O2 SUPPLEMENTATION DURING EXERCISE IN CF
tients who retained CO 2 during exercise
employed a relatively low mean VE/MVV
ratio (87 ± 15070), compared to a group
who did not retain CO 2 and had a higher
mean VE/MVV ratio (113 ± 9070).
In contrast, the peak PETC02 of one
patient at peak exercise was markedly
lower with hyperoxic than normoxic air
(30 versus 50 mm Hg, respectively), per­
haps because of better oxygen supply to
fatiguing ventilatory muscles. In fact, Bye
and associates (4) have demonstrated that
oxygen supplementation enabled some
adult patients with COPD to employ low­
er VE with less evidence of fatigue of the
diaphragm than they had in room air.
CO 2 retention was uncommon during
submaximal exercise with hyperoxic air.
Except for one patient, PETC02remained
below 41 mm Hg throughout the 10-min
submaximal exercise period with 30070
O 2. It appears that, for most patients,
CO 2 retention is not likely to occur with
oxygen supplementation during sub max­
imal exertion that might be encountered
with everyday physical activities.
The results of the present investigation
also confirm previous reports (1) that
oxyhemoglobin desaturation during ex­
ercise is associated with severe pulmonary
disease. Because we were interested in
evaluating the practical issue of the ef­
fects of oxygen supplementation during
exercise in patients with cystic fibrosis,
we chose to categorize our patients into
groups according to whether or not they
were logical candidates for such supple­
mentation, that is, whether they showed
oxyhemoglobin desaturation with exer­
cise in room air (21070 O 2). In this study,
11 of 34 patients exhibited substantial
oxyhemoglobin desaturation, reaching
an Sao; of 90070 or lower at peak exer­
cise. All 11 of these patients in the Low
Sat group had an FEV. < 50070 predict­
ed. No patient with an FEV. > 50070
predicted had an Sao; of 90070 or lower
at peak exercise. Of the 23 patients in the
High Sat group (Sao, at peak exercise re­
mained above 90070), six had an FEV. <
50070 predicted. These findings suggest
that patients with severe disease (FEV 1
< 50070 predicted) are at greater risk of
desaturating during exercisethan patients
with mild-to-moderate disease (FEV l >
50070 predicted). However, even in pa­
tients with severe obstruction, desatura­
tion does not invariably occur during
exercise.
Similarly, in spite of the strong corre­
lation between resting Sao 2 and Sao 2 at
peak exercise (r == 0.823), it was not pos­
sible to predict Sao, at peak exercise from
resting Sao, for an individual patient. For
example, two patients had resting Sao,
values of 94070. At peak exercise, one
desaturated by 16070 to a low value of 78070.
The other patient's Sao2 dropped 1070,
reaching 93070 at peak exercise. From these
examples, the need to employ exercise
testing to identify patients who desatu­
rate during exercise is readily apparent.
In conclusion, the results of this in­
vestigation indicate that supplemental O 2
(F102 == 30070) minimizes oxyhemoglobin
desaturation in patients with CF who
desaturate during exercise with normox­
ic air. It decreases ventilatory and car­
diovascular work during peak exercise in
all patients regardless of disease severity
but does notimprove maximum work ca­
pacity. Furthermore, supplemental O 2
causes transient CO 2 retention in some
patients during progressive maximal
exercise.
O 2 supplementation also reduces min­
ute ventilation during submaximal exer­
cise in most patients. It also reduces heart
rate and prevents desaturation during
submaximal exercise in patients who de­
saturate during progressive exercise with
normoxic air.
These results suggest that even relative­
ly healthy patients with CF who do not
demonstrate significant oxyhemoglobin
desaturation may benefit from supple­
mental O 2 during exercise. However, the
benefits are minimal in these patients,
and they are unlikely to outweigh the ex­
pense and inconvenience of supplemen­
tal O 2. In contrast, supplemental O 2 may
be very important for patients who
desaturate, not only for minimizing hyp­
oxemia and possibly preventing or delay­
ing the development of pulmonary
hypertension, but also for reducing ven­
tilatory and cardiovascular work during
exercise.
References
1. Henke KG, Orenstein DM. Oxygen saturatico
during exercise in cystic fibrosis. Am Rev Respc.
Dis 1984; 129:708-11.
2. Coates AL, Boyce P, Muller D, Mearns vu.,
Godfrey S. The role of nutritional status, airwav
obstruction, hypoxia and abnormalities in serum
lipid composition in limiting exercise tolerance 0
the children with cystic fibrosis. Acta Paediau Scand
1980; 69:353-8.
3. Cropp GJ, Silbiger LM, Pankow GA, Ford Ak,
Hirsch JA. Effects of oxygen breathing on exer­
cise tolerance in cystic fibrosis (CF). Am Rev Respcr
Dis 1987; 135(Suppl:A286).
4. Bye PT, Esau SA, Levy RD, et al. Venulatory
muscle function during exercise in air and oxygen
in patients with chronic air-flow limitation. Am Rev
Respir Dis 1985; 132:236-40.
5. Bradley BL, Garner AE, Billiu D, Mestas J~~.
Forman J. Oxygen assisted exercise in chronic oc­
structive lung disease. Am Rev Respir Di5 19-'~.:
118:239-43.
6. Swinburn CR, Wakefield 1M, Jones P\\". Rc~­
tionship between ventilation and breathlessness dur­
ing exercise in chronic obstructive airways disease
is not altered by prevention of hypoxaernia. C~
Sci 1984: 67:515-9.
7. Welch H. Hyperoxia and human performance:
a brief review. Med Sci Sports Ex 1982; 14::53~':.
8. Orenstein DM. Diagnosis of cystic fibross. Serr­
in Respir Med 1985; 6:252-60.
9. Block ER. Oxygen therapy. In: Fishman .-\..?~
ed. Pulmonary diseases and disorders. Nev Yore:
McGraw-Hill, 1988; 2321.
10. Godfrey S. Exercise testing in cr.ildrer..
Philadelphia: W.B. Saunders, 1974; 30.
11. Dodd DS, Brancatisano T, Engel LA ChCSl
wall mechanics during exercise in patients with se­
vere chronic airflow obstruction (CAO). :;..."11 R=",,'
Respir Dis 1984; 129:33-8.
12. Godfrey S, Mearns M. Pulmonary f~l.ctiCJ:l
and response to exercise in cystic fibrosis. Arch Dis
Child 1971; 46:144-51.
13. Godfrey S. Exercise testing in ctrldrer..
Philadelphia: W.B. Saunders, 1974; 146-52.
14. Marcotte JE, Grisdale RK, Levison E. et .a:
Multiple factors limit exercise capacity i:. 0~':
fibrosis. Pediatr Pulmonol 1986; 2:274-~-~.
15. Coates AL, Canny G, Zinman R, el :1. Tne
effects of chronic airflow limitation, increased de-..:d
space, and the pattern of ventilation on ~as 0:­
change during maximal exercise in advance; 0'~':
fibrosis. Am Rev Respir Dis 1988; 138:1::~:.L.