Distribution of ultrasound pachymetry central corneal thickness among188 patients in the study population (mean [SD], 554.5 [53.8]).
Customize your JAMA Network experience by selecting one or more topics from the list below.
Shih CY, Zivin JSG, Trokel SL, Tsai JC. Clinical Significance of Central Corneal Thickness in the Managementof Glaucoma. Arch Ophthalmol. 2004;122(9):1270–1275. doi:10.1001/archopht.122.9.1270
To evaluate the effect of central corneal thickness determination onthe clinical management of patients with glaucoma and glaucoma suspect.
A cross-sectional retrospective study was performed on 188 consecutivepatients. Mean ultrasound pachymetry measurements of central corneal thicknessand corresponding Goldmann applanation tonometry measurements were obtained.Intraocular pressures (IOPs) were corrected using linear and mathematical(Orssengo-Pye) algorithms. Measurement-significant outcomeswere defined as an IOP adjustment of 1.5 mm Hg or greater and outcomes-significant results as an IOP adjustment of 3.0 mm Hg or greater.Changes in therapy such as the use of eyedrops and addition or cancellationof laser therapy or surgery were then noted for those individuals with measurement-or outcomes-significant changes.
Using the linear correction scale, 105 (55.9%) of 188 patients had atleast a measurement-significant adjustment in their IOP measurements: 67 (35.6%)had adjustments between 1.5 and 3.0 mm Hg, while 38 (20.2%) had an outcomes-significantIOP adjustment (≥3.0 mm Hg). Among the 188 patients, 16 (8.5%) had a changein eyedrop therapy, 4 (2.1%) had a change regarding laser therapy, and 6 (3.2%)had a change in the decision regarding glaucoma surgery. Using the exponentialcorrection (Orssengo-Pye) scale, similar percentages were obtained.
Pachymetry-measured central corneal thickness has a significant effecton the clinical management of patients with glaucoma and glaucoma suspect.
Over the past 50 years, central corneal thickness (CCT) measurementhas been an important variable in the assessment of intraocular pressure (IOP)values in patients undergoing refractive and corneal transplant surgery, aswell as in contact lens wearers.1 Studies byEhlers and Hansen2 and Whitacre et al3 stressed that IOP measurements should be adjustedfor CCT. However, the incorporation of CCT-adjusted IOP measurements intodaily clinical practice was limited until recently, when the Ocular HypertensionTreatment Study4 reported that CCT was a strongpredictor for the development of primary open-angle glaucoma in patients withocular hypertension. In particular, this study demonstrated thatsubjects with decreased CCT measurements had an increased risk of developing primaryopen-angle glaucoma (for every 40-µm decrease in CCT, the relative riskwas 1.71). Moreover, individuals with CCTs of 555 µm or less had a 3times greater risk of developing glaucoma compared with patients with CCTsof greater than 588 µm. Moreover, with millions of individuals havingundergone laser in situ keratomileusis surgery,5,6 thereis a growing concern that the process of removing corneal tissue during thissurgery (with resulting thinner corneas) will lead to an increased difficultyin diagnosing glaucoma, because this surgery tends to alter IOP measurementsand may in turn require greater emphasis on the assessment of the optic discand visual fields for the diagnosis and treatment of glaucoma.7,8
As summarized in a meta-analysis by Doughty and Zaman1 onthe effect of CCT on IOP measurements, studies9,10 havehypothesized that CCT and IOP are related to or interdependent on one another,except at gross extremes. Furthermore, variations in mean CCT have been observedin patients with different types of glaucoma.11-13 Altogether,these studies suggest that IOP measurements are affected by CCT values. Infact, misdiagnosis often occurs when CCT is not considered, because generallypatients with normal-tension glaucoma have thinner corneas and those withglaucoma suspect have thicker corneas.11 Accordingto manometric data from Ehlers and colleagues,14 44%of patients with normal-tension glaucoma would be reclassified as having primaryopen-angle glaucoma, and 35% of patients with ocular hypertension would bereclassified as having normal IOPs. Furthermore, Herndon et al15 foundthat as many as 65% of patients with ocular hypertension could be reclassifiedas having normal IOPs.
