- Email: [email protected]
Positive correlation between Tono-pen intraocular pressure and central corneal thickness
Positive Correlation between Tono-Pen Intraocular Pressure and Central Corneal Thickness Ali A. Dohadwala, Rejean Munger, PhD, Karim F. Damji, MD Objective: To examine the relationship between intraocular pressure (IOP) readings taken by the Tono-Pen tonometer (Mentor O&O, Norwell, MA) and central corneal thickness (CCT). Design: Prospective cross-sectional population study. Participants: There were 651 eyes of 332 healthy subjects. Main Outcome Measures: A questionnaire was given to each subject requesting information on gender, age, race, and other factors that can influence IOP. The IOP then was measured using the Tono-Pen followed by measurements of CCT using an ultrasonic pachymeter. Results: The IOP was found to increase by 2.9 mmHg/100 m CCT in males and 1.2 mmHg/100 m in females. For males, CCT was found to be statistically significant in predicting IOP (P ⬍ 0.001 in the right and left eyes) and diabetes was of borderline significance (P ⫽ 0.012 in the right eye, P ⫽ 0.089 in the left eye). For females, CCT was of borderline significance (P ⫽ 0.064 in the right eye, P ⫽ 0.019 in the left eye). In females, a family history of glaucoma (P ⫽ 0.021 in the right eye, P ⫽ 0.022 in the left eye) and hypertension (P ⫽ 0.010 in the right eye, P ⫽ ⬍0.001 in the left eye) were also significant in the prediction of IOP. Race was found to be a significant predictor of CCT (P ⬍ 0.001 in both right and left eyes) for both males and females. Conclusion: Clinicians should be aware that, as with the Goldmann applanation tonometer, the Tono-Pen has a systematic error in IOP readings caused by its dependence on CCT. Tono-Pen IOP readings are positively correlated to CCT in males and, to a lesser extent, in females as well. The CCT measurements should be considered to ensure proper interpretation of IOP measurements in the diagnosis and management of disorders in which the CCT or IOP readings are outside normal limits. Ophthalmology 1998;105:1849 –1854 Intraocular pressure (IOP) is an important risk factor that has a significant influence in the diagnosis and management of ocular hypertension and various forms of glaucoma. It is therefore important to ensure that IOP readings are taken using a highly accurate method and that factors that can influence IOP are recognized. Goldmann applanation tonometry is widely used and considered the “gold standard” for clinical measurements of IOP. The IOP readings taken by this method have been shown to correlate with the central corneal thickness (CCT).1,2 This tonometer achieves its best correlation with manometric measurements at a CCT of approximately 0.5 mm, and it will tend to overestimate IOP in structurally thicker corneas and underestimate IOP in thinner corneas.3,4 To our knowledge, the relationship between measured IOP and CCT in a normal population has not been assessed with newer tonometers such as the Tono-Pen (Mentor O&O, Norwell, MA).5 The Tono-Pen is in widespread clinical use
Originally received: September 19, 1997. Revision accepted: May 15, 1998. Manuscript no. 97677. From the University of Ottawa Eye Institute, Ottawa General Hospital, Ottawa, Ontario, Canada. Address correspondence and reprint requests to Karim F. Damji, MD, University of Ottawa Eye Institute, 501 Smyth Rd, Ottawa, ON, K1H 8L6, Canada.
because of its electronic design with LCD display, its ease and flexibility of use,6,7 its portability, and its accuracy in cases of irregular corneas.8 This study examines the relationship between IOP and CCT readings taken by the TonoPen tonometer and discusses factors that predict IOP and CCT in a general population sample.
Methods A prospective study was conducted at four public buildings (3 malls, 1 hospital) between 9 AM and 4 PM as part of an eye health information and glaucoma-screening program. Participation was strictly voluntary, and all individuals who were willing to be tested were included. Initially, each participant filled out an anonymous questionnaire that asked for information pertaining to possible risk factors and exclusion criteria (Fig 1). Subjects were then given topical anesthesia (Alcaine; Alcon, Ft. Worth, TX) bilaterally a few seconds before being tested by qualified technicians. Intraocular pressure was measured with a Tono-Pen by one of six operators. The IOP reading was followed by a measurement of the CCT using a Mentor Advent ultrasonic pachymeter by one of five operators. Three acceptable readings within a range of ⫾2 mmHg for IOP and ⫾5 m for CCT were recorded for each eye. Both instruments were calibrated before use in each screening clinic. SPSS SigmaStat 2.0 software (SPSS, Chicago, IL) was used to perform the statistical analysis. Statistical significance was set at
1849
Ophthalmology Volume 105, Number 10, October 1998
Figure 1. Questionnaire used in the study.
