Evaluation of lactate and C-reactive protein in the assessment of acute myocardial infarction.

Purpose: The purpose of this study was to assess the usefulness of lactate and C-reactive protein as early markers of acute myocardial infarction (AMI) in conjunction with established markers, such as creatine phosphokinase (CPK)-MB and troponin I.

Methods: The study population consisted of all patients admitted with suspected AMI regardless of age, race, gender, or comorbid conditions. All patients in cardiac heart failure or cardiogenic shock were excluded. Lactate and C-reactive protein were drawn at admission, and then at 2 and 4 hours, respectively. The positive and negative predictive values of these markers were calculated in relation to the primary outcome measure.

Results: A total of 62 patients were enrolled in the study. Of those, 18 patients had documented AMI and all underwent thrombolysis, thrombolysis with angioplasty, or stent placement. The lactate was measured at 2 and 4 hours, and lactate levels correlated with CPK and CPK-MB, in addition to troponin I. In our cohort, lactate had a sensitivity of 75% and a specificity of 95.5%, with a positive predictive value of 85.7% and a negative predictive value of 91.3%.

Conclusion: As an inexpensive and readily available marker for AMI, lactate seems to be reliable and could be used in an emergency setting to facilitate the decision-making process for chest pain syndrome. It would be a helpful adjunct on whether the patient can be safely discharged or should be promptly admitted.

**********

Approximately 750,000 acute myocardial infarctions (AMIs) occur yearly in the United States. The most common clinical presentation is chest pain, which accounts for over 5 million annual visits to the emergency department (ED). Traditionally, clinicians base their judgment on whether to admit the patient on the patient's history, physical examination, imaging (electrocardiography [ECG], and the cardiac enzymes creatine phosphokinase (CPK), CPK-MB, and troponin I, (1), (2) all of which are nonspecific or not sensitive enough. (1) (3) (4) Unfortunately, this classic approach toward a "rule-out" diagnosis of AMI is, in many cases, very expensive and time consuming to the medical staff and patient, thus leaving a time-sensitive and inexpensive test still missing to the clinician. A recent study by Schmiechen et al (1) demonstrated serum lactate to be a very sensitive marker for myocardial injury in the acute evaluation of chest pain. Although they found the test to have high sensitivity, on the order of 96%, it was found to be only 55% specific. The purpose of this study is to evaluate the potential of using serum lactate to screen which patients require admission to the hospital for further investigation of their chest pain and which patients with reasonable certainty can be discharged from the ED. Patients presenting with elevated lactate levels, in conjunction with their medical history and ECG, can benefit from more urgent care and intervention to reduce significant myocardium damage. It is our hope that this strategy can help prevent costly admission to the hospital for rule-out AMI. (5) (6)

Patients and Methods

A prospective cohort of 62 patients who were seen at the ED of Memorial Health University Medical Center, a tertiary care teaching hospital serving southeast Georgia, with substernal chest pain and suspected AMI were enrolled into the study, regardless of age, race, gender, or comorbid conditions. Patients in cardiogenic shock; those who underwent cardiopulmonary resuscitation; patients with active seizure activity, alcohol intoxication, or acute or chronic renal failure; patients with smoke inhalation; or those who presented more than 24 hours after symptom onset were excluded from the study. We also excluded those patients with acute ST-segment elevation on ECG who either received thrombolytics or were taken to the cardiac laboratory for revascularization. Patients were further assessed for the confirmation of their AMI by the cardiology team of the medical center. All patients with substernal chest pain and suspected AMI received serial ECG, cardiac enzymes, echocardiography, and cardiac catheterization for confirming or refuting the diagnosis of AMI.

Once seen in the ED, patients were provided with an informed consent document about their participation in the study. In addition to administering ECG, CPK, CPK-MB, and troponin I, patients consented to two laboratory tests: serum lactate and C-reactive protein (CRP). Serum lactate levels were drawn and measured at three different time periods: on arrival, and then at 2 and 4 hours after their arrival to the ED. CRP was obtained once within the first 4 hours from arrival time to the ED. In addition to the laboratory tests, patient demographic information such as age, gender, and race was also collected, according to the protocol approved by Memorial Health University Medical Center's Institutional Review Board.

Statistical Analysis

The statistical analysis of the predictive power of serum lactate over the presence of AMI starts with a simple 2 x 2 table dichotomizing the serum lactate values into two groups: lower than normal and greater than or equal to the normal cutoff value of 18 mmol/dl. Alternatively, we can obtain more statistical power by using serum lactate values as a continuous variable within a regression model. At the time of arrival, whether an AMI has occurred, or the exact probability of such event, is not directly measurable by the ED staff. However, the measurement of serum lactate may help to rule out the AMI, by acting as a surrogate to detect the AMI. A logistic regression analysis of serum lactate values in predicting an AMI will be conducted to measure its potential for the initial triage of patients presenting to the ED complaining of chest pains. In a logistic regression analysis, we can assume that the logic transformation of the probability of AMI can be expressed as a linear function in the parameter of the covariates.

Results

Demographics

The study population was divided into two groups, as represented in Table 1. The [chi square] tests of the different demographic characteristics failed to indicate statistically significant differences between the AMI and the non-AMI patients. However, as expected, Student's t tests of the mean serum lactate values indicated large, statistically significant differences in each of the three time periods for which they had been drawn. In addition, there were no statistically significant differences in the level of CRP between patients with AMI and those without, and as such, no other statistical methods will be used in conjunction with this clinical laboratory test for the remainder of the article.

Bivariate Results

Results for serum lactate for the three time periods are presented in Table 2. Serum lactate drawn immediately after qualifying for the study has a sensitivity of 75%, a specificity of 95.5%, a positive predictive value of 85.7%, and a negative predictive value of 91.3%. The likelihood ratio, which is the number of times more likely a test result is found in diseased compared with nondiseased (ie, the ratio of the sensitivity to 1--specificity) is equal to 16.5, with a 95% confidence interval ranging from 4.14 to 65.8. Serum lactate, therefore, has good sensitivity and is very specific in detecting that an AMI had occurred. Similar results are also generated from blood drawn after 2 and 4 hours, although the sensitivity of the test is somewhat reduced after 4 hours from the time the patients were entered into the study. Three patients had serum lactate above 18 at time 0 (two at 21 and one at 20) but dropped below 18 (two at 12 and one at 16) at 4 hours. One patient had a serum lactate of 16 at time 0 and after 2 hours, but lactate elevated to 19 at 4 hours. Patients without a confirmed AMI did not have a large enough change in their serum lactate level to assign them to a different serum lactate level group, leaving specificity levels unchanged through time.

Logistic Regression Results

Results for three logistic regression models, on the basis of the time the serum lactate was measured, are presented in Table 3. We obtained for the model a parameter of 0.3321 (P = 0.0006) for serum lactate as an independent continuous covariate. By taking the exponential of this parameter value, we transform this result into the change in the odds of an AMI for a one-unit increase in the value of serum lactate. Therefore, the odds of suffering an AMI increases by 39% for each one-unit increase in the serum lactate value. For the model where serum lactate was drawn 2 hours after entry into the study, an increase of one unit in the serum lactate value increases the odds of having suffered an AMI by 29% (P = 0.0022). Finally, for the model where serum lactate was drawn 4 hours after entry into the study, an increase of one unit in the serum lactate value increases the odds of having suffered an AMI by 22% (P = 0.0051). Serum lactate appears to have better predictive power in detecting that an AMI has occurred the earlier it can be obtained from the patient. However, because of our relatively small sample size, this result in the timing of serum lactate is not statistically significant.

After obtaining estimates of the parameter for the serum lactate, we can use a logistic function to return to the predicted probabilities of AMI. In Figure 1, we present the plot of the predicted probabilities against the serum lactate values. At each lactate value presented in the data, we label each point on the graph with the underlying AMI data (ie, the number of patients with an AMI and the number of patients without an AMI). We observe that our model works reasonably well, as few AMI patients are present in the bottom part of the plot (lowest probability of AMI) and most AMI patients are present in the top part of the plot (higher probability of AMI). Figure 2 depicts the predicted probability curves for the three models based on the timing of the serum lactate. The figure supports our regression point estimate that serum lactate seems to have better predictive power of detecting that an AMI has occurred the earlier it can be obtained.

Figure 3A depicts the logistic regression results: the plot of the sensitivity against 1--specificity (ie, likelihood ratio). The best tradeoff between greater sensitivity and greater specificity occurs in the upper left corner of the plot, where specificity is at approximately 95% (1--specificity is approximately 5%) and sensitivity is approximately 76%--the same values obtained earlier in the bivariate analysis. To see this more clearly, we plotted Youden's index (the sum of sensitivity and specificity--1) at each specificity value, which is presented in Figure 3B. We can observe that for each of the three models, the misclassification minimization occurs at a specificity of 95%.

[FIGURE 1 OMITTED]

Discussion

A timely diagnosis of acute coronary syndrome and myocardial infarction presents a diagnostic dilemma to the admitting physician to appropriately triage patients to receive appropriate intervention including thrombolytic or revascularization treatment. It has been suggested that up to 8% of the patients presenting to the ED with acute coronary syndrome are initially sent home for lack of an established diagnosis. (7), (8)

[FIGURE 2 OMITTED]

Echocardiography, serial ECG, and serologic tests for myocardial injury have been used either for diagnosis of acute myocardial injury or to confirm the diagnosis as an adjunct modality. However, none of the modalities are sufficiently sensitive to rule out acute ischemic events in the first few hours of the ED evaluation. (9-12)

Toth et al (9) reported the usefulness of echocardiography in the first 72 hours of ischemic events. However, in most cases, the diagnosis should be reached within the first few hours to appropriately treat the patient. (13)

Schmiechen et al (1) examined the use of serum lactate level to evaluate patients with acute chest pain syndrome. They reported the institution of lactate level in diagnosis of AMI approaching 96% sensitivity, with a specificity of 55%. Their patient population consisted of all newcomers to the ED with a diagnosis of acute chest pain syndrome/myocardial infarction; however, they excluded those patients who arrived in cardiac arrest. It is known that both systemic and poor coronary perfusion could cause elevation of serum lactate in the absence of lactic acidosis. (14)

We chose to exclude all complicated AMIs, such as cardiac arrhythmia, congestive heart failure, pulmonary edema, and pulmonary embolism, because it is widely believed that these conditions might cause further elevation of serum lactate and produce a false-positive component. (13-15) All patients with a presumptive diagnosis of AMI and who entered the study subsequently underwent extensive investigation using serial ECG, cardiac enzymes (including CPK and CPK-MB), troponin, echocardiography, and cardiac catheterization that resulted in either thrombolytic treatment, angioplasty, or surgical revascularization. In all probability, the elevation in serum lactate is secondary to myocardial ischemia and poor systemic perfusion as a result of compromised systolic function and decreased cardiac output in the absence of lactic acidosis. (9-15)

The purpose of this study is not to promote replacing cardiac enzymes such as CPK, CPK-MB, or troponin with lactate measurement. However, we have demonstrated that the early increasing serum lactate has excellent positive predictive value in the first 6 hours of presentation. This process identifies a high-risk patient with evolving myocardial injury and perfusion deficit and, at times, might facilitate the decision-making process and the triaging of the patient to receive immediate attention for confirmatory tests and appropriate intervention. Noncardiac chest pain or atypical presentation in general and in younger patients, particularly younger women, still presents a significant challenge to the treating physician. Triage tools with both positive and negative predictive value, such as serum lactate, could be used as an early marker to reach a meaningful decision in the hectic environment of the ED.

[FIGURE 3 OMITTED]

In our experience, the serum lactate was easily obtainable and the turnover time was within minutes. Because we have excluded all the complicated myocardial infarctions, the increase in the serum lactate seems to be secondary to an acute ischemic event, myocardial dysfunction, and subsequent poor systemic perfusion. Moreover, we found no value for CRP in establishing or refuting AMI. (5)

The conclusion of our study is limited by the small sample size imposed by single-medical-center and protocol criteria. The results should be confirmed by a larger, multicenter investigation. If confirmed, the lactic acid level might even become more important in the first-step diagnosis and management of chest pain syndrome.

Conclusions

Serum lactate is both sensitive and specific in uncomplicated myocardial infarction. These findings support prior studies in which lactate was found to be sensitive. However, our data seem to indicate that serum lactate may be more specific than prior studies indicate. An analysis using a larger cohort may be able to confirm our results. We found no predictive value in measuring CRP for the screening of uncomplicated infarction. We recommend that serum lactate be incorporated into the decision-making process on patients presenting to the ED complaining of chest pains. Lactate is a simple, inexpensive laboratory test with a rapid turnaround time that could help clinicians in their decision regarding whether to admit a patient for costly workup or safely discharge the patient home. Further studies may be warranted to investigate the cost effectiveness and potential of including serum lactate readings as a tool to "rule in" AMI. Other tests can then follow to further investigate on an individual basis.

Key Points

* Patients with chest pain are challenging and constitute a diagnostic dilemma in an emergency setting.

* Accurate and timely screening and diagnosis of acute myocardial infarction is not only lifesaving but also cost effective and avoids unnecessary expenses.

* Serum lactate has high positive and negative predictive value and facilitates early diagnosis and is cost effective.

It is the greatest of all mistakes to do nothing Because you can only do a little. Do what you can.

--Sydney Smith

Table 1. Demographics and lactate values of the study population by
confirmed acute myocardial infarction status (a)

                             AMI
                     Mean or %  Standard   No. of Cases
                                deviation

Age (yr)               67.3       11.08         18

Gender                                          18

  Male                 50.0%                     9

  Female               50.0%                     9

Race                                            18

  Black                66.7%                    12

  White                33.3%                     6

C-reactive protein

   <0.4                94.4%                    17

   [greater than or     5.6%                     1
   equal to]0.4

Lactate (mmol/dl)

  0 h                  24.7       13.3          16

  2 h                  24.8       12.6          17

  4 h                  24.0       12.6          16

                            No AMI
                     Mean or %  Standard   No. of cases  P value
                                 deviation

Age (yr)               71.4       11.26         44        0.199

Gender                                          44

  Male                 36.4%                    16

  Female               63.6%                    28        0.321

Race                                            44

  Black                63.6%                    28

  White                36.4%                    16        0.821

C-reactive protein

   <0.4                97.7%                    41

   [greater than or     2.3%                     1        0.500
   equal to]0.4

Lactate (mmol/dl)

  0 h                  12.7        3.5          44        0.003 (b)

  2 h                  13.8        3.8          42        0.002 (b)

  4 h                  14.4        3.4          42        0.009 (b)

(a) AMI, acute myocardial infarction.
(b) Student's t tests were used for age and lactate to assess
statistically significant differences between the two study groups;
[chi square] tests were used for gender and race.
(b) Indicates statistically significant difference at the 0.05 level.

Table 2. Cross-tabulation of acute myocardial infarction status and
serum lactate dichotomized at the normal level of 18 mmol/dL (a)

                                           0 h
Drawn time of                         AMI       No AMI
serum lactate

Lactate [greater than or equal to]18   12          2

Lactate <18                             4         42

Sensitivity, % (95% CI)               75.0 (50.1-91.5)

Specificity, % (95% CI)               95.5 (85.8-99.2)

PPV (%)                                     85.7

NPV (%)                                     91.3

LR-positive                           16.5 (4.14-65.8)

LR-negative                             .26 (.11-.61)

OR (95% CI)                           63.0 (8.45-671.22)

                                           2 h
Drawn time of                         AMI       No AMI
serum lactate

Lactate [greater than or equal to]18   13          2

Lactate <18                             4         40

Sensitivity, % (95% CI)               76.5 (52.5-92.0)

Specificity, % (95% CI)               95.2 (85.2-99.2)

PPV (%)                                     86.7

NPV (%)                                     90.9

LR-positive                           16.1 (4.05-67.3)

LR-negative                             .25 (.10-.58)

OR (95% CI)                           65.0 (12.31-343.3)

                                          4 h
Drawn time of serum lactate           AMI      No AMI

Lactate [greater than or equal to]18   10        2

Lactate <18                             6       40

Sensitivity, % (95% CI)               62.5 (37.6-83.2)

Specificity, % (95% CI)               95.2 (85.2-99.2)

PPV (%)                                     83.3

NPV (%)                                     87.0

LR-positive                           13.1 (3.2-53.5)

LR-negative                             .39 (.21-.74)

OR (95% CI)                           33.3 (6.70-166.6)

(a) AMI, acute myocardial infarction; PPV, positive predictive value;
NPV, negative predictive value; LR, likelihood ratio; CR, odds ratio;
CI, confidence interval.

Table 3. Logistic regression results (a)

                          0 h               2 h               4 h
                       B     P value     B     P value     B     P value

Variable

  Constant          -6.3992  0.0001   -5.3172  0.0002   -4.4193  0.0004
  Lactate            0.3321  0.0006    0.2583  0.0022    0.1995  0.0051

Model [chi square]  27.848   0.0001   22.911   0.0001   16.453   0.0001

[R.sup.2]             0.54              0.46             0.36

(a) Age, gender, and race were found to be highly nonstatistically
significant in the occurence of an AMI and were therefore omitted as
predictors from the regression models.

From the Memorial Health University Medical Center, Savannah, GA.

Reprint requests to Ali R. Rahimi, MD, FACP, AGSF, Internal Medicine Education, Memorial Health University Medical Center, P.O. Box 23089, Savannah, GA 31403-3089. Email: [e-mail]

Accepted July 25, 2002.

Copyright [c] 2003 by The Southern Medical Association 0038-4348/03/9611-1107

References

1. Schmiechen NJ, Han C, Milzman DP. Emergency department use of rapid lactate to evaluate patients with acute chest pain. Ann Emerg Med 1997;30:571-577.

2. Roberts R, Henry PD, Witteeveen SA, et al. Quantification of serum creatine phosphokinase isoenzyme activity. Am J Cardiol 1974;33:650-654.

3. Adams JE III, Abendschein DR, Jaffe AS. Biochemical marker of myocardial injury: Is MB creatine kinase the choice for the 1900s? Circulation 1993;88:750-763.

4. Yasmineh WG, Ibrahim GA, Abbasnezhad M, et al. Isoenzyme distribution of creatine kinase and lactate dehydrogenase in serum and skeletal muscle in Duchenne muscular dystrophy, collagen disease, and other muscular disorders. Clin Chem 1978;24:1985-1989.

5. Ueda S, Ikeda U, Yamamoto K, et al. C-reactive protein as a predictor of cardiac rupture after acute myocardial infarction. Am Heart J 1996; 131:857-860.

6. Jaffe AS. More rapid biochemical diagnosis of myocardial infarction: Necessary? Prudent? Cost effective? Clin Chem 1993;39:1567-1569.

7. Lee TH, Rouan GW, Weisberg MC, et al. Sensitivity of routine clinical criteria for diagnosing myocardial infarction within 24 hours of hospitalization. Ann Intern Med 1987;106:181-186.

8. Hedges JR, Rouan GW, Toltzis R, et al. Use of cardiac enzymes identifies patients with acute myocardial infarction otherwise unrecognized in the emergency department. Ann Emerg Med 1987;16:248-252.

9. Toth C, Csomos M, Vadnay I. Significance of early echocardiography in acute myocardial infarct [in Hungarian]. Orv Hetil 1997;138:787-791.

10. Selker HP, Zalenski RJ, Antman EM, et al. An evaluation of technologies for identifying acute cardiac ischemia in the emergency department: A report from a National Heart Attack Alert Program Working Group. Ann Emerg Med 1997;29:13-87.

11. Roberts R, Gowda KS, Ludbrook PA, et al. Specificity of elevated serum MB creatine phosphokinase activity in the diagnosis of acute myocardial infarction. Am J Cardiol 1975;36:433-437.

12. Mair J, Artner-Dworzak E, Lechleitner P, et al. Cardiac troponin T in diagnosis of acute myocardial infarction. Clin Chem 1991;37:845-852.

13. Jaffe AS. New marker for myocardial infarction: Are they needed? Contemp Intern Med 1997;9:6-12.

14. Mizock BA. Lactic acidosis. Dis Mon 1989;35:233-300.

15. Bernstein WK, Aduen J, Bhatiani A, et al. Simultaneous arterial and venous lactate determination in critically ill patients. Crit Care Med 1994;22(Suppl):A227 (abstract).

Ali R. Rahimi, MD, FACP, AGSF, Peter M. Marzano III, MD, and Christian M. Richard, MSc