Believing corneal pachymetry to be essential to the care of patients,Tanaka stated that
. . . performing pachymetry may influence the management of allpatients by allowing clinicians to (1) observe or pursue less aggressive treatmentof patients with pseudo-ocular hypertension; (2) modify the target intraocularpressure in patients with glaucoma; and (3) detect an elevated intraocularpressure in otherwise normal patients with thin corneas and "normal" applanationreadings.16(p544)
There is growing consensus that routine measurement of CCT may be criticalfor the proper management of patients with glaucoma and glaucoma suspect notonly for cost reasons, but most important, for effective quality care. Therefore,we hypothesized that CCT has a significant effect on the clinical managementof patients with the diagnoses of glaucoma and glaucoma suspect. To evaluatethe effect of incorporating CCT measurements into daily clinical practice,we performed a cross-sectional retrospective study of consecutive patientsseen for glaucoma care.
A cross-sectional retrospective study was performed on 188 consecutivepatients seen at an academic medical center glaucoma practice between June20, 2002, and August 20, 2002. Patients with known corneal pathologic conditions(eg, with corneal edema or those who had undergone penetrating keratoplasty)were excluded from the study. However, patients who had undergone refractivesurgery were not excluded from the study. Three consecutive ultrasound pachymetry(Sonomed, Inc, Lake Success, NY) measurements of CCT were obtained from eacheye, and a mean value was then computed. Measurements were performed undertopical anesthesia. The corresponding Goldmann applanation tonometry (GAT)measurements were obtained at each of these visits with the use of fluoresceinsolution (in most cases, only one measurement was taken). These GAT measurementswere performed by one of us (J.C.T.). In most cases, the CCT measurementswere performed before applanation tonometry. In all cases, the CCT measurementswere done before gonioscopy.
Given the small number of patients in the study, variables such as thenumber and type of glaucoma medications taken by each patient were not consideredin subset analysis (eg, to determine whether differences were observed inCCT for patients taking topical carbonic anhydrase inhibitors). Analyses includedthe mean CCT and GAT values of the right eye of patients. If reliable CCTor GAT values were unobtainable from the right eye, then the correspondingmeasurements from the left eye were collected. Institutional review boardapproval was obtained for the retrospective data collection and analysis.
The recorded IOP measurements were then adjusted for CCT using 545 µmas the reference value (in a review1 of 80studies using ultrasound pachymetry, the mean ± SD CCT was 544 ±34 µm). The IOP data were corrected using a linear algorithm and a mathematicalmodel. The linear correction scale (based on extensive literature review)added or subtracted 2.5 mm Hg for every 50-µm difference in CCT fromthe reference value of 545 µm (ie, a 1.0-mm Hg change for every 20-µmdifference in CCT from the reference value of 545 µm). A correctionfactor (CF) of 2.5 mm Hg (for every 50-µm deviation from the referenceCCT value) was used, because various CFs in the literature have ranged fromapproximately 1.00 to 3.57 mm Hg for every 50-µm deviation. In fact,in the study by Ehlers and Hansen,2 the CFreported was 3.57 mm Hg per 50-µm deviation (ie, about 5 mm Hg for every70 µm), and the meta-analysis of 134 studies by Doughty and Zaman1 found that the slope of the regression line createdby the data resulted in 3.33 mm Hg per 50-µm deviation. Furthermore,in a cannulation study of 125 patients undergoing phacoemulsification cataractsurgery with corresponding manometric water column and Perkins tonometry measurements,Pillunat and colleagues17 calculated an approximate2.5-mm Hg change in IOP for every 50-µm difference in corneal thickness.
Adjustments of IOP were made according to the following linear formula:Corrected IOP = Measured IOP – (CCT – 545)/50 × 2.5 mm Hg.
Sensitivity analyses of the linear model were performed, substituting2.0 and 3.0 mm Hg as the CFs (for every 50-µm deviation from the referenceCCT of 545 µm).
To provide a comparison, a mathematical formula derived by Orssengoand Pye18 was used, because the formula accountsfor factors such as the anterior radius of curvature, thickness, and Poissonratio of the cornea. The formula is as follows: True IOP = Goldmann applanationIOP ÷ K, where K is a complex CF dependent on applanted area, radiusof curvature of the anterior cornea, center thickness of the cornea, and Poisson'sratio of the cornea. A standard radius of curvature of the anterior cornea(7.74 mm) was used for the purpose of this study as it was not calculatedfor individual patients.
