Initiating Medical Therapy for Glaucoma
Tony Realini, MD, MPH
Assessing the efficacy of newly-initiated intraocular pressure (IOP)-lowering therapy is surprisingly challenging. Consider that we usually initiate therapy when IOP has reached an unacceptably high threshold. By definition, IOP is at or near the peak of its range when we start treatment. Statistically speaking, IOP is likely to be lower at the next visit even if we do nothing. This is termed regression to the mean: values at the extreme of the range are likely to be followed by values closer to the mean just by chance alone. So if IOP is going to be lower next time whether or not we initiate treatment, how can we be confident that an IOP reduction is attributable to therapy and not just to regression to the mean?
The monocular drug trial was proposed to assist with this clinical dilemma. Taking advantage of the paired nature of the eyes, one eye is treated and the fellow eye is not. The IOP change in the treated eye should represent a combination of both therapeutic change and spontaneous change. The IOP change in the untreated eye should represent purely spontaneous change. The difference in IOP change between the two eyes should reveal the therapeutic IOP change attributable to therapy.
The monocular drug trial doesn't work
As elegant as this sounds, there is a problem. The monocular drug trial doesn't work. We conducted the first formal evaluation of the monocular trial in 2004. We asked whether the IOP reduction in the first-treated eye predicted IOP reduction in the second-treated eye. It didn't. This findings has been confirmed by several independent groups.
More recently, we asked a different question: does the monocular trial predict long-term IOP reduction in the trial eye? To address this question we conducted a prospective study in which the eye undergoing the monocular trial was selected at random and the study personnel conducting the IOP assessments were masked to the assigned eye. We found that the monocular trial poorly predicted long-term IOP reduction, where long-term IOP reduction was defined as the mean IOP reduction based on two pre-treatment IOP measurements taken during the month before treatment and two on-treatment IOP measurements taken over the six months after treatment was initiated. Our findings were promptly confirmed in a robust post hoc analysis conducted in the Ocular Hypertension Treatment Study data set. At least seven studies have now evaluated the monocular drug trial, and aside from one group finding a weak correlation between IOP reductions in first-treated and second-treated eyes of glaucoma suspects (but not in established glaucoma patients), the totality of evidence supports the unfortunate conclusion that the monocular drug trial is not a useful clinical tool.
In fact, there is potential harm to relying on the monocular drug trial. If the monocular trial leads one to the false conclusion that the drug worked, then an ineffective therapy is continued for some period of time until at future visits the true inefficacy of therapy is revealed; given the typical time frame between visits, this can leave patients on ineffective therapy for several months. Alternatively, the monocular trial can lead one to believe that a drug is ineffective when in fact it is effective. This is potentially harmful in that we typically start with prostaglandin therapy; if we erroneously conclude that prostaglandin therapy is ineffective, we may move on to less effective and less safe therapy unnecessarily.
There is potential harm to relying on the monocular drug trial
Why does the monocular trial fall short of its intended purpose? The underpinnings of the monocular trial are a set of assumptions that have not borne up under careful scrutiny. Space precludes a discussion of these assumptions and their shortcomings, but they include the assumptions that fellow eyes exhibit symmetric spontaneous IOP fluctuations and symmetric IOP responses to the same medication; that diurnal IOP curves are reproducible from day to day in a given eye; that medications exhibit no contralateral crossover effect; and that patients use their medications as prescribed. Numerous studies from our group and others have called each of these assumptions into question.
What, then, are we to do? First, because of its lack of efficacy and potential risks as described above, we should stop using the monocular drug trial.
We should stop using the monocular drug trial
The only valid reason for a monocular trial is in the setting of safety concerns. If both eyes require treatment and there are no safety concerns, both eyes should be treated simultaneously. Ideally, several pre-treatment IOP measurements will be obtained, as will several on-treatment measurements. Glaucoma is a slowly progressive disease and the vast majority of patients do not require urgent IOP reduction. It is reasonable to give newly-initiated therapy more than one chance to demonstrate its efficacy. Therapy should only be deemed ineffective if mean IOP is unchanged based on several pre-treatment and several on-treatment IOP measurements. The optimal timing and frequency of these measurements has not yet been established. Our group is working on it.
