- Email: [email protected]
Predischarge maximal exercise test identifies risk for cardiac death in patients with acute myocardial infarction
Predischarge Maximal Exercise Test Identifies Risk for Cardiac Death in Patients with Acute Myocardial Infarction Jens Rokkedal Nielsen, MD, Hans Mickley, MD, Else Marie Damsgaard, MD, and Anders Frgland, MD
A maximal exercise test was performed in 54 patients with acute myocardial infarction (AMI) before discharge and in 49 age-matched control subjects. The long-term prognosis was assessed after an average follow-up of 7.6 years in AMI patbnts and 5.6 years in control subjects. The maximal work capacity and systolic blood pressure imrease in AMI patients was 59% that of control subjects (p
R
isk stratification of survivorsof myocardial infarction offers 2 major benefits. High-risk patients can be identified for closer medical follow-up and further diagnostic and therapeutic procedurescan be instituted to alter their poor prognosis. Patients at low risk may be spared needlessinvasive studies and unwarranted physical restrictions. The prognostic value of early exercise testing after acute myocardial infarction (AMI) has not been adequately established.Only few investigators have shown a positive correlation betweenthe developmentof significant ST-segment depression and l-year’ as well as long-term mortality.2 Several investigators have demonstrated that ST-segment depression is significantly prognostic for subsequentcardiac events3-lo However, other investigators have not been able to find this relation.“-is Davidson and DeBusk7 found a maximal workload <4 METS predictive of subsequent cardiac events in patients without significant ST depressionor angina pectoris during exercisetesting after AMI. Other investigators have noted that patients with a low maximal workload have an increasedrisk of subsequent cardiac events15or long-term mortality.i6 However, Fioretti et ali4 found no difference in work capacity between survivors and nonsurvivorsof AMI. Finally, a low increase or decreasein systolic blood pressure during exercisetesting early after AM1 has been predictive of future cardiac event&l0 or mortality.3~14~17 Again, these results have been contradicted by others.2,9 This study aims to clarify the following questions about a maximal exercise test performed before discharge by patients with AMI: By how much do decreasesin maximal work capacity in AM1 patients differ from an age-matched randomly selected control group? What is the long-term prognostic value of maximal work capacity, systolic blood pressureincreaseand ST-segmentshifts? METHODS
From the Department of Internal Medicine, Fredericia Hospital, Fredericia, Denmark. This study was supported in part by grants from the Danish Heart Foundation, Denmark. Manuscript received June 20, 1989; revised manuscript received and accepted September 5,1989. Address for reprints: Jens Rokkedal Nielsen, MD, Department of Medicine B, Odense University Hospital, DK-5000 Odense, Denmark.
Patients: During a 15month period ending in 1978, 61 consecutive patients <67 years of age experienced AMI. Seven were excluded (3 did not want to participate, 1 was referred to another hospital and 3 were unable to exercisedue to orthopedic problems). Only 3 of the remaining 54 patients were women. Control subjeds: The control subjects were agematched men randomly selected from the municipal register. Of the 61 contacted by letter and telephone,49 were included and performed the exercise test during the months of January and March 1980. Twelve were
THE AMERICAN JOURNAL OF CARDIOLOGY JANUARY 15, 1990
149
excluded (6 did not want to participate, contact was not established with 3 and 3 were excluded for noncardiac reasons). Study procedure: AM1 patients and control subjects performed a maximal exercisetest on a bicycle ergometer (Siemens-Elema), beginning with 50 watts (3.3 METS) for 6 minutes, and with increments of 50 watts every 6 minutes until exhaustion or other reasons for stopping the test.is In women the workload given was half that of men. Cuff systolic blood pressure was ob Maximal
work
capacity. --
1.0-- -h.-l--L
-L-----
Max watt>72
(n=33)
0.8
p
a.> t 0.4 3
Max watts72
F z 0.2 z =)O
01
23456789 Observation _SyStolic
period
(years)
blood pressure
L --I--
increase.
A SBP> 30mmHg (n=40)
p< 0.0001
fj , , , , ,y;fyommHg 0123456769 Observation ST-%ment
period
(years)
changes.
ST< tmm (n =37) STZlmm (n=17)
fo.j, , , , , (, , , 0123456789 Observation
period
(years)
PIGURE 1. Kapbn-Meier mdysb demodratesthatbw maximalwerticqmcity(Maxwett)andbwsysteficbbed
presswehcrease(ASBP)atmaximdexercisetestpradbchargearerebtedtesigdcmt~~ef WdbCdei3th.ThiSW~llOttiWCS#WfWST-SegmWltdliftS.
