Ischemia induces aggravation of baseline repolarization abnormalities in left ventricular hypertrophy: a deleterious interaction

Aigars Rubulis, Jens Jensen, Gunilla Lundahl, Jari Tapanainen, Lennart Bergfeldt


Epidemiological studies show that left ventricular hypertrophy (LVH) and hypertension (HT) in coronary artery disease increases the risk for cardiovascular events including sudden cardiac death (SCD). According to experimental studies, myocardial hypertrophy is associated both with altered electrophysiological properties (including prolonged repolarization) and increased vulnerability to ischemia. However, human data to support a repolarization-related mechanism for the increased SCD risk has not been provided. We therefore studied 187 patients undergoing three-dimensional vectorcardiographic monitoring during coronary angioplasty. Eight parameters reflecting different aspects of ventricular repolarization were used: 1) the ST segment (ST-VM and STC-VM), 2) the T vector (QRS-T angle, Televation, and Tazimuth), and 3) the T vector loop (Tavplan, Teigenv, and Tarea). Data collection was performed at rest and at the time of maximum ischemia during coronary occlusion. The patients were divided into three groups: 33 patients with ECG signs of LVH (18 with HT), 54 with HT but without LVH signs, and 100 patients with neither. Coronary artery disease patients with LVH not only had the most abnormal baseline repolarization (as expected) but also a significantly more pronounced repolarization response during coronary occlusion, whereas HT patients had mean parameter values between LVH patients and those without neither HT nor LVH signs. Because there is a relation between increased SCD risk and repolarization disturbances in various clinical settings, the results of the present study are in agreement with animal data and epidemiological observations, although other factors than disturbed repolarization might be of importance.

  • hypertension
  • myocardial ischemia
  • coronary artery disease
  • electrocardiography
  • vectorcardiography

sudden cardiac death (SCD) is one of the major health problems in the industrialized part of the world. Even though its occurrence in high-risk individuals among post-myocardial infarction (MI) patients has been in focus for many years, SCD often strikes unexpectedly as the first manifestation of coronary artery disease (CAD) and in mildly symptomatic patients with hypertension and mild to moderate heart failure (22). Apart from CAD, a common denominator in SCD victims is some degree of myocardial hypertrophy, and hypertension-induced left ventricular hypertrophy (LVH) adds risk for arrhythmias and SCD in CAD (22). Furthermore, LVH per se has been recognized as an important risk factor for nonfatal and fatal cardiovascular events including SCD, presumably owing to ventricular arrhythmias (15, 18, 19, 21). According to experimental studies, this risk for malignant arrhythmias is most consistently linked to a prolonged and heterogeneous repolarization, although myocardial disarray, fibrosis, and gap junction heterogeneity might contribute (1, 10). Some patients with LVH, in particular those with hypertension, also have an increased propensity for developing CAD. Although experimental studies (see discussion) have shown reduced tolerance to ischemia and subsequent reperfusion in myocardial hypertrophy, including increased dispersion of repolarization, higher incidence of ventricular arrhythmias, and greater infarct size (6, 8, 29, 31, 33), the mechanistic interaction between LVH and ischemia has rarely been studied in humans. Because of these experimental observations, we focused on ventricular repolarization and its response to acute ischemia in a human model of coronary occlusion during percutaneous coronary intervention (PCI).

In situ global assessment of ventricular repolarization and its response to interventions is not easily obtained in humans. Invasive techniques do not provide continuous global assessment and also pose ethical and practical problems. Because the standard ECG has well-known limitations, three-dimensional vectorcardiographic (VCG) assessment of repolarization has emerged as a possible tool (2). This is mainly due to the proven relation between the degree of repolarization heterogeneity and the morphology of the T vector loop (from here: T loop) both at baseline and during acute ischemia (2, 17, 23, 26). Thus the normally elongated elliptical T loop in one preferential plane becomes more circular and bulgy during acute ischemia, which can be expressed in quantitative terms (see methods). Our hypothesis was that CAD patients with signs of LVH would show more pronounced repolarization abnormalities during acute ischemia than patients without.