1850
Dohadwala et al 䡠 Correlation of Tono-Pen IOP and Central Corneal Thickness Table 1. Descriptive Statistics of the Population* Medical History
Race
Female
Male
Total
Age [SD] (yrs) Family history of glaucoma Hypertension Diabetes Thyroid disease Myopia Total
44.9 [18.0] 36 20 7 20 56 165
45.4 [16.8] 15 27 8 2 51 162
45.2 [17.4] 52 48 15 22 109 332
IOP [SEM] (mmHg) CCT [SEM] (m)
16.3 [0.4] 545.3 [2.6]
15.2 [0.3] 548.7 [2.8]
15.6 [0.2] 547.3 [2.0]
White Black Asian Native Other Total
Female
Male
Total
113 16 9 2 13 165
111 16 15 3 6 162
227 32 24 5 19 332
* IOP and CCT means are for OD only. SD ⫽ standard deviation; SEM ⫽ standard error of the mean; IOP ⫽ intraocular pressure; CCT ⫽ central corneal thickness.
P ⬍ 0.05. For all statistical tests, the mean IOP and CCT of the three values within the preferred range was used. Because the eyes for one subject are not independent when considering ocular parameters such as IOP, the right and left eyes were always analyzed separately.9 The IOP dependency on operator and right eye/left eye was examined using a two-way analysis of variance (ANOVA). The prediction of IOP by certain factors was analyzed simultaneously using a backward stepwise multiple linear regression. The CCT, age, gender, race, hypertension, diabetes, thyroid disease, myopia, and a family history of glaucoma were taken into account in the multivariate analysis. As with IOP, the dependency of CCT measurements on the right eye/left eye and operator was determined by a two-way ANOVA. The aforementioned factors also were analyzed for their relationship with the dependent variable CCT using a backward stepwise multiple linear regression.
Results
A two-way ANOVA on gender and hypertension was performed to determine whether differences between these groups were statistically significant. The difference in mean IOP between males (15.2 mmHg ⫾ 0.3 [standard error of the mean, SEM]) and females (16.3 mmHg ⫾ 0.4 [SEM]) was statistically significant (P ⫽ 0.038), whereas the difference in IOP between hypertensives (14.5 mmHg ⫾ 1.1 [SEM]) and nonhypertensives (15.4 mmHg ⫾ 0.5 [SEM]) was not found to be statistically significant (P ⫽ 0.467). There was no statistically significant interaction between gender and hypertension (P ⫽ 0.349). Therefore, the only factors that had any predictive importance to IOP were gender (P ⫽ 0.038) and CCT (P ⬍ 0.001). Intraocular pressure was found to increase at a rate of 1.94 mmHg per 100 m of increase in CCT (Fig 2). Because gender was a significant predictor of IOP, the groups were separated according to gender. A multiple linear regression was performed for the data sets, taking into account the factors used for the whole population. As listed in Table 2, CCT was found to be statistically significant for predicting IOP in males (P ⬍ 0.001). Diabetes was of borderline significance (P ⫽ 0.012 in the right eye and P ⫽ 0.089 in the left eye) in
A total of 351 subjects were tested in the screening protocol. Subjects were excluded from the study if they were diagnosed with glaucoma (n ⫽ 7), were wearing contact lenses (n ⫽ 7), had no CCT or IOP measurements for both eyes (n ⫽ 3), had prior eye trauma (n ⫽ 1), or had pseudophakos (n ⫽ 1). In total, 19 of 351 subjects were excluded completely from the study. Of the remaining 332 subjects, 651 eyes were included in the study. Some eyes (5 right and 8 left) were excluded because prior trauma to the eye hindered measurement of IOP or CCT or both. Descriptive statistics for the population are presented in Table 1. If IOP or CCT data were missing from an eye, that eye was excluded from analysis and the other eye was still used in the analysis. However, if, for example, age, gender, and race were missing, the multivariate regression used eliminated the subject from the specific test that involved the missing information. Therefore, the sum of the individual subgroups may not reflect the sum of the total population. Unless otherwise mentioned, the results presented are for right eyes only; left eye results correlated highly with those of right eye. A majority (90%) of the IOP measurements were obtained by two operators and did not differ significantly (P ⫽ 0.460). The only factors found to have a statistically significant effect on the prediction of IOP were CCT (P ⬍ 0.001) and gender (P ⫽ 0.007). Hypertension was of borderline significance (P ⫽ 0.052).
