OHTS: corneal thickness is the newest risk factor for glaucoma

Data from the Ocular Hypertension Treatment Study (OHTS) was released in June to great acclaim. This centennial study has previously provided information on the bilateral response to the monocular use of timolol, baseline corneal thickness values, baseline visual field characteristics and the need to confirm visual field defects. The major question that the OHTS study addressed was: “Does treating ocular hypertension prevent or delay the development of primary open-angle glaucoma (POAG)?”

Should therapy for ocular hypertension become the de facto management for our patients with elevated IOP? If we should treat ocular hypertension, which individuals should we treat? Do we treat everyone with IOPs above 21 mm Hg, or are there risk factors in addition to elevated IOP that clarify those at greatest risk? Since the report’s release, I have been surprised at the different ways clinicians have interpreted the report. To some, it validates the concept of prophylactic therapy for ocular hypertension, while for others, it provides evidence that most individuals with elevated eye pressure do require therapy.

Management has evolved

During the past decade, the management of ocular hypertension has evolved. The standard IOP for starting medical therapy for a person with ocular hypertension has been reduced. While 30 mm Hg was a commonly used figure a decade ago, it is now not unusual to begin therapy for an individual with full visual fields, healthy appearing optic nerves and one or two risk factors (race, age, family history) when the IOP approaches the mid-twenties. The OHTS data support this approach to a certain extent, in essence validating the evolving clinical paradigm. The key is understanding which risk factors place an individual with elevated IOP at greatest risk.

Recruitment for OHTS began 8 years ago. Twenty-two centers throughout the United States were involved, with 1,636 participants between the ages of 40 and 80 years old enrolled.

For an individual to be considered, the IOP had to be between 24 mm Hg and 32 mm Hg in one eye and between 21 mm Hg and 32 mm Hg in the other. Other entry criteria included open anterior chamber angles and two normal visual fields for each eye, as determined by the visual field reading center. Normal optic discs as determined using stereo photography by the optic disc reading center were also needed. Individuals were randomized in equal proportions to either medical therapy or careful observation. Neither the patient nor clinician was masked to the randomization assignment.

Therapy goals

The goal for therapy was to reduce the IOP by at least 20% from the highest untreated pressure and to get the IOP below 24 mm Hg. Any medication could be used and was added or changed to obtain the 20% goal, until maximally tolerated medical therapy was reached. New medications could be used when they became available. While topical beta-blockers were the predominant medication initially, agents such as prostaglandins and alpha-agonists showed increasing usage over the life of the study.

Follow-up was done on a 6-month basis, with each visit consisting of a history, visual acuity, full-threshold Humphrey 30-2 visual field, slit-lamp examination, IOP and direct ophthalmoscopy. At the yearly visit, a dilated fundus examination and stereoscopic optic disc photography also were performed. The primary outcome measure for the development of POAG in one or both eyes was defined by a reproducible visual field abnormality or a reproducible clinically significant optic disc deterioration.

Demographics of participants

The medication group included 817 individuals, and the observation group included 819 individuals. From these groups, 702 in the medication group and 706 in the observation group completed the study. Twenty-five percent of both groups were African Americans, and approximately 70% were whites, with Native Americans, Hispanics and Asians making up the small remaining numbers.

The mean entering IOP for both groups was 24.9 mm Hg. Approximately 35% of individuals had a family history of glaucoma, and 34% were myopic. Eleven percent had a history of migraine headaches, and 12% had diabetes.

With therapy, the IOP in the medication group was reduced to 19.3 mm Hg (averaged across all visits), which was a 22.9% reduction. At the 60-month cutoff, when data were available for all participants, 4.4% of individuals in the medication group and 9.5% in the observation group developed glaucomatous damage. Optic disc endpoints were most commonly noted in both the medication group and the observation group: 18 (50%) optic nerve vs. 15 (41.7%) with visual field loss in the medication group and 51 (57.3%) vs. 29 (32.6%) with visual field loss in the observation group. Nine individuals (10.1%) in the observation group and three (8.3%) individuals in the medication group showed changes to the visual field and optic nerve concurrently.

Risk factors

Baseline factors that were associated with the development of POAG, as determined by univariate analysis, were older age, African American race, male gender, larger vertical cup-to-disc ratio, larger horizontal cup-to-disc ratio, elevated IOP, heart disease and thinner corneal thickness measurements. The increased size of the cup-to-disc ratio as a factor associated with glaucoma is not surprising, but is an interesting finding nonetheless.

