- A piece published yesterday in the Lancet advocates greater sensitivity to heart failure (HF) in long term outcomes trials. We think there is broad agreement to this that has already been established.
- That said, there are several complicating factors related to definition of heart failure, how heart failure is assessed, how to ensure consistency, etc.
In yesterday’s Lancet, Pfeffer et al. (a group of high profile colleagues include Dr. Herzel Gerstein, Dr. Rury Holman, and Professor John McMurray) provided a piece on heart failure and large cardiovascular outcomes trials. Their main message is that heart failure is a common problem for many people with diabetes and as such it should be formally assessed and reported as one of the outcomes in large outcomes trials. Until now it has been somewhat spotty and neither journals nor regulators were always asking to see the effect of interventions on that outcome. This piece should contribute to sensitizing readers further to its importance; this has already begun, of course, when SAVOR was announced and a potential HF signal became apparent. .
To date, HF has not necessarily been a pre-specified endpoint in all trials at least in part because it is challenging to measure and varies across healthcare systems/regions based on resources. That said, now that a potential signal has come up (with some or all DPP-4 inhibitors), many believe it should be added to the list of concerns that are monitored as new trials come out and should be proactively included as an endpoint in future trials.
We believe there is already broad agreement to this and that the devil will be in the details. Our preliminary questions are below.
Close Concerns Questions
1. Characterization and assessment: To date, there does not appear to be a systematic way that HF is assessed as an endpoint. HF is clearly not as easy to assess as MACE, for example. Strangely, HF as an endpoint was included in RECORD but not PROActive (even though HF was already a problem with TZDs). We would have assumed HF as an endpoint was in all the recent CVOTs, but we don’t believe it was. SAVOR and EXAMINE obviously have brought some big concern to this issue; that said, how to begin formal assessment is easier said that done.
One question is how to characterize HF imbalances and how they should be assessed. Some would say perhaps a small imbalance in HF events would not necessarily be a show stopper for a type 2 drug, although it would certainly have consequences for labeling and for post-approval requirements. On the other hand, a large imbalance would be worrying, particularly given other alternatives available. How to define large and small imbalances is an open question. .
Broadly speaking, how is systematic evaluation defined?
2. Global variability in assessment: As we understand it, heart failure is not easy to ascertain. To boot, in some parts of the world the threshold for HF is quite low, and in other areas, it is not. The implications for trials are hard to assess. We are not sure how most KOLs or safety boards would weight an HF admission compared to an MI or stroke admission. We will be looking into how HF is assessed globally, and how much assessments vary from one healthcare system to another and how that affect trials.
3. Mechanistic cause: With the TZDs there were few surprises about HF due to the edema and weight gain that was seen and later the macular edema. It appears people with diabetes are more prone to HF (we have heard as much as by six times), no different than pancreatic cancer. As opposed to TZDs, with DPP-4s, there isn't an obvious mechanistic reason to consider. This makes future data all the more valuable.
4. Cardiovascular disease? We are curious to see, outside CV disease, what else will be added. It might be argued, why just add heart failure, and why just CV safety? Perhaps there should be mandated safety outcome studies covering all safety issues, and not just those that have alarmed researchers in the past in the past (notably liver and MI, but also cancer, fractures and heart failure). Which way will the CV outcome pendulum swing?
5. Greater sensitivity: A call for greater sensitivity is going to draw broad agreement and in fact already has, after EXAMINE (and SAVOR). We are curious if there a working group or any papers that discuss in depth questions about assessment, variability, etc.
6. The eternal question. What are the implications for payers and regulators globally?
As we add to this piece as we learn more, we will include additional questions and answers.
--by Kelly Close