Figure 2. Intraocular pressure (IOP) vs. central corneal thickness (CCT) (right eye vs. left eye) regression for general population sample. IOPOD ⫽ CCT (m) *0.0194 (mmHg/m) ⫹ 6.37 (mmHg) IOPOS ⫽ CCT (m) *0.0198 (mmHg/m) ⫹ 5.63 (mmHg).
1851
Ophthalmology Volume 105, Number 10, October 1998 Table 2. Multiple Linear Backward Stepwise Regression Results (P)
Table 3. ANOVA Results: All Pairwise Multiple Comparison Test on Mean CCT for Different Races
IOP
Age Family history of glaucoma Hypertension Diabetes Thyroid disease Myopia Race CCT
Male
Female
CCT
0.515 0.799 0.598 0.012* 0.523 0.308 0.324 ⬍0.001*
0.385 0.021* 0.010* 0.112 0.986 0.623 0.479 0.064†
0.369 0.141 0.763 0.097 0.225 0.819 ⬍0.001*
IOP ⫽ intraocular pressure; CCT ⫽ central corneal thickness. * P ⬍ 0.05. † P ⫽ 0.019 (left eye).
males. For females, CCT was of borderline statistical significance (P ⫽ 0.064 in the right eye and P ⫽ 0.019 in the left eye), and a family history of glaucoma and hypertension was significant in the prediction of IOP (Table 2). Intraocular pressure was found to increase by 2.9 mmHg ⫾ 0.8 mmHg (SEM) per 100 m of CCT in males and 1.2 mmHg ⫾ 0.7 mmHg (SEM) per 100 m of CCT in females (Fig 3). For CCT, 95% of the measurements were taken by three operators, and there was a significant difference between the pachymetry measurements obtained from the different operators (P ⬍ 0.001). No significant difference existed between CCT measurements for the right and left eyes (P ⫽ 0.576). The race of the subject was the only statistically significant predictor of CCT (P ⬍ 0.001) in addition to operator of the pachymeter. A two-way ANOVA was performed on CCT to discover an interaction between the two factors. Race still was found significant (P ⬍ 0.001) even after taking into account the variability by operator. The races that had statistically significant differences in mean CCT from each other were whites versus blacks and whites versus Asians (Table 3).
Figure 3. Intraocular pressure (IOP) vs. central corneal thickness (CCT) regression for males and females (right-eye data only). IOPmale ⫽ CCT (m) *0.0290 (mmHg/m) ⫺ 0.69 (mmHg) IOPfemale ⫽ CCT (m) *0.0121 (mmHg/m) ⫹ 8.93 (mmHg).
1852
Race
Mean [SEM]
Significant Difference*
White Black Asian Native Other
552.5 [2.3] 529.7 [5.3] 532.8 [7.0] 527.8 [12.0] 539.1 [9.9]
— Yes Yes No No
CCT ⫽ central corneal thickness; SEM ⫽ standard error of the mean. * The mean value from the white population (P ⬍ 0.05).