Before anyone entered into the study, the optic nerves were carefully examined and stereoscopic photographs were evaluated by a separate reading center team. All individuals and their optic nerves were considered to be healthy, even though cupping was larger than average. Was glaucomatous damage present but difficult to detect or is a large cup-to-disc ratio a risk factor such that the nerve is more susceptible to damage? Using multivariate analysis, baseline predictive factors were older age, larger vertical or horizontal cup-to-disc ratio, higher IOP and thinner corneal measurements.

Thin corneas and IOP

The most surprising information derived from OHTS was the dramatic association between thinner corneal measurements and the development of glaucoma. At the 5-year time point, 36% of individuals who at presentation had IOPs higher than 25.75 mm Hg (mean 27.9) and a corneal thickness of less than 555 µm (mean 532) developed POAG, while only 6% of those with the same IOP but with corneas thicker than 588 µm (mean 613.4) developed POAG. Looking at the issue of corneal thickness in another way, at 5 years, 15% of individuals with a vertical cup-to-disc ratio of 0.30 (mean 0.16) and a corneal thickness of less than 555 µm (mean 532) developed POAG while only 4% of those entering with the same vertical cup-to-disc ratio but with corneas thicker than 588 µm (mean 613.4) developed POAG. The differences are enormous and are clinically significant.

The issue of IOP underestimation due to a thin cornea was brought to our attention several years ago with the advent of refractive surgery. After having refractive surgery, individuals have thinner corneas and lower IOPs. This measurement error is not new, and Goldmann in his original papers on Goldmann tonometry discussed the instrument being calibrated for a corneal thickness of 520 µm. Algorithms from Ehlers and Whitacre, among others, have attempted to compensate for corneal thickness with an adjustment of IOP (Acta Ophthalmol. 1975;53:34-43. Am J Ophthalmol. 1993;115:592-596).

Compensating for thin corneas

Their data suggest that subtracting from 2.0 to 7.5 mm Hg per 100 µm from the normal corneal thickness is appropriate. Doughty and Zaman performed a meta-analysis on papers that investigated this issue and found that 545 µm for mean corneal thickness should be used as a base (Surv Ophthalmol. 2000;44:367-408).

They recommend either adding or subtracting 2.5 mm Hg from the IOP for every 50 µm that the corneal thickness varies from 545 µm. This “fudge factor” is the one that most clinicians presently use to adjust IOP based upon corneal thickness.

In another paper, Orssengo and Pye describe a mathematical model to obtain IOP based on corneal factors (Bull Math Bio. 1999;61:551-572). Both individuals are from the Engineering and Optometry Schools at the University of New South Wales, and their model coincides with recent information. Brandt in his report on corneal thickness from OHTS found that African Americans tend to have thinner corneas, which would lead to an underestimation of IOP (Ophthalmology. 2001;108:1779-1788). Fourteen percent of African Americans and 27% of white individuals had increased corneal thickness (more than 600 µm). From this paper, using the most liberal corrective factor, it was calculated that 57% of white patients and 37% of black patients would have corrected IOPs in the normal range and would not have been eligible for OHTS.

Importance of pachymetry

The OHTS data make several points. First, pachymetry is an important test that allows clinicians to better understand the nature of IOP. Many individuals may be spared the years of close management if it is found that their elevated IOP is due to having thicker-than-average corneas. On the other hand, individuals with thinner-than-average corneas are at great risk for developing glaucoma when their IOP is even mildly elevated. The risk appears greater than even the most liberal corrective factor predicts, suggesting that thinner corneas may be associated with an overall connective tissue-like disorder that adversely impacts the integrity of the optic nerve and makes it more susceptible to damage.

Pachymetry has been rarely performed. Few clinicians even now have a pachymeter, let alone perform the test on ocular hypertensives or glaucomas. This should change quickly.

The OHTS data make a compelling argument that pachymetry should be performed on all individuals at risk for or diagnosed with glaucoma. One cannot tell if a cornea is thick based upon the slit-lamp examination unless it is significantly edematous. Also, there is little change in corneal thickness with age (approximately 10 µm per decade).

Corneal thickness measurements are useful, even for individuals diagnosed with glaucoma, for several reasons. The first is that the true untreated IOP is better understood. Also, corneal thickness values relate to the observed IOP reduction with medical therapy.

A report from the Association for Research in Vision and Ophthalmology meeting in 2001 using OHTS data showed a 3-mm difference in IOP reduction with beta-blockers when comparing groups with the thickest corneas to individuals with the thinnest corneas. (Invest Ophthalmol Vis Sci. 2001;42(4):S421.) The IOP reduction was not as great when therapy was initiated on individuals with thick corneas. It is not necessarily that the medication was not working as effectively, but rather that we are not able to measure what the actual IOP reading is. Individuals with thick corneas often do not appear to respond to medications as well, which can be artifactual and must be kept in mind in evaluating drug efficacy.

Lessons learned from OHTS

What lessons have we learned from the OHTS data? First, this study validates recent trends in the management of ocular hypertension. Therapy is indicated when IOP is elevated and multiple risk factors are present, such as race or age. Also, larger cupping is a risk factor that must be weighed. Corneal thickness is the newest risk factor, and its significance cannot be underestimated. Pachymetry is an important test that should be performed on all individuals at risk for or having glaucoma. Thin corneas pose a risk that appears to be comparable to race. The opposite is also true in that individuals with thicker corneas and elevated IOPs but without other risk factors do not require the careful management that was once part of their purview. At 5 years, 9.5% developed glaucomatous damage without therapy, which is not a surprising finding.

The fact that 4.4% of individuals with ocular hypertension progressed to glaucoma with preventive medical therapy makes one ask why. Were the target IOPs not aggressive enough? Are larger reductions in IOP needed? Were these individuals going to progress anyway, and their form of glaucoma is pressure-independent?

The gap between treated and untreated individuals (5.1% difference) expands when the 7-year data are reviewed, though the entire data set is not available. Approximately 13% of the untreated individuals developed damage at 7 years (84 months), while the number of conversions in the treatment group remained around 4.4%.

The National Eye Institute will continue funding, with OHTS II taking over from OHTS. The same centers will be involved and will attempt to enroll all currently enrolled individuals. The follow-up will be similar, but everyone will be treated (including those in the original observation group). The therapy and goals for treatment will be similar to OHTS. The objective for OHTS II is to determine if delayed therapy makes a difference in regard to the conversion of ocular hypertension to glaucoma. Finally, the clinicians and researchers involved in OHTS should be commended for a landmark treatise that has given us great insights into the management of individuals with elevated IOP.

Data from the Ocular Hypertension Treatment Study (OHTS) was released in June to great acclaim. This centennial study has previously provided information on the bilateral response to the monocular use of timolol, baseline corneal thickness values, baseline visual field characteristics and the need to confirm visual field defects. The major question that the OHTS study addressed was: “Does treating ocular hypertension prevent or delay the development of primary open-angle glaucoma (POAG)?”

Should therapy for ocular hypertension become the de facto management for our patients with elevated IOP? If we should treat ocular hypertension, which individuals should we treat? Do we treat everyone with IOPs above 21 mm Hg, or are there risk factors in addition to elevated IOP that clarify those at greatest risk? Since the report’s release, I have been surprised at the different ways clinicians have interpreted the report. To some, it validates the concept of prophylactic therapy for ocular hypertension, while for others, it provides evidence that most individuals with elevated eye pressure do require therapy.

Management has evolved

During the past decade, the management of ocular hypertension has evolved. The standard IOP for starting medical therapy for a person with ocular hypertension has been reduced. While 30 mm Hg was a commonly used figure a decade ago, it is now not unusual to begin therapy for an individual with full visual fields, healthy appearing optic nerves and one or two risk factors (race, age, family history) when the IOP approaches the mid-twenties. The OHTS data support this approach to a certain extent, in essence validating the evolving clinical paradigm. The key is understanding which risk factors place an individual with elevated IOP at greatest risk.

Recruitment for OHTS began 8 years ago. Twenty-two centers throughout the United States were involved, with 1,636 participants between the ages of 40 and 80 years old enrolled.

For an individual to be considered, the IOP had to be between 24 mm Hg and 32 mm Hg in one eye and between 21 mm Hg and 32 mm Hg in the other. Other entry criteria included open anterior chamber angles and two normal visual fields for each eye, as determined by the visual field reading center. Normal optic discs as determined using stereo photography by the optic disc reading center were also needed. Individuals were randomized in equal proportions to either medical therapy or careful observation. Neither the patient nor clinician was masked to the randomization assignment.

