Non-invasive Risk Stratification Early After a Myocardial Infarction - The Risk Estimation Following Infarction Non-invasive Evaluation (REFINE) Study

Register or Login to View PDF Permissions
Permissions× For commercial reprint enquiries please contact Springer Healthcare:

For permissions and non-commercial reprint enquiries, please visit to start a request.

For author reprints, please email
Average (ratings)
No ratings
Your rating
Copyright Statement:

The copyright in this work belongs to Radcliffe Medical Media. Only articles clearly marked with the CC BY-NC logo are published with the Creative Commons by Attribution Licence. The CC BY-NC option was not available for Radcliffe journals before 1 January 2019. Articles marked ‘Open Access’ but not marked ‘CC BY-NC’ are made freely accessible at the time of publication but are subject to standard copyright law regarding reproduction and distribution. Permission is required for reuse of this content.

Why Do We Need Risk Stratification Tools?

Sudden death accounts for between 300,000 and 500,000 deaths each year in North America.1 Patients with a history of myocardial infarction (MI) have a four-fold higher risk of sudden death than those without such a history. Most sudden deaths in ambulatory populations result from life-threatening ventricular arrhythmias that lead to a cardiac arrest.2 Since survival from an out-of-hospital cardiac arrest is typically poor,1 identifying patients prior to its development is essential.

While the implantable cardioverter–defibrillator (ICD) effectively treats life-threatening ventricular arrhythmias, our current approach of identifying patients who may benefit from a prophylactic ICD (by identifying a low ejection fraction [EF]) is hampered by poor sensitivity and specificity.3,4 Multiple non-invasive markers have been developed to identify patients at risk of a cardiac arrest. These tools can be categorised broadly as either assessing autonomic tone or evaluating underlying electrical substrate. While individual markers of impaired autonomic tone identify patients at risk, when used as single measures they are hampered by poor sensitivity.5–7 These markers of impaired autonomic tone typically identify a three- to five-fold higher risk of serious events after MI, but fewer than one-third of at-risk patients are identified. Markers of electrical substrate have focused on beat-to-beat changes in cardiac repolarisation or T-wave alternans.8–11

While these techniques have merit, individually they are hampered by low positive accuracy8 or poor sensitivity.10,11 Thus, methods that identify most patients at risk of a cardiac arrest and provide reasonable positive accuracy are required. Studies conducted prior to the contemporary era of aggressive post-MI care suggest that combining measures of autonomic tone with electrical substrate may aid in patient identification.12 However, there are no contemporary data to support this concept. Recent refinements assessing autonomic tone5–7 and electrical substrate8–11 have provided the opportunity to validate this concept in patients receiving optimal post-MI management.

Optimal Risk Assessment After Myocardial Infarction

The Risk Estimation Following Infarction Non-invasive Evaluation (REFINE) study was designed with two main goals:3 to determine when non-invasive test results provide the most reliable information on future risk, and to derive an optimal combination of non-invasive parameters to identify patients at risk of serious outcomes. A group of 322 patients underwent a battery of non-invasive assessments in the acute (two to four weeks) and non-acute (10 to 14 weeks) periods after MI to determine the best time to assess risk. A typical post-MI population was enrolled (see Table 1). During an average follow-up of four years, 30 patients died and 24 patients suffered a fatal or near-fatal cardiac arrest. Non-invasive testing of patients in the acute and non-acute early post-MI periods was performed, since significant left ventricular remodelling typically occurs during the 12 weeks after MI.13 The non-invasive assessments included state-of-the art measures of autonomic tone (heart-rate turbulence, heart-rate variability, baroreflex sensitivity) and electrical substrate (T-wave alternans, signal-averaged electrocardiogram [ECG]). Most patients in the REFINE study (75%) underwent revascularisation in the initial post-MI period. The majority of patients also received aggressive medical therapy throughout follow-up, with 80% receiving beta blockers, angiotensin-converting enzyme (ACE) inhibitors or angiotensin receptor blockers, antiplatelet agents and statins at three years of follow-up.

When Should We Test Post-myocardial Infarction and Which Tests Are Best?

Testing in the non-acute phase provided more reliable information on risk of all of the non-invasive parameters than compared with testing in the initial four weeks post-MI. This was not surprising, as significant improvements in EF were observed over the initial eight to 10 weeks post-MI. In fact, the average relative improvement in EF was 18% (see Table 1). The REFINE study also demonstrated that autonomic measures obtained from a 24-hour ambulatory ECG recording or Holter monitor provided similar information to more complex testing. Similar results were also observed for T-wave alternans measured using an exercise treadmill protocol and T-wave alternans measured by a Holter monitor immediately after a submaximal exercise test. Thus, reliable non-invasive test results were achieved using a single, relatively simple testing approach.

What Is the Optimal Combination of Parameters?