Discussion Previous investigators have reported on the relationship between CCT and measured IOP for Goldmann applanation tonometry.1– 4,10,11 The Goldmann tonometer has been calibrated for a CCT of approximately 0.52 mm and will tend to overestimate IOP in structurally thicker corneas and underestimate IOP in thinner corneas. In corneas with a CCT greater than 0.52 mm, this systematic error in IOP measurement is related to more corneal rigidity due to the increased CCT. This increase in rigidity results in more force being applied to applanate the cornea. The additional force is registered by the instrument as an increase in IOP. Therefore, for thick corneas, the instrument will usually report IOPs greater than the true IOP. In thinner corneas, the rigidity decreases and the instrument will report pressures lower than the true IOP. To our knowledge, this issue has not been addressed with regard to the Tono-Pen. The Tono-Pen electronic tonometer is used extensively to measure IOP and is considered fairly accurate when compared to manometric readings in normal eyes.5,12–14 In addition, studies comparing IOP measured by the Goldmann tonometer and the Tono-Pen have shown a high degree of correlation between the two instruments12,15–17 with a correlation coefficient as high as 0.97.5 This correlation is not surprising, considering that both instruments depend essentially on applanation of the cornea, although the Tono-Pen is a tonometer based on the Mackay–Marg principle, whereas Goldmann tonometry is based on a modified version of the Imbert–Fick Law. However, some studies have indicated that the Tono-Pen overestimates Goldmann tonometry at low IOP range (⬍9 mmHg) and underestimates it at high IOP range (⬎30 mmHg).5,15,16,18 Frenkel et al15 concluded that Tono-Pen measurements correspond well to Goldmann in the 11- to 20-mmHg interval. In our population, 85.3% (n ⫽ 279) of the IOP measurements were within this interval. The remaining measurements were evenly distributed above and below this range with 7.3% (n ⫽ 24) of the IOP measurements above this interval and 7.3% below it. Therefore, we do not believe that the Tono-Pen’s inherent inaccuracy at extreme values was a significant factor in biasing our results. Our population-based study shows that there is a direct correlation between the CCT and IOP measured by the
Dohadwala et al 䡠 Correlation of Tono-Pen IOP and Central Corneal Thickness Tono-Pen tonometer with an increase in IOP of approximately 2 mmHg per 100 m of CCT increase. Ehlers et al4 and Whitacre et al3 have shown similar systematic errors with Goldmann applanation tonometry in manometric studies. Ehlers et al reported an approximate mean error in measured IOP of 5 mmHg for every 70 m CCT change at a true IOP of 20 mmHg, whereas Whitacre et al predicted a mean error of 3.5 mmHg for every 70 m change in CCT. Interestingly, the dependence of IOP on CCT was different for males and females. Males showed a stronger dependence on CCT than did females. The strength of this correlation in females was of borderline statistical significance (partially due to lack of power of the performed tests). This weak association between CCT and measured IOP in females may be because of an effect caused by confounding factors not considered in this study (e.g., the onset of menopause or the stage of menstrual cycle).19,20 The difference in the IOP–CCT correlation between the genders needs to be confirmed and, if reproduced, could have important implications for the adjustment of IOP in the clinical setting. The dependency of current tonometry on CCT is an important issue. For example, many patients with ocular hypertension have structurally thick corneas, and after adjustment for CCT, the IOP may fall into the “normal” range. Hence, patients avoid having to undergo costly testing and follow-up examinations, as well as being labeled as having a disease that in some cases simply is an artifact of measurement. For instance, Argus11 states than in his study of 96 patients, 30% of ocular hypertensives with a measured IOP of 21 mmHg or greater had IOPs of 18 mmHg or lower when adjusted for CCT. Furthermore, corneas may be iatrogenically thinned such as after myopic photorefractive keratectomy. In such cases, IOP readings may be falsely lower and need to be adjusted upward to reflect the true IOP and, hence, avoid the diagnosis of glaucoma from being missed. Factors that affect IOP have been studied extensively.19,21 Because of the significant difference in IOP between the genders found in our study and in others20,22,23 in which females tended to have higher IOPs than males, the population was separated to examine the risk factors for elevated IOP between the genders. Hypertension is known to cause increased IOP,24 –26 although it is unclear why its significance seems to differ between the sexes. Even though persons with a family history of glaucoma have been found to have a higher IOP than is common,20,27 the fact that it is significant in predicting IOP only in females in our population is somewhat of an anomaly. This apparent disparity may be misleading, because it may be that the female population was more aware of its family medical history than its male counterparts. Studies have found a positive correlation between IOP and those 40 years of age and older.20,28 Although no such relationship reached statistical significance in our study, there was a trend for IOP to increase with age in females older than 40 years but not in males. This increase in IOP in females older than 40 years of age may possibly be attributed to the onset of menopause.20 Although it has been shown that CCT seems to be independent of many factors in ocular physiology,29 our
study finds that there is a significant difference in the CCT found among Asians, blacks, and whites. However, the difference between the mean CCT of the races (⬍25 m) suggests that the variance is most likely not clinically significant. Moreover, our study found no statistically significant difference among the mean IOP of whites, blacks, and Asians, although the sample sizes for the black and Asian populations were small. Our findings show that IOP readings by the Tono-Pen, like Goldmann tonometry, are directly correlated with the CCT. Thus, clinicians should be aware that there might be a systematic error in Tono-Pen readings. The CCT measurements should be considered to ensure proper interpretation of Tono-Pen IOP in the diagnosis and management of disorders in which the measured IOP does not correspond to other clinical findings, or in cases in which CCT may have been altered such as in myopic photorefractive keratectomy. In conclusion, there is a need for a follow-up multicenter, prospective study evaluating the effect of CCT on IOP in patients with and without glaucoma to see whether there is any significant difference among these groups. Acknowledgment. The authors thank Mentor Canada for supplying the Tono-Pens used in this study.
