Microvolt T-Wave Alternans in an Unselected Heart Failure Population: Pros and Cons

This editorial refers to ‘Profile of microvolt T-wave alternans testing in 1003 patients hospitalized with heart failure’, by C.E. Jackson et al., published in this issue on pages 377–386. Currently, reduced left ventricular ejection fraction (LVEF) is the strongest predictor of life-threatening arrhythmias and is used to select patients for primary prophylactic implantable cardioverter-defibrillator (ICD) implantation. However, using this measure of mechanical impaired function to predict an electrophysiological substrate will result in a relatively high number of patients to treat those susceptible. Two trials for primary prophylactic ICD implantation, Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) and Multicenter Automatic Defibrillator Implantation Trial II (MADIT II), report a percentage of 76% and 79% of patients that did not use their device during the follow-up of 21 and 45.5 months, respectively.1,2 On the other hand, a substantial number of sudden cardiac deaths (SCDs) occur in patients with a preserved LVEF, who are not eligible for an ICD on the basis of this risk stratifier. However, it is known that in patients with a cardiac resynchronization therapy device with defibrillator function (CRT-D), improvement of LVEF >35% reduces the number of ICD interventions.3 To optimize patient selection for primary prophylactic ICD implantation, electrophysiological tools for risk stratification have been tested. Microvolt T-wave alternans (MTWA), defined as an alternation in T-wave morphology in every other heart beat reflecting spatiotemporal heterogeneity of repolarization, has been demonstrated to be a significant predictor of arrhythmic and cardiovascular death in patients with ischaemic and non-ischaemic cardiomyopathies. It is regarded as a class IIa index for stratification of fatal ventricular arrhythmias in the guidelines (ACC/AHA/ESC) 2006.4 Although negative predictive values were high, ranging from 91.7% to 100%, the hazard ratio of a positive MTWA test seems to vary depending on the different populations studied (hazard ratio 1.24–6.53).5–12 For detection of MTWA, the spectral method requires a graded increase in heart rate (HR) usually achieved by exercise.13 It is not possible to perform an MTWA test in patients who are unable to exercise or who have atrial fibrillation (AF). As a consequence, the studies, which investigated MTWA have been conducted in highly selected cohorts. The number of patients excluded based on these limitations is often not mentioned, making it difficult to determine its applicability in clinical practice. Therefore, it would be interesting to investigate the patients' eligibility for MTWA testing in an unselected heart failure (HF) population, as Jackson et al.14 have done in this issue of the journal. In this study,14 2361 patients with decompensated HF were screened, and 1003 were prospectively enrolled independently of LVEF. After 1 month, only 648 patients were able to perform an MTWA test. Of these 648 patients, 242 (37%) patients had to be excluded due to AF, inability to exercise [n = 43 (6.6%)], or continuous ventricular pacing [n = 33 (5.1%)]. In only 330 patients (47%) was an MTWA test performed: 100 (30.3%) were positive, 152 (46.1%) indeterminate, and 78 (23.6%) negative. In the 47% of patients in which MTWA testing could be performed, a remarkably high number of indeterminate test results (49%) were reported,14 considerably higher than in previous studies, which show a range of 12–41% (median 21%) (Table 1). An indeterminate test result can be due to patient factors (failure to maintain HR between 105 and 110 b.p.m. for ≥1 min, frequent ectopy, unsustained MTWA) or technical factors (excessive noise on electrocardiogram, rapid rise in HR through the target exercise HR range of 105–110 b.p.m.). Inability to achieve the target HR is reported as the main reason for an indeterminate result14–16 (see Table 2). The proportion of patients on beta-blockers used in this unselected HF population is comparable with previous studies and does not seem to explain the discrepancy in the number of indeterminate tests. The presented study provided further information concerning the reasons for being unable to achieve target HR: chronotropic incompetency and physical limitations. The first category consisted of patients who attained a maximum HR below 110 b.p.m. which failed to rise with further exercise. The second group discontinued exercise because of physical limitations without meeting the criteria for a positive or negative result (i.e. 59 of the total number of 192 patients).14 Patients with an indeterminate test seemed to be characterized by a poorer clinical status, as shown by independent predictors such as older age and a lower peak HR during exercise. Furthermore a higher log brainnatriuretic peptide (logBNP) and lower peak oxygen consumption (peak VO2) was observed.12,14 Therefore, the risk of dying from causes other than arrhythmic death is probably higher, which may explain how MTWA reached significance in relation to total mortality but not to arrhythmic death.8 In the results of most trials, positive and indeterminate tests are grouped together. Therefore, it is not possible to assess to what extent the number of indeterminate tests can be attributed to these results. In a comparable cohort of 303 patients with congestive HF,12 after adjusting for peak VO2, the predictive value of non-negative MTWA was no longer statistically significant (hazard raio 1.18, 95% confidence interval 0.64–2.17, P < 0.06). The number of indeterminate tests accounted for this, since peak VO2 was lower in the indeterminate group compared with the MTWA-positive group (15.6 vs. 19.2, P < 0.001). This suggests that combining the results may obscure the possible predictive value of a positive result. It would therefore be interesting to assess if, and to what extent, the number of indeterminate tests influences the predictive value of MTWA in these HF patients. The current study14 did not show any results of mortality and incidence of SCD during follow-up. Therefore, the predictive value of MTWA cannot be assessed. Nevertheless, it is important to discuss what the value of the prevention of SCD is in these patients with severe HF. In the SCD-HeFT trial, patients with an admission for HF in the previous 3 months were excluded and a subanalysis of this trial showed no benefit of ICD implantation in patients with severe HF New York Heart Association (NYHA) class III (5 year mortality rate 48% vs. 45% in the placebo group). In risk stratification for primary prophylactic ICD, three types of HF patients are of interest: patients with (i) LVEF ≤35%, eligible for ICD, who will not need ICD therapy; (ii) LVEF ≥35%, not eligible for ICD implantation, who would benefit from this; and (iii) a poor prognosis more likely to die from causes other than arrhythmic death but do fulfil ICD criteria. Clinical applicability of MTWA testing is limited to selected groups. In an unselected HF population, MTWA testing results in a substantial high number of indeterminate tests. Therefore, it is questionable whether MTWA alone is able to be of predictive value and for what patient population. With an increasing number of primary prophylactic ICD implants (LVEF of ≤35%), concerns rise about cost-effectiveness and the benefit/complication ratio. Investigations to improve risk stratification remain worthwhile, although many fail to reach a significant predictive value. To embrace the complexity of the various cardiac substrates underlying sudden cardiac death, combining different methods of risk stratification could be the solution. An ongoing multicentre trial, EU-TrigTreat, sponsored by the European Commission, enrols patients eligible for ICD implantation to compare a number of invasive and non-invasive electrophysiological risk markers (including MTWA and short-term variability) and explore their (combined) predictive value for life-threatening arrhythmias.17 This work received funding from the European Community's Seventh Framework Program FP7/2007-2013 under grant agreement No. HEALTH-F2-2009-241526, EUTTrigTreat. Conflict of interest: none declared.