clinical investigations

An Algorithm for the Interpretation of

Cardiopulmonary Exercise Tests*
William L. Eschenbacher, M.D., EC.C.P.; and Anthony .\lannina, M.D.

We have developed and are using an algorithm for the straightforward technique for arriving at an interpretation
interpretation of cardiopulmonary exercise tests that are of these tests has resulted in a more consistent approach to
performed in our Pulmonary Diagnostic Service Depart- interpretation and an excellent teaching guide for physicians
ment. As its decision points, this algorithm uses routinely and technicians. (Chest 1990; 97:263-67)
obtained measurements from the results of these exercise
tests, such as Vo., Vco,, VJo:, SaO,, HR, and AT. Using the
algorithm results in an objective determination of limitation
to exercise and allows for the differentiation between MV\' =maximal voluntary ventilation; AT= anaerobic thresh-
old; VR=ventilatory reserve; HRR=heart rate response;
pulmonary and cardiac or circulatory limitation. This

E xercise testing has been used by cardiologists and

pulmonologists for the evaluation of the heart
HR, and Sa02 • With this algorithm, we can evaluate
a patient's response to exercise and determine whether
and lungs under the physiologic stress of increasing there was limitation to exercise lwcause of the lungs,
external workloads. Cardiac stress tests were devel- the heart, or both.
oped to evaluate possible ischemic changes that may
MATERIALS AND 1\IETIIOI>S
occur as the pressure-pulse product or the modified
tension time index is increased. 1 During a cardiac Cardiopulmonary t•xerdst' h'sts wt•n· ordered hy pulmonary
physicians in our division and pt•rlimned on patients with primary
stress test, changes in the ECG and blood pressure
pulmonary dist'ases: chronic obstrnl'livt' pulmonary diSt'aSt': dilfust'
are monitored during increasing workloads. Cardio- intt'rstitial fibrosis: pulmonary vaseular diseast•; <ll't·upational pul-
pulmonary exercise tests in addition can help deter- monary dist•ase: primary pulmonary hypPrtension: t'k. Each pa-
mine other potential limiting factors to exercise: the tit•nt had a full st't of st:mdard pulmonary funl'lion tt'sts lll'fmt• tilt'
lungs; or the heart; or both. Thus, in addition to ext'rdSt· tt•st (flow/volumt' loop: lung volumes: Dt"; 15-st'l111UI MVV
using a pneumotaehograph-bast•d pulmonary analyzt•r (Medit·al
monitoring the ECG and blood pressure, cardiopul-
Craphics syslt•m 1070) and a body plethysmo~raph (l\lt•dical Craph-
monary exercise tests monitor exhaled oxygen, carbon ics systt'm H»l5).
The cardiopulmonary t'Xerdse tPsts wert• J>t'rlimm•d using a
For editorial comment see page 257 treadmill (Marquette Electronil's st•rit•s 182.5) and an ext•rl'ise
system analyzer li>r the analysis of exhaled gases and t•xhaled
wntilation (Medil'al Graphics system 2001). Tlw protoc.11l uSt•d IC>r
dioxide, VE, and arterial blood gas tensions or oxygen
thest• studit•s was t•ither tlw pulmonary pmt<K.1ll or tlw low
saturation (or both). perli>rmant·t• prot<K.1>1 as outlim•d in tlw manual li>r tlw tn•admill
Although much has been written regarding the (Marquetlt• Electronic.-s, hll'). In addition, t•lt•l'lnK.·ardiographil'
basic physiologic responses to exercise, little infor- monitoring (Eaton Medical Croup nuKit•l C-2700) and noninvasive
mation has been provided for a routine and consistent oximetry (llt•wlett-Pat·kanlnuKiei47201A) wen• pt•rlin·nwd on t•ach
palit•nt during the ext•rl'ise protcK·ol. During tlw test. hlocKI pn•ssurt·
approach to the interpretation of a cardiopulmonary
was also nwasurt'd nsin~ a portabl .. sphygmomanomt'lt·r. Tlw Vt:,
exercise test. One example of an interpretative tech- rPspiratory rate. tidal volunw, \·o,, and \'co, wt•n· mt•asurt'd on a
nique is shown in a recent book by \Vasserman and breath-by-breath basis.
others. 2 We have developed and are currently using a Tt•stin~ was terminatt'd wht•n tht• patient si~nallt·d exhaustion,
simple algorithm for the interpretation of cardiopul- fatigut•, shortness of breath. It•!( pain. or du•st pain or wht•n ST-
St'!(ment ehanges or a eardiae arrhythmia was noted on tlw 12-lead
monary exercise tests. This algorithm is based upon
ECC.
the routinely obtained parameters of \:E, Vo2 , Vco 2 , The fi>llowing paramell'rs wt•re determined li>r t•aeh test and
nSt•d fi>r tlw interpretation of tlw n•sults: \·o,; \·co,; \•t:: VR (VH =I
*From the Pulmonar,· and Critical Care 1\ledidne Dh·ision, llni- - [\'Elnax/prl'dicted l\1\\']): n·ntilatory e<tnival.. nts (\'I·:No,: Vt:A'
versity of Mk·higan f.tedkal Cenlt•r, Ann Arhor.
co,); ox~w·n saturation: IIHH (dist·usst•d subsequt•ntly): ami AT. The
Manuscript re<:eived April HJ; revision aceeptt·d Jmu• 13.
Reprint requests: Dr. f:sclu•nbaclwr; Pulmcmar!J l'rmcticm l~1h, AT was t•xpn•ssed as tht• oxygt•n l'Onsumption at whkh till' v~:t\·o,
Unitlt'rsit!l of Michigan Mc•dical Center; Ann Arbor .J/J/()9-0026 ratio inen•ast•s and was dl'lt•rmined graphic.-ally fi>r t•ach t•xt•rcist•