References 1. Johnson M, Kass MA, Moses RA, Grodzki WJ. Increased corneal thickness simulating elevated intraocular pressure. Arch Ophthalmol 1978;96:664 –5. 2. Hansen FK. A clinical study of the normal human central corneal thickness. Acta Ophthalmol (Copenh) 1971;49:82–9. 3. Whitacre MM, Stein RA, Hassanein K. The effect of corneal thickness on applanation tonometry. Am J Ophthalmol 1993; 115:592– 6. 4. Ehlers N, Bramsen T, Sperling S. Applanation tonometry and central corneal thickness. Acta Ophthalmol (Copenh) 1975; 53:34 – 43. 5. Hessemer V, Rossler R, Jacobi KW. Tono-Pen, a new tonometer. Int Ophthalmol 1989;13:51– 6. 6. Khan JA, LaGreca BA. Tono-Pen estimation of intraocular pressure through bandage contact lenses. Am J Ophthalmol 1989;108:422–5. 7. Khan JA, Davis M, Graham CE, et al. Comparison of Oculab Tono-Pen readings obtained from various corneal and scleral locations. Arch Ophthalmol 1991;109:1444 – 6. 8. Rootman DS, Insler MS, Thompson HW, et al. Accuracy and precision of the Tono-Pen in measuring intraocular pressure after keratoplasty and epikeratophakia and in scarred corneas. Arch Ophthalmol 1988;106:1697–700. 9. Ederer F. Shall we count numbers of eyes or numbers of subjects? Arch Ophthalmol 1973;89:1–2. 10. Whitacre MM, Stein R. Sources of error with use of Goldmann-type tonometers [review]. Surv Ophthalmol 1993;38:1– 30. 11. Argus WA. Ocular hypertension and central corneal thickness. Ophthalmology 1995;102:1810 –2. 12. Boothe WA, Lee DA, Panek WC, Pettit TH. The Tono-Pen. A manometric and clinical study. Arch Ophthalmol 1988;106: 1214 –7. 13. Hessemer V, Ro¨ssler R, Jacobi KW. Comparison of intraocular pressure measurements with the Oculab Tono-Pen vs
1853
Ophthalmology Volume 105, Number 10, October 1998
14.
15. 16. 17. 18. 19. 20. 21.
manometry in humans shortly after death. Am J Ophthalmol 1988;105:678 – 82. Hessemer V, Ro¨ssler R, Jacobi KW. Tono-pen versus manometer. Vergleichende Augeninnendruckmessungen an menschlichen Leichenaugen [Eng abstr]. Klin Monatsbl Augenheilkd 1989;195:23–7. Frenkel RE, Hong YJ, Shin DH. Comparison of the Tono-Pen to the Goldmann applanation tonometer. Arch Ophthalmol 1988;106:750 –3. Kao SF, Lichter PR, Bergstrom TJ, et al. Clinical comparison of the Oculab Tono-Pen to the Goldmann applanation tonometer. Ophthalmology 1987;94:1541– 4. Minckler DS, Baerveldt G, Heuer DK, et al. Clinical evaluation of the Oculab Tono-Pen. Am J Ophthalmol 1987;104: 168 –73. Kooner KS, Cooksey JC, Barron JB, et al. Tonometry comparison: Goldmann versus Tono-Pen. Ann Ophthalmol 1992; 24:29 –36. Colton T, Ederer F. The distribution of intraocular pressures in the general population. Surv Ophthalmol 1980;25:123–9. Armaly MF. On the distribution of applanation pressure. I. Statistical features and the effect of age, sex, and family history of glaucoma. Arch Ophthalmol 1965;73:11– 8. Shields MB. Textbook of Glaucoma, 3rd ed. Baltimore: Williams & Wilkins, 1992;53– 83.
