ocular hypertension and central corneal thickness
Post on 01-Jan-2017
Embed Size (px)
Ocular Hypertension and CentralCorneal Thickness
William A. Argus, MD
Purpose: To determine whether ocular hypertension (OHT) is over diagnosed inpatients with increased corneal thickness.
Methods: Thirty-six patients with OHT were compared with 29 control subjectsand 31 patients with glaucoma. Corneal thickness was determined by ultrasonic pa-chymetry and measured at a single point, 1.5 mm temporal to the corneal light reflex.
Results: The corneal thickness (mean standard deviation) in patients with OHTwas 0.610 0.033 mm. This was significantly greater than in patients with glaucoma(0.557 0.039 mm) and control subjects (0.567 0.036 mm).
Conclusion: It is well documented that increased corneal thickness leads to arti-ficially high estimations of intraocular pressure. This study confirms that a significantnumber of patients with OHT have a normal intraocular pressure if corneal thickness istaken into account. Corneal pachymetry is clinically helpful in est imating intraocularpressure, determining the risk of visual loss, and establishing a target pressure.Ophthalmology 1995; 102: 1810-1812
Goldmann I understood that corneal rigidit y opposedindentation. He assumed that the resistance to ind en-tation offered by the cornea was counterbalanced bythe surface tension drawing the tonometer tip onto thecornea. Thi s assumption is not true for all patients.Johnson et al,2 Whitacre et al,? and Ehlers et al" haveshown that central corneal thickness affects the esti-mation of intraocular pressure (lOP) with applanationtonometry. Whitacre et ai, in a study of 15 eyes, foundthat in eyes with thin corneas lOPs were underestimatedby as mu ch as 4.9 mmHg, and that in eyes with thi ckcorneas lOP was overestimated by as mu ch as 6.8mmHg. The purpose of this study is to determine ifcorneal th ickness pla ys a clinically significant role inclassifying patients as having ocular hypertension(OHT), glaucoma, or a normal lOP.
Originally received: April 21, 1995.Revision accepted: August 9, 1995.
Fort Wayne Glaucoma Center, Fort Wayne, Indiana.
The author states that he has no proprie tary interest in the developm entor marketing of this article or any of the instruments or method s (com-peting or otherwise) used in this article.
Reprint requests to William A. Argus, MD, Fort Wayne Glaucoma Cen-ter, 7030 Pointe Inverness Way, Suite 240, Fort Wayne, IN 46804.
Materials and Methods
Patients were recruited from a privat e referral practice.Patients who were referr ed for OHT or glaucoma weregiven a thorough ophthalmologic evaluation, includingapplanation tonometry, gon ioscopy, Humphrey 24-2static threshold perimetry, and disc photography. Pa-tients with prior intraocular surgery, corneal disease,or corneal edema were excluded. Thirty-six patients hadOHT with normal optic nerves , normal visual fields,and applanation tensions of 21 mmHg or greater.Thirty-on e pati ents had glaucoma (25 had high-tensionand 6 had low-tension glaucoma). These patients hadevidence.of optic nerve cupping and corresponding vi-sual field loss. Patients with equi vocal findings (glau-coma suspects) were excluded from the study to estab-lish more homogeneous cohorts. There were 29 controlsubject s with no history of elevated lOP or suggestionof glauc oma. I examined and classified all patients inthe study. Central corneal thickness then was measuredwith a KMI RK-5000 (Alcon , Ft. Worth, TX) ultrasonicpach ymeter, 1.5 mm temporal to the corneal light reflex,as described by Casebeer (The Keratorefractive System ,Chiron Educational Systems, Claremont, CA). Theprocedure was performed on both eyes. Measurementswere recorded to 1/1000 of a mm . The corneal th icknessmeasurements were deemed reliable enough to set a
Argus' Ocular Hypertension and Central Corneal Thickness
OHT = ocular hypertension; SD = standard deviation; lOP = intraocularpressure; N/A = not applicable.
The mean lOP for the glaucoma group was not calculated becausemany patients were receiving treatment. Pretreatment levels ofIOP wereunknown.
ControlSubjects(n = 29)
Glaucoma (n= 31)
ORT(n = 36)
Table 1. Summary of Data
Mean SD 0.610 0.033 0.557 0.0390.567 0.036Range 0.539-0,710 0.414-0.610 0.505-0.637
diamond keratotomy knife (DuoTrac, Rapid City, SO)to 100% corneal thickness.
In a given individual, corneal thickness measurementsbetween right and left eyes are not independent of eachother. Therefore, for statistical analysis, only the result ofone eye per patient should be used. There was no signif-icant differencebetween the population of right eyes versusleft eyes. lt was elected at random to present the resultsof the right eyes.
Mean corneal thickness in the OHT group was statisticallysignificantly higher (P < 0.001) than the glaucoma groupand the control group. The corneal thickness (mean standard deviation) for OHT was 0.610 0.033 mm. Themean corneal thickness for the glaucoma group was 0.557 0.039 mm and for the control subjects was 0.567 0.036 mm (Table I).
Intraocular pressure has been shown to have a poor pos-itive predictive value (poor sensitivity and specificity) forglaucoma in screening tests. Population-based studiessuggest that one sixth to one half of all patients with pri-mary open-angle glaucoma may have lOP levels consis-tently below 22 mmHg. 5,6 Half of all patients with open-angle glaucoma will have lOPs below 22 mmHg at a singlescreening. Additionally, many individuals with an elevatedlOP do not have optic nerve damage." In only 2% to 5%of these patients, demonstrable optic nerve damage willdevelop over a 5-year period. 8 Although there is a definitedose-response relation between lOP and glaucomatousvisual fieldloss, the causal link between lOP and glaucomais not always clear. Some cases of normal-tension glau-coma can be mimicked by anterior ischemic optic neu-ropathy,? by optic nerve ischemia from hemodynamiccrises,'? or by neurologic disease. II Some authors haveproposed that the elevation of lOP occurs as a conse-quence rather than a cause of optic nerve deterioration. 12
lt is possible that the poor predictive positive value ofapplanation tonometry in glaucoma is related in part toa systematic error in measurement dependent on cornealrigidity and thickness. lt would seem possible to explainat least some cases of low-tension glaucoma and OHTbased on an error in estimating lOP caused by abnormalcorneal thickness. Johnson et al2 published the case of a17-year-old girl who had lOPs of 30 to 40 mmHg in botheyes, with normal visual fields and optic nerve heads.Medical treatment was unsuccessful in lowering the lOPssubstantially. The central corneal thickness was 0.90 mmin each eye in the absence ofcorneal edema. Cannulationof the left anterior chamber showed an lOP of II mmHg,whereas lOPs on the Perkins' and Schiotz' tonometerswere 35 and 34 mmHg, respectively. This suggested thatcentral corneal thickness should be measured in patientsin whom lOPs do not correspond to other clinical findings.
Whitacre et al' correlated lOP with Perkins tonometryin vivo in 15 eyes. The eyes were cannulated before in-traocular surgery, and lOPs were controlled manometri-cally to 10, 20, and 30 mmHg. There was a statisticallysignificant relation between corneal thickness and the errorof Perkins tonometry. Thin corneas produced underes-timations ofIOP by as much as 4.9 mmHg, whereas thickcorneas produce overestimations by as much as 6.8mmHg.
Ehlers et al" performed simultaneous manometry andPerkins or Draeger hand-held tonometry on 29 eyesscheduled for cataract or glaucoma surgery. All the eyeshad normal corneas and none had astigmatism greaterthan 1.50 diopters. Three tonometer readings were takenat manometrically determined lOPs of 10 and 30 mmHg.A systematic error in the accuracy of applanation tonom-etry, proportional to the true lOP and the central cornealthickness, was present with a highly significant correlationcoefficient (P < 0.001) of 0.707 at 10 mmHg and 0.737at 30 mmHg. The authors interpolated that at a true lOPof20 mmHg, a corneal thickness of0.45 mm would pro-duce a mean underestimation of 5.2 mmHg, and a cornealthickness of 0.59 mm would produce a mean overesti-mation of4.7 mmHg. The Goldmann tonometer was be-lieved to give accurate readings when the central cornealthickness was 0.52 mm. Whitacre et al' state that Perkinstonometry appears to be most accurate when the cornealthickness was approximately 0.54 to 0.55 mm when thecorneal thickness was measured by the overlap methodof optical pachymetry.
This study used ultrasonic pachymetry as opposed tooptical pachymetry. Chisholm et aI,13 using ultrasonic pa-chymetry, found a corneal thickness (mean standarddeviation) of 0.559 0.038 mm in normal eyes and 0.55 I 0.045 mm in glaucomatous eyes. Their data are con-
Ophthalmology Volume 102, Number 12, December 1995
sistent with our study (eyeswith glaucoma, 0.557 0.039mm; normal eyes, 0.567 0.036 mm).
An implicit assumption in postulating a relationbetween corneal thickness and the systematic error inapplanation tonometry is that central corneal thicknessis related to corneal rigidity and to the resistance ofthe cornea to indentation. In most studies, applanationtonometry is most accurate at mean corneal thickness.For ultrasonic pachymetry, mean corneal thickness isapproximately 0.560 mm. However, there is a widerange of corneal thickness in the healthy population.In this study, corneal thickness ranged from 0.414 to0.710 mm. Ehlers et al" interpolated that applanationtonometry is over/underestima