CHEST I 97 I 2 I FEBRUARY. 1990 263

 I. PULMONARY LIMITAnON TO EXERCISE

HGFEDCBA J KL MN 0 P

II. CARDIAC LIMITAnON

TO EXERCISE >50

FIGURE 1. Algorithm for interpretation of cardiopul-

monary exercise tests. VE, Ventilatory equivalent for
carbon dioxide; S, change in Sa02 ; IS, ischemic
symptoms (chest pain, s:r-segment changes, etc); and
AT%, ratio of AT to Vo2max. For explanation of
Q R S T uv w interpretations, A through W, see Table 1.

test.' Vo.max; he or she will still exhibit some VR. One report suggests
The algorithm that we used for the actual interpretation of the that without pulmonary limitation, this reserve should be greater
results of cardiopulmonary exercise tests is shown in Figure 1. In than 30 percent.• Patients with pulmonary disease may have no
this approach, there were certain parameters that were used as ventilatory reserve left at their vo.max. Patients with less than 30
decision points for the evaluation of pulmonary and cardiac or percent VR are said to have a ventilatory mechanical limitation.
circulatory limitation to exercise. The first parameter that was The next parameter that we used in our algorithm is the VEmaxl
examined was the Vo2 max. This value for the patient should have Vco2 • This value is a good overall determinant of the efficiency of
been more than 90 pen:ent of the predicted maximal value for that the lung as a gas exchange unit. Normally, at maximal exercise, this
patient. The predicted values for vo.max were determined by value will be 25 to 35,' and values above 40 represent an excessive
regression equations and used by Medical Graphics Corp in their ventilation that is necessary to overcome the inability of the lung to
equipment.' When using separate equations for underweight and excrete carbon dioxide due to gas exchange problems. Patients with
normal individuals vs obese individuals, Wasserman et al<.a have values of VEmaxlVco2 of greater than 40 are said to have a gas
shown that subjects without pulmonary or cardiac limitation to exchange abnormality; however, another possibility for an increased
exercise should achieve approximately 100± 10 percent of the ventilatory equivalent is any abnormal drive to ventilation such as
predicted Vo2 max. Therefore, we have arbitrarily set 90 percent of anxiety. Usually, anxiety at the beginning of an exercise study can
predicted Vo2 max as the lower limit of normal. We realize that a result in a ventilatory equivalent that is greater than 40; however,
more statistically appropriate guide could be used (eg, 95 percent as exercise proceeds, this value will decrease as the drive to
confidence interval); however, we have found this limit of90 percent ventilation becomes more dependent upon the factors associated
to be simple to use and easy to teach. If an individual is able to with exercise metabolism. This value will also increase again towards
achieve 90 percent or greater of his or her predicted vo.max, they maximal exercise. The cutoff value that was chosen (40) should take
may still have some pulmonary or cardiac limitation, but obviously, into account the normal increases in this value with maximal
it would be mild in quality to allow them to achieve close to their exercise.'
predicted Vo2rnax. On the other hand, if the patient is not able to Finally, for the determination of pulmonary limitation to exercise,
achieve 90 percent of his or her predicted vo.max, then the we examined the change in Sa02 either by arterial blood gas
pulmonary or cardiac limitation, if present, would be either determinations or by noninvasive oximetry. The acccuracy of
moderate or severe in quality. oximeters in measuring a change in oxygen saturation is ± 2.5 to
The next decision parameter that we used was the VR: ± 3.5 percent (95 percent confidence limits). 7 Therefore, a decrease
VR=[l- (VEmaxlpred MVV)] X 100% in oxygen saturation of more than 4 percent is considered to be
abnormal. In the context of exercise testing, desaturation can occur
where the VEmax is the maximal minute ventilation achieved with most <.'Ommonly in patients with diffusion limitations,• although
exercise, and the predicted MVV is determined by 41 X FEV,.• other pulmonary abnormalities, such as shunts or ventilation-
When an individual without pulmonary limitation exercises to a perfusion mismatching, may result in exercise-associated desatura-