1854
22. Qureshi IA, Xi XR, Huang YB, et al. Distribution of intraocular pressure among healthy Pakistani. Chin J Physiol 1996; 39:183– 8. 23. Hirvela H, Tuulonen A, Laatikainen L. Intraocular pressure and prevalence of glaucoma in elderly people in Finland: a population-based study. Int Ophthalmol 1994;18:299 –307. 24. Dielemans I, Vingerling JR, Algra D, et al. Primary openangle glaucoma, intraocular pressure, and systemic blood pressure in the general elderly population. The Rotterdam Study. Ophthalmology 1995;102:54 – 60. 25. Shiose Y. The aging effect on intraocular pressure in an apparently normal population. Arch Ophthalmol 1984;102: 883–7. 26. Tielsch JM, Katz J, Sommer A, et al. Hypertension, perfusion pressure, and primary open-angle glaucoma. A populationbased assessment. Arch Ophthalmol 1995;113:216 –21. 27. Seddon JM, Schwartz B, Flowerdew G. Case– control study of ocular hypertension. Arch Ophthalmol 1983;101:891– 4. 28. Loewen U, Handrup B, Redeker A. Ergebnisse einer Glaukomreihenuntersuchung (Statistische Auswertung) [Eng Abstr]. Klin Monatsbl Augenheilkd 1976;169:754 – 66. 29. Ehlers M, Hansen FK. Further data on biometric correlations of central corneal thickness. Acta Ophthalmol (Copenh) 1976; 54:774 – 8.
Originally received: September 19, 1997. Revision accepted: May 15, 1998. Manuscript no. 97677. From the University of Ottawa Eye Institute, Ottawa General Hospital, Ottawa, Ontario, Canada. Address correspondence and reprint requests to Karim F. Damji, MD, University of Ottawa Eye Institute, 501 Smyth Rd, Ottawa, ON, K1H 8L6, Canada.
because of its electronic design with LCD display, its ease and flexibility of use,6,7 its portability, and its accuracy in cases of irregular corneas.8 This study examines the relationship between IOP and CCT readings taken by the TonoPen tonometer and discusses factors that predict IOP and CCT in a general population sample.
Methods A prospective study was conducted at four public buildings (3 malls, 1 hospital) between 9 AM and 4 PM as part of an eye health information and glaucoma-screening program. Participation was strictly voluntary, and all individuals who were willing to be tested were included. Initially, each participant filled out an anonymous questionnaire that asked for information pertaining to possible risk factors and exclusion criteria (Fig 1). Subjects were then given topical anesthesia (Alcaine; Alcon, Ft. Worth, TX) bilaterally a few seconds before being tested by qualified technicians. Intraocular pressure was measured with a Tono-Pen by one of six operators. The IOP reading was followed by a measurement of the CCT using a Mentor Advent ultrasonic pachymeter by one of five operators. Three acceptable readings within a range of ⫾2 mmHg for IOP and ⫾5 m for CCT were recorded for each eye. Both instruments were calibrated before use in each screening clinic. SPSS SigmaStat 2.0 software (SPSS, Chicago, IL) was used to perform the statistical analysis. Statistical significance was set at
1849
Ophthalmology Volume 105, Number 10, October 1998
Figure 1. Questionnaire used in the study.
1850
Dohadwala et al 䡠 Correlation of Tono-Pen IOP and Central Corneal Thickness Table 1. Descriptive Statistics of the Population* Medical History
Race
Female
Male
Total
Age [SD] (yrs) Family history of glaucoma Hypertension Diabetes Thyroid disease Myopia Total
44.9 [18.0] 36 20 7 20 56 165
45.4 [16.8] 15 27 8 2 51 162
45.2 [17.4] 52 48 15 22 109 332
IOP [SEM] (mmHg) CCT [SEM] (m)
16.3 [0.4] 545.3 [2.6]
15.2 [0.3] 548.7 [2.8]
15.6 [0.2] 547.3 [2.0]
White Black Asian Native Other Total
Female
Male
Total
113 16 9 2 13 165
111 16 15 3 6 162
227 32 24 5 19 332
* IOP and CCT means are for OD only. SD ⫽ standard deviation; SEM ⫽ standard error of the mean; IOP ⫽ intraocular pressure; CCT ⫽ central corneal thickness.
