Seeking the truth in the EGPS requires a critical review

August 1, 2005

Boston—The surprising outcomes of the European Glaucoma Prevention Study (EGPS) can be accounted for by a number of explanations and should not be interpreted as meaning that lowering IOP has no benefit for glaucoma, said Harry A. Quigley, MD.

Boston-The surprising outcomes of the European Glaucoma Prevention Study (EGPS) can be accounted for by a number of explanations and should not be interpreted as meaning that lowering IOP has no benefit for glaucoma, said Harry A. Quigley, MD.

"There is conclusive evidence from the results of the Ocular Hypertension Treatment Study (OHTS), the Collaborative Normal Tension Glaucoma Study, and the Early Manifest Glaucoma Treatment Study that lowering IOP by 20% to 30% using medication, laser treatment, or surgery decreases the rate of disease progression by more than 50%," said Dr. Quigley, director, Glaucoma Service and Dana Center for Preventive Ophthalmology, Wilmer Eye Institute, Johns Hopkins University School of Medicine, Baltimore.

Like the OHTS, the EGPS was a randomized, controlled clinical trial aiming to evaluate whether lowering IOP in patients with ocular hypertension slows the appearance of "initial" damage as assessed by optic disc change and/or visual field loss. However, there were major differences in their designs.

In OHTS there was no placebo control, patients assigned to the treatment group were aware of their therapy, and they could receive any medication alone or in combination to achieve the IOP target of a 20% lowering.

In contrast, EGPS was a placebo-controlled trial, investigators and patients were masked to treatment assignment, and dorzolamide t.i.d. was the only medication used.

"No target pressure can be set in a placebo-controlled, double-masked trial because the opportunity to modify treatment in nonresponders could introduce active treatment in the placebo group," Dr. Quigley explained.

The progression rate analysis from the EGPS showed that an initial field defect developed in 2.2% of patients annually in both the treated and untreated groups. That result was similar to the 1.9% rate occurring in the OHTS controls, whereas the progression rate was cut in half among patients assigned to the OHTS treatment group.

"Other studies have also reported an approximate 2% annual progression rate, suggesting that EGPS was essentially a natural history study," Dr. Quigley said.

IOP regression to the mean is one potential confounder that might account for the results of the EGPS. Patients entered into the EGPS had to have a screening IOP between 22 and 29 mm Hg, but there was no requirement to establish that the IOP was consistently elevated in that range at another visit, and about half of the population had an IOP in the range of 22 to 23 mm Hg.

"In a person whose IOP measures 23 mm Hg on one day, there is a good chance that with retesting on another day, the IOP may be only 21 mm Hg simply due to regression to the mean," Dr. Quigley said. "In fact, the authors of the EGPS admitted in their paper's discussion that phenomenon may account for the initial drop in IOP observed in both groups."

At the baseline visit in the EGPS, both the dorzolamide and placebo groups had a mean IOP of about 23.5 mm Hg, and by their first visit at 6 months, both groups experienced a similar decrease of >2 mm Hg. Between 6 months and 5 years, there was a progressive further decline in IOP in both the treated and control groups.