Combining markers of autonomic tone with measures of electrical substrate better identified patients at risk of serious outcomes than any of the tests alone. Depending on the combination of parameters used, a six- to eight-fold increased risk of cardiac arrest was observed for patients who were categorised as having abnormal test results (see Figure 1). Compared with prior studies using one parameter, the combination of parameters in the REFINE study yielded substantially higher sensitivity. While previous studies have demonstrated sensitivities of <30%, the REFINE approach doubled this. Importantly, the combination of parameters identified in the REFINE study also resulted in high positive accuracy (21%), meaning that a patient identified as at risk of a cardiac arrest had a greater than one in five chance of developing a cardiac arrest in follow-up. These results indicate that a simple combination of non-invasive tests appears to provide powerful and reliable risk prediction.

Should Non-invasive Tests Be Used to Predict Implantable Cardioverter–Defibrillator Efficacy in 2008?

The results of the REFINE study suggest that patients most likely to benefit from an ICD can be identified using a combination of non-invasive parameters. Of those tested, heart-rate turbulence and T-wave alternans performed best. Despite these very positive results, there is insufficient evidence to use these tests to decide whether any given patient should receive a prophylactic ICD. For example, it was previously thought that an abnormal signal-averaged ECG reliably predicted patients at risk of a cardiac arrest. However, the Coronary Artery Bypass Graft (CABG)-Patch trial showed no benefit of prophylactic ICD therapy in patients with a low EF and an abnormal signal-averaged ECG undergoing cardiac bypass surgery.14 Likewise, abnormal heart rate variability was also thought to identify post-MI patients at risk of a cardiac arrest. However, the Defibrillator in Acute Myocardial Infarction Trial (DINAMIT) demonstrated no benefit of prophylactic ICD therapy for patients with a low EF and an abnormal heart-rate variability measured early after MI.15 Finally, in the recently reported Microvolt T-wave Alternans Testing for Risk Stratification of Post-MI Patients (MASTERS) I study, exercise T-wave alternans failed to predict a higher risk of sudden death or life-saving ICD therapies in a group of patients with low EF values undergoing prophylactic ICD therapy.16

Thus, while the combination of non-invasive tools identified in the REFINE study appears to predict which patients will versus will not benefit from prophylactic ICD therapy, there is currently insufficient evidence to make clinical decisions based on these test results. Prospective trials are required to prove that these tests do predict which patients are likely to benefit from prophylactic ICD therapy. Two large studies specifically addressing the utility of the combination of parameters shown to be useful in the REFINE study are expected to commence in the next six to 12 months. Until these data are available it is premature to use any of these non-invasive tests to guide ICD therapy.

Should the REFINE Study Results Affect Patient Management?

The ability of heart-rate turbulence and T-wave alternans to predict a high or low risk of serious outcomes after MI has clinical implications. The four out of five post-MI patients with normal test results can be reassured, as they are at very low risk of problems (see Figure 1). In contrast, the one in five higher-risk patients need to be closely monitored. Therapies proved to reduce mortality post-MI should be prescribed. These medications include beta blockers, antiplatelet agents, statins, ACE inhibitors, angiotensin receptor blockers and aldosterone antagonists. Patients should also be assessed for residual ischaemia and treated as per current standards. Whether prophylactic ICD therapy can further reduce mortality in patients with impaired heart rate turbulence and abnormal T-wave alternans is unknown, but will be answered by upcoming trials.


  1. Zipes DP, Can J Cardiol, 2005;21(Suppl A):37–40.
  2. Bayes de Luna A, et al., Am Heart J, 1989;117(1):151–9.
    Crossref | PubMed
  3. Exner DV, et al., J Am Coll Cardiol, 2007;50(24):2275–84.
    Crossref | PubMed
  4. Germano JJ, et al., Am J Cardiol, 2006;97(8):1255–61.
    Crossref | PubMed
  5. Bauer A, et al., Lancet, 2006;367(9523):1674–81.
    Crossref | PubMed
  6. La Rovere MT, et al., Lancet, 1998351(9101):478–84.
    Crossref | PubMed
  7. Schmidt G, et al., Lancet, 1999;353(9162):1390–96.
    Crossref | PubMed
  8. Gehi AK, et al., J Am Coll Cardiol, 2005;46(1):75–82.
    Crossref | PubMed
  9. Myles RC, et al., Circulation, 2007;116(25):2984–91.
    Crossref | PubMed
  10. Nieminen T, et al., Eur Heart J, 2007;28(19):2332–7.
    Crossref | PubMed
  11. Couderc JP, et al., Europace, 2007.
  12. Reinhardt L, et al., Am J Cardiol, 1996;78(6):627–32.
    Crossref | PubMed
  13. Savoye C, et al., Am J Cardiol, 2006;98(9):1144–9.
    Crossref | PubMed
  14. Bigger JT Jr, N Engl J Med, 1997;337(22):1569–75.
    Crossref | PubMed
  15. Hohnloser SH, et al., N Eng J Med, 2004;351(24):2481–2488.
    Crossref | PubMed
  16. Chow T, et al., J Am Coll Cardiol, 2007;49(1):50–58.
    Crossref | PubMed