In regard to CCT-adjusted IOP values, measurement-significant adjustments were defined as IOP corrections of 1.5 mm Hg or greater(in either direction). Although there may be differences in opinion regardingthe value of key CFs, our analysis suggests that 1.5 and 3.0 mm Hg are appropriatecutoff values for determining potential outcome effects. Although resultsof the Early Manifest Glaucoma Trial19 andthe Ocular Hypertension Treatment Study4 suggesta 10% difference in outcome for each millimeter of mercury, this figure wasderived from extrapolated data. As a conservative measure, we opted to setour first cutoff slightly higher at 1.5 mm Hg.
The 1.5-mm Hg figure was arrived at because a change as small as 1.0mm Hg is noted to be measurement significant in clinical trials; the EarlyManifest Glaucoma Trial reported that "each higher (or lower) millimeter ofmercury of IOP on follow-up was associated with an approximate 10% increased(or decreased) risk of progression" in patients with early manifest glaucoma.19(p48) Furthermore, Parrish et al20 defined a 1.5-mm Hg difference as significant intheir randomized clinical trial comparing the IOP-lowering effects of topicalprostaglandin analogues. Other authors have defined similar differences inIOP (eg, 1.75 mm Hg) as being significant between study groups.21
Any CCT-associated IOP adjustments of 3.0 mm Hg or greater (in eitherdirection) were designated as outcomes significant. A3.0-mm Hg decrease in the initial IOP corresponded to an almost 50% reductionin the risk of glaucoma observed in the population-based Baltimore Eye Survey.22 Moreover, per findings in the Early Manifest GlaucomaTrial19 (in which each 1.0-mm Hg reductionin IOP corresponded to a 10% decreased risk of progression of glaucoma), a3.0-mm Hg reduction in IOP would result in an approximate 30% decreased riskof glaucoma progression.
The medical charts were reviewed, and notations that indicated the effectof CCT evaluation on the treatment plan were evaluated. Alterations in theglaucoma treatment plan were then noted for patients with measurement- andoutcomes-significant IOP adjustments (ie, ≥1.5 and ≥3.0 mm Hg, respectively).These changes in therapy included (1) addition or discontinuation of antiglaucomamedications, (2) recommendation or deferment of glaucoma laser procedures,or (3) recommendation or deferment of glaucoma incisional surgery.
Linear regression calculations were performed to ascertain the effectof variables such as age, race, and sex on adjustments in IOP following CCTmeasurements.
One hundred eighty-eight subjects were included in the study. The CCTand GAT were analyzed from the right eye; when neither of these readings wasreliable, the CCT and GAT from the left eye were used (3/188 [1.6%]). Table 1 gives the demographic characteristicsof the study population. Sixty-nine men (36.7%) and 119 women (63.3%) wereenrolled. The mean ± SD age of the subjects was 71.4 ± 13.8years. One hundred fifty-one (80.3%) of the patients were white, 26 (13.8%)were African American, and 11 (5.9%) were Asian. The mean ± SD CCTwas 553.0 ± 53.6 µm for women and 557.0 ± 54.0 µmfor men (P>.50). The mean ± SD CCT for AfricanAmericans was 539.3 ± 53.0 µm, and it was 562.2 ± 53.5µm and 572.4 ± 56.5 µm for whites and Asians, respectively(P>.15).
The mean ± SD CCT measurement for the entire sample was 554.5± 53.8 µm (Figure 1).The mean ± SD IOP recorded was 15.8 ± 4.6 mm Hg. The most commoninitial diagnosis was primary open-angle glaucoma, in 87 patients (46.3%).The next most prevalent diagnoses were chronic angle-closure glaucoma in 26patients (13.8%), normal-tension glaucoma in 21 patients (11.2%), glaucomasuspect in 18 patients (9.6%), and ocular hypertension in 11 patients (5.9%).
Using the Orssengo-Pye model, 119 patients (63.3%) had at least a measurement-significantadjustment in their IOP after CCT assessment(Table 2). Thirty-nine patients (20.7%) had adjustments between 1.5and 3.0 mm Hg, while 22 (11.7%) had outcomes-significant IOP changes (≥3.0-mmHg adjustment made to IOP). Of those patients who had at least a measurement-significantchange to their IOP, 28 patients (14.9%) had a change in their treatment planas a result of adjustments made to their IOP following CCT correction. Ofthese patients, 19 (10.1%) had a change in their medication regimen, 4 (2.1%)had a change in whether they had laser surgery, and 5 (2.7%) had a changein whether they had glaucoma surgery.
