I learned a new term: seeding trial.1 A seeding trial is a marketing tool developed by pharmaceutical companies to change the prescribing habits of physicians to sell more of their drug. They do this by running a controlled trial. However the trial is not designed to learn more about the drug’s benefits and harms. Instead the trial is used as a vehicle to get physicians comfortable with prescribing a drug for a condition it was not originally intended for.
Consider the following example of a recently published study named PARAGLIDE-HF.2 This study looked into using sacubitril-valsartan in patients with heart failure with preserved ejection fraction (HFpEF). As context, sacubitril-valsartan is a drug that has an FDA indication for the treatment of heart failure with reduced ejection fraction (HFrEF), another form of heart disease that differs significantly from HFpEF despite the similar names.
PARAGLIDE-HF sounds impressive at first. It is a multicenter, double blinded, randomized controlled trial. The study drug, sacubitril-valsartan is tested against valsartan, the control group in patients who were hospitalized for heart failure with preserved ejection fraction (HFpEF). The problem, however lies in its primary endpoint: change in NT-proBNP.
NT-proBNP is a surrogate biomarker for heart failure symptoms and outcomes. It is not a validated marker for any clinical concerns patients may have, namely living longer or living better. Because it’s not validated, a study that shows improvements in the NT-proBNP following an intervention is useless, because it remains unknown if changes in NT-proBNP levels affect overall survival or quality of life. If the results of this study does not add any knowledge about whether or not patients actually benefit from this drug, what was the point of running this trial in the first place?
One explanation could be that this was a seeding trial. This study was not designed to answer a scientific question e.g. do patients who were hospitalized for HFpEF benefit from taking this new drug? Perhaps the main intention of the study was to recruit doctors who typically don’t prescribe this drug for HFpEF as trial investigators, so that they normalize the use of this drug for this indication.
In other words, physicians who typically don’t use sacubitril-valsartan for the treatment of HFpEF were invited to participate. Throughout the trial, these physicians have grown comfortable prescribing sacubitril-valsartan for HFpEF. The hope is this habit will continue once the trial is over. Then these doctors will normalize this practice to colleagues around them, and even to medical trainees. Drug sales will increase with no proven clinical benefit.
Some other characteristics of a seeding trial beside using a non-useful primary endpoint are: the recruited physicians had prescribed a similar competitor drug prior to the study, the study is sponsored by the company’s marketing division rather than their research division, there are minimal data collection requirements, and the data collected are of little use for the company.
There are incentives to extend the indication of this drug for HFpEF. HFpEF is a disease with much higher prevalence than heart failure with reduced ejection fraction (HFrEF), the condition for which this drug has an FDA indication. So finding a way to have its use normalized for a broader indication would certainly increase its sales.
When I consider this seeding trial strategy, I couldn’t help but look back and wonder if there are medications I prescribe regularly which have no data for clinical benefit, but rather made it’s way into routine clinical practice as a result of marketing campaigns by pharma guised as science.
Here are some medications that come to mind:
( This is purely educated guesses, I did not do much more searching)
SSRIs. SSRIs must be one of the drugs with the broadest number of uses. From major depressive disorder to menopausal symptoms and fibromyalgia, I’d be curious to see the strength of the studies that were done for its off-label uses.
The gabapentinoids are not too far off from the SSRI drug class in the amount of conditions its used for. What was once used for seizures, is now virtually a panacea for pain of any etiology, characteristic and duration.
I’m also reminded of all the deliberation in internal medicine residency about the varying strengths and side effect profiles of angiotensin receptor blockers, a class of drug for the treatment of hypertension. I wonder if seeding studies were done to get doctors to consider a different brand name based on varying patient factors e.g. ethnicity, resistance to other meds, inability to tolerate certain meds, kidney disease, etc.
Other drugs that make me skeptical for their broad uses include proton pump inhibitors and tricyclic antidepressants. I’m sure there are many more I haven’t thought of.
Pharma will do what pharma does. From a marketing perspective, designing a seeding trial is ingenious though borderline unethical. I’m learning how important it is for physicians to be able to critically appraise the literature themselves.
Studies like this further reinforce my suspicion that most medications we use are likely to have weak evidence supporting its use. Directionally, I’d guess that > 90% of what is done in primary care is based on zero to low level evidence, with marketing ploys such as the seeding trial obscuring the evidence pool further. Staying up to date and navigating biotechnology and medicine to improve health may seem complex and limitless, but I’m starting to think once you remove the junk there isn’t all that much left that actually work.
https://www.nejm.org/doi/10.1056/NEJM199411173312007?url_ver=Z39.88-2003&rfr_id=ori:rid:crossref.org&rfr_dat=cr_pub%20%200www.ncbi.nlm.nih.gov
https://www.jacc.org/doi/10.1016/j.jacc.2023.04.019
This is great. I needed some context about seeding trials, I had no idea such practices existed. This is a sharp condemnation of status quo medical science. Prompts me to wonder how such a massively influential and respected profession has become so duped. Physicians have lots of learning and catching up to do, if we're going to do best by patients.
All this is to say, I appreciate substack and the potential is has for catalyzing much needed critical discourse.
I also was unaware of this approach, thanks!