Experts debate, reach consensus on parameters for myopia control trials

SILVER SPRING, Md. – Patient selection, eligibility criteria, efficacy endpoints, study design and patient-reported outcomes are all crucial factors in a randomized controlled clinical trial designed to evaluate the effectiveness of a device to control myopia progression, according to expert panelists.

Representatives from the major optometry and ophthalmology organizations co-sponsored a workshop held at the FDA to suggest guidelines for future clinical trials involving contact lenses with this indication in children.

Sponsoring organizations included the American Academy of Optometry, American Optometric Association (AOA), American Academy of Ophthalmology, American Society of Cataract and Refractive Surgery (ASCRS), the American Association for Pediatric Ophthalmology and Strabismus (AAPOS), the Contact Lens Association of Ophthalmologists and the FDA.

Image: Hemphill N

Panel discussion on age, refractive error

A panel that was moderated by Natalie A. Afshari, MD, FACS, chief of cornea and refractive surgery at the University of California San Diego and chair of the ASCRS FDA Committee, and Christie L. Morse, MD, a private practitioner in New Hampshire and executive vice president of AAPOS, debated the appropriate age for participants, refractive error, type of refraction, amount of astigmatism, progression, previous myopia control treatment and ethnicity.

The panelists differed slightly on their recommendations for the age range.

Jane Gwiazda, PhD, FAAO, director of research and a professor of vision science at the New England College of Optometry, recommended a range of 8 to 12 years.

Gwiazda, who chaired the Correction of Myopia Evaluation Trial (COMET), said the 6- and 7-year-olds in that study “were harder to find because they needed to have at least 1.25 D of myopia.”

Donald O. Mutti, OD, PhD, FAAO, a professor at the Ohio State University, said the Bifocal Lenses in Nearsighted Kids (BLINK) study included 7- to 11-year-olds.

Panelist Pauline Cho, BOptom, MEd, PhD, FAAO, a professor in the school of optometry at PolyU in Hong Kong, said combined data from a number of studies involves children 6 to 12 years old.

“If we recruit younger children we need better care,” she said. “We have strict after care and see the children every 3 months.”

Afshari asked the panel if they believed more complications occurred in younger children.

Mutti said that the CLAY [Contact Lens Assessment in Youth] study showed that older teens and college students have higher risk behaviors, while younger children are “still in the home environment and accept instruction from the parent.”

Cho noted that in her research, “We haven’t seen a single case of microbial keratitis. That’s because we really stress compliance on the parents.”

Gwiazda and Mutti agreed that 6-year-olds were more difficult to examine.

Regarding the range of refractive error allowed, Gwiazda suggested -1 D to -4 D.

“We don’t want to get into the realm of pathologic myopia,” she said. “There seemed to be a larger treatment effect in the lower treatment range in previous studies.”

Mutti offered a slightly different recommendation: -0.75 D to -5 D.

“The considerations on the lower end are, as Jane said, you need enough myopia to have an unambiguous diagnosis of myopia,” he said. “The problem with -0.25 D or -0.50 D is that there’s a certain variability in technique and response, but -0.75 D is beyond measurement error. The AOA said -1 D is certainly worthy of treatment, but -0.75 is my personal lower level.”

Mutti said it would be valuable if evidence showed that higher forms of myopia can still be changed by environmental influences.

“It’s the group that’s most in need of devices that would control the correction of myopia, and -5 D was our dividing line in our randomized clinical trials,” he said.

Cho recommended increasing it to -6 D.

Gwiazda also suggested allowing up to 1.5 D of astigmatism.

Mutti noted that the BLINK study allowed only up to 1 D of astigmatism.

“I regretted that decision a bit, because in recruitment we saw children not qualify with 1.03 D of astigmatism,” he said. “We had to turn away subjects eager to participate.

“I would say past 1.5 D of astigmatism you’re probably highly unlikely to get acceptable acuity,” he continued.

The panelists had a strong consensus regarding the type of refraction used in these clinical trials, but differing opinions on the agent used.

Gwiazda noted that “cycloplegic autorefraction seems to be the gold standard, with an open field of view refractor. And cyclopentolate is probably the better choice.”

Mutti agreed that cycloplegic autorefraction is used “almost universally. It’s nearly impossible with an autorefractor to inject bias.”

However, he would use tropicamide.

“I chose cyclopentolate in an early study when we wanted maximum cycloplegia,” Mutti said. “It will also wipe out your follow-up. We have to reduce subject burden and maximize retention potential.”

Cho noted that tropicamide does not work well on Asian children due to their dark irises.

Gwiazda said she would be comfortable with tropicamide.

“It is not necessary to completely wipe out accommodation in order to get a child to relax enough to give you a valid measurement of distance refractive error,” Mutti added. “Tropicamide coaxes them into relaxation rather than flattening them with cyclopentolate cycloplegia.”

The panelists also addressed the issue of anisometropia.

“We allowed 1 D of spherical equivalent refractive difference between the two eyes in the COMET study,” Gwiazda said, and it did not change over the course of the study.

“I would be willing to go up to 1.5 D,” she added.

Mutti said he sees 1 D used most often, although the BLINK study allowed 2 D.

“Our rationale was that anisometropia may not be much of a consideration as long as each eye qualifies for the study and there’s good acuity,” he said.

Cho said: “If we can correct the eye properly, we don’t see why there should be a restriction on anisometropia. I would say restrict it at 2 D.”

The panelists were charged with determining if progression of myopia should affect a patient’s eligibility for a study of myopia control devices.

“Ideally, you would enroll children in the study and follow them yourself for a period of time to gauge the progression of myopia before randomizing them to treatment, but that will add more time,” Gwiazda said.

A number of studies, including COMET, saw children progress twice as fast in the winter as in the summer, she noted.

“If you’re only doing a 6-month period of observation, and you did it in the summer, you won’t have as much progression,” Gwiazda said.

“There’s no question that requiring a certain amount of progression to enroll will increase the study power,” Mutti said, “but the negatives outweigh the positives, and the biggest negative is effect on generalizability. You’ll understand efficacy in a certain subclass of myopes. If I open it to everyone, I’m studying the world of myopes and can make a more valuable statement about the efficacy of devices.”

Cho also recommended against using progression to determine eligibility.

“If you are going to offer this treatment to a patient, the practitioner is unlikely to wait and see how the child will progress before offering it,” she said. “Parents are concerned once kids become myopic.”

The panelists agreed that ethnicity is an issue in such trials.

“Asians in Asia and Asians in America have faster progression rates,” Mutti said. “In the U.S., it’s not ethically reasonable to restrict enrollment based on heritage or ethnicity.”

This panel finally reached a consensus that study subjects should be between the ages of about 7 and 12 years old with a refractive error of about -1.0 D to about -4.0 D, spherical equivalent, achieved through cycloplegic autorefraction using tropicamide. The panelists agreed that 1.5 D of astigmatism should be allowed, along with a maximum amount of 1.5 D of anisometropia.

Regarding additional eligibility requirements, progression of myopia would not be a factor, but subjects may not have participated in any previous myopia control treatment (such as atropine or bifocal contact lens wear), and the ethnicity of the trial should reflect the U.S. population.

Panel addresses study design, parameters

A panel led by AOA President-Elect Christopher J. Quinn, OD, who is also president of Omni Eye Services, Essex Specialized Surgical Institute and the New Jersey Center for Cornea and Refractive Surgery, and Michael X. Repka, MD, MBA, the American Academy of Ophthalmology’s medical director for governmental affairs, debated the appropriate study design and parameters for clinical outcomes in a trial to evaluate the effectiveness of a device designed to control myopia progression.

The panel was first charged with determining the control.

“If we’re looking at contact lenses, I don’t see how you can’t make the control group a single vision soft lens,” panelist David A. Berntsen, OD, PhD, FAAO, an associate professor at the University of Houston College of Optometry, said.

Panelist Helen A. Swarbrick, PhD, FAAO, a professor of optometry specializing in contact lenses at the School of Optometry and Vision Science at the University of New South Wales, agreed with Berntsen, particularly in order to provide study masking, and moderator Quinn declared this the consensus.

The panel next addressed whether refractive error change or axial elongation should be the primary effectiveness endpoint.

“Although axial elongation appeals as a primary measure, the understanding of how that’s measured needs to be taken into account,” Swarbrick said.

“Both are equally important,” Berntsen said. “What will resonate most with practitioners is refractive error ... refractive error is the language we use and that parents will understand. We absolutely need to have axial elongation, but I have a hard time saying you don’t need both [refractive error and axial length].”

“The main purpose of controlling myopia is to make sure we control the growth of the eyeball,” Cho added. “Controlling refractive error is a plus.”

Panelist Jodhbir S. Mehta, MD, BSc(Hons), MBBS, FRCOphth, FRCS(Ed), who conducts a broad range of laboratory-based and clinical research, agreed that axial elongation and refractive error are both important.