Similar percentages were obtained using the linear model(Table 2).One hundred five patients (55.9%) had at least a measurement-significantadjustment to their IOP, with 67 (35.6%) having less than a 3.0-mm Hg correctionand 38 (20.2%) having an outcomes-significant correction (≥3.0 mm Hg ineither direction). Among the 188 patients, 26 (13.8%) had a change in theirtreatment plan as a result of CCT-associated IOP correction. Sixteen patients(8.5%) had a change in medication therapy, 4 patients (2.1%) had a defermentor addition of laser surgery, and 6 patients (3.2%) had a change in whetherthey would receive glaucoma surgery.
Sensitivity analyses were performed following substitution of the linearCF (with 2.0-, 2.5-, and 3.0-mm Hg CFs; Table 3). With the more conservative CF of 2.0 mm Hg per 50-µmdeviation in CCT, 43.6% of patients experienced at least a measurement-significantadjustment in their IOP measurements. With the less conservative adjustmentof 3.0 mm Hg per 50-µm deviation, the percentage of patients with measurement-significantchanges in IOP was 63.3%. Although the percentage of patients with measurement-significantchanges less than 3.0 mm Hg remained relatively similar (33.5%-35.6%), thepercentage of patients with outcomes-significant changes varied greatly dependingon the adjustment used; 10.1% of the patients experienced outcomes-significantchanges with the 2.0-mm Hg CF, compared with 29.4% of patients when usingthe 3.0-mm Hg CF.
The percentage of patients who experienced changes in medical or surgicaltherapy varied slightly based on the CF used (Table 3).The percentage of patients who had a change in medicationtherapy ranged between 8.0% and 9.1%(with the 2.0- and 3.0-mm Hg CFs, respectively),while the percentage of patients with a change in the recommendation of lasersurgery ranged from 1.6% (with the 2.0- and 3.0-mm Hg CFs) to 2.1% (with the2.5-mm Hg CF). The percentage of patients with a change in recommendationof glaucoma incisional surgery ranged between 2.7% (with the 2.0-mm Hg CF)and 3.2% (with the 2.5- and 3.0-mm Hg CFs) (Table 3).
Based on the Orssengo-Pye mathematical model, we compared the effectsof IOP adjustment in regard to clinical management between different racialgroups (Table 4). Measurement-significantIOP corrections were found in 69.3% of African Americans, 62.9% of whites,and 54.6% of Asians (P = .62). Of those patientswith measurement-significant IOP adjustments, 30.6% of African Americans hada change in their therapy, compared with 13.3% of whites and 0% of Asians(P>.05). Furthermore, 15.3% of African Americanshad a change in medication therapy, compared with 9.9% of whites; 11.5% ofAfrican-Americans had a change in whether they had laser therapy, comparedwith 0.7% of whites; and 3.8% of the African Americans had a change in whetherthey had glaucoma surgery, compared with 2.7% of whites (P>.05 for all). The only statistically significant difference observedwas that African Americans and whites were more likely to experience outcomes-significantIOP changes (≥3.0 mm Hg) than Asians (P<.05).
There were significant sex differences seen when using the Orssengo-Pyemodel (Table 5). As such, 73.9%of the men had a measurement-significant change in their IOP, compared with57.1% of the women (P = .03). Furthermore, 14.5%of the men had a change in their medication therapy, compared with 7.6% ofthe women (P = .14). Among men, 4.3% had a changein whether they would have glaucoma surgery, compared with 1.8% of the women.The only category in which women had a greater change in therapy was lasersurgery, in which 2.8% of the women had this change vs 1.4% of the men.
Although some authors11,23 havereported that patients may be misdiagnosed because of the absence of CCT determinationor the subsequent adjustment of IOP, we are not aware of any studies thathave assessed the effect of CCT-associated IOP adjustments on glaucoma clinicalmanagement.
In our study, approximately half of the 188 eyes examined required anadjustment in IOP measurement of 1.5 mm Hg or greater. We chose 1.5 mm Hgas a key correction end point, because this value is often cited as a significantdifference in clinical studies assessing IOP efficacy between medications.Furthermore, it is not known whether these calculated adjustments in IOP arerelevant to clinical management. Based on our analysis, there appears to beclinical usefulness in these IOP corrections, as approximately 8% to 10% ofthe patients had a change in their medication therapy, about 2% had a changein the recommendation (or deferment) of laser procedures, and about 3% hada change in the recommendation (or deferment) of glaucoma incisional surgery.However, because our study was cross-sectional and there are no long-termdata to support the clinical implications of these changes, it is impossibleto extrapolate the true effect of these IOP adjustments and the resultantchanges in clinical decisions on patient outcomes.
Because there is controversy regarding the IOP correction algorithm,we performed the CCT-associated adjustments based on the most prominent formulasused, a linear model derived based on a literature search and a mathematicalmodel developed by Orssengo and Pye.18 A comparisonof the 2 correction formulas yielded similar results. A sensitivity analysisyielded results that were nonmonotonic in 2 instances (ie, the percentageof patients with measurement-significant changes <3.0 mm Hg and the percentageof patients with changes in laser surgery recommendation). Because the category-significantchanges less than 3.0 mm Hg were an intermediate category between no changeand changes greater than 3.0 mm Hg, as one moved from more to less conservativeCFs (ie, 2.0-3.0 mm Hg per 50-µm change), there was an initial increasefollowed by a slight decrease as more patients were shifted into the 3.0-mmHg range or greater (eg, the large increase to 29.4% with the 3.0-mm Hg CF).With regard to the recommendation for laser surgery, the move to less conservativeCFs (ie, 3.0 mm Hg per 50-µm change) may have caused a shift in theneed for glaucoma filtration surgery (rather than laser surgery) in the onepatient affected. Regardless of the models and correction algorithms studied,adjustments for IOP based on CCT are critical for clinical management.
The only statistically significant difference between races was thatAsians were less likely to experience outcomes-significant changes comparedwith whites and African Americans (when the Orssengo-Pye model was used).The lack of greater statistical significance between groups may be due toseveral factors, including that our patient sample (from a tertiary care practice)may not be representative of the general population. The patients analyzedwere already diagnosed as having glaucoma or glaucoma suspect. Furthermore,sample sizes of the different races were not large enough to generate statisticalpower, with only 26 African Americans and 11 Asians included in the study.For example, the population-based Barbados eye studies of a predominantlyblack community (1142 participants) reported that black participants tendedto have thinner corneas than white participants.24
Limitations of this study include its retrospective analysis and itsshort duration. Patients were seen in an academic medical center glaucomapractice, with most having well-controlled glaucoma (mean IOP, approximately16.0 mm Hg). Moreover, there was no way to confirm the true IOP of patients,because they were not taken to the operating room to have their anterior chamberscannulated for IOP assessment.25 Finally, wedo not know whether our definitions of measurement-significant (≥1.5 mmHg) and outcomes-significant (≥3.0 mm Hg) results are useful for accuratelyassessing the extent of the IOP adjustments.
Another limitation of this study is that it addresses CCTs obtainedusing ultrasound pachymetry, which is only one method of measuring CCT. Studies26,27 have shown that, depending on themethod used, statistically different CCTs may be obtained. Ventura et al13 noted that optical low coherence reflectometry representsthe most precise pachymetric method available, with its measurements reproducibleto 1 µm. In the future, appropriate refinement of the correction formulamay need to be undertaken, depending on the corneal pachymetric method adopted.
This study presents preliminary data regarding the effects of CCT-adjustedIOP on clinical management of patients with glaucoma and glaucoma suspect.Because this was a cross-sectional study, further randomized controlled studiesare needed to elucidate the effects of CCT on clinical management and consequentlong-term patient outcomes. Given these future studies, it may be easier todelineate clear and effective treatment guidelines. Additional studies areneeded to better test the consistency of these results across different racialand ethnic groups.
Correspondence: James C. Tsai, MD, Edward S. Harkness Eye Institute,Department of Ophthalmology, College of Physicians and Surgeons, ColumbiaUniversity, 635 W 165th St, New York, NY 10032 (email@example.com).
Submitted for publication July 16, 2003; final revision received January2, 2004; accepted March 4, 2004.
This study was supported by the Columbia University Homer McK. ReesScholar Award (Dr Tsai) and Eye Surgery Fund (Dr Tsai) and by an unrestricteddepartmental grant from Research to Prevent Blindness, New York.
Create a personal account or sign in to: