Arterial Blood Gases and Pulmonary Function

Testing in Acute Bronchial Asthma

Predicting Patient Outcomes
Richard M. Nowak, MD; Michael C. Tomlanovich, MD; Diane D. Sarkar, MD;
Paul A. Kvale, MD; John A. Anderson, MD

Pretreatment and posttreatment arterial blood gas and pulmonary drawn anaerobically from the radial
function testing measurements were prospectively compared as to their artery into a heparinized glass syringe

ability to assess asthma severity accurately and, thus, predict the outcome in while the patient was at rest and breath¬
102 episodes of acute bronchial asthma initially seen in the emergency ing room air. It was then analyzed for
department. The Pao2, Paco2, or pH was unable to separate these patients Pa02, PaC02, and pH.
Uniform therapy consisted of terbuta-
requiring admission from those that could be confidently discharged, while line sulfate, 0.25 mg subcutaneously, and,
the 1-s forced expiratory volume (FEV,) and peak expiratory flow rate (PEFR) if clinically necessary, intravenous amino-
did so both before and after treatment. Furthermore, virtually all patients phylline with a 5.6-mg/kg loading dose
with hypercarbia (Paco2 >42 mm Hg) and/or severe hypoxemia (Pao2 <60 (downwardly adjusted with recent oral
mm Hg) had a PEFR below 200 L/min, or an FEV, below 1.0 L. Thus, selective theophylline therapy) and a 0.9-mg/kg/hr
use of arterial blood gas analysis should substantially decrease both maintenance infusion.'0 Any further thera¬
diagnostic cost and patient discomfort without jeopardizing health care. py was left to the discretion of the emer¬
(JAMA 1983;249:2043-2046) gency medicine physician. Repeated
PEFR, FEV„ and arterial blood gas mea¬
surements while breathing room air were
MEASUREMENTS of arterial blood possible, specific pulmonary function obtained before the patient's discharge or
admission to the hospital. In virtually all
gas tensions are widely recommended testing guidelines for a more selective
cases, three PEFR and FEV, measure¬
in the assessment and treatment of useof blood gas analysis to identify ments were obtained and the best value
asthmatic patients in the emergency hypercarbia and/or severe hypoxe¬ was recorded.
department13 despite a rather poor mia. Decisions to admit or discharge patients
correlation between gas tensions and were based on clinical assessment and,
asthma severity as measured by pul¬ PATIENTS AND METHODS
occasionally, FEV, measurements." We
monary function testing.4'8 Patients between the ages of 16 and 40 did not use the PEFR in decision making.
This study was undertaken (1) to years who were seen in the emergency All percent-of-predicted normal values
prospectively compare the specificity department with acute bronchospasm were calculated after the study was com-

of pretreatment and posttreatment
arterial blood gas analysis with pul¬
monary function testing in predicting
patient outcome and (2) to develop, if
From the Divisions of Emergency Medicine (Drs
Nowak, Tomlanovich, and Sarkar), Pulmonary Medi-
cine (Dr Kvale), and Allergy and Immunology (Dr
Anderson), Henry Ford Hospital, Detroit.
Reprint requests to Division of Emergency Medi-
cine, Henry Ford Hospital, 2799 W Grand Blvd,
Detroit, MI 48202 (Dr Nowak).
when one of the investigators was present
and who fulfilled the criteria for asthma
as defined by the American Thoracic Soci¬
ety' were uniformly included into the
study. Any patient with any cardiac or
other lung disease was excluded. Peak
expiratory flow rate (PEFR) measure¬
ments, using a mini-Wright peak flow-
meter, and then spirometry (1-s forced
expiratory volume [FEV,]), using a single-
wedge bellows spirometer, were performed
before any treatment. Arterial blood was
pleted.'2'1 At the conclusion of the study,
the patients were divided into three cate¬
gories. Group 1 included all those who
were admitted to the hospital, while
groups 2 and 3 included those who were
sent home. Each discharged patient was
interviewed in the emergency department
or by telephone 48 hours later, and a
questionnaire (Fig 1) was completed to
determine the outcome. Group 2 patients
had two or more affirmative answers,
while discharged patients with fewer than

Downloaded from www.jama.com by guest on November 25, 2009

two affirmative answers were placed in similarly poor correlation (correla¬ for individual percent-of-predicted
group 3. The objective accuracy of this tion coefficient, .4426) was found in FEV, values. Note that C02 retention
questionnaire has been previously veri¬ comparing the posttreatment Pao, (>42 mm Hg) occurred only with an
fied.14 and the percent-of-predicted FEV,.
The data were analyzed using two-
FEV, below 25% of the predicted
The pretreatment Paco2 is com¬ value. A posttreatment comparison of
sample t tests to examine the pretreat- with the percent-of-predicted
ment and posttreatment differences pared these variables showed a far worsen¬
among the three groups. Relationships FEV, for all patients in Fig 2 (corre¬ ing correlation (correlation coeffi¬
between the PEFR or FEV, and the Pao; lation coefficient, —.5337). There was cient, -.0199).
and PacOj measurements were examined likewise a wide variation in the Paco2 All patients with a pretreatment
by means of correlation coefficients.

RESULTS Is your asthma worse now than when you left the emergency room?
Eighty-six patients, 54 women and Yes No
32 men, were treated for 102 episodes
of acute asthma. The mean age was
25.1 years and the mean duration of Have you had to return to any emergency room or see another doctor?
therapy in the emergency department
was 4.8 hours, with no statistically Yes No
significant group differences. Thirty-
two patients (31.4%) were admitted Has your asthma kept you awake at night since you left our emergency room?
(group 1); 20 patients (19.6%) were
discharged but with further respira¬ Yes No
tory problems (group 2); and 50
(49.0% ) were discharged without sub¬ Has your breathing prevented you from resuming your usual activities?
sequent problems (group 3). Some
Yes No
patients refused to allow drawing of
blood for repeated determination of
arterial blood gas values at the cessa¬ Fig 1.—Questionnaire used in studying patients 48 hours after discharge from emergency
tion of treatment, and the mini- department.
Wright peak flowmeter was unavail¬
able for a short portion of the study. Table 1.—Comparison of Mean Pretreatment Arterial Gas Analysis
These factors explain the small dif¬ and Pulmonary Function Testing Measurements
ference in the number of patients
with FEV, and PEFR measurements No. of
1-s Forced Expiratory Volume
Pao„ Paco,,
and in the number in the pretreat- Group Cases mm Hg* mm Hg* pH* Absolute, Lt % of Predicted!
ment and the posttreatment phases of 1 32 68.4 ±8.8 36.3 ±5.9 7.39 ±0.05 0.68 20.1
this study. The mean pretreatment 2 20 71.7±8.7 34.5±7.2 7.40±0.05 0.86 25.3
3
Pao2, Paco2, pH, FEV,, and PEFR for 50 71.5± 12.0 35.8±6.9 7.41 ±0.05 1.12 33.4
all groups are shown in Table 1. There Peak Expiratory Flow Rat«
was no statistically significant group Absolute, L/ mint % of Predicted!
difference between any of the arterial 1 30 68.4±8.8 36.3±5.9 7.39±0.05 134.3 30.3
blood gas variables, whereas both the 2 20 71.7±8.7 34.5±7.2 7.40±0.05 128.0 29.3
3 48 71.9±12.1 35.5±6.6 7.41 ±0.05
FEV, and PEFR were able to separate 176.2 39.7

accurately group 1 from group 3 *Mean±SD. All P>05, two-sample /test, in comparison of groups.
tAll P > 05. two-sample /test, except as indicated.
patients. Similarly, the mean post- $P<05.
treatment Pao2, PaC02, pH, FEV,, and
PEFR are shown in Table 2. Again,
there was no statistically significant Table 2.—Comparison of Mean Posttreatment Arterial Gas Analysis
group difference between any of the and Pulmonary Function Testing Measurements
arterial blood gas measurements,
1-s Forced Expiratory Volume
whereas there were statistically sig¬ No. of Pao„ PacOj,
nificant differences in the pulmonary Group Cases mm Hg* mm Hg* pH* Absolute, Lt % of Predlctedt
1 28 75.3±11.6 30.9 ±3.8 7.45 ±0.04 1.02
function tests (both the FEV, and 32.1
2 13 80.2±10.9 31.5±4.5 7.44±0.03 1.60 49.6
PEFR were able to distinguish accu¬
3 44 78.0±12.7 32.3±4.5 7.43±0.04 2.16 62.5
rately groups 1 and 2 patients from Peak Expiratory Flow Rate
group 3 patients).
A comparison of the pretreatment Absolute, L/mlnt % of Predlctedt
1 28 75.3±11.6 30.9±3.8 7.45±0.04 186.6 42.9
Pa02 and the percent-of-predicted 2 81.0±10.9 31.7±4.7 216.7 52.2
12 7.44±0.02
FEV, for all patients showed a corre¬ 3 41 77.5± 13.0 32.4±4.6 7.43±0.04 322.7 71.7
lation coefficient of .4134, with wide
variations in the Pao2 for individual *Mean±SD. All P>05, two-sample ttest, in comparison of groups.
tAII P<05, two-sample ftest, except as indicated.
percent-of-predicted FEV, values. A *P>05.

Downloaded from www.jama.com by guest on November 25, 2009

 55
PaCOj above 42 mm Hg and/or a Pa02
below 60 mm Hg are shown in Table
3. As might be expected, a higher
50
percentage of these patients were in
group 1. It is noteworthy, however,
that 16% of all group 3 patients had 45 I
pretreatment hypercarbia and/or se¬
vere hypoxemia. Virtually all of these
patients could be detected by specific 40 • MS *|

pulmonary function testing guide¬

lines (Table 4). These included an
absolute FEV, below 1.0 L or a PEFR 35
below 200 L/min.
The pretreatment-to-posttreatment .
• • •

change in Pao, as compared with that 30

for FEV, is seen in Fig 3 (correlation
coefficient, .1824). There was a wide
fluctuation in the degree of improve¬ 25 • •

ment in the Pa02 with specific

improvements in the FEV,. Also, 20 I
there were a number of patients (13/
79, or 17%) with a variable decrease 10 20 30 40 50 60 70 80 90
in the Pao2 accompanied by an obvi¬ %-of-Predicted FEV,
ous rise in the FEV,.
The pretreatment-to-posttreatment Fig 2.—Comparison of pretreatment Paco2 and percent-of-predicted 1-s forced expiratory
volume (FEV,) in groups 1 (squares), 2 (triangles), and 3 (circles). Numbers indicate more than
one value at same location.

Table 4.—Pulmonary Function Criteria Identifying All Cases

Table 3.—Pretreatment Hypercarbia of Hypercarbia* and Severe Hypoxemiatt
and/or Severe Hypoxemia
%-of-Pr»dlct«d
No. (%) of Patients
_Absolute Value_Normal Value
1-s forced expiratory volume <1.0L <25
Paco, Pao,
Group >42 mmHg <60 Hgmm Peak expiratory flow rate_<200 L/min_<30
1 8/32(25) 5/32(16)
2 2/20(10) •Paco,, greater than 42 mm Hg.
2/20(10) fPao,, less than 60 mm Hg.
3 8/50(16) 8/50(16) tone group 3 patient was the exception, exceeding these guidelines but with a Pao, of 56 mm Hg.

Fig 3.—Relationship of pretreatment-to-posttreatment change in Pao, Fig 4.—Relationship of pretreatment-to-posttreatment change in Paco,
and 1-s forced expiratory volume (FEV,). Numbers indicate more than and 1 -s forced expiratory volume (FEV,). Numbers indicate more than
one value at same location. one value at same location.

35
10
30 x
X 6 XX X
25 X XXX
X
x.. x
20 XX
.Sx_ X X XXX
I XXX
I X XX
15 52XX xx x x x X X
s -6
XX X2
5
10 i 2 X
XX
XXX
X
XX

X X
-10
22 X x

5 X X X XX X
! xx Sx -14
Jexxx X
0
-T~"","~*—-x-x--
x
-18
I im
I X
-5t X
-22
X
-10 I »
-26

-15 -30 -I
I I I
-0.6 0 0.6 1.4 2.2 3.0 -0.6 0 0.6 1.4 2.2 3.0
-1.0 -0.2 0.2 1.0 1.6 2.6 -1.0 -0.2 0.2 1.0 1.8 2.6

Change in FEV,, L Change in FEV,, L

Downloaded from www.jama.com by guest on November 25, 2009

change in PaC02 is compared with
that for FEV, in Fig 4 (correlation
coefficient, —.0821). There was an
inconsistent increase or decrease in
the Paco2 with specific improvements
in the FEV,. Also, 17 patients had
improved their FEV, (0 to 1.8 L) but
had virtually unchanged Paco2 val¬
ues.
COMMENT
Many have advocated the useful¬
ness of arterial blood gas analyses in
acute bronchial asthma to objectively
stage severity," despite the lack of
verification of their value. This stag¬
ing system grades asthma severity
based on the degree of hypoxemia and
hypocarbia or hypercarbia. Our data,
and those of others,"8 have consistent¬
ly shown a relatively poor correlation
between the percent-of-predicted
FEV, and the Pao2 or the Paco2.
Furthermore, we have shown (Tables
1 and 2) that the mean Pa02, PaC02,
and pH in each of our three clinical
outcome groups were not statistically
different. The PEFR and the FEV,,
however, have consistently discrimi¬
nated among these same clinical
groups. Thus, arterial blood gases are
not absolutely reflective of asthma
severity when compared with pulmo-
1. Nardell EA, Slate JL, Westphal DM, et al:
Asthma, in Wilkins EW, Dineen JJ, Moncure AC
(eds): Massachusetts General Hospital Textbook
of Emergency Medicine. Baltimore, Williams &
Wilkins Co, 1979, pp 138-145.
2. Ungar JR: Respiratory emergencies associ-
ated with bronchial asthma, in Schwartz GR,
Safar P, Stone JH, et al (eds): Principles and
Practices of Emergency Medicine. Philadelphia,
WB Saunders Co, 1978, pp 862-867.
3. McKenzie SA, Edmunds AT, Godfrey S:
Status asthmaticus in children. Arch Dis Child
1979;54:581-586.
4. McFadden ER, Lyons HA: Arterial blood
gas tension in asthma. N Engl J Med 1968;
278:1027-1032.
5. Tai E, Read J: Blood-gas tensions in bron-
chial asthma. Lancet 1967;1:644-646.
6. Rebuck AS, Read J: Assessment and man-
agement of severe asthma. Am J Med 1971;
Downloaded from www.jama.com by guest on November 25, 2009
nary function testing and clinical out¬
come.
Extreme airway obstruction and/or
fatigue, however, may be associated
with eventual elevation of the Paco2
and a correspondent worsening of
arterial hypoxemia. The specific pul¬
monary function criteria (Table 4)
that will indicate those patients at
risk include an absolute FEV, below
1.0 L or a PEFR below 200 L/min.
Sixteen percent of patients who did
well after discharge from the emer¬
gency department had come there
with hypercarbia and/or severe hy¬
poxemia. There is no question that
these patients had initially severe
asthma, but the most important man¬
agement consideration was the degree
of reversibility of the attack. If these
patients have specific improvement in
their pulmonary function testing
(posttreatment PEFR above 300 L/
min or FEV, above 2.1 L), they may be
confidently discharged from the
emergency department.1,1 Also, once
the FEV, is above 1.0 L or the PEFR
above 200 L/min, there is no need for
repeated arterial blood gas analysis.
Furthermore, there were 43 pa¬
tients with severe impairment in pul¬
monary function (FEV,, <25% of
predicted value, Fig 2) who did not
References
51:788-798.
7. Kelsen et al:
SG, Kelsen DP, Fleegler BF,
Emergency room assessment and treatment of
patients with acute asthma. Am J Med 1978;
64:622-628.
8. Murray AB, Hardwick DF, Pirie GE, et al:
Assessing severity of asthma with Wright peak-
flow meter. Lancet 1977;1:708.
9. American Thoracic Society: Chronic bron-
chitis, asthma, and pulmonary emphysema: A
statement by the Committee on Prognostic Stan-
dards in Non-tuberculous Respiratory Disease.
Am Rev Respir Dis 1962;85:762-768.
10. Piafsky KM, Ogilvie RI: Dosage of theo-
phylline in bronchial asthma. N Engl J Med
1975;292:1218-1222.
11. Nowak RM, Gordon KR, Wroblewski DA,
Spirometric evaluation of acute bronchial
et al:
asthma. J Am Coll Emerg Physicians 1979;
8:9-12.
have elevation of the Paco2. Sole
reliance on arterial blood gas analysis
in these patients would underesti¬
mate the severity of the degree of
airway obstruction.
The change in Pao2 or PaC02 does
not accurately reflect change in pul¬
monary function testing. Indeed,
some patients, despite improved air
flow with treatment, exhibit a tran¬
sient worsening of their hypoxemia.
This is thought to be secondary to the
dilation of previous constricted pul¬
monary vessels produced by sympa-
thomimetics and/or theophylline.15'6
Thus, repeated arterial blood gas
analysis in those patients, improving
to an FEV, above 1.0 L or a PEFR
above 200 L/min, is of no value in
further assessment of asthma severi¬
ty.
In summary, arterial blood gas
analysis correlates poorly with pul¬
monary function testing and is thus
of little predictive value in determin¬
ing patients' outcome. Hypercarbia
and/or severe hypoxia is seen only in
patients exhibiting specific pulmo¬
nary function testing abnormalities.
These include an FEV, below 1.0 L
(25% of predicted value) or a PEFR
below 200 L/min (30% of predicted
value).
12. Morris JF, Koski A, Johnson LC: Sprio-
metric standards for healthy nonsmoking adults.
Am Rev Respir Dis 1971;103:57-67.
13. Cherniack RM, Raber MB: Normal stan-
dards for ventilatory function using an auto-
mated wedge spirometer. Am Rev Respir Dis
1972;106:38-46.
14. Nowak RM, Pensler ML, Sarkar DD, et al:
Comparison of peak expiratory flow and FEV,
admission criteria for acute bronchial asthma.
Ann Emerg Med 1982;11:64-69.
15. Rees HA, Millar JS, Donald KW: A study
of the clinical course and arterial blood gases of
patients in status asthmaticus. Q J Med 1968;
148:541-561.
16. Read J, Tai E: Response of blood gas
tensions to aminophylline and isoprenaline in
patients with asthma. Thorax 1967;22:543-549.