Therapy goals

The goal for therapy was to reduce the IOP by at least 20% from the highest untreated pressure and to get the IOP below 24 mm Hg. Any medication could be used and was added or changed to obtain the 20% goal, until maximally tolerated medical therapy was reached. New medications could be used when they became available. While topical beta-blockers were the predominant medication initially, agents such as prostaglandins and alpha-agonists showed increasing usage over the life of the study.

Follow-up was done on a 6-month basis, with each visit consisting of a history, visual acuity, full-threshold Humphrey 30-2 visual field, slit-lamp examination, IOP and direct ophthalmoscopy. At the yearly visit, a dilated fundus examination and stereoscopic optic disc photography also were performed. The primary outcome measure for the development of POAG in one or both eyes was defined by a reproducible visual field abnormality or a reproducible clinically significant optic disc deterioration.

Demographics of participants

The medication group included 817 individuals, and the observation group included 819 individuals. From these groups, 702 in the medication group and 706 in the observation group completed the study. Twenty-five percent of both groups were African Americans, and approximately 70% were whites, with Native Americans, Hispanics and Asians making up the small remaining numbers.

The mean entering IOP for both groups was 24.9 mm Hg. Approximately 35% of individuals had a family history of glaucoma, and 34% were myopic. Eleven percent had a history of migraine headaches, and 12% had diabetes.

With therapy, the IOP in the medication group was reduced to 19.3 mm Hg (averaged across all visits), which was a 22.9% reduction. At the 60-month cutoff, when data were available for all participants, 4.4% of individuals in the medication group and 9.5% in the observation group developed glaucomatous damage. Optic disc endpoints were most commonly noted in both the medication group and the observation group: 18 (50%) optic nerve vs. 15 (41.7%) with visual field loss in the medication group and 51 (57.3%) vs. 29 (32.6%) with visual field loss in the observation group. Nine individuals (10.1%) in the observation group and three (8.3%) individuals in the medication group showed changes to the visual field and optic nerve concurrently.

Risk factors

Baseline factors that were associated with the development of POAG, as determined by univariate analysis, were older age, African American race, male gender, larger vertical cup-to-disc ratio, larger horizontal cup-to-disc ratio, elevated IOP, heart disease and thinner corneal thickness measurements. The increased size of the cup-to-disc ratio as a factor associated with glaucoma is not surprising, but is an interesting finding nonetheless.

Before anyone entered into the study, the optic nerves were carefully examined and stereoscopic photographs were evaluated by a separate reading center team. All individuals and their optic nerves were considered to be healthy, even though cupping was larger than average. Was glaucomatous damage present but difficult to detect or is a large cup-to-disc ratio a risk factor such that the nerve is more susceptible to damage? Using multivariate analysis, baseline predictive factors were older age, larger vertical or horizontal cup-to-disc ratio, higher IOP and thinner corneal measurements.

Thin corneas and IOP

The most surprising information derived from OHTS was the dramatic association between thinner corneal measurements and the development of glaucoma. At the 5-year time point, 36% of individuals who at presentation had IOPs higher than 25.75 mm Hg (mean 27.9) and a corneal thickness of less than 555 µm (mean 532) developed POAG, while only 6% of those with the same IOP but with corneas thicker than 588 µm (mean 613.4) developed POAG. Looking at the issue of corneal thickness in another way, at 5 years, 15% of individuals with a vertical cup-to-disc ratio of 0.30 (mean 0.16) and a corneal thickness of less than 555 µm (mean 532) developed POAG while only 4% of those entering with the same vertical cup-to-disc ratio but with corneas thicker than 588 µm (mean 613.4) developed POAG. The differences are enormous and are clinically significant.

The issue of IOP underestimation due to a thin cornea was brought to our attention several years ago with the advent of refractive surgery. After having refractive surgery, individuals have thinner corneas and lower IOPs. This measurement error is not new, and Goldmann in his original papers on Goldmann tonometry discussed the instrument being calibrated for a corneal thickness of 520 µm. Algorithms from Ehlers and Whitacre, among others, have attempted to compensate for corneal thickness with an adjustment of IOP (Acta Ophthalmol. 1975;53:34-43. Am J Ophthalmol. 1993;115:592-596).

Compensating for thin corneas

Their data suggest that subtracting from 2.0 to 7.5 mm Hg per 100 µm from the normal corneal thickness is appropriate. Doughty and Zaman performed a meta-analysis on papers that investigated this issue and found that 545 µm for mean corneal thickness should be used as a base (Surv Ophthalmol. 2000;44:367-408).

They recommend either adding or subtracting 2.5 mm Hg from the IOP for every 50 µm that the corneal thickness varies from 545 µm. This “fudge factor” is the one that most clinicians presently use to adjust IOP based upon corneal thickness.

In another paper, Orssengo and Pye describe a mathematical model to obtain IOP based on corneal factors (Bull Math Bio. 1999;61:551-572). Both individuals are from the Engineering and Optometry Schools at the University of New South Wales, and their model coincides with recent information. Brandt in his report on corneal thickness from OHTS found that African Americans tend to have thinner corneas, which would lead to an underestimation of IOP (Ophthalmology. 2001;108:1779-1788). Fourteen percent of African Americans and 27% of white individuals had increased corneal thickness (more than 600 µm). From this paper, using the most liberal corrective factor, it was calculated that 57% of white patients and 37% of black patients would have corrected IOPs in the normal range and would not have been eligible for OHTS.

Importance of pachymetry

The OHTS data make several points. First, pachymetry is an important test that allows clinicians to better understand the nature of IOP. Many individuals may be spared the years of close management if it is found that their elevated IOP is due to having thicker-than-average corneas. On the other hand, individuals with thinner-than-average corneas are at great risk for developing glaucoma when their IOP is even mildly elevated. The risk appears greater than even the most liberal corrective factor predicts, suggesting that thinner corneas may be associated with an overall connective tissue-like disorder that adversely impacts the integrity of the optic nerve and makes it more susceptible to damage.

Pachymetry has been rarely performed. Few clinicians even now have a pachymeter, let alone perform the test on ocular hypertensives or glaucomas. This should change quickly.

The OHTS data make a compelling argument that pachymetry should be performed on all individuals at risk for or diagnosed with glaucoma. One cannot tell if a cornea is thick based upon the slit-lamp examination unless it is significantly edematous. Also, there is little change in corneal thickness with age (approximately 10 µm per decade).

Corneal thickness measurements are useful, even for individuals diagnosed with glaucoma, for several reasons. The first is that the true untreated IOP is better understood. Also, corneal thickness values relate to the observed IOP reduction with medical therapy.

A report from the Association for Research in Vision and Ophthalmology meeting in 2001 using OHTS data showed a 3-mm difference in IOP reduction with beta-blockers when comparing groups with the thickest corneas to individuals with the thinnest corneas. (Invest Ophthalmol Vis Sci. 2001;42(4):S421.) The IOP reduction was not as great when therapy was initiated on individuals with thick corneas. It is not necessarily that the medication was not working as effectively, but rather that we are not able to measure what the actual IOP reading is. Individuals with thick corneas often do not appear to respond to medications as well, which can be artifactual and must be kept in mind in evaluating drug efficacy.

Lessons learned from OHTS

What lessons have we learned from the OHTS data? First, this study validates recent trends in the management of ocular hypertension. Therapy is indicated when IOP is elevated and multiple risk factors are present, such as race or age. Also, larger cupping is a risk factor that must be weighed. Corneal thickness is the newest risk factor, and its significance cannot be underestimated. Pachymetry is an important test that should be performed on all individuals at risk for or having glaucoma. Thin corneas pose a risk that appears to be comparable to race. The opposite is also true in that individuals with thicker corneas and elevated IOPs but without other risk factors do not require the careful management that was once part of their purview. At 5 years, 9.5% developed glaucomatous damage without therapy, which is not a surprising finding.

The fact that 4.4% of individuals with ocular hypertension progressed to glaucoma with preventive medical therapy makes one ask why. Were the target IOPs not aggressive enough? Are larger reductions in IOP needed? Were these individuals going to progress anyway, and their form of glaucoma is pressure-independent?

The gap between treated and untreated individuals (5.1% difference) expands when the 7-year data are reviewed, though the entire data set is not available. Approximately 13% of the untreated individuals developed damage at 7 years (84 months), while the number of conversions in the treatment group remained around 4.4%.

The National Eye Institute will continue funding, with OHTS II taking over from OHTS. The same centers will be involved and will attempt to enroll all currently enrolled individuals. The follow-up will be similar, but everyone will be treated (including those in the original observation group). The therapy and goals for treatment will be similar to OHTS. The objective for OHTS II is to determine if delayed therapy makes a difference in regard to the conversion of ocular hypertension to glaucoma. Finally, the clinicians and researchers involved in OHTS should be commended for a landmark treatise that has given us great insights into the management of individuals with elevated IOP.