The patients in this study were initially enrolled for a VCG study on the ST segment response during PCI, approved by the Ethics Committee of Karolinska Institutet. One hundred ninety-two consecutive patients gave their informed consent to participate and underwent successful PCI (14). Inclusion criteria were angina pectoris, a positive exercise test, or both, and ≥1 significant angiographic stenosis (≥60% vessel diameter reduction). Exclusion criteria were MI within 48 h, ongoing ischemia or angina pectoris at the start of the procedure, atrial fibrillation, bundle branch block, and pacemaker rhythm. Five patients had to be excluded for technical reasons. For the specific purpose of this study, the remaining 187 patients were allocated to three subgroups on the basis that they would represent a spectrum of increasing degree of LVH. Thus one subgroup included 33 patients with LVH according to ECG criteria (30); in 20 (87%) of the 23 patients with available echocardiography records, the presence of myocardial hypertrophy was supported. A history of hypertension (HT) was present in 18, and 5 more patients had blood pressure measurements suggesting at least intermittent HT. Ventricular remodeling after MI was the probable cause of LVH in six, whereas one patient each had aortic stenosis and obesity. Two patients had no identifiable cause except their CAD. Fifty-four were allocated to the “hypertensive” subgroup, on the basis of a history of arterial HT, and 100 to the “normotensive” subgroup, because they had neither a history of HT nor ECG signs of LVH. Patient characteristics are presented in Table 1.

View this table:
Table 1.

Clinical characteristics of the 187 patients subdivided into normotensive, hypertensive, and LVH subgroups

VCG monitoring of PCI.

This procedure has been described in detail elsewhere, including the two VCG parameters applied for ischemia monitoring in patients with acute coronary syndromes (14, 26, 28). Thus ST-VM expresses the vector magnitude of the ST segment deviation from the isoelectric level, and STC-VM is the ST-VM corrected for baseline deviation.

The T vector and loop morphology was analyzed off-line from the averaged QRST complexes via customized software (23). The reproducibility of the different parameters used in this study and their ability to separate CAD patients from healthy controls have been reported elsewhere (26, 32). In brief, for each 10-s period, an averaged three-dimensional T loop and maximum QRS and T vectors were determined. The electrical forces in the heart show interindividual differences in direction. To achieve the best anatomical representation of these forces, the spatial characteristics of the maximum T vector were expressed by the T vector elevation (Tel), the T vector azimuth (Taz), and the QRS-T angle (26, 32) (Fig. 1). Tel is the angle between the maximum T vector and an axis perpendicular to the horizontal plane, zero being the vector directed downward (a scale from 0° to 180°). This angle increases during ischemia induced by left anterior descending artery (LAD) occlusion (26). Taz is the angle between the projected maximum T vector on the horizontal plane (XZ) and the left extremity of the X-axis, zero being the vector pointed to the left. Forward vector motions are defined as 0 to 180°, and backward vector motions as 0 to −180°. Also this angle increases during LAD occlusion (26). The QRS-T angle reflects the angular difference between the maximum QRS vector and maximum T vector (on a scale from 0 to 180°). In other words, it represents the simplified spatial deviation between the dominant ventricular depolarization and repolarization waves, which increases in pathological disorders like MI, LVH, and left bundle branch block (3, 16).

Fig. 1.

T elevation (Tel) (A) is the angle between the maximum T vector and an axis perpendicular to the horizontal plane, 0° being the vector directed downward. T azimuth (Taz) (B) is the angle between the projected maximum T vector on the horizontal plane (XZ) and the left extremity of the X axis, 0° being the vector pointed to the left. Forward vector motions are defined as 0 to 180° and backward vector motions as 0° to −180°, respectively. The QRS-T angle (C) reflects the angular difference between the maximum QRS and T vectors (on a scale from 0 to 180°). Reprinted from Heart Rhythm, vol. 1, Rubulis et al. T vector and loop characteristics in coronary artery disease and during acute ischemia, p. 28–34, Copyright 2005, with permission from Heart Rhythm Society (26).

The T loop was characterized by three parameters: Tarea, Tavplan, and Teigenv (23, 26) (Fig. 2). Tarea (measured in μVs) is the “three-dimensional” area between the curve of the T wave and the isoelectric line from the QRS end (=the J point) to T end in X, Y and Z Frank leads, calculated as Tx2 + Ty2 + Tz2. Tavplan (measured in μV) expresses the bulginess of the loop defined as the mean distance from the preferential plane for the sample values in the T loop. Considering the sample values of the T loop as mass points, the matrix of inertia can be defined (7). Tarea and Tavplan both increase during acute ischemia (23, 26). Teigenv (unitless) expresses the form and geometry of the T loop and was calculated as the quotient between the two highest and perpendicular eigenvalues (approximate diameters) of the matrix of inertia = (d1/d2)2, where d1 ≥ d2, where d1 and d2 are diameters 1 and 2, respectively. For the special case of a circle the Teigenv = 1. The more elliptical (“healthier”) the loop, the higher is the Teigenv. The Teigenv thus has an inverse relation to repolarization heterogeneity (2, 17, 23) and decreases during LAD occlusion, when QT and JT dispersion increase (23, 26). [Note that if the largest “diameter” is chosen as denominator (2) an elliptical form will have a low value; still the more circular the T loop the closer to 1 this value becomes].