iso
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 65
tained before exercise,in the last minute of each exercise stage,at maximal exerciseand at 1, 2 and 10 minutes in the recovery phase. An electrocardiogram (Vi, V=,, V,) was recorded simultaneously with the blood pressuremeasurements.The exercisetests were carried out when the patients were ready for discharge.All prescribed medications were maintained during the exercise tests. Maximal work capacity was expressed as maximal watts calculated as previously described.i9 Low work capacity of AM1 patients was defined as up to half the mean value of control subjects.Cardiovascular responsewas expressedas an increase in systolic blood pressure.An increasein systolic blood pressureof <30 mm Hg was defined as a low blood pressure response.An ischemic responsewas defined as ST-segment shifts of I1 mm. The predictive value of a positive and a negative test was evaluated for eachvariable, and for the combined variables. Follow-up period: The follow-up period was 7.6 f 0.4 years for the AM1 patients and 5.8 f 0.1 years for control subjects. The endpoints were cardiac and noncardiac death. Data handling and statWeal m Data were collected on precededforms and entered into a computer file. The Student t test was used to compare mean values and a chi-square test was used to compare incidences of discrete variables between groups. KaplanMeier analyses were used to demonstrate if exercise variables were correlated with a significant cumulative probability of cardiac death. A p value <0.05 was deemedsignificant. RESULTS Acute myocardial
PREDISCHARGE: The infarction: mean age of the patients was 55 f 8 years. Thirteen patients (24%) had had a previous myocardial infarction. The mean duration of hospitalization was 20 f 6 days, with a maximal lactic dehydrogenase(LDH) of 1,441 f 981 U/liter. The AM1 was complicated by ventricular fibrillation in 13%,heart failure in 9%, angina pectoris in 6%, atrioventricular block in 4% and supraventricular tachycardia in 6%. No significant differencesbetween the AM1 subgroupswere found in incidenceof previous infarction, duration of hospitalization, maximal LDH and complications. Twenty-one patients were treated with cardioactive drugs: digoxin (13 patients), ,&blocking agents (8 patients) and calcium antagonists (3 patients). Twenty-nine percent of survivors had been treated with at least 1 of theseagents;in those with the endpoints of cardiac and noncardiac death 69 and 14%, respectively, were treated with at least 1 of these agents. This difference was not significant (p
<0.08). Follow-up:
Sixty-six cardiac eventsoccurred in 33 of the AM1 patients during 7.6 f 4 years of follow-up. Thus, a cardiac death occurred in 16, noncardiac death in 7, 16 episodesof reinfarction in 14, 20 episodesof unstable angina pectoris in 15 and 7 episodesof cardiac failure in 5 patients. Exercise tesk The exercisevariables for the control subjectsand for the AM1 subgroupsspecified for survivors and those who died from cardiac or noncardiac
TABLE
I Exercise
Variables
from
AMI Patients,
AMI Subgroups
and Control
Subjects
AMI Subgroups
No. of subjects Maximal work capacity (watts) Systolic blood pressure increase (mm Hg) No. of subjects with STsegment shifts (%) * p
Total
Cardiac Death
54
16
84 f 32* 45 f 20*
17 (31)
t patients
Noncardiac Death
Surwors
7*
319
62 f 277
90f34
94f29
143 f 42
30f
5oi30
50*20
75 f 25
15
8 (50)
8
l(l4)
wth cardiac death in the follow-up
period YS survivors;
deaths during follow-up are listed in Table I. The AM1 patients had a maximal work capacity 59% that of the control subjects, and a systolic blood pressureincrease 59% of the control subjects. Thirty-one percent of the AM1 patients had significant ST-segment shifts compared to 6% of the control subjects. Those who died from cardiac reasonsat follow-up had a maximal work capacity 43% of the control subjects, and a systolic blood pressureincrease 38% of the control subjects.Fifty percent of the patients had significant ST-segment shifts. The exercisevariables were almost equal between AM1 survivors and those who died from noncardiac reasons(Table I). Reasonsfor stopping the exercisetest for AM1 patients and control subjectsare listed in Table II. Serious complications such as VT, VF and death were not observed.Ninety-six percent of the control subjects had a maximal work capacity >lOO watts. No significant correlation was found between maximal work capacity and maximum LDH (r = 0.01, difference not significant, n = 49). Prediction of long-term mortality: In patients with a maximal work capacity 572 watts, cardiac deaths occurred in 60%, and in patients with maximal work ca-
TABLE Patients
* 6 men and 1 woman;
49
(26)
II Reasons for Stopping and Control Subjects
Fatigue Dyspnea Anglna pectoris Arrhythmias Electrocardiographtc Other
3 (6)
3 29 men and 2 women.
changes
the Exercise
Test in AMI
AMI 1%)
Control Subjects (%)
54 24 9 5 4 4
67 29 2 0 0 2
pacity >72 watts cardiac deaths occurred only in 11% (Table III). In patients with a systolic blood pressure increase <30 mm Hg cardiac deaths occurred in 64% and in patients with a blood pressureincrease 230 mm Hg cardiac deaths occurred only in 18%. In patients with significant ST-segment shifts cardiac deaths occurred in 47%,comparedto 23% in those with no significant ST changes.Kaplan-Meier analysis demonstrated that low maximal work capacity and low systolic blood pressure responsewere correlated with significant cumulative probability of cardiac death (Figure 1). This was not the casefor ST-segmentshifts (Figure 1). Fig-
0
0 FIGURE 2. The accuracy efapesRivezmdnegalive predlscltarge exercise ted formti-csdiacmortaliiinacute myocadal infarction (AM) paBe&.
