| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 1996;61:1040-1041
© 1996 The Society of Thoracic Surgeons
Oxford Heart, Centre John, Radcliffe Hospital, Headington, Oxford, Ox3, 9du, Uk
To the Editor:
We read with interest the article by Pelletier and associates [1] in which they used serum levels of total creatine kinase, catalytic activity of its MB isoenzyme, creatine kinase-MB mass concentrations, and troponin-T levels to compare the myocardial protection obtained with intermittent antegrade warm versus cold blood cardioplegia. Interpretation of these results, however, with regard to statistical analysis and biochemical release curves requires further consideration.
We have also measured the serum troponin-T levels in a series of 120 adult patients undergoing coronary artery grafting or aortic valve replacement. Blood samples for serum levels of troponin-T were obtained before anesthesia and 1, 6, 24, and 72 hours after the release of the aortic cross-clamp. The distribution of serum troponin-T levels differed significantly from the normal distribution by both Kolmogorov-Smirnov and 
2 tests (p < 0.0001) (Fig 1
). Therefore, mean +/- standard deviation, Student's t test, and two-way analysis of variance cannot be used for the comparison of serum troponin-T levels between groups. Logarithmic transformation of the original values or Mann-Whitney U test are the methods of choice in such analysis.
|
Edt (y mL-1 kg), where Q is total liberation of troponin-T, BW is body weight (kg), Kd is the fractional rate of disappearance of troponin-T from the blood, Edt is the integral function of troponin-T serum levels calculated in each patient from the surface under the curves of serial determinations, and y is the specific unit of measurement (µg). The fractional disappearance rate of first-order processes follows a characteristic pattern, with an initial decay followed by a sustained monoexponential decay, and the fractional disappearance rate is calculated from the slope of the monoexponential portion of the curve [3]. This process demands multiple measurements during the phase of the marker's elimination from the plasma. In our study 20% of the patients (24/120) had either rising plasma troponin-T levels or a second peak 72 hours after the operation. It seems, therefore, unlikely that only five measurements, limited to a period of 48 postoperative hours, could provide complete and reliable fractional disappearance rate values for troponin-T. One possible solution to this would be simply to calculate the areas under the curves, as this makes no assumption regarding the monoexponential portion of the curve and hence fractional disappearance rates.
References
Department of, Surgery, Montreal Heart, Institute, 5000 E Belanger, Montreal, Pq H1t, 1c8, Canada
To the Editor:
Two questions were raised regarding our analysis of the changes in serum levels of troponin-T to evaluate myocardial protection with warm or cold blood cardioplegia, namely, the validity of using Student's t test and the fact that we measured the changes only during the first two postoperative days.
We have referred the issue of the statistical analysis with Student's t test to our consultants in biostatistics (A. Ciampi, PhD, and A. Couturier, MSc). As pointed out by Taggart and associates, serum troponin-T level is not normally distributed, and the t test may not be allowed. However, when the distribution of population ``is not normal, for large sample sizes, the t analysis is rescued by the Central Limit Theorem'' [1], and therefore the analysis with the t test is still valid because of the large sample size (200 patients) in our study. This is because when the sample size is large, the asymptotic distribution of the sample mean is normally distributed [2]. Nevertheless, for the sake of certainty, we performed the two nonparametric statistical tests they have suggested, namely, the logarithmic transformation and the Mann-Whitney U test, and both resulted in a significant difference in the changes in troponin-T, with p values (p < 0.00001) even smaller than that obtained with Student's t test (p < 0.0002) in our original analysis.
As for the second issue, we agree that troponin-T release may continue for up to 5 days after operation. However, as the release curves clearly show, the difference between the two groups occurs during the first 24 hours and the two curves become parallel and close together after 24 hours. Measurements for a longer period of time would not have altered this relationship and may have introduced other confounding factors such as late myocardial damage, which was not the purpose of the study. Total troponin-T release therefore represents that of the first 48 hours, during which six measurements were obtained. Because this test is costly and time-consuming, we chose to perform a limited number of samplings. We appreciate the interest expressed by Taggart and associates in our study.
References
This article has been cited by other articles:
|
|
 |
|
  I. Birdi, G. D. Angelini, and A. J. Bryan Biochemical Markers of Myocardial Injury During Cardiac Operations Ann. Thorac. Surg., March 1, 1997; 63(3): 879 - 884. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