Fig. 2.

Schematic illustration of the principles in calculating the T vector loop parameters. A: Tarea is the “3-dimensional” area under the curve from QRS end to T end in X, Y, and Z Frank leads, calculated as Formula. B: Tavplan expresses the bulginess of the loop as the mean difference from the preferential plane. C: Teigenv expresses the form of the loop. For an ellipse, the eigenvalues are proportional to the squares of the major and minor axes of the ellipse. The Teigenv parameter will therefore be Teigenv = (d1/d2)2. For the special case of a circle the Teigenv = 1. Reprinted from J Internal Med, vol. 248, Nowinski et al. Changes in ventricular repolarization during percutaneous transluminal angioplasty in humans assessed by QT interval, QT dispersion and T vector loop morphology, p. 126–136, Copyright 2000, with permission from Blackwell Publishing (23).

For each patient, four measurements on the trend curves were obtained. 1) The baseline for each parameter was defined before the contrast injection. Calculation of mean and SD was based on 10-s averages obtained during a 3-min recording (≤18) with the least noise level. 2) A single value was obtained at the end of the first balloon inflation. 3) Another single value was obtained at the time for the maximum STC-VM, serving as the reference for maximum ischemia (12). 4) The last series of values was obtained starting 2 min after the last deflation; the mean and SD were calculated as for the baseline value.

Deflections in the trend curve that were obvious artifacts or due to arrhythmias were discarded and not used in the analysis.

Statistical methods.

Mean and standard deviation or median (with the 25th and 75th percentiles) were used for descriptive statistics. Between-groups comparisons of data with approximately normal distribution (Shapiro-Wilks W-test) were performed by analysis of variance, whereas data with skewed distribution were analyzed with Kruskal-Wallis statistic. Data with an extreme positive skewness (Teigenv) were logarithmically transformed before analysis. For categorical variables, χ2 analysis of contingency table was applied. Within-group analyses of data with approximately normal distribution were performed by repeated-measures ANOVA, whereas skewed data were tested with the Friedman two-way ANOVA by ranks. Analysis of covariance (ANCOVA) was applied to test the response to ischemia, taking into account the baseline differences between subgroups. A backward stepwise multiple linear regression analysis, with inclusion at the F 4.00 level and exclusion at the F 3.00 level, was used to evaluate the influence of age, gender, diabetes mellitus, HT, LVH, previous MI, presence of visible collaterals, and treatment with beta-blockers on every VCG parameter, which served as dependent variables. A P value <0.05 was considered statistically significant with the exception for ANOVA in between-group comparison, where Bonferroni correction was applied. All analyses were performed using the program Statistica for Windows, release 6.0 (StatSoft, Tulsa, OK).


Patients in the LVH subgroup were older, but had less widespread CAD. HT patients received treatment more often with calcium antagonists and angiotensin-converting enzyme inhibitors. Otherwise there were no significant differences between the groups (Table 1).

Baseline comparisons.

Baseline values of all VCG parameters are presented in Table 2 and Fig. 3. The LVH patients had increased ST-VM, a wider QRS-T angle, a more anterior and cranial T vector, and a rounder T loop than normotensive patients, all reflecting a more abnormal repolarization at baseline as expected. The HT patients differed statistically significantly from the normotensive only by a wider QRS-T angle. There was, however, a general pattern where the mean values of most parameters in the HT group, such as Tel, Taz, Teigenv, and ST-VM, were between those of the normotensive and LVHsubgroups, probably reflecting a continuous spectrum of myocardial hypertrophy.

Fig. 3.

Relative differences between the baseline values of different vectorcardiographic parameters are shown. Mean values of the parameters of the normotensive group serve as reference (100%) to calculate means ± SE (boxes) and standard deviations (whiskers) in all groups. Hypertensive and left ventricular hypertrophy (LVH) groups are compared with the normotensive: *P < 0.05; **P < 0.01. ST-VM, vector magnitude of ST segment deviation from isoelectric level.

View this table:
Table 2.

VCG parameters at baseline, at the end of the 1st balloon inflation, and at maximum ischemia (max STC-VM)

In multiple linear regression analysis, the width of the QRS-T angle was increased in the presence of LVH, HT, and previous MI, respectively (P < 0.001; P < 0.01; P < 0.05; R2 = 0.15).

