Monday September 1 2008
The participants took capsules of omega-3 oils as well as their usual medication
“Daily dose of fish oil 'could help thousands with heart failure',” is the headline in the Daily Mail. It reports on a study suggesting that a single capsule of omega-3 oil could help to keep some people with heart failure out of hospital and “reduce the threat [of death] by up to 14%".
The well-conducted study underpinning this story suggests there is a small benefit in terms of hospital admissions and mortality for patients with heart failure who take omega-3 oils. However, the absolute benefits are small and the oils in this study were taken at the same time as standard treatments for heart failure. These treatments (including β-blockers and ACE-inhibitors) have proven clinical benefit of a magnitude far greater than that seen with omega-3 oils here.
Where did the story come from?
Researchers from the GISSI-Prevention trial carried out this study. There are many professionals in this research group, primarily from medical and research institutions across Italy. Luigi Tavazzi is noted as the Chairman of the GISSI steering committee. The study was funded by Societa Prodotti Antibiotici, Pfizer, Sigma Tau and AstraZeneca. It was published in the peer-reviewed medical journal The Lancet.
What kind of scientific study was this?
The study behind these stories is a randomised controlled trial in which the researchers were exploring the effects of omega-3 fatty acids, also called n-3 polyunsaturated fatty acids (n-3 PUFA) in patients with chronic heart failure. Omega-3 fatty acids are found naturally in oily fish, such as sardines, mackerel and salmon; however, in this study, the researchers provided omega-3 capsules.
Researchers recruited patients from 326 cardiology centres and 31 internal medicine centres across Italy. Patients were men and women over 18 years who had chronic heart failure (with symptoms rated as II–IV on the New York Heart Association classification – a scale for determining severity of heart failure based on physical symptoms such as breathlessness and angina). Patients who were already taking or who had been told not to take (contraindicated) n-3 PUFA were excluded from the trial, as were people with non-heart related illness, such as cancer. Patients who had another experimental treatment in the month before the study started, planned heart surgery or were pregnant, were also excluded.
In total, 7,046 patients were randomly assigned to receive one of two treatments: a daily capsule of n-3 PUFA or a placebo pill. Patients and people involved in carrying out the study were not aware of what treatment the patients were receiving, that is, they were blinded. Patients were followed up at regular intervals (at one month, three months, six months and then every six months after that) for an average of four years. During their visits an ECG was performed, blood samples were taken, compliance with treatment was assessed and any adverse effects were recorded. Patients were encouraged to continue taking their regular treatments for heart failure – these are proven, effective treatments and would have included ACE-inhibitors, β-blockers, diuretics, spirinolactone and digitalis.
At the end of the study, the researchers compared the time to death and the time to death or hospitalisation for cardiac reasons between the group taking n-3 PUFA capsules and the placebo group. They also grouped patients to see whether treatment was having different effects across age groups, by the severity of their heart failure (NYHA symptoms and left ejection fraction), the cause of heart failure, whether they had diabetes or not and their total cholesterol at baseline. In their analyses, the researchers accounted for factors that were significantly different between the placebo and treatment group at baseline.
What were the results of the study?
Data from 6,975 patients were analysed (71 patients were excluded after randomisation). The researchers note that after two years of treatment, differences were evident in all-cause mortality and hospitalisation for cardiac reasons between the two groups. Death from any cause occurred in 27% taking n-3 PUFA and in 29% taking the placebo. Absolute risk of death was reduced by 1.8% (95% CI 0.3%-3.9%). There was a difference in rate of all-cause death or admission to hospital for cardiac causes between the groups: 57% with n-3 PUFA versus 59% with placebo – this reduction in risk was only of borderline statistical significance. There were no differences in rates of sudden cardiac death between the groups, in numbers with heart attack or stroke or in presumed arrhythmic death or from worsening heart failure. The researchers also note that by the end of the study about 30% of patients in each group were not taking their assigned medication (treatment and placebo).
What interpretations did the researchers draw from these results?
The researchers conclude that their study shows a moderate benefit of n-3 PUFA in reducing all-cause mortality and admissions to hospital for cardiac reasons. They acknowledge that the benefit they saw was smaller than expected, which may be because it was in a population continuing to take prescribed treatments for their heart failure. Minor adverse events confirm the safety of the drug.
What does the NHS Knowledge Service make of this study?
This multi-centre randomised controlled trial was well conducted and as it took place in so many centres across Italy, the results and measures of adherence may reflect what would happen in practice. The following points are worth highlighting:
- Patients continued to take their prescribed medication for heart failure. The benefits of n-3 PUFA are in addition to these treatments.
- The absolute benefit of the n-3 PUFA was small. In total there were 955 (27.3%) deaths in the treatment group compared with 1,014 (29.1%) in the placebo group: a difference of 1.8% that was only just statistically significant (95% CI 0.3 to 3.9). This ‘small’ effect size must be kept in mind when interpreting the headlines. When the researchers limited their analysis to people who were more compliant with the treatment (i.e. those who took it for at least 80% of the time), although the hazard of all-cause mortality was reduced by 14%, this was still only an absolute reduction of 2% (i.e. 26% in the n-3 PUFA group died compared with 29% in the placebo group).
- A third of patients in each group were not taking their medication by the end of the study. Given the wide reach of this research, such compliance figures probably reflect what would happen in practice if n-3 PUFA was routinely given. Of the 3% of patients who stopped the trial because of adverse reactions, most were minor gastrointestinal reactions (tummy upsets) and these occurred equally frequently for patients on the active drug and placebo.
The researchers’ conclusions that n-3 PUFA is a simple and safe treatment providing a small beneficial advantage is a balanced one, which reflects the results of this study. Given that adverse effects were of minor clinical significance, n-3 PUFA is a safe treatment. Taking multiple medications can be a problem for patients with heart failure, but for these patients, n-3 PUFA could be added to their standard treatment regimen that usually includes drugs that have more of a clinical effect.
Sir Muir Gray adds...
Even I'm eating more sardines now - particularly ones in olive oil.