Treatment with Entresto (the combination of sacubitril and valsartan, formerly known as LCZ 696) instead of an ACE inhibitor will add approximately one to two years of life for most people taking the drug, according to new estimates published in the New England Journal of Medicine.
Because of the time and population limitations of clinical trials it is impossible to state with confidence the precise long-term effect of new drugs in various populations in real-world clinical settings. Now the investigators in the PARADIGM-HF trial have used data from the trial to estimate the effect of Entresto in people of different age. They estimated survival curves based on the age of trial subjects at the time of their randomization in the trial. The effort was modeled on an earlier study that performed similar calculations on patients in the SOLVD-Treatment trial.
Sanjay Kaul (Cedars-Sinai Hospital) said that “this statistical approach offers a model-free, practical, easy-to-understand way for assessing clinically important treatment differences and deserves greater attention in the design and interpretation of randomized controlled trials.”
The investigators calculated a consistent survival benefit of 1-2 years across a broad range of patient ages. According to their estimates, the life expectancy of a 55-year-old patient would be extended by 1.4 years (CI -0.1-2.8), from 11.6 years for enalapril to 12.9 years with sacubitril-valsartan. For freedom from cardiovascular death or hospitalization for heart failure, the primary endpoint of the trial, sacubitril-valsartan was associated with a mean benefit of 2.1 years (CI 1.0-3.3) in the same age group.
For 65-year-old patients life expectancy would be extended by 1.3 years (CI 0.3-2.4), from 10 years for enalapril to 11.4 years with sacubitril-valsartan, and an additional 1.6 years for freedom from the combined endpoint.
The benefit was still evident for 75-year-old patients, who had an increased life-expectancy of 1.1 years. By 80 years of age, however, the benefit had been reducted to 0.2 years.
Kaul said the results were especially impressive since PARADIGM-HF was not a placebo-controlled trial but an active controlled trial. “It is important to keep in mind that the treatment benefit relative to placebo would be greater,” Kaul said.
Kaul pointed to one important limitation of the study, which “as the authors themselves acknowledge, is the assumption that sacubitril–valsartan treatment benefit remains consistent with long-term use. The PARADIGM-HF trial was stopped earlier (median follow-up of 27 months) than the prespecified follow-up of 34 months. So the authors are extrapolating from a less than 3-year time-exposure to a 30-40 year time-exposure.”
I asked Milton Packer, who recently moved to Baylor University Medical Center and is one of the authors of the paper, whether it would be fair to conclude that sacubitril-valsartan would not confer a lot of benefit in 80-year-olds. His response:
Actually, there is no way of knowing the answer to that question. Remember how the analysis is done. It is actuarial; the same way as life insurers predict lifespan and write policies. So to answer a question about someone who is 85 years old, you would need a meaningful number of 85-year-old patients. As you can imagine, we had very few of those, so we can’t answer the question. I know you would be tempted to extrapolate from the curves, but the data point at 80 years is already based on sparse data.
I asked an outside heart failure expert, Larry Allen (University of Colorado), for his thoughts about how an analysis like this can be used to inform clinical practice.
While analyses like these require a certain degree of statistical
gymnastics, the goal seems noble: to provide some crude estimate of
absolute lifetime benefit of taking the medication. That can be helpful
to individual patients and to facilitate shared decision-making: i.e. “If
you take this new medication for the rest of your life rather than taking
the old one you are already on, you are likely to live about 1-2 years
longer and with less/later adverse health events – the hassle of changing
medications and higher copay should be weighed against that.” It can be
helpful for payers as well: i.e. The roughly $4000 annual cost of the
drug, over an average of XX years survival on drug, leads to 1-2 years of
increased survival (and the other health events estimated), as compared to
enalapril at approximately $48 annual cost.
Cardiology Versus Oncology
Packer offered another observation, pointing out that this sort of analysis is generally not done for cardiology trials, as “they generally do not have a sufficiently wide range of ages with a meaningful number of events at each age. But PARADIGM-HF did have that, so we could do the analyses.”
“Also,” said Packer, “remember that if you do this for oncology trials, the duration of prolongation of life is generally weeks, not years.”