For most patients, 30 days without medication is unlikely to lead to clinically relevant progression.
A similar problem arises when a patient new to our office is already using a glaucoma medication. In that situation, we have to begin by determining how well the current medication is working. If we have documentation of several pretreatment and on-treatment measurements in the patient’s records, we can figure out whether or not the primary therapy is doing its job. But if we don’t have access to records stating what the patient’s pressure was before treatment, then we have no idea how effective the medication is. We only know what the pressure is with the patient on the medication. That doesn’t tell us how well it’s working.
The Monocular Drug Trial
The monocular drug trial was proposed as a way to determine whether our treatment is doing what we want it to do, taking into account spontaneous IOP fluctuation. For example, if both eyes are 26 mmHg, you can treat one eye but not the other. Then, when the patient comes in next time, the change in the untreated eye will tell you what the spontaneous IOP fluctuation is; the change in the treated eye will be a combination of spontaneous and therapeutic IOP change. If you subtract the change in the untreated eye from the change in the treated eye, then you’re subtracting out the spontaneous component, and you should be left with the therapeutic effect caused by the medication. You’re basically using the fellow eye as the control eye for the treated eye. This idea sounds great: It’s elegant, simple, safe, free and only requires one extra office visit.
The problem is, it doesn’t actually work. That’s because a lot of things have to be true in order for it to work, and none of those things have turned out to be true. For instance:
• Both eyes would have to fluctuate symmetrically. If you’re going to believe that the untreated eye tells you about the treated eye, then you have to believe that they spontaneously fluctuate up and down together. The data does not support that. We published a paper almost a decade ago categorizing the occurrence and frequency of asymmetric IOP fluctuations, where the IOP fluctuates significantly more in one eye than the other.1 We found that this is very common in treated glaucoma patients, as well as in normal subjects. Robert Weinreb, MD, also looked at left vs. right eye fluctuations using his own (arguably cleaner) data; he also found that there was poor correlation between the two.2-4
• Pressures measured at the same time of day would have to be fairly consistent. We know that there can be a huge intraday change in IOP, so we try to neutralize that by doing our IOP check at the same time of day when we start the medicine and when we bring the patient back a month or so later to assess its efficacy. The idea is that if we standardize the time of day, we minimize the diurnal component in IOP variability. That would make sense if IOP fluctuation follows the same rhythm every day.
This is another assumption that has turned out not to be true. We conducted an NIH-funded study to evaluate this.5,6 We brought in glaucoma patients and healthy subjects and checked their pressure every two hours from 8 a.m. to 8 p.m. on five different days over the course of a year to see whether their IOP rhythms were conserved from day to day. In fact, they were not conserved in either treated glaucoma patients or in normal subjects. In other words, what your pressure does today does not adequately characterize what it does on other days.
• The medicine would have to work as well in one eye as in the other. Our group and other researchers have done studies looking at this, and the data shows that there is actually fairly poor correlation between the IOP responses of fellow eyes to the same meds.7
• There would have to be no crossover effect. If one believes that the untreated eye is displaying purely spontaneous IOP variability unaffected by the therapy, then the medicine put in the treated eye can’t be lowering pressure in the other eye. That would throw everything off.
For the most part, this is not an issue with prostaglandins because they’re rapidly metabolized once they’re systemically absorbed. But that’s not true as soon as you get into adjunctive therapy and add beta blockers. It’s well-known that beta blockers have a significant crossover effect, on the order of 1.5 to 2 mmHg in the other eye.8 So that’s 1 to 2 mmHg of medication effectiveness that gets masked if you do a monocular drug trial with a beta blocker.
• The patient would have to be following your directions. Patients are notorious for not following instructions. If they have trouble putting drops in both eyes the way we ask them to, what are the odds they’ll remember to only put the drop in one eye, and do so on schedule?
The Proof in the Pudding
Given all of these reasons to doubt the validity of monocular trials, we decided to conduct a retrospective study of a monocular drug trial to see if it worked. The question we asked was: Does the reduction that we see in the trial eye predict the IOP response in the fellow eye when we treat it? The answer was no.9 Others have replicated that study as well, probably a dozen times now.
The idea that monocular drug trials don’t work was not well-received at first. Since this was hard for many clinicians to believe, we decided to conduct a prospective, randomized, investigator-masked study of a monocular drug trial. It also occurred to me that asking whether the treated eye predicts the IOP in the fellow eye was asking the wrong question. The real question is: Does the monocular trial predict long-term response to that medication? So, we set out to try to answer that question.
In this study, one eye was randomly assigned to treatment. The study personnel assessing IOP at baseline and on treatment were masked as to which eye was being treated, and we did three pretreatment measurements and three on-treatment measurements going out to six months to determine the mean IOP change. The difference between the pretreatment mean and the on-treatment mean was our gold standard; that told us how well the medicine worked long-term. So, we looked to see how well the monocular trial predicted that. The answer was: very poorly.10
Again, this outcome was not well-received by clinicians, many of whom had come to rely on monocular drug trials to evaluate the efficacy of their treatments. So at this point, the Ocular Hypertension Treatment Study investigators decided to analyze their data to see whether or not they agreed. They had initiated treatment in their treatment group using the monocular drug trial, and they had multiple pretreatment and on-treatment pressure readings. So they were able to perfectly recapitulate our methodology in an ad hoc analysis of their OHTS database.
They found exactly what we found.11 And after our paper and the OHTS paper came out, the monocular trial finally fell out of favor.
Making the Decision
Evaluating the impact of your treatment is crucial to helping your patients—and I suspect that a lot of medications are not, in fact, helping patients as much as their doctors think. But if a monocular drug trial won’t give you the information you need, how do you decide whether the medicine you’ve prescribed is working?
Our group is trying to answer that question. With NIH funding, we’ve conducted an evaluation of numerous candidates—clinical tools that might provide better short-term estimates of long-term IOP reduction than the monocular trial does. At this point, we’re analyzing the data from that trial, so an evidence-based answer is still in the offing. (We hope to publish our findings with the next year.)
In the meantime, the gold standard is to get multiple pretreatment measurements so you have a reliable idea of where you’re starting from; then, get several on-treatment measurements to characterize the efficacy of the newly added IOP lowering therapy. This is not as easy as conducting a monocular drug trial, but there’s no shortcut at this time.
Here are a few strategies that will help ensure you make a good decision:
• Don’t be afraid to have a washout period. If a patient comes in already on therapy and past records are not available, it’s entirely reasonable to have a washout period and reestablish the pretreatment baseline on at least two occasions before restarting therapy. That will allow you to assess whether or not the patient is getting benefit from the primary therapy. I would avoid doing this in those few patients who have advanced glaucoma—i.e., advanced cupping and visual field loss encroaching within the central 10 degrees. But for most patients, 30 days without medication is extremely unlikely to lead to clinically relevant progression.
Of course, a washout period isn’t necessary if you have data from the previous provider that you’re comfortable with—data that clearly documents pretreatment IOP, taken using a tonometer that you trust, and with more than one baseline pressure measurement. In that situation it’s entirely reasonable to operate on the assumption that you know the real baseline pressure.
• If you can get previous medical records, do so. However, specify that you want the original disc photos, not a copy. Unfortunately, it’s not uncommon to have new patients arrive without their medical records. It’s not on most people’s to-do list to tell their doctor that they’re moving when they relocate. I’m happy to forward a patient’s records, but if one of my patients moves away, I may not even know until he fails to show up for three appointments in a row.
When you’re the one taking on the patient, it’s always worthwhile to formally request the release of previous medical records. Those records are a gold mine, and not just for baseline IOP information. On the records-release form I specifically ask for all clinic notes, all visual fields and any imaging or disc photos that were acquired. The disc photos are particularly valuable, but the person forwarding the records will sometimes make a photocopy of them and send me the copies. Photocopies are totally worthless. To avoid that, I often put in parentheses: “Please send me the original disc photos; I will return them to you at your request.”
• Resist the urge to switch medications on the first post-treatment visit. If a new patient starts at 26 mmHg and then comes in at 24 or 25 mmHg after using a prostaglandin for a month, it’s very tempting to conclude that the patient didn’t respond to the prostaglandin, and therefore switch the patient to your next favorite first-line agent in a different drug class. However, there’s a reasonable chance you’re wrong; the patient may have responded but had an IOP fluctuation that masked the response. So you really owe it to the patient not to write in her chart: “prostaglandin non-responder.” If you do, you’ve denied her the most effective, safest and most convenient class of meds available—and possibly erroneously. She’ll have to go on to less safe, less effective, less convenient therapy.
Before you do that, get at least one more on-treatment pressure check a week or two later. (If you get wildly different measurements the first two times, a third might help you decide which measurement was more likely the fluctuation.) If the medication truly isn’t working, the patient will almost certainly not be any the worse for wear.
At this point, all we can say for sure is that we have not yet optimally characterized the proper frequency and timing of IOP measurements necessary to characterize pretreatment and on-treatment IOP. However, it’s definitely more than one measurement.
So: Add or Switch?
Suppose a patient comes in with a pressure of 30 mmHg, and my target pressure is 13 mmHg. If I didn’t have existing records, I’d begin by taking multiple pretreatment measurements to ensure that this is the real baseline pressure. Then I’d start him on a prostaglandin. If he returns and his IOP is 26 mmHg, and that is confirmed by a second measurement a few weeks later, this patient only got 4 mmHg from the treatment. Some clinicians might add another drug, but that outcome is far less than I would expect from a prostaglandin, based on the results of studies that characterize its IOP-lowering profile. Rather than add a second medication, in this situation it might be worth discontinuing the prostaglandin and switching the patient to a different class of drug. (I’m all for switching to another drug class instead of reflexively adding another drug in a situation like this.)
On the other hand, suppose the patient returns and two consecutive measurements show that his new pressure is 20 mmHg. Some clinicians might see that as too far above the target of 13 mmHg and try switching to another medication. However, that’s a reasonable pressure drop to expect from a prostaglandin, and if you need a 17-point drop in pressure, no single medication is going to give it to you. So if the drug was tolerated and I got that significant a change, I’d continue the prostaglandin and add something else. The prostaglandin is doing all it can; it just needs a little help to reach the target pressure.
The reality is, any given medication may not be working as well as we think, or it may be working better than it appears to be working at the first on-treatment exam. As clinicians, we rarely worry about whether our measurements are really telling us what we think they’re telling us. As a result, it’s very tempting to keep adding medications until the in-clinic measurement is low enough to meet our target. But to truly help our patients, we need to make sure we have an accurate pre-treatment baseline pressure, and then check the reliability of the first IOP measurement on the drug by taking a second measurement. That will give us a much better reason to conclude that a drug is working—or conclude that it isn’t working and convince us to try something different. REVIEW
Dr. Realini is an associate professor of ophthalmology at West Virginia University Eye Institute in Morgantown.
1. Realini T, Barber L, Burton D. Frequency of asymmetric intraocular pressure fluctuations among patients with and without glaucoma. Ophthalmology 2002;109:7:1367-71.
2. Sit AJ, Liu JH, Weinreb RN. Asymmetry of right versus left intraocular pressures over 24 hours in glaucoma patients. Ophthalmology 2006;113:3:425-30. Epub 2006 Jan 10.
3. Liu JH, Sit AJ, Weinreb RN. Variation of 24-hour intraocular pressure in healthy individuals: Right eye versus left eye. Ophthalmology 2005;112:10:1670-5.
4. Liu JH, Weinreb RN. Asymmetry of habitual 24-hour intraocular pressure rhythm in glaucoma patients. Invest Ophthalmol Vis Sci. 2014 Oct 16. pii: IOVS-14-14464. doi: 10.1167/iovs.14-14464. [Epub ahead of print]
5. Realini T, Weinreb RN, Wisniewski S. Short-term repeatability of diurnal intraocular pressure patterns in glaucomatous individuals. Ophthalmology 2011;118:1:47-51.
6. Realini T, Weinreb RN, Wisniewski SR. Diurnal intraocular pressure patterns are not repeatable in the short term in healthy individuals. Ophthalmology 2010;117:9:1700-4.
7. Liu JH, Realini T, Weinreb RN. Asymmetry of 24-hour intraocular pressure reduction by topical ocular hypotensive medications in fellow eyes. Ophthalmology 2011;118:10:1995-2000.
8. Piltz J, Gross R, Shin DH, Beiser JA, Dorr DA, Kass MA, Gordon MO. Contralateral effect of topical beta-adrenergic antagonists in initial one-eyed trials in the ocular hypertension treatment study. Am J Ophthalmol 2000;130:4:441-53.
9. Realini T, Fechtner RD, Atreides SP, Gollance S. The uniocular drug trial and second-eye response to glaucoma medications. Ophthalmology 2004;111:3:421-6.
10. Realini TD. A Prospective, randomized, investigator-masked evaluation of the monocular trial in ocular hypertension or open-angle glaucoma. Ophthalmology 2009;116:7:1237-42.
11. Bhorade AM, Wilson BS, Gordon MO, Palmberg P, Weinreb RN, Miller E, Chang RT, Kass MA; Ocular Hypertension Treatment Study Group. The utility of the monocular trial: Data from the ocular hypertension treatment study. Ophthalmology 2010;117:11:2047-54.