Turning Lemons into Lemonade: Learning from Disappointing Trials

When a potential new treatment is tested for the first time in people, everyone hopes for the best—that this will be the one that really works and really makes a difference for patients. But the reality is that most drugs aren’t breakthroughs. In fact only about one in five of the drugs that are promising enough to begin tests in humans turn out to be safe and effective enough to make it to market.

However, just because a drug isn’t quite good enough doesn’t necessarily mean that the trial is a total loss. A clearly negative outcome at least provides a straightforward answer. It allows researchers to stop putting time, energy and resources into a doomed project, and move on to other projects that are more likely to succeed. “Sometimes this is the best answer that a trial can give: That road has been explored and it’s closed,” says Cristina Sampaio, MD, PhD, the Chief Clinical Officer at CHDI. “We should not put more investments there.”

“Any time you can do a trial, you’re generating data from humans, and that’s invaluable to drug hunters.”
—Robert Pacifici

Much worse is a trial that ends with no definitive answer. That often happened in the past in HD research because it was just too hard to find enough people to sign up for the study, says Bernhard Landwehrmeyer, MD, a neurologist at the University of Ulm in Germany and the principal investigator of the Enroll-HD study. “In the 1950s through the 1980s, most HD studies were performed with a number of participants in the 2-digit range— that is, less than 100,” he says. “The result was that many of the questions were not answered. You were as uncertain [at the end] as you were before you conducted the study.” This was one of the major inspirations behind Enroll-HD—to make it easier for researchers who want to test a new drug to find enough volunteers who are available and at the right stage in their disease to join the trial.

Learning from setbacks in Alzheimer’s disease

Even if a drug doesn’t help, a trial can be important if researchers learn something new about the disease they’re trying to treat and they can build upon those insights. Some of the most effective drugs of the last 20 years, such as the cholesterol-lowering statins and the anti-TNF drugs for arthritis, followed a series of abandoned or negative studies. “In general, any time you can do a trial, you’re generating data from humans, and that’s invaluable to drug hunters,” says Robert Pacifici, PhD, CHDI’s Chief Scientific Officer.

To take one example, researchers trying to develop a drug for Alzheimer’s disease (AD) have been repeatedly disappointed when drugs that seem to work in animal models or small clinical trials are proven ineffective once they’re tested in a large group of people. Recently, three drug companies spent billions of dollars to test two drugs, bapineuzumab and solanezumab, that are supposed to eradicate protein deposits that form in the brains of people with AD. Major trials of both drugs wrapped up in 2012 after years of study. But neither successfully improved cognition or memory.

However, there was a glimmer of hope: Some people with milder symptoms seemed to benefit a little bit from one of the drugs. This was the first experimental evidence to back a theory that’s been floating around AD research for a long time: that the best time to treat the disease is right at the beginning, before much damage has been done to the brain. “It’s intuitive to say we should treat as early as possible,” says Sampaio. “But people have never been able to demonstrate this in practice. This is the first time there is a hint.” The evidence of a slight benefit is something that AD researchers have been seeking for a long time.

This evidence of a slight benefit is something AD researchers have been seeking.

One of these two drugs is now being tested again, to see if it can stop early-stage AD from getting worse. Several other trials are also underway or in the works with a similar idea. Based on evidence from these and other drug trials, researchers are also now thinking that successfully treating AD may require combining several drugs that target different aspects of the disease—an approach that has been effective in treating cancer and HIV infection.

With these AD studies, “there has been progress because people are learning about how the disease truly is,” says Sampaio. “It is not just dementia—it’s a chronic, protracted and long-term disease that starts almost 15 years before the dementia starts.”

What went wrong?

Getting useful information from a trial even when the drug has been shown to be ineffective is tricky, but since successful trials are relatively rare it’s essential to find out why a drug didn’t work the way it was expected to, says Pacifici. That way, researchers can make an educated decision about what to do next—whether that’s to increase the dose, try the same drug in another group of people, try a similar drug that has slightly different properties, or abandon the whole idea. “A negative trial can truly make a difference by providing extremely useful information for the next step,” says Sampaio.

In HD, one of the first things to find out is that an experimental drug can even get into the brain at a high enough concentration to be effective (the brain has protective mechanisms that block many chemicals, including drugs, so you can’t take it for granted that the drug actually reaches the brain.) If it doesn’t, “you know right from the get-go that this compound had no chance of working—it wasn’t a shot on goal,” says Pacifici. “You can stop at $100,000 and 10 people, as opposed to three years and $10 million.”

Each trial builds on the ones that came before it.

It’s also important to try to find out early in the testing process that the drug changes the function of the brain in the way it’s supposed to. For example, many therapies currently under development for HD are intended to lower the amount of mutant huntingtin protein in the brain. To get the most value out of these trials, researchers want to measure how much of the protein is there at the beginning and whether the drug changes those levels. Right now that’s not easy to do, but a coalition of HD researchers is developing a new test that can accurately measure the protein (Read more about this assay).

A trial that shows a drug is not effective is always disappointing. “We should do our best to minimize” the number of trials that don’t work out, says Sampaio. But the process of drug development is often cumulative—each trial builds on the ones that came before it. There can be a lot to learn from disappointment.
This story was originally published in the March 2014 issue of Enroll!