Control Subjects
Maximal
work
capacity
Systolic increase
blood pressure
X ST changes @ Variables
8 40 Y 8 0 1
combined
I’j , , , , , , , , , 0
Predictive
20
40
60
80
1
accuracy of a positive test THE AMERICAN
JOURNAL
OF CARDIOLOGY
JANUARY
15. 1990
151
TABLE
III
Predischarge
Maximal watts 572 >72 Systolic blood pressure increase(mm <30
Exercise
Variables
Predict
Long-Term
Mortality*
Cardiac Death
Noncardiac Death
Survivors
p Value
12 4
2 5
6 25
9
2
3
7
5
28
8
1
8
8
5
22
15
4
15
1
3
16
Hg) <0.005
230 ST-segment 11
changes (mm) NS
<0.001 Negative test * Anegative test occurred shiis. NS = not significant.
when none of thefollowngvariables
were present: maximal
watts 572 watts; systolic blood pressure increase
ST-segment
accordance with the findings of 96% in our control group. Our study showedthat AM1 patients had a maximal work capacity of 59% that of the control subjects. However, patients with endpoints of cardiac death had a maximal work capacity of 43% of control subjects.In survivors and those with an endpoint of noncardiac death, maximal work capacity was 66 and 63%, respeu tively. These results imply that a low maximal work capacity before discharge may be of prognostic value. When a low maximal work capacity was defined as 150% that of control subjectsa Kaplan-Meier analysis showed that low maximal work capacity actually was correlated with a significant cumulative probability of cardiac death. Thus, a predischargemaximal work capacity may be of great clinical value in identifying patients with low and high risk of future cardiac death. A low maximal work capacity may be the result of a de creasedleft ventricular function. This is in accordance with several other studies showing that a low left venDISCUSSION Although maximal exercise testing has rarely been tricular ejection fraction after AM1 is correlated with and mortality as sole endused before hospital discharge, this type of exercisetest future coronary events26*27 has been shown to be safe in this and other studies.20y21point. 13,14,17 In some studies patients who could not exercisefor the Myocardial function can also be judged by exerciseexpected duration of their protocol had an increased induced blood pressure response. In accordance with mortality risk over the next year.15,22,23 If, however,they Fioretti et a1,t4our results suggestthat inability to incould go the full duration of a submaximal workload, crease systolic blood pressure by at least 30 mm Hg the risk was very low. Although these studies seemable expressespoor myocardial function, and significantly to define low- and high-risk groups, no information is predicts long-term mortality (Figure 1). In severalother given concerning the maximal work capacity of the studies the lack of adequate blood pressure response AM1 patients, and their work capacity is not compared early after AM1 was associated with an increased to the normal population. Knowledge of thesedata may l-year mortality13J5J7 and an increase in coronary have important psychological benefits for patients and events.6J0 help cardiologists in the development of postdischarge The bad prognosisassociatedwith exerciseparameguidelines for exercise prescription. Finally, maximal ters indicating a poor left ventricular function may be work capacity might elicit further prognostic informa- caused by large infarcts or may be due to myocardial tion. ischemia causedby multivessel disease.28 In other studies of normal populations that were In our study 6% of control subjectsand 26% of surviage-matchedto AM1 patients, a maximal work capacity vors had significant ST-segmentshifts and a low risk of >I00 watts was found in 92%24and 95%.25This is in cardiac death. The ST-segment changes in this group ure 2 shows the predictive accuracy of a positive and negative test when maximal work capacity, systolic blood pressure increase and ST-segment shifts were evaluated separately or in combination. The predictive accuracy of a negative test for the exercisevariables is fairly high (77 to 88%), especially for the combined exercise variables (95%). Such a low risk was found in 37% of AM1 patients in this study (Table III). Contd subjectz INITIAL: The mean age of the control subjects was 55.8 years. Previous AM1 was reported in 2, angina pectoris in 3, hypertension in 5 and chronic pulmonary disease in 5 of them. Three were treated with digoxin and 3 with /3-blocking agents. Six subsequentcardiac events occurred in 5. Two subjects had a nonfatal AMI, 1 was hospitalized due to angina pectoris and cardiac failure and 2 died from noncardiac reasons.
152
THE AMERICAN
JOURNAL
OF CARDIOLOGY
VOLUME
65
may therefore be causedby myocardial ischemia due to and death. Am J Cardiol 1980;46:909-914. 7. Davidson DM, DeBusk RF. Prognostic value of a single exercise test 3 weeks minor coronary stenosesor good collateralization. Al- after uncomplicated myocardial infarction. Circulation 1980;61:236-242. though when compared to survivors maximal work ca- 6. Jespersen CM, Kassis E, Edeling C-J, Madsen JK. The prognostic value of pacity and systolic blood pressureresponsewere signifi- maximal exercise testing soon after first myocardial infarction. Eur Heart J cantly decreasedin patients with the endpoint of cardiac 9.1985,6:769-772. Krone RJ, Gillespie JA, Weld FM, Miller JP, Moss AJ and The Multicenter death, significant ST-segmentshifts were not. However, Postinfarction Research Group. Low-level exercise testing after myocardial inusefulness in enhancing clinical risk stratification. Circulation 1985; 50% of these patients had significant ST-segmentshifts. farction: 71:80-89. This indicates that at least some of the patients with 10. Murray DP, Salih M, Tan LB, Derry S, Murray RG, Littler WA. Which bad left ventricular function variables during exercise exercise test variables are of prognostic importance post-myocardial infarction? J Cardiol 1988;20:353-363. may have a bad prognosisnot due only to larger infarc- Int 11. Koppes GM, Kruyer W, Beckmann CH, Jones FG. Response to exercise tions, but also due to multivessel disease.In accordance early after uncomplicated acute myocardial infarction in patients receiving no with this thesis no significant correlation was found be- medication: long time follow-up. Am J Cardiol 1980;46;764-769. 12. Jelinek VM, McDonald IG, Rvan WF. Ziffer RW. Clemens A. Gerloff J. tween maximal LDH and maximal work capacity. Assessment of cardiac risk 10 days after uncomplicated myocardial infarction. Br However, some of the patients with ST-segment Med J 1982:284:227-230. 13. Jennings K, Reid DS, Hawkins T, Julian DJ. Role of exercise testing early changes had increased ST elevations. These changes after myocardial infarction in identifying candidates for coronary surgery. Br Med may be proportional to the size of the infarction,29*30 J 1984;288:185-187. and ST elevation appearsto be a bad prognostic mark- 14. Fioretti P, Brewer RW, Simoons ML, Das SK, Bos RJ. Wijns W, Reiber JHC, Lubsen J, Hugenholtz PG. Prediction of mortality in hospital survivors of er,29,30 even though this conclusion has been anticipated myocardial infarction. Comparison of predischarge exercise testing and radionuby others.9J0 elide ventriculography at rest. Br Heart J 1984;52;292-298. This study has shown that in consecutivesurvivors of 15. Weld FM, Bigger JT. Risk stratification by exercise testing and 24-hour electrocardiogram 2 weeks after acute myocardial infarction (abstr). AM1 a predischargemaximal exercisetest is a safe and ambulatory Circulation 1978:58(suppl II):II-198. effective diagnostic procedure for identifying patients 16. Saunamtiki KI, Andersen JD. Prognostic significance of the ST-segment response during exercise test shortly after acute myccardial infarction. Compariwith low and high risk of cardiac death. When the exer- son with other exercise variables. Eur Heart J 1983;4:752-760. cise variables are combined, the predischarge maximal 17. Waters DD, Bosch X, Bouchard A, Moise A, Roy D, Pelletier G, ThCroux P. exercisetest seemsto be of great clinical value, especial- Comparison of clinical variables and variables derived from a limited predischarge exercise test as predictors of early and late mortality after myocardial infarction. ly in identifying AM1 patients at low risk, the predictive JACC 1985:5:1-8. accuracy of a negative test being 95%. Such a low risk 16. Ellestad MH. Stress testing. Principles and practice. Philadelphia: Davis, was found in 37% of AM1 patients in this study. These 1980. 19. Pedersen PK, Nielsen JR. Absolute or relative work load in exercise testingpatients may not require either coronary angiography or significance of individual differences in working capacity. Scan J Clin Lab lmest P-blocking agents. Furthermore, the predischargeexer- 1984:44:635-642. Constant J. Prognostic information from early post-infarction exercise testcise test can be used to guide cardiac rehabilitation and 20. ing. Am J Med 1986;81:655-660. bring to the patient important psychological benefits, 21. Ibsen H, Kjqller E, Styperek J, Pedersen A. Routine exercise ECG three decreasethe period of convalescenceand improve the weeks after acute myocardial infarction. Acta Med Stand 1975;198:463-469. Samtam& KI, Andersen JD. Early exercise test in assessment of long-term return-to-work rate. The AM1 patients at high risk 22. prognosis after acute myocardial infarction. Acta MedScand 1981;209:185-191. should be followed in outpatient clinics and may require 23. Madsen EB, Gilpin E, Ahnves, Henning H, Ross J. Prediction of functional various types of medical treatment such as p blockers capacity and use of exercise testing for predicting risk after acute myocardial infarction. Am J Cardiol 1985;56:839-845. and even coronary angiography with a view of revascu- 24. Grimby G, Wilhelmsen L, Ekstrom-Jodal B. Aerobic power and related larization procedures, factors in a population study of men aged 54. Stand J Clin Lab fmmt REFERENCES 1. ThCroux P, Waters DD, Halphen C, Debaisieux J-C, Mizgala HF. Prognostic value of exercise testing soon after myccardial infarction. N Eng/ J Med 1979;301:341~345. 2. Velasco I, Tormo V, Ferrer LM, Ridocci F, Blanch S. Early exercise test for evaluation of long-term prognosis after uncomplicated myocardial infarction. Eur Heart J 1981;2:401-407. 3. Markiewicz W, Houston N, DeBusk RF. Exercise testing soon after myocardial infarction. Circulation 1977;56:26-31. 4. Smith JW, Dennis CA, Gassmann A, Gaines JA, Staman M, Phibbs B, Marcus FI. Exercise testing three weeks after myocardial infarction. Chest 1979:75:12~16. 5. Sami M, Kraemer H, DeBusk RF. The prognostic significance of serial exercise testing after myocardial infarction. Circulation 1979,60:1238~1246. 6. Starling MR. Crawford MH, Kennedy GT, O’Rourke RA. Exercise testing early after myocardial infarction: predictive value for subsequent unstable angina
1970:26:287-294. 25. Hagerup L. Coronary heart disease. Risk factors in men and women. Acta Med Stand 1974zuppl 557(thesis). 26. De Feyter PJ, Eenige MJ, Dighton DH, Visser FC, Jong J, Roes JP. Prognostic value of exercise testing. Coronary angiography and left ventriculography 6-8 weeks after myocardial infarction. Circulation 1982,66:527-536. 27. Norris RM, Barnaby PF, Brandt PWT, Geary GG, Whitlock RML, Wild CJ, Barrat-Boyes BG. Prognosis after recovery from first acute myocardial infarction: determinants of reinfarction and sudden death. Am J Cardiol1984;53;408413. 26. Sanz G, Gastaner A, Betriu A. Determination of prognosis in survivors of myocardial infarction. A prospective clinical angiographic study. N Engl J Med 1982:306:1065~ 1070. 29. Manner@ D, Bennett ED, Ward DE, Dawkins K, Dancy M, Valantine H, Mehta N. Accurate detection of triple vessel disease in patients with exercise induced ST segment depression after infarction. Br Heart J 1987;57:133138. 30. Grande P, Pedersen A. Myocardial infarct size and cardiac performance at exercise soon after myocardial infarction. Br Heart J /982;47:44-50.
THE AMERICAN JOURNAL OF CARDIOLOGY JANUARY 15, 1990
153
A maximal exercise test was performed in 54 patients with acute myocardial infarction (AMI) before discharge and in 49 age-matched control subjects. The long-term prognosis was assessed after an average follow-up of 7.6 years in AMI patbnts and 5.6 years in control subjects. The maximal work capacity and systolic blood pressure imrease in AMI patients was 59% that of control subjects (p
R
isk stratification of survivorsof myocardial infarction offers 2 major benefits. High-risk patients can be identified for closer medical follow-up and further diagnostic and therapeutic procedurescan be instituted to alter their poor prognosis. Patients at low risk may be spared needlessinvasive studies and unwarranted physical restrictions. The prognostic value of early exercise testing after acute myocardial infarction (AMI) has not been adequately established.Only few investigators have shown a positive correlation betweenthe developmentof significant ST-segment depression and l-year’ as well as long-term mortality.2 Several investigators have demonstrated that ST-segment depression is significantly prognostic for subsequentcardiac events3-lo However, other investigators have not been able to find this relation.“-is Davidson and DeBusk7 found a maximal workload <4 METS predictive of subsequent cardiac events in patients without significant ST depressionor angina pectoris during exercisetesting after AMI. Other investigators have noted that patients with a low maximal workload have an increasedrisk of subsequent cardiac events15or long-term mortality.i6 However, Fioretti et ali4 found no difference in work capacity between survivors and nonsurvivorsof AMI. Finally, a low increase or decreasein systolic blood pressure during exercisetesting early after AM1 has been predictive of future cardiac event&l0 or mortality.3~14~17 Again, these results have been contradicted by others.2,9 This study aims to clarify the following questions about a maximal exercise test performed before discharge by patients with AMI: By how much do decreasesin maximal work capacity in AM1 patients differ from an age-matched randomly selected control group? What is the long-term prognostic value of maximal work capacity, systolic blood pressureincreaseand ST-segmentshifts? METHODS
From the Department of Internal Medicine, Fredericia Hospital, Fredericia, Denmark. This study was supported in part by grants from the Danish Heart Foundation, Denmark. Manuscript received June 20, 1989; revised manuscript received and accepted September 5,1989. Address for reprints: Jens Rokkedal Nielsen, MD, Department of Medicine B, Odense University Hospital, DK-5000 Odense, Denmark.
Patients: During a 15month period ending in 1978, 61 consecutive patients <67 years of age experienced AMI. Seven were excluded (3 did not want to participate, 1 was referred to another hospital and 3 were unable to exercisedue to orthopedic problems). Only 3 of the remaining 54 patients were women. Control subjeds: The control subjects were agematched men randomly selected from the municipal register. Of the 61 contacted by letter and telephone,49 were included and performed the exercise test during the months of January and March 1980. Twelve were
THE AMERICAN JOURNAL OF CARDIOLOGY JANUARY 15, 1990
149
excluded (6 did not want to participate, contact was not established with 3 and 3 were excluded for noncardiac reasons). Study procedure: AM1 patients and control subjects performed a maximal exercisetest on a bicycle ergometer (Siemens-Elema), beginning with 50 watts (3.3 METS) for 6 minutes, and with increments of 50 watts every 6 minutes until exhaustion or other reasons for stopping the test.is In women the workload given was half that of men. Cuff systolic blood pressure was ob Maximal
work
capacity. --
1.0-- -h.-l--L
-L-----
Max watt>72
(n=33)
0.8
p
a.> t 0.4 3
Max watts72
F z 0.2 z =)O
01
23456789 Observation _SyStolic
period
(years)
blood pressure
L --I--
increase.
A SBP> 30mmHg (n=40)
p< 0.0001
fj , , , , ,y;fyommHg 0123456769 Observation ST-%ment
period
(years)
changes.
ST< tmm (n =37) STZlmm (n=17)
fo.j, , , , , (, , , 0123456789 Observation
period
(years)
PIGURE 1. Kapbn-Meier mdysb demodratesthatbw maximalwerticqmcity(Maxwett)andbwsysteficbbed
presswehcrease(ASBP)atmaximdexercisetestpradbchargearerebtedtesigdcmt~~ef WdbCdei3th.ThiSW~llOttiWCS#WfWST-SegmWltdliftS.
iso
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 65
tained before exercise,in the last minute of each exercise stage,at maximal exerciseand at 1, 2 and 10 minutes in the recovery phase. An electrocardiogram (Vi, V=,, V,) was recorded simultaneously with the blood pressuremeasurements.The exercisetests were carried out when the patients were ready for discharge.All prescribed medications were maintained during the exercise tests. Maximal work capacity was expressed as maximal watts calculated as previously described.i9 Low work capacity of AM1 patients was defined as up to half the mean value of control subjects.Cardiovascular responsewas expressedas an increase in systolic blood pressure.An increasein systolic blood pressureof <30 mm Hg was defined as a low blood pressure response.An ischemic responsewas defined as ST-segment shifts of I1 mm. The predictive value of a positive and a negative test was evaluated for eachvariable, and for the combined variables. Follow-up period: The follow-up period was 7.6 f 0.4 years for the AM1 patients and 5.8 f 0.1 years for control subjects. The endpoints were cardiac and noncardiac death. Data handling and statWeal m Data were collected on precededforms and entered into a computer file. The Student t test was used to compare mean values and a chi-square test was used to compare incidences of discrete variables between groups. KaplanMeier analyses were used to demonstrate if exercise variables were correlated with a significant cumulative probability of cardiac death. A p value <0.05 was deemedsignificant. RESULTS Acute myocardial
PREDISCHARGE: The infarction: mean age of the patients was 55 f 8 years. Thirteen patients (24%) had had a previous myocardial infarction. The mean duration of hospitalization was 20 f 6 days, with a maximal lactic dehydrogenase(LDH) of 1,441 f 981 U/liter. The AM1 was complicated by ventricular fibrillation in 13%,heart failure in 9%, angina pectoris in 6%, atrioventricular block in 4% and supraventricular tachycardia in 6%. No significant differencesbetween the AM1 subgroupswere found in incidenceof previous infarction, duration of hospitalization, maximal LDH and complications. Twenty-one patients were treated with cardioactive drugs: digoxin (13 patients), ,&blocking agents (8 patients) and calcium antagonists (3 patients). Twenty-nine percent of survivors had been treated with at least 1 of theseagents;in those with the endpoints of cardiac and noncardiac death 69 and 14%, respectively, were treated with at least 1 of these agents. This difference was not significant (p
<0.08). Follow-up:
Sixty-six cardiac eventsoccurred in 33 of the AM1 patients during 7.6 f 4 years of follow-up. Thus, a cardiac death occurred in 16, noncardiac death in 7, 16 episodesof reinfarction in 14, 20 episodesof unstable angina pectoris in 15 and 7 episodesof cardiac failure in 5 patients. Exercise tesk The exercisevariables for the control subjectsand for the AM1 subgroupsspecified for survivors and those who died from cardiac or noncardiac
TABLE
I Exercise
Variables
from
AMI Patients,
AMI Subgroups
and Control
Subjects
AMI Subgroups
No. of subjects Maximal work capacity (watts) Systolic blood pressure increase (mm Hg) No. of subjects with STsegment shifts (%) * p
Total
Cardiac Death
54
16
84 f 32* 45 f 20*
17 (31)
t patients
Noncardiac Death
Surwors
7*
319
62 f 277
90f34
94f29
143 f 42
30f
5oi30
50*20
75 f 25
15
8 (50)
8
l(l4)
wth cardiac death in the follow-up
period YS survivors;
deaths during follow-up are listed in Table I. The AM1 patients had a maximal work capacity 59% that of the control subjects, and a systolic blood pressureincrease 59% of the control subjects. Thirty-one percent of the AM1 patients had significant ST-segment shifts compared to 6% of the control subjects. Those who died from cardiac reasonsat follow-up had a maximal work capacity 43% of the control subjects, and a systolic blood pressureincrease 38% of the control subjects.Fifty percent of the patients had significant ST-segment shifts. The exercisevariables were almost equal between AM1 survivors and those who died from noncardiac reasons(Table I). Reasonsfor stopping the exercisetest for AM1 patients and control subjectsare listed in Table II. Serious complications such as VT, VF and death were not observed.Ninety-six percent of the control subjects had a maximal work capacity >lOO watts. No significant correlation was found between maximal work capacity and maximum LDH (r = 0.01, difference not significant, n = 49). Prediction of long-term mortality: In patients with a maximal work capacity 572 watts, cardiac deaths occurred in 60%, and in patients with maximal work ca-
TABLE Patients
* 6 men and 1 woman;
49
(26)
II Reasons for Stopping and Control Subjects
Fatigue Dyspnea Anglna pectoris Arrhythmias Electrocardiographtc Other
3 (6)
3 29 men and 2 women.
changes
the Exercise
Test in AMI
AMI 1%)
Control Subjects (%)
54 24 9 5 4 4
67 29 2 0 0 2
pacity >72 watts cardiac deaths occurred only in 11% (Table III). In patients with a systolic blood pressure increase <30 mm Hg cardiac deaths occurred in 64% and in patients with a blood pressureincrease 230 mm Hg cardiac deaths occurred only in 18%. In patients with significant ST-segment shifts cardiac deaths occurred in 47%,comparedto 23% in those with no significant ST changes.Kaplan-Meier analysis demonstrated that low maximal work capacity and low systolic blood pressure responsewere correlated with significant cumulative probability of cardiac death (Figure 1). This was not the casefor ST-segmentshifts (Figure 1). Fig-
0
0 FIGURE 2. The accuracy efapesRivezmdnegalive predlscltarge exercise ted formti-csdiacmortaliiinacute myocadal infarction (AM) paBe&.
Control Subjects
Maximal
work
capacity
Systolic increase
blood pressure
X ST changes @ Variables
8 40 Y 8 0 1
combined
I’j , , , , , , , , , 0
Predictive
20
40
60
80
1
accuracy of a positive test THE AMERICAN
JOURNAL
OF CARDIOLOGY
JANUARY
15. 1990
151
TABLE
III
Predischarge
Maximal watts 572 >72 Systolic blood pressure increase(mm <30
Exercise
Variables
Predict
Long-Term
Mortality*
Cardiac Death
Noncardiac Death
Survivors
p Value
12 4
2 5
6 25
9
2
3
7
5
28
8
1
8
8
5
22
15
4
15
1
3
16
Hg) <0.005
230 ST-segment 11
changes (mm) NS
<0.001 Negative test * Anegative test occurred shiis. NS = not significant.
when none of thefollowngvariables
were present: maximal
watts 572 watts; systolic blood pressure increase
ST-segment
accordance with the findings of 96% in our control group. Our study showedthat AM1 patients had a maximal work capacity of 59% that of the control subjects. However, patients with endpoints of cardiac death had a maximal work capacity of 43% of control subjects.In survivors and those with an endpoint of noncardiac death, maximal work capacity was 66 and 63%, respeu tively. These results imply that a low maximal work capacity before discharge may be of prognostic value. When a low maximal work capacity was defined as 150% that of control subjectsa Kaplan-Meier analysis showed that low maximal work capacity actually was correlated with a significant cumulative probability of cardiac death. Thus, a predischargemaximal work capacity may be of great clinical value in identifying patients with low and high risk of future cardiac death. A low maximal work capacity may be the result of a de creasedleft ventricular function. This is in accordance with several other studies showing that a low left venDISCUSSION Although maximal exercise testing has rarely been tricular ejection fraction after AM1 is correlated with and mortality as sole endused before hospital discharge, this type of exercisetest future coronary events26*27 has been shown to be safe in this and other studies.20y21point. 13,14,17 In some studies patients who could not exercisefor the Myocardial function can also be judged by exerciseexpected duration of their protocol had an increased induced blood pressure response. In accordance with mortality risk over the next year.15,22,23 If, however,they Fioretti et a1,t4our results suggestthat inability to incould go the full duration of a submaximal workload, crease systolic blood pressure by at least 30 mm Hg the risk was very low. Although these studies seemable expressespoor myocardial function, and significantly to define low- and high-risk groups, no information is predicts long-term mortality (Figure 1). In severalother given concerning the maximal work capacity of the studies the lack of adequate blood pressure response AM1 patients, and their work capacity is not compared early after AM1 was associated with an increased to the normal population. Knowledge of thesedata may l-year mortality13J5J7 and an increase in coronary have important psychological benefits for patients and events.6J0 help cardiologists in the development of postdischarge The bad prognosisassociatedwith exerciseparameguidelines for exercise prescription. Finally, maximal ters indicating a poor left ventricular function may be work capacity might elicit further prognostic informa- caused by large infarcts or may be due to myocardial tion. ischemia causedby multivessel disease.28 In other studies of normal populations that were In our study 6% of control subjectsand 26% of surviage-matchedto AM1 patients, a maximal work capacity vors had significant ST-segmentshifts and a low risk of >I00 watts was found in 92%24and 95%.25This is in cardiac death. The ST-segment changes in this group ure 2 shows the predictive accuracy of a positive and negative test when maximal work capacity, systolic blood pressure increase and ST-segment shifts were evaluated separately or in combination. The predictive accuracy of a negative test for the exercisevariables is fairly high (77 to 88%), especially for the combined exercise variables (95%). Such a low risk was found in 37% of AM1 patients in this study (Table III). Contd subjectz INITIAL: The mean age of the control subjects was 55.8 years. Previous AM1 was reported in 2, angina pectoris in 3, hypertension in 5 and chronic pulmonary disease in 5 of them. Three were treated with digoxin and 3 with /3-blocking agents. Six subsequentcardiac events occurred in 5. Two subjects had a nonfatal AMI, 1 was hospitalized due to angina pectoris and cardiac failure and 2 died from noncardiac reasons.
152
THE AMERICAN
JOURNAL
OF CARDIOLOGY
VOLUME
65
may therefore be causedby myocardial ischemia due to and death. Am J Cardiol 1980;46:909-914. 7. Davidson DM, DeBusk RF. Prognostic value of a single exercise test 3 weeks minor coronary stenosesor good collateralization. Al- after uncomplicated myocardial infarction. Circulation 1980;61:236-242. though when compared to survivors maximal work ca- 6. Jespersen CM, Kassis E, Edeling C-J, Madsen JK. The prognostic value of pacity and systolic blood pressureresponsewere signifi- maximal exercise testing soon after first myocardial infarction. Eur Heart J cantly decreasedin patients with the endpoint of cardiac 9.1985,6:769-772. Krone RJ, Gillespie JA, Weld FM, Miller JP, Moss AJ and The Multicenter death, significant ST-segmentshifts were not. However, Postinfarction Research Group. Low-level exercise testing after myocardial inusefulness in enhancing clinical risk stratification. Circulation 1985; 50% of these patients had significant ST-segmentshifts. farction: 71:80-89. This indicates that at least some of the patients with 10. Murray DP, Salih M, Tan LB, Derry S, Murray RG, Littler WA. Which bad left ventricular function variables during exercise exercise test variables are of prognostic importance post-myocardial infarction? J Cardiol 1988;20:353-363. may have a bad prognosisnot due only to larger infarc- Int 11. Koppes GM, Kruyer W, Beckmann CH, Jones FG. Response to exercise tions, but also due to multivessel disease.In accordance early after uncomplicated acute myocardial infarction in patients receiving no with this thesis no significant correlation was found be- medication: long time follow-up. Am J Cardiol 1980;46;764-769. 12. Jelinek VM, McDonald IG, Rvan WF. Ziffer RW. Clemens A. Gerloff J. tween maximal LDH and maximal work capacity. Assessment of cardiac risk 10 days after uncomplicated myocardial infarction. Br However, some of the patients with ST-segment Med J 1982:284:227-230. 13. Jennings K, Reid DS, Hawkins T, Julian DJ. Role of exercise testing early changes had increased ST elevations. These changes after myocardial infarction in identifying candidates for coronary surgery. Br Med may be proportional to the size of the infarction,29*30 J 1984;288:185-187. and ST elevation appearsto be a bad prognostic mark- 14. Fioretti P, Brewer RW, Simoons ML, Das SK, Bos RJ. Wijns W, Reiber JHC, Lubsen J, Hugenholtz PG. Prediction of mortality in hospital survivors of er,29,30 even though this conclusion has been anticipated myocardial infarction. Comparison of predischarge exercise testing and radionuby others.9J0 elide ventriculography at rest. Br Heart J 1984;52;292-298. This study has shown that in consecutivesurvivors of 15. Weld FM, Bigger JT. Risk stratification by exercise testing and 24-hour electrocardiogram 2 weeks after acute myocardial infarction (abstr). AM1 a predischargemaximal exercisetest is a safe and ambulatory Circulation 1978:58(suppl II):II-198. effective diagnostic procedure for identifying patients 16. Saunamtiki KI, Andersen JD. Prognostic significance of the ST-segment response during exercise test shortly after acute myccardial infarction. Compariwith low and high risk of cardiac death. When the exer- son with other exercise variables. Eur Heart J 1983;4:752-760. cise variables are combined, the predischarge maximal 17. Waters DD, Bosch X, Bouchard A, Moise A, Roy D, Pelletier G, ThCroux P. exercisetest seemsto be of great clinical value, especial- Comparison of clinical variables and variables derived from a limited predischarge exercise test as predictors of early and late mortality after myocardial infarction. ly in identifying AM1 patients at low risk, the predictive JACC 1985:5:1-8. accuracy of a negative test being 95%. Such a low risk 16. Ellestad MH. Stress testing. Principles and practice. Philadelphia: Davis, was found in 37% of AM1 patients in this study. These 1980. 19. Pedersen PK, Nielsen JR. Absolute or relative work load in exercise testingpatients may not require either coronary angiography or significance of individual differences in working capacity. Scan J Clin Lab lmest P-blocking agents. Furthermore, the predischargeexer- 1984:44:635-642. Constant J. Prognostic information from early post-infarction exercise testcise test can be used to guide cardiac rehabilitation and 20. ing. Am J Med 1986;81:655-660. bring to the patient important psychological benefits, 21. Ibsen H, Kjqller E, Styperek J, Pedersen A. Routine exercise ECG three decreasethe period of convalescenceand improve the weeks after acute myocardial infarction. Acta Med Stand 1975;198:463-469. Samtam& KI, Andersen JD. Early exercise test in assessment of long-term return-to-work rate. The AM1 patients at high risk 22. prognosis after acute myocardial infarction. Acta MedScand 1981;209:185-191. should be followed in outpatient clinics and may require 23. Madsen EB, Gilpin E, Ahnves, Henning H, Ross J. Prediction of functional various types of medical treatment such as p blockers capacity and use of exercise testing for predicting risk after acute myocardial infarction. Am J Cardiol 1985;56:839-845. and even coronary angiography with a view of revascu- 24. Grimby G, Wilhelmsen L, Ekstrom-Jodal B. Aerobic power and related larization procedures, factors in a population study of men aged 54. Stand J Clin Lab fmmt REFERENCES 1. ThCroux P, Waters DD, Halphen C, Debaisieux J-C, Mizgala HF. Prognostic value of exercise testing soon after myccardial infarction. N Eng/ J Med 1979;301:341~345. 2. Velasco I, Tormo V, Ferrer LM, Ridocci F, Blanch S. Early exercise test for evaluation of long-term prognosis after uncomplicated myocardial infarction. Eur Heart J 1981;2:401-407. 3. Markiewicz W, Houston N, DeBusk RF. Exercise testing soon after myocardial infarction. Circulation 1977;56:26-31. 4. Smith JW, Dennis CA, Gassmann A, Gaines JA, Staman M, Phibbs B, Marcus FI. Exercise testing three weeks after myocardial infarction. Chest 1979:75:12~16. 5. Sami M, Kraemer H, DeBusk RF. The prognostic significance of serial exercise testing after myocardial infarction. Circulation 1979,60:1238~1246. 6. Starling MR. Crawford MH, Kennedy GT, O’Rourke RA. Exercise testing early after myocardial infarction: predictive value for subsequent unstable angina
1970:26:287-294. 25. Hagerup L. Coronary heart disease. Risk factors in men and women. Acta Med Stand 1974zuppl 557(thesis). 26. De Feyter PJ, Eenige MJ, Dighton DH, Visser FC, Jong J, Roes JP. Prognostic value of exercise testing. Coronary angiography and left ventriculography 6-8 weeks after myocardial infarction. Circulation 1982,66:527-536. 27. Norris RM, Barnaby PF, Brandt PWT, Geary GG, Whitlock RML, Wild CJ, Barrat-Boyes BG. Prognosis after recovery from first acute myocardial infarction: determinants of reinfarction and sudden death. Am J Cardiol1984;53;408413. 26. Sanz G, Gastaner A, Betriu A. Determination of prognosis in survivors of myocardial infarction. A prospective clinical angiographic study. N Engl J Med 1982:306:1065~ 1070. 29. Manner@ D, Bennett ED, Ward DE, Dawkins K, Dancy M, Valantine H, Mehta N. Accurate detection of triple vessel disease in patients with exercise induced ST segment depression after infarction. Br Heart J 1987;57:133138. 30. Grande P, Pedersen A. Myocardial infarct size and cardiac performance at exercise soon after myocardial infarction. Br Heart J /982;47:44-50.
THE AMERICAN JOURNAL OF CARDIOLOGY JANUARY 15, 1990
153