Sympathetic Dysinnervation and QT Variability
Sympathetic Dysinnervation and QT Variability
Diabetes mellitus affects an estimated 250 million patients worldwide and cardiac autonomic neuropathy is a very common complication, again estimated at 2.5% in the general diabetic population and up to 90% in long-standing type I diabetes mellitus. The diagnosis of cardiac autonomic neuropathy is of utmost importance as it affects significantly the prognosis. Both the cardiac involvement and confounding disorders need to be identified to prevent the unfavorable future of these patients. Cardiac involvement can be accomplished by mapping of the heart using I-MIBG scanning. Since the early 1980s, it was discovered that metaiodobenzylguanidine scanning detects cardiac autonomic dysinnervation. The washout rate is increased and the distribution is altered in patients with congestive heart disease and dilated cardiomyopathy. Abnormalities in MIBG uptake was encountered in dilated cardiomyopathy, myocardial infarction, and long QT patients. Subsequently this scanning was used to reveal cardiac sympathetic nervous dysfunction in diabetic patients. Diabetic patients had more extensive MIBG uptake than normal subjects or diabetic patients without sympathetic dysinnervation. MIBG competes with norepinephrine on the uptake into the sympathetic nerve terminals. If this innervation is damaged the uptake will be disturbed and the washout will be quick because the substance is not linked to the nerve terminal.
In the study published in this issue of the Journal, MIBG was used to detect sympathetic dysinnervation in diabetic patients. QT variance was the second measurement used and correlation between the 2 was intended. QT variance is an electrophysiological measurement and used in studies of noninvasive risk assessment. Although the pathophysiological basis of QT variance is not yet fully elucidated, it is certainly linked to the repolarization phase of the cardiac cycle.
During the last 3 decades the QT interval has been intensively studied. In 1990, the QT dispersion was introduced with great enthusiasm, for only little more than 10 years to be abandoned. It was never implemented in the clinical practice and remained merely a study tool. There were 3 main reasons for this failure: the difficulty to explain the pathophysiological basis and fully correlate the QT dispersion with the repolarization process, the failure to localize the defect to regions of the heart and methodological difficulties. The main reason for the methodological difficulties was the inaccuracy of the T wave ending detection. Low-amplitude, shallow T waves had the most hidden end because artifacts and filtering.
This difficulty is also applicable to QT variance. Suggested in the late 1980s, QT variance is not meant to localize the physiological or pathological process to certain region of the heart, but still requires accurate detection of the QT interval. It involves recording several consecutive QT intervals and to calculate the standard deviation of the differences between these consecutive QT intervals. QT variability is not a typical linear function, but a fractal one and belongs to more advanced group of tests. To overcome the inaccuracy of the T off detection, it was proposed to avoid using the whole length of QT and to limit the test to peak R-to-peak T. Detection of peak R is accurate, but occasionally determination of peak T is as difficult as the determination of T off. The other methodical limitations of QT variability are the need to define the lead used for recording (1 out of the 15 classical and orthogonal leads), the interobserver and intraobserver reproducibility, the duration of ECG recording and definition of the cutoff points between low and high risk. All these limitations made it difficult to compare studies and to agree upon accepted technical guidelines. The automatic recording also has technical limitations because the need to choose a lead, to define the sample and the duration of recording.
Berger et al. suggested a QT variance index (QTVI) in 1997. This index intended to normalize the QT variability to heart rate. This normalization has both positive and negative effects. Any combination and integration of different fractal tests amplifies the positive predictive value without affecting the already high negative predictive value. However, adding another proved risk predictor can alter the results not necessarily due to the effect of the tested QT variability. In 13 published major studies during the last 16 years a total of 3,014 patients were tested and although the tests had very different designs and methods, all demonstrated high predictive value of QTVI, especially in the highest quartile (<75%) group. Recently, Khandoker et al. demonstrated the correlation between QTVI and severity of cardiovascular autonomic neuropathy. This adds to the assumption that QTVI represents, at least partially, the sympathetic activity in the heart. However, QT represents the repolarization phase, and for this reason it depends on several membrane channels. These channel distributions and density are an internal cellular property and not necessarily affected by any autonomic nervous system activation.
The Sacre et al. study, published in this issue of the Journal, intended to correlate between QT variance, QTVI and MIBG scanning measured sympathetic dysinnervation. In the supine phase, although the QTVI was significantly higher in patients with sympathetic dysinnervation, this difference was contributed mainly by the heart-rate variability and not by the QT variability. As mentioned earlier, this is one of the disadvantages of integrated tests. In the standing phase, the heart-rate variability, as expected, was not changed significantly in the dysinnervation group. The QTVI had a trend toward a higher value due to QT variance change. Taking in consideration all the limitations of the test mentioned before, it is crucial that in a process so easily biased there is a need for a large cohort to achieve an acceptable result. Unfortunately, this interesting study cannot provide the desired answer because the very small patient population (occasionally bellow 10 and even 0 in 1 group). The statistical evaluation is unacceptably limited if the study population is so small.
A more standard method, accepted by all, is urgently required. Standardization of the sample choosing, duration of the recording, the cutoff point between high and low risk and length of QT interval used for analysis are primary prerequisites for large comparable studies. Only in this way QT variance and QTVI can be finally tested in large prospective studies needed to implement the test in clinical practice. Otherwise, the fate of QT variance will be similar to the fate of QT dispersion.