Response to balloon inflation.

All three subgroups showed a response in VCG parameters to balloon inflation, but the between-group comparisons showed that this response was subgroup dependent (Table 2). Thus during coronary occlusion and myocardial ischemia, the T loop became more bulgy (Tavplan increased) and more circular (Teigenv decreased), reflecting more abnormal (heterogeneous) repolarization (Fig. 4).

Fig. 4.

T vector loops of 3 patients are presented at baseline (B), and at maximum ischemia (I) during balloon inflation in the right coronary artery. T loops are shown in 3 dimensions/planes (columns). The T loop scale was determined automatically by the software and vary. Top row (Normotensive, B) shows the most normal T loop with an elongated, elliptical morphology (Teigenv 76) in 1 preferential plane (Tavplan 0.4 μV, here approximately perpendicular to the transversal plane). The most bulgy loop (Tavplan 1 μV) is seen in one of the middle rows (Hypertensive, I). The most circular loop (Teigenv 1) is seen in the bottom row. See text for definitions.

Generally, the LVH patients had the most altered mean values of VCG parameters at maximum ischemia (reference = maximum STC-VM). Thus, on top of the abnormal baseline repolarization, there was a significantly more pronounced reaction to ischemia in the LVH subgroup as judged by the response of the T vector (Taz, ANCOVA, P < 0.05) and the changes in STC-VM (Kruskal-Wallis, P < 0.05) (Fig. 5).

Fig. 5.

Between-group differences at maximum ischemia are shown for Tazimuth (A) and ST-VM corrected for baseline deviation (STC-VM) (B). See also Table 2.

In hypertensive patients, most parameter mean values during ischemia were between the values observed in LVH and those in normotensive patients, just as at baseline. However, a statistically significant difference was observed only in the Tarea, which increased significantly more in HT compared with normotensive patients (ANCOVA, P < 0.05).

When looking at the relation between different parameters, there was a correlation between the ST segment changes and changes in Tarea (Spearman correlation coefficient: r ≈ 0.80, P < 0.00001), but otherwise no clear relations appeared. This overall lack of correlation might be due either to the fact that the other repolarization parameters were poor measures of ischemia and/or to the fact that any correlation was confounded by the alterations observed already at baseline in the HT and LVH groups. Support for the latter interpretation was obtained from the within-group comparisons. Thus, when looking at the normotensive subgroup with the least baseline alterations significant changes were observed both in the T vector and (all 3) T loop parameters. In contrast, with increasingly more pronounced baseline abnormalities there were fewer significant changes; 2 and 1 T loop parameters/s, respectively, changed significantly in the hypertensive and LVH subgroups (Table 2, see Methodological aspects and limitations). Overall there was a significant inverse relation between the baseline values of T vector and loop parameters and the changes at maximum ischemia, or, turned the other way around, an abnormal baseline repolarization had a limited range within which to become even more abnormal during coronary occlusion. The exception was Taz (see Response to balloon inflation, second paragraph). However, ST-VM and STC-VM were sensitive to ischemia regardless of the subgroup, and STC-VM thus showed a significantly more pronounced response in the LVH group (Table 2, Fig. 5B). Because there was a difference in the extent and distribution of CAD a main effect ANOVA as well as multiple regression analysis was applied. The presence or absence of LVH came out as the most important factor in this analysis, whereas the number of the diseased vessels only had minor influence on the response to ischemia (as reflected by STC-VM).