However, Swarbrick added: “If one set of my records got destroyed, let’s have it be refractive error. Axial elongation is what we’re concerned about.”

Berntsen said he if had to choose one that would allow for comparison across different devices (soft contact lenses and orthokeratology), it would be axial length.

“Axial length would be more useful,” Mehta added, “because of reproducibility of the data.”

After a lengthy debate, the preference went to axial elongation, despite the fact that, “patients and parents understand refractive error change better than axial elongation,” Repka said as he summarized the panel’s consensus.

The panel next debated the clinically meaningful differences for endpoints between or within study arms.

“This is one of the hardest questions,” Berntsen said. “In the U.S., 0.50 D [treatment effect] per year means you’re stopping myopia progression. Say the treatment sees a 45% reduction in myopia progression, but we set a cut point of 50%. It was working really well for a good number of people, but it didn’t meet that magic line.”

“If you had myopia control therapy that works 100% in 20% of children, is that good?” Swarbrick added. “Would we approve that? This is a complex question.”

“If you have a child who would have been a -6 D and is now a -3 D, I think that’s a slam dunk,” Berntsen said.

The panel was unable to reach consensus on the clinically meaningful differences for the endpoints. General suggestions from the panelists and audience members included: 50% treatment effect between groups, less than 0.75 D of progression over 3 years or 30% reduction in myopia progression between groups.

Swarbrick, Berntsen and Cho agreed that the minimum duration of a premarket study should be at least 2 years, preferably 3 years.

“If there were a requirement that you have to look at what happens if you abstain from using the device, you get into more complex issues,” Berntsen said. “However, if you’re doing a study to find out that something is working, I have a problem with the ethics of telling the patient that you need to take them out of it. It seems like a hard conversation to have with the parent.”

The panelists said that if it was decided that abstention from device wear to ensure stability of the refractive outcomes was necessary, 6 months was long enough, although 1 year “was preferred because of seasonal impacts,” Repka summarized.

The panel also discussed what would be an acceptable rate of microbial keratitis (MK) in this study population.

Robin L. Chalmers, OD, FAAO, an independent clinical trial consultant and adjunct professor at Indiana University School of Optometry, who served on a different panel, shared her expertise.

“Except for the optics, it will be similar to the types of lenses that will be used now,” she said. “To me, there’s no reason to suspect that the rates with these other lenses would be higher, especially in your clinical trials.”

Jeffrey J. Walline, OD, PhD, FAAO, associate dean for research at the Ohio State University, who also served on another panel, agreed with Chalmers, saying MK was not included in the definition of previous studies on overnight ortho-K.

“Why make this a higher standard of burden than other contact lenses that are approved by the FDA?” he said.

Deepinder K. Dhaliwal, MD, Lac, an associate professor of ophthalmology at the University of Pittsburgh School of Medicine and director of the Center for Integrative Eye Care, who moderated a different panel, added: “Perhaps in the interest of protocols, if there was any red eye or mild redness, the lens would have to be removed. All MKs start out small. The people that come in right away have a very small, typical non-vision threatening complication. When care is delayed, that’s when it becomes an issue. With proper vigilance we would not have a very high rate at all.”

However, “it’s important that we have an approved definition of MK,” Repka said in his summary of the panel consensus. “Without that, it would not be a comparable rate. One suggestion would be the definition taken from more recent publications.”

He said combining that with registry information and previous clinical studies should be an option.

Patient-centric outcomes

The FDA believes patient-centric outcomes are “of paramount importance” when evaluating today’s medical devices, and this was the last topic debated by a panel led by Mutti and Dhaliwal.

The panel was charged with determining what outcomes should be collected and how.

“Patient-reported outcomes should be looking at myopia control and how it affects the patient,” panelist Walline said, and they should “come from the patient only. The only ones the parent should report are those that the parent can observe.”

“The main questions would be about patient health,” panelist Chalmers noted. “You don’t want the parent to be speaking for the child. But the one thing that’s a bit tricky in terms of gathering the symptoms from a patient is that an 8-year-old’s ability to elicit or describe their halos would be very different than a 12-year-old’s.”

In his summary of the panel’s recommendations, Mutti said, “The consensus was that the patient-reported outcomes should be centered upon symptoms.”

Symptoms of interest would come from the patient and be related to wearing the myopia control device, he said, such as wear time, sense of visual disability, sense of visually related quality of life and sense of responsibility of the subject.

“Questions need rigorous development and need to be in proper terminology and at the cognitive level of the age of the child,” he said.

The panel was also charged with addressing whether patient preference studies would be informative for the benefit-risk determination of myopia control medical devices.

“The simple answer to our question has to be yes,” Mutti said. “Anyone having to make a determination about putting their child into such a device has to have an idea of the typical benefit and risk.”

However, the studies should be ancillary, “framed in a language parents and children can understand and in terms of both benefit and risk,” he said.

The panel discussed a wide range of methods to improve enrollment and retention.

Panelist Thomas (Tim) L. Steinemann, MD, a corneal and external eye disease specialist and professor of ophthalmology at the Case School of Medicine, Case Western Reserve University, Cleveland, said it would be effective to learn why people dropped out of past studies.

Walline said he and his fellow investigators “get a public records request; a public school is required to provide names and addresses of parents. That’s our second most successful way to recruit. Handing out five flyers to every subject that comes into our study and saying to give them to five friends has been our most successful way. The art of retention is providing comprehensive care and free treatment (product). We had almost no one drop out.”

Other tactics suggested were using social media, “deputizing” care providers, assisting with logistics such as providing transportation and approaching sports clubs, the panel concluded.

Parents, children and potential advocacy groups could be better engaged by understanding what parents want through focus groups, understanding why parents might not want to participate, and involving pediatricians and local providers as sub-investigators, they said. – by Nancy Hemphill, ELS, FAAO

• References:
• Walline JJ, et al. Bifocal Lenses in Nearsighted Kids. https://clinicaltrials.gov/ct2/show/NCT02255474. Updated July 2016. Accessed October 20, 2016.
• Chalmers RL, et al. Invest Ophthalmol Vis Sci. 2011;52(9).
• Gwiazda J, et al. Correction of Myopia Evaluation Trial. https://clinicaltrials.gov/ct2/show/NCT00000113. Updated April 14, 2016. Accessed October 20, 2016.

• For more information:
• Natalie A. Afshari, MD, FACS, can be reached at DrAfshari@ucsd.edu.
• David A. Berntsen, OD, PhD, FAAO, can be reached at dberntsen@uh.edu.
• Robin L. Chalmers, OD, FAAO, can be reached at chalmers2097@gmail.com.
• Pauline Cho, BOptom, MEd, PhD, FAAO, can be reached at pauline.cho@polyu.edu.
• Deepinder K. Dhaliwal, MD, Lac, can be reached at dhaliwaldk@upmc.edu.
• Jane Gwiazda, PhD, FAAO, can be reached at gwiazdaj@neco.edu.
• Jodhbir S. Mehta, MD, BSc(Hons), MBBS, FRCOphth, FRCS(Ed), can be reached at jodhbir.s.mehta@snec.com.
• Christie L. Morse, MD, can be reached at cmorse@crhc.org.
• Donald O. Mutti, OD, PhD, FAAO, can be reached at mutti.2@osu.edu.
• Christopher J. Quinn, OD, can be reached at quinn@omnieyeservices.com.
• Michael X. Repka, MD, MBA, can be reached at mrepka@jhmi.edu.
• Thomas (Tim) L. Steinemann, MD, can be reached at tsteinemann@metrohealth.org.
• Helen A. Swarbrick, PhD, FAAO, can be reached at h.swarbrick@unsw.edu.au.
• Jeffrey J. Walline, OD, PhD, FAAO, can be reached at walline.1@osu.edu.

Disclosures: Afshari, Gwiazda, Morse, Quinn and Steinemann reported no relevant financial disclosures. Berntsen has received research materials from Bausch + Lomb but reports no financial interests. Chalmers has received support from or is a consultant for Alcon Research, CooperVision and Johnson & Johnson Vision Care. Dhaliwal reported she is a consultant for Bausch + Lomb and Novabay and that she has conducted research or has spoken on behalf of Abbott Medical Optics, Imprimis, Ocular Therapeutix, Sightlife and Staar. Mehta has lectured for Carl Zeiss Meditec, Ziemer, Moria, Santen and Revision Optics. Mutti reported that he has received research materials from Bausch + Lomb and is an unpaid consultant to Johnson & Johnson Vision Care and Welch Allyn. Repka reported no relevant financial disclosures. He has received grant support from the National Eye Institute and the American Academy of Ophthalmology. Swarbrick has received research support from Bausch + Lomb, BE Enterprises, Capricornia Contact Lens, Contex and Paragon Vision Sciences and has consulted for CooperVision and Johnson & Johnson. Walline has received research materials from Bausch + Lomb.