HONcode Certified

This website is certified by Health On the Net Foundation. Click to verify.

This site complies with the HONcode standard for trustworthy health information: verify here.

Slow and steady wins the race when it comes to heart failure

Ami Banerjee
Last edited 30th August 2010

Heart failure is a major cause of death all over the world, but also causes a lot of disability as a chronic condition, especially with ageing populations. 2-3% of the population suffer with heart failure. Heart failure patients are often prescribed a whole range of medicines to treat their blood pressure, kidney disease and many other conditions. One more tablet called IVABRADINE (or Procoralan) looks set to join the list. I just saw the results of the SHIFT trial presented at the European Congress of Cardiology in Stockholm today and they were simultaneously published in the Lancet online.

50% of heart failure patients have a high heart rate (defined as greater than 70 beats per minute). Beta-blockers reduce the heart rate and have been shown to reduce mortality in heart failure. However, they are not always tolerated well, partly because they also cause a drop in blood pressure. Ivabradine is a new drug which reduced heart rate without much effect on blood pressure, and so may be a new option to treat heart failure.

The SHIFT trial was a double-blinded, placebo-controlled, randomised controlled trial of ivabradine in 6500 patients with moderate to severe heart failure and a regular heart rhythm. This trial specifically looked at the heart rate of patients at the start of the trial (“baseline”) and throughout the trial. The main or primary outcome of the trial was death or hospitalisation due to heart failure. Double blinding means that neither the patients nor the researchers knew which treatment the patients received. Controlling with a placebo allows the researchers to estimate the effect of the drug beyond no treatment. Randomisation means that patients randomly received placebo or the drug (in this case, ivabradine), and removed bias in the selection of patients. The trial lasted for just under 2 years.

Basically, ivabradine reduced death and hospitalisations by 18%, and the drug was very well tolerated, with few side effects of unduly low heart rate (bradycardia) or low blood pressure. The authors concluded that for every 1 beat per minute increase in heart rate, there was a 3% increase in mortality in a continuous relationship. They also found that baseline heart rate predicted the degree of risk of death, and interestingly, patients with the highest heart rate at baseline had the greatest reduction in heart rate with the drug, ivabradine. High heart rate has been shown to be a “risk marker” for outcome of patients with heart failure. This trial seems to suggest that a high heart rate may also be a “risk factor” for heart failure, i.e. it may have a role in causing the disease. Either way, “the slower, the better” seems to be the motto for the heart when it is failing.


Hi Helen

You’re not being stupid. You’ve calculated the NNT in exactly the same way as we’ve all been taught.

The explanation is hidden away in the statistical analysis section of the Lancet paper. At the top of page 878 it says that the NNT was calculated using the “estimated probability of having an event at 1 year in the Kaplan-Meier curves.”

So we can’t check their working out.

Best wishes




I have been looking at this study and reassuringly calculated the same NNTs as you using the figures in the Lancet paper.

However, please can someone explain the following quote by the authors, found in the results section:

"On the basis of this absolute risk reduction, 26 patients would need treatment for 1 year to prevent one cardiovascular death or hospital admission for heart failure".

Am I missing something? I can't see how they have arrived at this figure. For this primary endpoint, I calculated a NNT of 20. Also, I don't understand where the 1 year came from, as the study duration was 22 months.

Apologies if I am being very stupid and missing something obvious, clarification will be gratefully received,



Numbers needed to treat

Thanks Carl. Using the numbers from the Lancet paper, I got the following information:

6558 patients were randomised. 3268 were in the ivabradine arm and 27 were excluded, leaving 3241 patients who were included in the analyses (2 lost to follow up and 73 withdrew consent). 3290 were in the placebo arm and 26 were excluded, leaving 3264 in the analysis (1 lost to FU, 58 withdrew consent). The results were intention-to-treat.

The experimental event rate (EER) is the rate of events in the experimental arm of the study (patients receiving ivabradine in this case). The control event rate (CER) is the rate of events in the placebo or control arm of the study. The authors give figures for all-cause mortality, cardiovascular mortality, death from heart failure, all-cause hospital admission and hospital admission for worsening heart failure. The absolute risk reduction (ARR) is the difference between the AER and the CER. The number needed to treat (NNT) is the number of patients who need to receive the treatment to avoid one event and equals 1/ARR.

For all-cause mortality, the EER was 503/3241 or 16%, and the CER was 552/3264 or 17%. Therefore, the ARR is 0.17-0.16=0.01, and NNT is 1/0.01=100. In other words, 100 people need to be treated with ivabradine to avoid one death. For cardiovascular mortality, the EER is 14% and the CER is 15%. Therefore the ARR is again 0.01 and the NNT is again 100. For death from heart failure, the EER is 3% and the CER is 5%. Therefore, the ARR is 0.05-0.03=0.02 and the NNT is 1/0.02=50.

The more impressive results are for hospital admission for worsening heart failure and the combination of all of these outcomes (death, worsening heart failure or hospital admission). For hospital admission for worsening heart failure, the EER is 16% and the CER is 21%. Therefore, the ARR is 0.21-0.16=0.05 and the NNT is 1/0.05=20. For the combined outcome, the EER is 30% and the CER is 25%. The ARR is therefore 0.30-0.25=0.05 and the NNT is 1/0.05=20. This means that you only have to treat 20 patients with ivabradine to avoid death, hospital admission or worsening heart failure.

Slow and steady wins the race when it comes to heart failure

Be interested to know the absolute event rates, and the NNT, as well as the numbers lost to follow up and the intention to treat results

Can you supply these

Cheers Carl

Twitter TrustTheEvidence.net


Search the TRIP Database

TRIP Database


Recent Comments