284 Interpretation ol Cardiopulmonary Exercise Tests (Eschenbac~ Mtullllna)

 250
Table 1-Interpretative Results Using Algorithm
ABNORMAL HEART RATE RESPONSE
~ I. Pulmonary limitation In exercise
200
A. No pulmonary limitation or dt•crt'ast'd t•lli•rl or cardiac
limitation
150 B. Mild dilfusion·lype limitation
C. Mild gas exchange ahnorrnality
100
D. Mild gas exchange almormality and dilfnsion·lypt· limitation
E. Mild ventilatory mechanical limitation
F. Mild ventilatory mt·chanieal limitation and dilfnsinn·type
50 limitation
C . Mild ventilatory mechanical limitation and ~as exchange
abnorrnality
0 1000 2000 3000 4000 5000 6000 II. Mild ventilatory mt•chanieal limitation and gas exd1ange
abnormality and dilfnsiorHype limitation
Vo2,mllmln
I. Decrt•ased t•lli•rl nr cardiac limitation
Frct rnE 2. Normal and ahnormal responst·s fo.r increase in IIR J. McKie rale or St'Vert• dillirsiorHypt• limitation
during exercist• when plottt•d against increasing Vo,max. K. McKit·raiP or severt• gas exdlangP ahnorrnality
L. McKlt•rah' or St'H'rt' gas t•whange abnorrnality and dillilsion·
lion . Therel<1re, palit•nls with desatnration with exerdse an• said to type limitation
have a dilfnsion·lype limitation. !\1. l\lcKierate or spven· n•ntilatory nwehanieallirnitatiun
Afte r estahlishing whether the patie nt either dot•s or does not N. McKierale or St'\'t'rt' n •ntilatory mt'l·hanical limitation and
han• a pulmonary limitation to t•xt•rcise , wt• tht•n determined dilfnsiorHypt• limitaticm
whether a t•ardiac or circulatory limitation to exe rcise may t•xist. 0. McKit•ralt• or st•n•rt• n •ntilatnry rnt•chanicallirnilation and gas
Onr first dt'cision paranlt'lt·r li1r this determination was the IIRR. t•xdlange ahnorrnality
Normally. tlwre should he a linear relationship l)t'tween oxygen P. MtKierale or seven• ventilatory mt•chanieallimilalion and gas
consumption and IIR .'' If tllt'rt' art· problems with tlw heart as a t'xchangt• almorrnality and dill'nsiorHypt• limitation
pump (eg, a cardiomyopathy), the IIRH at any oxygen consumption II. Cardiac or circulatory limitation to exerdSt·
may he inappropriately increased. As shown in Figure 2. the normal Q. McKieralt• or st•n•n• cardiac "pump" limitation (cardiomyopa·
increases in II H with t•xercist• are shown in tlw lowt•r shad .. d area. thy ; deconditioning)
An ahnormal increase in II H with t•xercist• is shown in tlw shadt•d R. Cardiac "pnrnp" limitation (eanliomyopathy: det,mditioning)
area ahow tht· normal area. \\{> can caknlalt• IIRH hy the li1llowing S. Cardiac "pump" limitation and circulatory limitation (pulmo·
limnnla: nary vaseular or peripheral vaseular tliSt•ase. or " pump" limitation)
T. l\lcKieralt· or S<·ven· pulmonary limitation (S<'t' J through I') or
IIRR = (IIHmax -IIHresl)~ ' o , max- Vo,resl)
IJC~•r elli•rl
where IIHmax is tlw heart rate at m;tximal t•xercise, HHresl is tht• L1 . 1\o oll\·ious eardiac or cireulatory limitation
heart ralt• at rest. Vo,max is till' maximal oxyge n l1>11S11mption in V. Circulatory limitation (pulmonary \·ast·ular or periplu·ral ,·as·
liters Jlt'r minnie. and \•o,resl is the oxygt•n l1lllsnmplion at rest in eular dist'aSt', or "pump" limitation)
liters per minnie. Normally, this ratio (IIHH) will he 25 to 3.5 lin· \\' Isdlt'mic heart diseast•
trained indh·idnals and 3.5to 451i•r sedentary or untrained subjects.
Patients with a cardiomyopathy or det1mditioning nr otlwr "cardiac
pump" prohlem may have a IIRH of more than 50. This value of 50
as a cntolf point hetween normal individuals and individuals with
ht'arl dist•ase or det'lnditioning was chosen arbitrarily hnt was Table 2-Use of Algorithm for 52-Year-Old Woman u·ith
based upon the reported data li•r normal individuals.'' Tht'St' valtlt's S11spected P11lrrwnary Vasc11lar Di•ease*
appl)· only to patients who are nul taking a medication that t1mld
hlnt:k the IIRR In t'xercist·. Fc•r example , if a patient is receiving Maximal
~-adrenergic blockade. his nr llt'r HHH In ext•rcise can he dt•· Data lkst EXt•rl'iS<•
creased;"' hm\·en•r, using this paramt'lt•r. we can also identil~ · \ ·o,. mVmin 2SO I. OSI
patients who did nul give a maximal dli•rl and had a rednt·l'd \'co,, rnVmin 240 1.2.5H
\ 1o,max. Their IIHH will still he normal. Also, patit•nts who did not
VE . IJmin H.6 62.H
achieve tlwir Vo,max because nf only Vt'ntilatory limitation shnnld \ · E~ ' co, 411.:3 .')().2
also have a normal HHH; howt•\'t'r, thnst' patients with hoth a SaO, % H.5 SH
ventilatory and cardiac nr circulatory limitation can he identifit'd IIR , hprn 9.'3 15.'3
hy having an increased IIHH and a pulmonary limitation In exerdse AT, mVmin 421;
identified hy the parameters listed previnnsl):
If the H RR is normal (<50), ,,.,. the n dl'lermined if the t'Xercist• *Baseline pulmonary function lt•sts: FEV,. 2.61; lis (1021lt.'rcenl of
lest was stopped hecanst• of electrot:ardiographic changes (de- prediclt'd); FVC, 3.R6 L (Ill pt'rt't'nl of prt•dieted); and Dt", 12.24
pressed ST segments, arrhythmias, l'lc) or l)t'('anse of dwsl pain nr rnVrnin/rnm llg (46 pe rcent of pn•dided). Caleulations: Vo,maxl
hypotension. These all can suggest an ischemic cardiac limitation Vo,nutx predielt'd =I ,OS I/I ,457 = 74 pt•rePnt: VH = 1- (\'nnaxlpn·-
to exercise (IS= ischemit· symptoms). diclt'd MVV)={I-(62.H/(2.61lx4l)l}x 100 = 41.3 J)l'rct'nl; IIHH=
Finally, we examined tlw AT. We delt•rmined the ratio of the (IIHmax -IIHt•st)No,rnax- Vo, rest) = (15:3- H3)/(I.Oill- .2SO) =
oxygen consumption at the AT with the actual Vo,nuLx achieved or 74.9; AT%= (\ ' o,AT~·o,max) or (\'o,ATA'o,m<LX prt•didt•d) X 100
the predicted Vo,rmLx. If this ratio is less than 40 percent, this J>t'rcenl = (421;/l,llll) X 100 or (42S/1.457) X J()() = :3!).6 or 2H.4 pt•r-
suggests a circulatory nr "pump" limitation to t•xerdse. Normally. eenl. l ' sing algorithm: Land S: mcKit•ratt• or St'\'t'rt' gas exchangt•
this ratio is 55 to 60 percent.' It can l>t' decreased because the and dilfusion·tnw pulmonary limitation and eardiac " pnmp" and
exercising mnsdes have switched over to anaerobic nwtahnlism at l'irculatory limitation ("•mpatihle with pulmonary vascular diS<•aS<·).