Myocardial hypertrophy is, independent of its cause, a condition that electrophysiologically is characterized by prolonged repolarization (1, 10) and has a “reduced repolarization reserve” (27), essentially meaning reduced compensatory mechanisms for counteracting perturbations of repolarization. The clinical consequence is an increased risk for ventricular arrhythmias, including ventricular fibrillation and SCD (1, 10, 19, 29). HT not only is the most common cause of LVH, but it also accelerates the development of CAD and increases the oxygen demand, creating a highly life-threatening combination that is reflected by the associated increased incidence of SCD (34). According to experimental studies, there is a decreased tolerance to induced ischemia and reperfusion in myocardial hypertrophy (6, 8, 29, 31, 33). Several mechanistic factors have been identified, although their relative importance remains to be clarified. Thus myocardial hypertrophy is associated with 1) impaired coronary vasodilator reserve; 2) decreased capillary density and increased diffusion distance, owing to increased myocardial cell diameter without proportional proliferation of capillary vessels; 3) decreased high-energy phosphate content; 4) impaired fatty acid oxidation leading to increased dependence on glucose metabolism; and 5) reduced myocardial glucose transport into the cell (6, 8, 29, 33). The net result is an increased ischemic zone on top of a reduced repolarization reserve and hence an increased risk for ventricular fibrillation and SCD (6, 8, 29, 33). It is therefore logical to postulate that also in this human model the patients with LVH would have decreased tolerance to ischemic events. Our results provide evidence that this hypothesis is correct, at least in relation to electrophysiological changes. LVH patients indeed showed most abnormal repolarization, both at baseline (as expected) and during coronary occlusion, and thus were most sensitive to ischemia, followed by the HT subgroup. In addition, experimental evidence demonstrates good correlation between ST-VM, STC-VM, and the size of the ischemic area (13, 24). There is a relation between increased SCD risk and repolarization disturbances in various clinical settings. The results of the present study are therefore in line with animal data and epidemiological observations, although other factors than disturbed repolarization might be of importance. They are also in line with the early clinical observations of a so-called “stone heart” after cardiac surgery in patients with myocardial hypertrophy (5). Malignant arrhythmias are rare during PCI, and it was therefore not possible to relate the degree of repolarization abnormality to the occurrence of procedure-related arrhythmias. An alternative approach would be to assess the prognostic value of these parameters in relation to cardiovascular events during follow-up.

Methodological aspects and limitations.

ECG criteria for LVH have high specificity (>95%) but low sensitivity (20). Therefore, systematic high-quality echocardiography or magnetic resonance imaging would have been preferable. The available echocardiography data, although not complete, arrived at a very similar result and basically confirmed the presence of LVH. Also the HT and normotensive subgroups presumably included patients with some degree of myocardial hypertrophy, e.g., as a consequence of post-MI remodeling. In fact, our results very much agree with the notion that there was a more or less continuous spectrum of increasing degree of myocardial hypertrophy when going from the normotensive (55% with previous MI and probable ventricular remodeling) via the hypertensive to the LVH subgroup, rather than sharp borders between them. This would rather reduce than augment the between-group differences and supports the validity of our results. Furthermore, at the time for recruitment, the diagnosis of hypertension was based on repeated measurements of blood pressure >160/95 mmHg, which is too conservative a criterion according to a recent classification (9). However, because both blood pressure and left ventricular mass are continuous variables, the entire patient group probably reflects continuity in these aspects, and the main results of our study should therefore be valid. In fact, this overlap should also act to reduce rather than augment any differences between the subgroups.

Unequal number of patients in the three subgroups affects statistical power and may also contribute to underestimation of subgroup differences. Patients with HT more often received angiotensin-converting enzyme inhibitors and calcium antagonists, which also might have influenced the results.

The data, obtained at maximum STC-VM and at the end of the procedure, could have been influenced by different degrees of preconditioning. Therefore, the end of the first inflation served as the reference for comparisons with the data obtained at maximum STC-VM (Table 2). Additionally, the sudden and short-lasting coronary occlusion during PCI is not a perfect model of the “naturally” appearing ischemia. However, it is probably the best experimental situation available in humans, and its validity and reliability have been shown previously (25).


First of all, the method we have used can be applied to monitor and assess the effects of both short- and long-term therapeutic or modifying interventions, as has been shown for pacing-induced ventricular remodeling (32). Furthermore, the QRS-T angle might become an important tool for epidemiological studies and risk stratification: because it appears to be a sensitive marker of altered repolarization in LVH and HT, it correlates with the severity of LVH (11), which itself is a prognostic factor for future cardiovascular events (15, 18, 19), and it has shown prognostic value in post-MI and general populations (3, 16).

Conclusion and implications.

In a population of CAD patients, not only was the presence of LVH associated with more abnormal repolarization at baseline, but there was also a markedly aggravated repolarization in response to ischemia induced by coronary occlusion. These results are in line with the animal experiments, showing a decreased tolerance to ischemia in myocardial hypertrophy, and available epidemiological data, demonstrating an increased risk for nonfatal and fatal cardiovascular events (including SCD) in the presence of LVH. To reduce this risk, intense therapeutic efforts should be directed against both coronary atherosclerosis and HT, as well as against their structural sequels.


This work was supported by a scholarship (A. Rubulis) from the Swedish Institute via the New Visby Programme, by grants from the Seraphimer Foundation, Margaretha af Ugglas’ Foundation, and the Swedish Heart-Lung Foundation.


Gunilla Lundahl was, at the time of this study, an employee of Ortivus AB, Täby, Sweden, and the VCG systems, which are commercially available, were a loan from Ortivus AB. There are no other disclosures to be made.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract