The prostaglandin wars

Abstract

T HEY HAVE BEEN CALLED “THE PROSTAGLANDIN Wars.” The stakes are high—perhaps $1 billion a year and growing. The battlefields are ophthalmologists’ offices, medical meetings, ophthalmic tabloids, courts of law, Internet chat groups, and now the peerreviewed medical literature. The weapons are advertisements, sales forces, market researchers, lawyers, and— increasingly—data from peer-reviewed medical studies. Battles have been and are being fought over the validity and infringement of patents, and classification, nomenclature, and metabolism of compounds. They have at times become ugly, not least because the adversaries are some of the companies that we know, respect, and have relied on for decades to make the drugs we use for our glaucoma and other ophthalmic patients and that have been generous in their support of academic meetings, societies, educational activities, and research. The latest episodes have been associated with two recent publications in THE JOURNAL.1,2 At issue is which of the prostaglandin (PG) analogs (and yes, in the opinion of this writer all four of the clinically available prostanoidlike ophthalmic drugs are functional analogs of PGF2 , all working by a similar biochemical and physiologic mechanism) is the most effective and best-tolerated ocular hypotensive agent. The article by Noecker and associates purports to show that bimatoprost (Lumigan) 0.03% is more effective at reducing intraocular pressure (IOP) than latanoprost (Xalatan) 0.005% when each is given once nightly to patients with ocular hypertension or glaucoma having baseline 8:00 AM IOP of approximately 25 mm Hg. The article by Parrish and colleagues purports to show that in head-to-head comparisons of bimatoprost 0.03%, latanoprost 0.005%, and travoprost (Travitan) 0.004% given once nightly to glaucoma or ocular hypertensive patients with similar starting IOP, all three agents produce similar IOP reduction. The authors of both papers, and others, agree that latanoprost produces less conjunctival hyperemia than does bimatoprost, and the Parrish article shows the same finding for latanoprost vs travoprost. The clinical significance of the hyperemia data is itself a matter of intense debate, but the most controversy relates to the efficacy comparisons. It is not my intent to re-review specific studies or the reporting thereof, but rather to propose some possible explanations for the apparent discrepancies. Let us assume that all three drugs have about the same efficacy. If one drug is in fact very slightly more effective, it might require an enormous sample size to find the difference. If the sample sizes are too small, some studies might find the difference, some not, based on chance alone. Further, differences in the populations of the different studies– be they in receptor characteristics, anterior segment enzymes in the case of prodrugs, or a thousand other things—could either mask a small difference or make it more apparent. If there were a larger difference in efficacy between the drugs, it would be apparent in nearly all the studies, even if the magnitude of the difference varied because of these other factors. An example of this can be seen in latanoprost vs timolol studies. In most such trials, latanoprost was demonstrably more effective, by approximately 1 to 2 mm Hg, and a significantly higher proportion of patients attained specific target IOPs; in only an occasional trial was there not a significant difference. Another example in the PG-IOP world is the effect of PGF2 and its analogs on fluorophotometrically measured aqueous humor flow. The conventional wisdom is that these compounds, in standard ocular hypotensive doses in humans or nonhuman primates, do not affect aqueous humor flow. Yet in most studies, there is slight “trend” (a nightmare term for statistical purists) for aqueous humor flow to be slightly but not statistically significantly higher in the PG-treated eye as opposed to the contralateral control eye or control subjects. The “effect” or “trend” is approximately 10%, and although no one study has enough subjects to prove the hypothesis, many ocular physiologists working in the area believe that the effect may be real but small and highly variable—and rather unimportant in terms of IOP given the much larger effect on uveoscleral outflow. Accepted for publication June 11, 2003. InternetAdvance publication at ajo.com June 11, 2003. From the Department of Ophthalmology and Visual Sciences, University of Wisconsin-Madison, Madison, Wisconsin. Inquiries to Paul L. Kaufman, MD, Department of Ophthalmology and Visual Sciences, University of Wisconsin Hospital and Clinics, F4/328 CSC 600 Highland Ave, Madison, WI 53792; fax: (608) 263-1466; e-mail: kaufmanp@mhub.ophth.wisc.edu