CHEST I 97 I 2 I FEBRUARY. 1990 265

an earlier workload because of the inability of the heart or the nary faculty in several instances. An example of the
cin:ulation to provide the net:essary oxygen for aerobic metabolism. use of the algorithm is shown by interpreting the
If the AT is not reached at all, this suggests that there is either a
pure moderate to severe pulmonary limitation to exercise, a mixed results of an exercise test for a patient with suspected
pulmonary and cardiac limitation, or poor effi1rt by the patient. pulmonary vascular disease (Table 2). In this example,
Using the algorithm shown in Figure 1, we were able to arrive at the woman with possible pulmonary vascular disease
the different interpretative diagnoses that are listed in Table 1. Each has normal pulmonary mechanics as part of her
letter, A through W, represents the end of a pathway in the baseline pulmonary function tests but a decrease in
algorithm.
her value for Dco. By exercise testing, we determine
RESULTS that she was not able to achieve her predicted
Vo2 max, but she had a normal VR; however, the
Using this algorithm, we have interpreted more
patient had an increased VEmaxlVco2 , and she had
than 20 cardiopulmonary exercise tests that were
desaturation with exercise. In addition, she had an
performed in our laboratory. We have found that the
abnormal HRR (74.9), and her AT was less than 40
interpretation of the tests when using this algorithm
percent of either her Vo 2 max or her predicted
not only gave a more consistent result but also was an
Vo2max. These factors taken together suggest that she
improvement upon the interpretation by our pulmo-
has both pulmonary and cardiac limitation to exercise.
We have compared our algorithm with the algorithm
Table 3-Compariaon of Use of Our Algorithm with
Algorithm ofWoasennan et fill* provided by Wasserman and others. 2 We have taken
11 representative studies from our laboratory and used
Interpretation of our algorithm and the algorithm of Wasserman and
Case Our Interpretation Wasserman et al• associates2 to obtain interpretations of these tests. The
1 Normal or decreased effort Normal results of this comparison is shown in Table 3 (the
2 Normal or decreased effort Obesity usually with low example shown in Table 2 is patient 11). We were
breathing reserve pleased to find that for nine of the 11 examples, the
3 Decreased effort Obesity usually with low
two algorithms gave quite similar interpretations,
breathing reserve
4 Mild ventilatory mechanical Obesity usually with low especially in suggesting the same organ system that
limitation breathing reserve could be limiting. In the two exceptions (patients 8
5 Moderate or severe Obstructive pulmonary and 9), we had suggested that some deconditioning or
ventilatory mechanical disease cardiac "pump" limitation could have existed, whereas
limitation by the algorithm of Wasserman et al2 the limitation
6 Moderate or severe Obstructive pulmonary
ventilatory mechanical disease
could have been due to obesity, poor effort, or
limitation and gas musculoskeletal disorder. It is possible that both
exchange abnormality and algorithms may be right in these two patients, in that
diffusion-type limitation these diagnoses are not mutually exclusive.
7 Cardiac "pump" limitation Early cardiovascular
(cardiomyopathy; disease DISCUSSION
demnditioning)
8 Cardiac ''pump" limitation Obesity usually with low We have described the use of an algorithm that will
(canliomyopathy; breathing reserve simplify the interpretation of a cardiopulmonary ex-
deconditioning) ercise test. As its decision points, this algorithm uses
9 Cardiac "pump" limitation Poor effort or musculo-
parameters that are routinely obtained during the
(canliomyopathy; skeletal disonler
demnditioning)
performance of an exercise test that includes the
10 Moderate or severe gas Pulmonary vascular measurement of exhaled gases. This method allows
exchange abnormality disease without right- for the interpretation of a cardiopulmonary exercise
with cardiac "pump" to-left shunt test in a straight-forward manner, so that an objective
limitation and circulatory
assessment of limitation to exercise can be deter-
limitation (pulmonary
vascular or peripheral mined. Using this method, we can also differentiate
vascular disease or between a pulmonary and a cardiac or circulatory
"pump" limitation) limitation to exercise. Previously, other than the algo-
11 Moderate or severe gas Early pulmonary rithm from Wasserman and colleagues, 2 only general
exchange abnormality and disease; pulmonary
guidelines were available for the interpretation of
diffusion-type limitation vascular disease
with cardiac "pump" these exercise tests. 11 - 13 Use of this algorithm can result
limitation and circulatory in a specific interpretation for the patient based upon
limitation (pulmonary objective test results. We believe that this approach
vascular or peripheral will be quite useful for the evaluation of patients with
vascular disease, or known pulmonary or cardiac disease (or both) or for
"pump" limitation)
the patient with unexplained dyspnea, especially when

