How Bad Is Fraud in Alzheimer’s Research? Q&A With Neurologist-Sleuth Matthew Schrag

A recent Science investigation found that Eliezer Masliah, one of the world’s leading Alzheimer’s and Parkinson’s researchers, engaged in research misconduct over the course of more than two decades. 

Masliah, who worked at the U.S. government’s National Institute of Health and served as chief adviser to the director at the National Institute on Aging, had substantial influence over how other researchers studied neurological conditions. His research has helped greenlight human clinical trials for a monoclonal antibody drug for Parkinson’s, prasinezumab, which ultimately failed to move the needle for patients. 

Masliah didn’t deny or challenge any of the examples of alleged misconduct in the Science investigation, nor did the drug companies he worked with, the universities that he worked at while doing this research, nor the federal agencies that funded the work. The same day the investigation was published, the NIH released a statement saying Masliah engaged in research misconduct. 

Dr. Matthew Schrag, a neurologist and assistant professor at Vanderbilt University, studies how the brain’s blood vessels affect cognitive decline. Outside of his lab at Vanderbilt, he has taken up an interesting hobby: investigating scientific misconduct. Some of these investigations are done in conjunction with journalists investigating fraud and sometimes as a paid consultant for legal cases. 

Schrag is one of the sleuths who investigated Masliah as well as several other companies and neuroscientists that have been accused of scientific fraud in the past few years: Cassava Sciences (where fraud allegations led to indictments and resignations), former Stanford president Marc Tessier-Lavigne (who stepped down after an investigation exonerated him, finding he had no knowledge of data manipulation by others in his lab), Berislav Zlokovic (on “indefinite leave” from USC after a 2023 investigation raised misconduct concerns), Sylvain Lesné and Karen Ashe (who retracted a widely cited paper after an admission of manipulated images), and even Schrag’s own scientific mentor (now the subject of an onoing university investigation with federal involvement).

“We need to be able to say that research integrity is a serious problem for this field, and we need to focus on it quite a lot more,” Schrag told Being Patient. “I think it’s very easy to be dismissive and say that these are rare, isolated problems, but we’ve seen over and over again how it’s altered the way we perceive this disease. I really believe that it has slowed down the quest for effective treatment.” 

Schrag’s hopeful solution? Come together as a field to acknowledge and attack the problem, head on. “We need to culturally engrain the sanctity of our data so that we’re really going to anchor on getting high-quality data first and not try to force it to align with preconceived ideas.”

Being Patient spoke with Schrag about the state of research in dementia, how sleuths spot fishy data, and why these potential acts of fraud can be so hard to spot. Responses were edited for clarity.

Being Patient: Tell us about yourself and how you got involved in investigating research fraud.

Matthew Schrag: I’m a neurologist and have been interested in studying Alzheimer’s disease for a long time. Anyone who’s been around this space any length of time is painfully aware that our success in treating Alzheimer’s disease is far less than we’d like it to be. 

This came to a head for me a number of years ago looking at the case of Cassava Sciences and their drug simufilam. I was asked by the attorneys of the whistleblowers [for a petition written to the FDA to halt clinical trials] to review their concerns and help them investigate the image data in the clinical and preclinical studies supporting the development of that drug. 

I was really quite surprised by the extensive problems we were able to find. 

My initial work on Cassava Sciences also led to studying the papers from Sylvain Lesné and Karen Ashe at the University of Minnesota, which were central to reformulating and supporting the amyloid hypothesis and then driving the development of the next generation of anti-amyloid therapies. 

What are some of the signs that there might be problems with the data in a study?

Schrag: I tend to look for one of three things. 

Occasionally you’ll look at an image, and you’ll just get the impression it looks funny. It’s exactly the same as walking through the grocery store and seeing a glossy magazine and looking at an image and saying “That looks like it’s airbrushed.”

“I was really quite surprised by the
extensive problems we were able to find.”

Another is when the data in a study are just too perfect. Often, biology is messy, and real datasets tend to have outliers. 

The third thing is that sometimes [data] aligns too perfectly with a hypothesis. 

When we see headlines of researchers being accused of fraud, how certain are the sleuths that something fishy is going on?

Schrag: “Something fishy” is a good non-legal term. 

We’re working from published public documents. You’re not looking at complete datasets. 

First of all, you always need to hear from the people who’ve produced the data. Science should work in a very open way, and people should be very willing to share their data with you, and to listen to any questions about it, and correct any errors. The first step is that the entire process should be very open and very respectful. 

“Often, biology is messy, and real
datasets tend to have outliers.”

We can very often be approaching certainty that an image has been altered from its original form. But, people can modify images for inappropriate but relatively benign reasons. The onus is on the individuals responsible for the data to demonstrate when that was [and clear up why the image was changed]. Part of being a scientist is that you have certain requirements for record-keeping, and you should be able to prove your case right.

Is some of the potential misconduct an open secret in the field, where researchers who haven’t been able to recreate the results of an experiment already suspect something is off?

Schrag: I think [it’s] a bit of both. In all of the high-profile cases that I’ve worked with, we’ve always had individuals, either during the investigation or after approach us and say, “I always suspected that”, or “I heard that from so and so [about the data problems].” 

Sometimes people say, “Oh, everybody knows that, so it doesn’t have much of an impact.” I don’t think that’s true. 

“Science should work in a very open way,
and people should be very willing to
share their data with you, and to listen
to any questions about it, and
correct any errors.”

It’s a very big academic community, and there may be niches where [potential fraud] is well known, but I think a lot of this data is being actively used, and you can tell that by how it’s being cited [by other researchers in their studies]. 

Why don’t scientists on the boards of drug companies or scientific peers spot these problems sooner?

Schrag: Those roles are not explicitly designed to regulate fraud. Sometimes peer reviewers do catch this stuff, and more often than not, journals turn a blind eye to it. 

It’s not primarily what people are being asked to do when they peer-review an article. Usually when peer reviewers are looking at it, they’re not doing forensic analysis, they’re addressing the science and how convincing the results are. 

It may well be that this needs to become a part of what peer reviewers do, or even that professionals who work for the journals [to look for fraud] should be part of the peer review process. But that’s not how the system works now.

Are journals and publishers receptive to these concerns?

Schrag: Journals are unaccountable. Journals and publishers, if they decide not to react to concerns, there’s usually very little recourse. These are often private entities. They are not responsive to federal directives or to institutional investigations. 

“[Journals] are often private entities.
They are not responsive to federal
directives or to institutional investigations.”

There are some journals that definitely take it more seriously than others, but many journals don’t. You can get a sense of that by simply pulling any journal and checking how often they retract a paper, and especially how often they retract one when a federal agency or institution or the authors [of the study] haven’t requested it. 

You’ll find that I would think the majority of journals simply won’t retract [in response to concerns over fraud].

With Cassava Sciences Phase 3 trials ongoing, when should regulators step in?

Schrag: Cassava Sciences’ Phase 3 trials began after a large portion of these concerns were already in the public domain and communicated to the NIH or the FDA. From my viewpoint, those trials should never have started and continue to experiment on patients when there is so much serious concern regarding the underlying data in the preclinical phase two trials. It’s highly unethical.

I think that any responsible regulatory agency involved in this process should be looking to take steps to protect these patients immediately. It should have happened many months ago. To me, this is unconscionable.

What are the ethical concerns with running trials of drugs based on incorrect or potentially fraudulent data?

Schrag: Even if the drug is safe, patients who have Alzheimer’s disease only have a timeframe where they’re going to [be a] fit for trial criteria. By enrolling them in a trial based on corrupt data, you’re still depriving them of the opportunity to participate in the trial where they might actually experience some benefit. 

“Any responsible regulatory agency involved
in this process should be looking to take
steps to protect these patients immediately.
It should have happened many months ago.”

If patients end up not trusting us as scientists and physicians in these roles, we’re not going to cure the disease. It’s going to take people signing up for these trials and putting their health and potentially lives on the line to bring a treatment to market that’s going to help people. When we violate our ethical responsibilities to these patients, we’re undermining the future of clinical trials in this space. 

Have the findings of fraud challenged how we understand the biology of Alzheimer’s?

Schrag: Totally. I think it’s time for a major rethink in this space. That process has started. We’re seeing funding, especially on the basic science level, of a more diverse range of topics now than what was happening 10 years ago.

That process can accelerate, and I think particularly the clinical research is still pretty entrenched in the notion that amyloid is driving all the problems in Alzheimer’s disease. It’s time for us to move past that idea and look at a broader range of topics. 

More than anything, we need new intellectual leadership in this space.

What are you excited about in the research space?

Schrag: Small blood vessels in Alzheimer’s disease are involved in the process. I don’t think most people think of Alzheimer’s disease as a blood vessel disease, but it definitely is.

What we’re seeing in the anti-amyloid trials is that a lot of the benefits of the treatment are being washed out because of injury to the blood vessels, with ARIA. 

Our objective in the work we’re pursuing in the lab has been to try to find ways to improve the plaque and tangle processes while also improving the function of the blood vessels. I think that if we can find ways to go after multiple nodes, multiple facets of what’s going on in Alzheimer’s disease, and not pigeonhole on just amyloid is where I’m hopeful that we’ll have more success.

UPDATED December 8, 2024, 4:23 P.M. ET: This article was updated with additional details of the outcomes of the various fraud investigations in which Schrag has taken part.