REFOCUS

I already wrote on FOCUS but it is such a key word for today so let’s go with REFOCUS.

Will an implantable cardioverter-defibrillator (ICD) be life-saving in patients with nonischemic systolic heart failure? If you have nonischemic systolic heart failure this is likely an important question for you.

Recently the New England Journal of Medicine reported a randomized trial of ICD implantation in 1,116 patients with nonischemic systolic heart failure and presented the conclusion: “In this trial, prophylactic ICD implantation in patients with symptomatic systolic heart failure not caused by coronary artery disease was not associated with a significantly lower long-term rate of death from any cause than was usual clinical care.”

The DynaMed Plus summary of this trial was “prophylactic ICD reduces sudden cardiac death (level 1 [likely reliable] evidence), but may not reduce all-cause mortality (level 2 [mid-level] evidence) in adults with nonischemic heart failure with reduced ejection fraction”

Both conclusions are accurate. Which conclusion do you find more closely captures the clinically relevant findings from this trial?

Both sources of course provide the details if you read beyond the conclusion. For example the results related to mortality include:

comparing ICD plus usual care vs. usual care alone

• all-cause mortality in 21.6% vs. 23.4% (hazard ratio 0.87, 95% CI 0.68-1.12)
• sudden cardiac death in 4.3% vs. 8.2% (p = 0.005, NNT 26)
• other cardiovascular death in 9.5% vs. 8.8% (not significant)

Assuming we are neither trying to encourage nor discourage use of prophylactic ICD therapy and we are purely wanting to report the findings of this trial so clinicians and patients can be most informed for decision-making which is the more CLINICALLY USEFUL interpretation and presentation?

A)  The difference in all-cause mortality was not statistically significant. Therefore we consider all-cause mortality unlikely to be affected by this therapy. Reductions in sudden cardiac death are likely offset by increases in other causes of mortality.  The most objective understanding of the most relevant findings for patient care is that the ultimate target of all-cause mortality was not achieved.

B)  The difference in sudden cardiac death was statistically significant and clinically meaningful, and this is a likely result of ICD use. The difference in all-cause mortality was not statistically significant but consistent with a reduction in sudden cardiac death not being fully offset by an increase in other causes, and the absence of statistical significance is more likely the result of not having an adequate sample size for the magnitude of effect. (For a 4% absolute difference the sample size needed is 500-600 when the effect occurs in 4%-8% people and is 1600-1700 when the effect occurs in 19%-23%.) The most objective understanding of the most relevant findings for patient care is that the ICD will reduce sudden cardiac death and we have insufficient data statistically to know if the result persists for measures of all-cause mortality.

CRITICAL APPRAISAL of medical evidence takes many considerations.  Sometimes the most important consideration is to REFOCUS on what matters.