Interpretation ol cardiopulmonary Exen:ise Tests (Eschenbache( Mannina)

the results are combined with routinely obtained valve and chronic obstructive pulmonary diseases. Chest 1983;
83:446-53
pulmonary function studies. From a teaching perspec-
7 Ries AL, Farrow JT, Clausen JL. Accuracy of two ear oximeters
tive, this algorithm has resulted in a greater under- at rest and during exercise in pulmonary patients. Am &v
standing and more consistent interpretations of car- Respir Dis 1985; 132:685-89
diopulmonary exercise tests by our medical residents, 8 Owens GR, Bogers RM, Pennock BE, Levin D. The diffusing
pulmonary fellows, and pulmonary faculty. capacity as a predictor or arterial oxygen desaturation during
exercise in patients with chronic obstructive pulmonary disease.
N Eng) J Med 1984; 310:1218-21
REFERENCES 9 Legge BJ, Banister EW. The Astrand-Ryhming nomogram
1 Ellestad MH. Stress testing: principles and practices. Philadel- revisited. J Appl Physiol1986; 61:1203-09
phia: FA Davis, 1986:22,136 10 Joyner MJ, Freund BJ, Jilka SM, Hetrick GA, Martinez E, Ewy
2 Wasserman K, Hansen JE, Sue DY, Whipp BJ. Principles of GA, et al. Effects of beta-blockade on exercise capacity of
exercise testing and interpretation. Philadelphia: Lea and Febi- trained and untrained men: a hemodynamic comparison. J Appl
ger, 1987 Physiol1986; 60:1429-34
3 Wasserman K, Whipp BJ. Exercise physiology in health and 11 Wait J. Cardiopulmonary stress testing: a review of noninvasive
disease. Am Rev Respir Dis 1975; 112:219-49 approaches. Chest 1986; 90:504-10
4 Hansen JE, Sue DY, Wasserman K. Predicted values for clinical 12 Weber KT, Janicki JS, McElroy PA, Reddy HK. Concepts and
exercise testing. Am &v Hespir Dis 1984; 129:S49-55 applications of cardiopulmonary exercise testing. Chest 1988;
5 Miller WF, Scacci R, Gast LR. Laboratory evaluation of 93:843-47
pulmonary function. Philadelphia: JB Lippincott Co, 1987:300 13 Neuberg GW, Friedman SH, \\\!iss MB, Herman MV. Cardio-
6 Nery LE, Wasserman K, French W. Oren A, Davis JA. Contrast- pulmonary exercise testing: the clinical value of gas exchange
ing cardiovascular and respiratory responses to exercise in mitral data. Arch Intern Med 1988; 148:2221-26

Plan to Attend ACCfPs

56th Annual Scientific Assembly

Toronto, Ontario, canada

October 22-26, 1990

CHEST I 97 I 2 I FEBRUARY, 1990 267