Editorial Comment
Diabetes mellitus affects an estimated 250 million patients worldwide and cardiac autonomic neuropathy is a very common complication, again estimated at 2.5% in the general diabetic population and up to 90% in long-standing type I diabetes mellitus. The diagnosis of cardiac autonomic neuropathy is of utmost importance as it affects significantly the prognosis. Both the cardiac involvement and confounding disorders need to be identified to prevent the unfavorable future of these patients. Cardiac involvement can be accomplished by mapping of the heart using I-MIBG scanning. Since the early 1980s, it was discovered that metaiodobenzylguanidine scanning detects cardiac autonomic dysinnervation. The washout rate is increased and the distribution is altered in patients with congestive heart disease and dilated cardiomyopathy. Abnormalities in MIBG uptake was encountered in dilated cardiomyopathy, myocardial infarction, and long QT patients. Subsequently this scanning was used to reveal cardiac sympathetic nervous dysfunction in diabetic patients. Diabetic patients had more extensive MIBG uptake than normal subjects or diabetic patients without sympathetic dysinnervation. MIBG competes with norepinephrine on the uptake into the sympathetic nerve terminals. If this innervation is damaged the uptake will be disturbed and the washout will be quick because the substance is not linked to the nerve terminal.
In the study published in this issue of the Journal, MIBG was used to detect sympathetic dysinnervation in diabetic patients. QT variance was the second measurement used and correlation between the 2 was intended. QT variance is an electrophysiological measurement and used in studies of noninvasive risk assessment. Although the pathophysiological basis of QT variance is not yet fully elucidated, it is certainly linked to the repolarization phase of the cardiac cycle.
During the last 3 decades the QT interval has been intensively studied. In 1990, the QT dispersion was introduced with great enthusiasm, for only little more than 10 years to be abandoned. It was never implemented in the clinical practice and remained merely a study tool. There were 3 main reasons for this failure: the difficulty to explain the pathophysiological basis and fully correlate the QT dispersion with the repolarization process, the failure to localize the defect to regions of the heart and methodological difficulties. The main reason for the methodological difficulties was the inaccuracy of the T wave ending detection. Low-amplitude, shallow T waves had the most hidden end because artifacts and filtering.
This difficulty is also applicable to QT variance. Suggested in the late 1980s, QT variance is not meant to localize the physiological or pathological process to certain region of the heart, but still requires accurate detection of the QT interval. It involves recording several consecutive QT intervals and to calculate the standard deviation of the differences between these consecutive QT intervals. QT variability is not a typical linear function, but a fractal one and belongs to more advanced group of tests. To overcome the inaccuracy of the T off detection, it was proposed to avoid using the whole length of QT and to limit the test to peak R-to-peak T. Detection of peak R is accurate, but occasionally determination of peak T is as difficult as the determination of T off. The other methodical limitations of QT variability are the need to define the lead used for recording (1 out of the 15 classical and orthogonal leads), the interobserver and intraobserver reproducibility, the duration of ECG recording and definition of the cutoff points between low and high risk. All these limitations made it difficult to compare studies and to agree upon accepted technical guidelines. The automatic recording also has technical limitations because the need to choose a lead, to define the sample and the duration of recording.
Berger et al. suggested a QT variance index (QTVI) in 1997. This index intended to normalize the QT variability to heart rate. This normalization has both positive and negative effects. Any combination and integration of different fractal tests amplifies the positive predictive value without affecting the already high negative predictive value. However, adding another proved risk predictor can alter the results not necessarily due to the effect of the tested QT variability. In 13 published major studies during the last 16 years a total of 3,014 patients were tested and although the tests had very different designs and methods, all demonstrated high predictive value of QTVI, especially in the highest quartile (<75%) group. Recently, Khandoker et al. demonstrated the correlation between QTVI and severity of cardiovascular autonomic neuropathy. This adds to the assumption that QTVI represents, at least partially, the sympathetic activity in the heart. However, QT represents the repolarization phase, and for this reason it depends on several membrane channels. These channel distributions and density are an internal cellular property and not necessarily affected by any autonomic nervous system activation.
The Sacre et al. study, published in this issue of the Journal, intended to correlate between QT variance, QTVI and MIBG scanning measured sympathetic dysinnervation. In the supine phase, although the QTVI was significantly higher in patients with sympathetic dysinnervation, this difference was contributed mainly by the heart-rate variability and not by the QT variability. As mentioned earlier, this is one of the disadvantages of integrated tests. In the standing phase, the heart-rate variability, as expected, was not changed significantly in the dysinnervation group. The QTVI had a trend toward a higher value due to QT variance change. Taking in consideration all the limitations of the test mentioned before, it is crucial that in a process so easily biased there is a need for a large cohort to achieve an acceptable result. Unfortunately, this interesting study cannot provide the desired answer because the very small patient population (occasionally bellow 10 and even 0 in 1 group). The statistical evaluation is unacceptably limited if the study population is so small.
A more standard method, accepted by all, is urgently required. Standardization of the sample choosing, duration of the recording, the cutoff point between high and low risk and length of QT interval used for analysis are primary prerequisites for large comparable studies. Only in this way QT variance and QTVI can be finally tested in large prospective studies needed to implement the test in clinical practice. Otherwise, the fate of QT variance will be similar to the fate of QT dispersion.
Source...