A New Vision for Dry Eye Clinical Trials w/ Dr. Joseph Tauber
In my thirty years supporting and running anterior segment ophthalmology trials, a lot has changed – especially for dry eye disease. Trials have become significantly more complex. Instead of the classic one-symptom-one variable, the FDA is requiring more evidence that therapies work. Companies must now measure co-primary variables and show replication across trials. Not to mention, trials still face operational challenges in areas like patient recruitment, site selection, symptom assessments, and much more.
So why, despite the increasing rigor of clinical trials, do we still struggle to treat this disease effectively?
For one, we’re still figuring out the best ways to measure dry eye – and its symptoms – with precision.
The tools for measuring dry eye have changed drastically over the last few decades – but even today, we lack a uniform and mutually agreed upon approach. Originally, doctors measured the classic seven symptoms – dry, gritty, sandy, burning, sensitivity to light, itching, discharge, and blurriness – before moving to the Ocular Surface Disease Index (OSDI). Now, companies use a variety of tools, from the eye dryness score (popularized by Shire for Lifitegrast) to less common Schirmer testing. In many cases, companies are defining their own parameters; like Allergan, who used the first responder concept in their successful Restasis trial. And while some companies achieve amazing success with this approach, it’s unclear whether the FDA will accept these different measures.
The tools and methods for corneal fluorescence staining have undergone a similar evolution. We’ve moved from the simple but imprecise Oxford System to the NEI scale. Now, we mostly use a modified NEI scale, which involves tracking and counting a series of dots across the five corneal zones. Although much more comprehensive than past tools, this method is also more subjective. How do we define dots? How do we define the location of a dot? What does it mean if one dot is bigger than others? When we go back to basics and scrutinize staining on a fundamental level, we don’t actually agree on what staining represents. There’s a commonly held but incorrect assumption that staining always shows inflammation. Staining is a demonstration of ineffective permeability barriers but does not necessarily indicate inflammation – other pathophysiological abnormalities can also compromise permeability.
Given we still don’t understand the relationship between staining and inflammation, it’s no surprise that the tools that do test this symptom fall short. Currently, many doctors look at MMP-9 levels to identify inflammation. But these tests don’t provide data on the grade of inflammation – just whether any inflammation is present. These results aren’t precise enough considering inflammation and accompanying discomfort vary dramatically. When it comes to cohort selection for a drug that targets severe cases, we can’t easily identify or recruit patients with high inflammation. Often, clinicians resort to guesswork when choosing patients, risking the success of the trial.
We still don’t fully understand dry eye and its causes.
To further complicate this, there’s a common misconception that evaporative dry eye results from insufficient lipid in the tears. But this isn’t always the case. Typically, there are two main types: lipid deficient evaporative dry eye, where tears evaporate and expose nerves – and oil retentive dry eye, where capped off glands and abscesses in the lids build up, causing symptoms.
Unfortunately, most trials don’t differentiate between the two, even though they’re clinically unique. It’s therefore not surprising that we lack tools for measuring these unique causes. For example, when it comes to lipid deficient evaporative dry eye, we can’t easily determine lipid layer thickness or describe the family of lipids in the tears.
We also need to better understand meibomian gland function, which is challenging with today’s tools. I once treated a patient who had five visible stripes on a Meibography reading. But when I investigated and began manually probing the eye, I discovered a total of 23 glands – the Meibography tool didn’t produce a clear reading. It’s a big problem that we can’t easily measure meibomian gland functioning, a core indicator of lipid deficient evaporative dry eye.
Most patients’ experience with dry eye is multifaceted.
Even when we can properly measure and diagnose a specific cause or symptom of dry eye, we have to pay attention to additional factors. Clinicians typically focus on what they can easily treat; however, most dry eye patients have multiple pathogenic processes. In addition to abnormal lipids or water production, many also experience neuropathic pain, whereby nerves get retrained to communicate a message of discomfort. Sometimes, patients continue to feel pain after the physiological abnormality is corrected.
Clinicians are just starting to understand how common neuropathic pain truly is. As a result, it’s difficult to diagnose. While we do have specialized confocal microscopes, these instruments require clinicians to conduct lengthy evaluations – and only some have access to them. We don’t yet know how to incorporate neuropathic pain into clinical trials, but there’s hope. Recently, multiple companies have begun exploring this symptom and investigating drugs that target it directly.
Technology continues to play a central role in improving dry eye trials and treatments.
As we continue to evolve dry eye trials and our understanding of this disease, technology can help address these challenges. Some companies are already using machine-driven, non-invasive breakup time assessments in place of stopwatches to reduce human error and cut out reaction time. While the reproducibility of this method remains uncertain, it’s a great testament to how these new tools can support trials. In the future, we’ll likely see technology play a big part in helping find, educate, and place study coordinators. And, technology can accelerate tedious trial tasks like data collection, reducing monitoring visits and improving the patient experience.
Software will continue to help us manage electronic source documents and speed up information transfer as more companies take advantage of the cloud. For diagnostics, machine learning has the potential to greatly uplevel testing. In the world of corneal fluorescence staining, for example, we can train models to assign dots and count pixels much faster – and more accurately – than humans. While ML hasn’t yet explored these areas, I’m confident that we’ll get there.
Despite the promise of technology, we must always make sure these advancements benefit patients. One example is iPads and smart watches: some trials have begun using these tools to measure symptom ontology and quickly collect data. However, devices are tedious for patients, who must constantly monitor and record symptoms. They also require additional education, and can pose a challenge for less tech-savvy patients. Technology already consumes much of our daily lives. We must use technology to reduce the burden on patients – not add to it. Asking patients to meticulously track their symptoms on an iPad is no better than handing them a 45-minute survey to complete. We must simplify the patient experience, and only collect the data we truly need to run a successful study.
As our understanding of dry eye changes, we need to continuously educate doctors, sponsors, and clinicians on the latest approaches
When we develop new technology, we also need to train doctors on these new-and-improved methods. For example, some doctors have begun using Osmolarity tests to detect early dry eye. These tests, which measure the difference in dryness between eyes, show great promise – but aren’t yet widely accepted.
It can take years for the status quo to catch up with the latest evidence. As we continue to make progress on dry eye, we can’t forget the importance of continuous education. It’s a really exciting time in dry eye clinical trials, and I’m hopeful that we’re moving in the right direction. Through a combination of new technology, education, and innovative trial design, we’ll continue to evolve how we treat this disease.