Usefulness of blood lactate as a predictor of shock development in acute myocardial infarction

Usefulness of Blood Lactate as a Predictor Shock Development in Acute Myocardial Infarction

of

Zarko MavriC, MD, Luka ZaputoviC, MD, Davorka zagat-, MD, Ante Matana, MD, and Davor Smokvina, DCC

Data were obtained and analyzed in 229 patients admitted to the coronary care unit from November 1966 through July 1969. The patients were classified into 2 groups: patients without or with only mild left ventricular failure (Killip class I or II) during their hospital stay (group I), and patients who were in Killip class I or II on admission but developed cardiogenic shock during hospitalization (group II). Discriminant function analysis was performed using the following variables: patients’ age, history of previous myocardial infarction, diabetes mellitus, blood lactate, urea, creatinine, creatine kinase, aspartate aminotransferase, lactate dehydrogenase concentrations, and chest x-ray cardiothoracic ratio. Variables that were found to significantly discriminate the 2 groups of patients were age, previous infarction, x-ray cardiothoracic ratio, blood urea and lactate concentrations. The risk index was computed, and blood lactate was the variable with the greatest predictive power for shock development. The sensitivity, specificity and predictive value of the risk index, taking various cutoff points, were cakulated. With a cutoff value of 1, sensitivity was 65%, specificity 91%, positive predictive value 36% and negative predictive value 97%. With a cutoff value of 2, sensitivity was 53% specificity 99%, positive predictive value 92% and negative predictive value 96%. (Am J Cardiol 1991;67:569-566)

From the Department of Internal Medicine, Division of Cardiology (Coronary Care Unit), Clinical Hospital Center Rijeka, and University of Rijeka School of Medicine, 5 1000 Rijeka, Croatia, Yugoslavia. Manuscript received August 3, 1990; revised manuscript received and accepted November 13, 1992. Address for reprints: Zarko Mavri6, MD, Department of Internal Medicine, Division of Cardiology (Coronary Care Unit), Clinical Hos-

ardiogenic shockis a leading causeof death in patients with acute myocardial infarction (AMI). First clinical signs of cardiogenic shock in AM1 are manifestations of poor peripheral perfusion, but these signs need not be the earliest indicator of inadequacy of cardiac function. Shepsl and Kessler2 and their co-workers demonstrated that peripheral blood lactate determination can be a useful prognostic marker for risk of recurrent cardiac arrest; an increasein blood lactate may be related to tissue hypoperfusion. In this study we tested the hypothesesthat elevated peripheral blood lactate concentration precededclinical manifestations of shock in patients with AM1 and that blood lactate could be used to predict the occurrence of shock.

C

METHODS

This study comprised 291 consecutivepatients with AM1 admitted to the coronary care unit from November 1988 through July 1989. The first 229 patients formed the study group, and the next 62 patients were used for cross-validation of the results obtained through discriminant function analysis. The diagnosis of AM1 was basedon the presenceof 12 of the following criteria: characteristic history of prolonged chest pain, elevation in serial cardiac enzymes-creatine kinase, aspartate aminotransferase, lactate dehydrogenase-and characteristic electrocardiographic ST-T changes with or without development of new Q waves. When there were only ST-T changes on the electrocardiogram (without new Q waves), elevation of cardiac enzymes was required for confirmation of AMI. The patients were further classified into 2 groups. Group I consisted of patients without or with only mild left ventricular failure (Killip classI and 113)during their hospital stay. Group II consistedof patients who were in Killip classI or II on admittance but developed cardiogenic shock during hospitalization. Patients with cardiogenic shock already present on arrival to the coronary care unit and those who developed shock during the first 24 hours were not included in the study. Patients with Killip class III left ventricular failure, patients who died from causesother than cardiogenic shock, and patients who were resuscitated from cardiac arrest also were not included. Cardiogenic shock was defined as systolic blood pressure<90 mm Hg for >30 minutes, accompaniedby the clinical signs of peripheral hypoperfusion and a decreasein urinary output (urine flow <20 ml/hour).

THE AMERICAN JOURNAL OF CARDIOLOGY MARCH 15, 1991

565

TABLE I Comparison Cardiogenic Shock

Between Patients With and Without Group I (n=212)

Group I/ (n = 17)

78 (37%) 134 (63%)

0 17 (100%)

177 (84%) 35(16%) 70 (33%) 35 (16%) 42 (20%)

10 (59%) 7 (41%) 7 (41%) 2(12%) 6 (35%)

p Value

Age <60 years 260 years Previous AMI No Yes Diabetes mellitus Thrombolytic therapy Killip class II Cardiothoracic ratio (%) <55 255 <60 260 Blood urea (mmol/liter) Blood creatinine (f.4mol/liter) Creatine kinase (IU/liter) Blood lactate (mmol/liter)


<0.05 NS NS NS

without deproteination,4using a commercial kit (Boehringer-Manheim GMBH, Germany). Statistical analysis was accomplishedthrough SPSS computer software.5 Multivariate analysis of variance and discriminant function analysis were performed on selectedvariables in the study group, and the discriminant function was cross-validatedon the secondgroup of patients. Continuous data are expressedas mean f standard error of the mean, and discrete variables as frequencies.Differenceswere consideredsignificant at p <0.05. RESULTS

Of 229 patients who formed the study group, 212 remained in Killip classI or II during their hospital stay (group I), and 17 developedcardiogenic shock (group
-4.50

138 (65%) 74 (35%) 202 (95%) 10 (5%) 7.81 f0.16 108k 12

6 (35%) 11(65%) 12(71%) 5 (29%) 13.32 f 2.23 171 f25

<0.05

-

-

i 5 E

4.00

1

3.50

-

3.00

-

2.50

-

5 3

FIGURE 1. Bbod lactate values obtained on successive days in 2 groups of patients. Values are expressed as mean f standard error of the mean. Group I = patients without shock; Group II = patients who developed shock during hospi-

cl z g -0 : m

2.00 0

566

I Suckessive

I lohate

valuk I

THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 67

I 4

TABLE II Results of Discriminant Selected Variables Fl Blood lactate Blood urea Previous AMI Cardiothoracic ratio

45” 44” 6+ 17*

Age

10’

TABLE Ill Distribution of Various Values of Risk Index in Patients With and Without Cardiogenic Shock

Function Analysis of F2

Coeff.

Con-.

29”

0.62 0.57 0.31 0.17 0.18

0.67 0.66 0.26 0.41 0.31

22’ 7r 2 2

Group II

Group I

* p co.oo1; + p <0.05; * p
Index

No. of Pts.

%

Cum. %

No. of Pts.

%

<-1 -14 (rl l-2 2-3 >3

31 106 55 18 2 0

15 50 26 8 1 0

15 65 91 99 100 -

0 3 3 2 3 6

0 18 18 12 18 34

Cum. % = cumulative

groups are shown in Figure 1. Although there was a rather wide variation of individual results in group II, 60% of patients in this group had blood lactate levels >3.6 mmol/liter on the day preceding the appearance of shock. Other variables (diabetes mellitus, blood creatinine, cardiac enzymes)were not found to have significant discriminating power, and there was no difference in proportions of patients with Killip class I or II between 2 groups. Regarding thrombolytic therapy, there was no significant difference between2 groups either in the incidence of its administration or in the blood lactate concentrations between those who did and did not receive this treatment. We did not identify any possible extracardiac causesfor hyperlactatemia.6,7All patients with non-Q-wave AM1 were in group I, and there was no difference between their lactate levels and those of patients with Q-wave AM1 belonging to group I. With use of stepwise discriminant analysis, the following function for risk index (I) was computed: I = 0.64 X L + 0.174 X U + 0.81 X PI + 0.711 X RTG + 0.387 X A - 3.61 where L = blood lactate (mmol/liter, mean value for group I, and the value obtained on the day before shock appearancefor group II); U = peak blood urea (mmol/liter); PI = previous infarction (0 = no, 1 = yes); RTG = cardiomegaly on chest x-ray (0 = cardiothoracic ratio of <60%, 1 = cardiothoracic ratio of 160%); and A = age (0 = age <60 years, 1 = age 160 years). The function had highly significant discriminating power: chi-square = 82, p 0.30). Inclusion of blood lactate into discriminant function analysis added significantly to the predictive power of the function: chisquare = 82 with lactate included versus chi-square = 54 when only other variables were entered into the analysis. Distribution of various values of the risk index I in our study population is shown in Table III. The sensitiv-

Cum. % 18 35 47 65 100

percent.

ity, specificity and predictive value of the index I, taking various cutoff points, were analyzed. Using a cutoff level of 1, we found that I I1 detected 65% of patients who were to developcardiogenic shock (sensitivity), and that the proportion of patients developing shock among those with I I1 was 36% (positive predictive value). A value of I
Many investigators have tried to identify high-risk patients with AM1 and have describedprognostic indexes using various clinical, biochemical, electrocardiographic, radiologic, invasive hemodynamic and radionuelide variables8-l* The study by Hands et alI9 focused specifically on prediction of shock development in patients with AM1 using patients’ age, left ventricular ejection fraction, serum creatine kinase, diabetes mellitus and history of previous AMI, and they found that shock occurrence could be predicted in >50% of cases. Although there is no evidence that early identification of high-risk patients would lead to a better prognosis, there is a theoretical possibility that early implementation of interventions, such as invasive hemodynamic monitoring or some new therapeutic modality, might decreasethe incidence of manifest shock. As theseinterventions could be potentially hazardous,it would be important not to exposelow-risk patients to the possible risk. The utility of such an index is even greater if it is based on simple clinical or laboratory findings that are

THE AMERICAN JOURNAL OF CARDIOLOGY MARCH 15, 1991

567

readily available for every patient. The value of periph- those obtained by crossvalidation of the function on the eral blood lactate concentration as a predictor for new secondgroup of patients. cardiac events has been demonstrated by Sheps,’ and Kessler* and their associates,who have found it to be a useful prognostic marker for risk of recurrent cardiac arrest. Although the precise mechanism of elevated REFERENCES 1. Sheps DS, Conde C, Cameron B, Lo WC, Appel R, Castellanos A, Harkness blood lactate is not clear, it could be related to peripher- DR, Myerburg RJ. Resting peripheral blood lactate elevation in survivors of al tissue hypoperfusion. prehospital cardiac arrest: correlation with hemodynamic, electrophysiologic and In our study we found that blood lactate concentra- oxyhemoglobin dissociation indexes. Am J Cardiol 197944:1276-l 282. Kessler KM, Kozlovskis P, Trohman RG, Myerburg RJ. Serum lactate: progtion is a significant predictor of shock development in 2.nostic marker for recurrent cardiac arrest? Am Hearr J 1987;113:1540-1544. patients with AMI, and that it can be used in conjunc- 3. Killip T III, Kimball JT. Treatment of myocardial infarction in a coronary care tion with other simple clinical or laboratory data (pa- unit. A two year experience with 250 patients. Am J Cardiol 1967;20:457-464. F. Lactate determination with LDH, GPT and NAD. In: Bergmeyer HU, tient’s age, history of previous AMI, blood urea concen- 4.ed.Noll Methods of Enzymatic Analysis. New York: Academic Press, 1974:1475. tration and x-ray cardiothoracic ratio) to predict shock 5. Tabachnick BG, Fidel1 LS. Multivariate Statistics. New York: Harper and 1983. occurrence. Mean blood lactate concentration during Row, 6. Huckabee WE. Abnormal resting blood lactate. Am J Med 1961;30:833-839. the period from admission to the appearance of first 7. Foster DW. Lactic acidosis. In: Braunwald E, Isselbacher KJ, Petersdorf RG, clinical signs of shock was significantly higher in pa- Wilson JD, Martin JB, Fauci AS, eds. Harrison’s Principles of Internal Medicine. McGraw-Hill, 1987:1797-1800. tients who developedshock (group II) than in patients 6.Hamburg: Norris RM, Brandt PWT, Caughey DE, Lee AJ, Scott PJ. A new coronary who remained hemodynamically stable (group I). prognostic index. Lmcet 1969;1:274-278. However, the difference was even greater and discrimi- 9. Peel AA, SempleT, Wang I, Lancaster WM, Dal1 JLG. A coronary prognostic for grading the severity of infarction. Br Heart J 1962;24:745-760. nation better when lactate value obtained on the day index 10. Kitchin AH, Pocock SJ. Prognosis of patients with acute myocardial infarcjust preceding occurrenceof shock in group II was com- tion admitted to a coronary care unit. I. Survival in hospital. Br Heart J 1977;39:1163-1166. pared with mean lactate concentration in group I (Ta- 11. Henning H, Gilpin EA, Covell JW, Swan EA, G’Rourke RA, Ross J. ble I). This value resulted in greater predictive power Prognosis after acute myocardial infarction: a multivariate analysis of mortality than any other variable included in the discrimination and survival. Circulation 1979;59:1124-1136. 12. Gheorghiade M, Anderson J, Rosman H, Lakier J, Velardo B, Goldberg D, analysis (Table II). Figure 1 showsthe daily lactate lev- Friedman A, Schultz L, Tilley B. Risk identification at the time of admission to els in the 2 groups; blood lactate in group II patients coronary care unit in patients with suspected myocardial infarction. Am Heart J can be seen to increase 2 days before shock develop- 1988;116:1212-1217. 13. Mulley AC, Thibault GE, Hughes RA, Barnett GO, Reder UA, Sherman ment, with further increase on the following day, dem- EL. Thecourseof patients with suspected myocardial infarction: the identification onstrating that the most recent value of blood lactate of low-risk patients for early transfer from intensive care. N Engl J Med 1980;302:943-948. should be used for calculation of the risk index in every- 14. Fuchs R, Scheidt S. Improved criteria for admission to cardiac care units. day practice. J,4MA 1981;426:2037-3041. Despite significant univariate F ratios, cardiotho- 15. Goldberg RJ, Gore JM, Gurwitz JH, Alpert JS, Brady P, Strohsnitter W, 2, Dalen JE. The impact of age on the incidence and prognosis of initial racic ratio and patients’ age did not maintain statistical Chen acute myocardial infarction: The Worcester Heart Attack Study. Am Heart J significance as multivariate predictors after adjustment 1989;117:543-549. was made for other variables (Table II). Nevertheless, 16. Brush JE, Brand PA, Acampera P, Chalmer B, Wadeers FJ. Use of the initial to predict in-hospital complications of acute myocardial infarcwe retained them in the analysis becauseof significant electrocardiogram tion. N Engl J Med 1985;312:1137-1141. pooled-within-group correlations between them and ca- 17. Battler A, Karliner JS, Higgins CB, Slutsky R, Gilpin EA, Froelicher VF, Ross J. The initial chest x-ray in acute myocardial infarction. Prediction of early nonical discriminant function, and becausevarious stud- and late mortality and survival. Circulation 1980;61:1004-1009. ies have suggestedthat age10J5J9-25 and heart size on x- 16. Silverman KJ, Becker LC, Bulkley BH, Burow RD, Mellits ED, Kallman ray17are important prognostic factors in AMI. Besides, CH, Weisfeldt ML. Value of early thallium-201 scintigraphy for predicting morin patients with acute myocardial infarction. Circulation 1980;61:996regarding small number of patients with cardiogenic tality 1003. shock in our study, there is a greater tendency to a type 19. Hands ME, Rutherford JD, Muller JE, Davies G, Stone PH, Parker C, Braunwald E, the MILIS Study Group. The in-hospital development of cardioII statistical error. shock after myocardial infarction: incidence, predictors of occurrence, outOur risk index for shock development exhibited a genie come and prognostic factors. J Am Coil Cardioi 1989;14:40-46. satisfactory potential for classifying patients with AMI 20. Peel AAF. Age and sex factors in coronary artery disease. Br Heart J 1955;17:319-326. who had no or only mild signs of left ventricular dys- 21. Harris R, Piracha AR. Acute myocardial infarction in the aged prognosis function (Killip class I and II) during early stages of and management. J Am Geriat Sot 1970;18:893-904. AMI. With a cutoff value of 1 for the risk index, its 22. Kincaid DT, Botti RE. Acute myocardial infarction in the elderly. Chest 1973;64:170-172. sensitivity was 65% and specificity 91%, with positive 23. Berman ND. The elderly patients in the coronary care unit. I. Acute myocarpredictive value 36% and negative predictive value 97%. dial infarction. J Am G&at Sot 1979;27:145-151. 24. Latting CA, Silverman ME. Acute myocardial infarction in hospitalized With a cutoff point of 2, sensitivity was 53%,specificity patients over age 70. Am Heart J 1980;100:31 l-318. 99%, positive predictive value 82% and negative predic- 25. Wilcox RG, Hampton JR. Influence of age in prehospital and hospital tive value 96%. These results are highly consistent with mortality of heart attacks. Br Heart J 1980;44:503-507.

568

THE AMERICAN

JOURNAL

OF CARDIOLOGY

VOLUME

67