P ⬍ 0.05. For all statistical tests, the mean IOP and CCT of the three values within the preferred range was used. Because the eyes for one subject are not independent when considering ocular parameters such as IOP, the right and left eyes were always analyzed separately.9 The IOP dependency on operator and right eye/left eye was examined using a two-way analysis of variance (ANOVA). The prediction of IOP by certain factors was analyzed simultaneously using a backward stepwise multiple linear regression. The CCT, age, gender, race, hypertension, diabetes, thyroid disease, myopia, and a family history of glaucoma were taken into account in the multivariate analysis. As with IOP, the dependency of CCT measurements on the right eye/left eye and operator was determined by a two-way ANOVA. The aforementioned factors also were analyzed for their relationship with the dependent variable CCT using a backward stepwise multiple linear regression.
Results
A two-way ANOVA on gender and hypertension was performed to determine whether differences between these groups were statistically significant. The difference in mean IOP between males (15.2 mmHg ⫾ 0.3 [standard error of the mean, SEM]) and females (16.3 mmHg ⫾ 0.4 [SEM]) was statistically significant (P ⫽ 0.038), whereas the difference in IOP between hypertensives (14.5 mmHg ⫾ 1.1 [SEM]) and nonhypertensives (15.4 mmHg ⫾ 0.5 [SEM]) was not found to be statistically significant (P ⫽ 0.467). There was no statistically significant interaction between gender and hypertension (P ⫽ 0.349). Therefore, the only factors that had any predictive importance to IOP were gender (P ⫽ 0.038) and CCT (P ⬍ 0.001). Intraocular pressure was found to increase at a rate of 1.94 mmHg per 100 m of increase in CCT (Fig 2). Because gender was a significant predictor of IOP, the groups were separated according to gender. A multiple linear regression was performed for the data sets, taking into account the factors used for the whole population. As listed in Table 2, CCT was found to be statistically significant for predicting IOP in males (P ⬍ 0.001). Diabetes was of borderline significance (P ⫽ 0.012 in the right eye and P ⫽ 0.089 in the left eye) in
A total of 351 subjects were tested in the screening protocol. Subjects were excluded from the study if they were diagnosed with glaucoma (n ⫽ 7), were wearing contact lenses (n ⫽ 7), had no CCT or IOP measurements for both eyes (n ⫽ 3), had prior eye trauma (n ⫽ 1), or had pseudophakos (n ⫽ 1). In total, 19 of 351 subjects were excluded completely from the study. Of the remaining 332 subjects, 651 eyes were included in the study. Some eyes (5 right and 8 left) were excluded because prior trauma to the eye hindered measurement of IOP or CCT or both. Descriptive statistics for the population are presented in Table 1. If IOP or CCT data were missing from an eye, that eye was excluded from analysis and the other eye was still used in the analysis. However, if, for example, age, gender, and race were missing, the multivariate regression used eliminated the subject from the specific test that involved the missing information. Therefore, the sum of the individual subgroups may not reflect the sum of the total population. Unless otherwise mentioned, the results presented are for right eyes only; left eye results correlated highly with those of right eye. A majority (90%) of the IOP measurements were obtained by two operators and did not differ significantly (P ⫽ 0.460). The only factors found to have a statistically significant effect on the prediction of IOP were CCT (P ⬍ 0.001) and gender (P ⫽ 0.007). Hypertension was of borderline significance (P ⫽ 0.052).
Figure 2. Intraocular pressure (IOP) vs. central corneal thickness (CCT) (right eye vs. left eye) regression for general population sample. IOPOD ⫽ CCT (m) *0.0194 (mmHg/m) ⫹ 6.37 (mmHg) IOPOS ⫽ CCT (m) *0.0198 (mmHg/m) ⫹ 5.63 (mmHg).
1851
Ophthalmology Volume 105, Number 10, October 1998 Table 2. Multiple Linear Backward Stepwise Regression Results (P)
Table 3. ANOVA Results: All Pairwise Multiple Comparison Test on Mean CCT for Different Races
IOP
Age Family history of glaucoma Hypertension Diabetes Thyroid disease Myopia Race CCT
Male
Female
CCT
0.515 0.799 0.598 0.012* 0.523 0.308 0.324 ⬍0.001*
0.385 0.021* 0.010* 0.112 0.986 0.623 0.479 0.064†
0.369 0.141 0.763 0.097 0.225 0.819 ⬍0.001*
IOP ⫽ intraocular pressure; CCT ⫽ central corneal thickness. * P ⬍ 0.05. † P ⫽ 0.019 (left eye).
males. For females, CCT was of borderline statistical significance (P ⫽ 0.064 in the right eye and P ⫽ 0.019 in the left eye), and a family history of glaucoma and hypertension was significant in the prediction of IOP (Table 2). Intraocular pressure was found to increase by 2.9 mmHg ⫾ 0.8 mmHg (SEM) per 100 m of CCT in males and 1.2 mmHg ⫾ 0.7 mmHg (SEM) per 100 m of CCT in females (Fig 3). For CCT, 95% of the measurements were taken by three operators, and there was a significant difference between the pachymetry measurements obtained from the different operators (P ⬍ 0.001). No significant difference existed between CCT measurements for the right and left eyes (P ⫽ 0.576). The race of the subject was the only statistically significant predictor of CCT (P ⬍ 0.001) in addition to operator of the pachymeter. A two-way ANOVA was performed on CCT to discover an interaction between the two factors. Race still was found significant (P ⬍ 0.001) even after taking into account the variability by operator. The races that had statistically significant differences in mean CCT from each other were whites versus blacks and whites versus Asians (Table 3).
Figure 3. Intraocular pressure (IOP) vs. central corneal thickness (CCT) regression for males and females (right-eye data only). IOPmale ⫽ CCT (m) *0.0290 (mmHg/m) ⫺ 0.69 (mmHg) IOPfemale ⫽ CCT (m) *0.0121 (mmHg/m) ⫹ 8.93 (mmHg).
1852
Race
Mean [SEM]
Significant Difference*
White Black Asian Native Other
552.5 [2.3] 529.7 [5.3] 532.8 [7.0] 527.8 [12.0] 539.1 [9.9]
— Yes Yes No No
CCT ⫽ central corneal thickness; SEM ⫽ standard error of the mean. * The mean value from the white population (P ⬍ 0.05).
Discussion Previous investigators have reported on the relationship between CCT and measured IOP for Goldmann applanation tonometry.1– 4,10,11 The Goldmann tonometer has been calibrated for a CCT of approximately 0.52 mm and will tend to overestimate IOP in structurally thicker corneas and underestimate IOP in thinner corneas. In corneas with a CCT greater than 0.52 mm, this systematic error in IOP measurement is related to more corneal rigidity due to the increased CCT. This increase in rigidity results in more force being applied to applanate the cornea. The additional force is registered by the instrument as an increase in IOP. Therefore, for thick corneas, the instrument will usually report IOPs greater than the true IOP. In thinner corneas, the rigidity decreases and the instrument will report pressures lower than the true IOP. To our knowledge, this issue has not been addressed with regard to the Tono-Pen. The Tono-Pen electronic tonometer is used extensively to measure IOP and is considered fairly accurate when compared to manometric readings in normal eyes.5,12–14 In addition, studies comparing IOP measured by the Goldmann tonometer and the Tono-Pen have shown a high degree of correlation between the two instruments12,15–17 with a correlation coefficient as high as 0.97.5 This correlation is not surprising, considering that both instruments depend essentially on applanation of the cornea, although the Tono-Pen is a tonometer based on the Mackay–Marg principle, whereas Goldmann tonometry is based on a modified version of the Imbert–Fick Law. However, some studies have indicated that the Tono-Pen overestimates Goldmann tonometry at low IOP range (⬍9 mmHg) and underestimates it at high IOP range (⬎30 mmHg).5,15,16,18 Frenkel et al15 concluded that Tono-Pen measurements correspond well to Goldmann in the 11- to 20-mmHg interval. In our population, 85.3% (n ⫽ 279) of the IOP measurements were within this interval. The remaining measurements were evenly distributed above and below this range with 7.3% (n ⫽ 24) of the IOP measurements above this interval and 7.3% below it. Therefore, we do not believe that the Tono-Pen’s inherent inaccuracy at extreme values was a significant factor in biasing our results. Our population-based study shows that there is a direct correlation between the CCT and IOP measured by the
Dohadwala et al 䡠 Correlation of Tono-Pen IOP and Central Corneal Thickness Tono-Pen tonometer with an increase in IOP of approximately 2 mmHg per 100 m of CCT increase. Ehlers et al4 and Whitacre et al3 have shown similar systematic errors with Goldmann applanation tonometry in manometric studies. Ehlers et al reported an approximate mean error in measured IOP of 5 mmHg for every 70 m CCT change at a true IOP of 20 mmHg, whereas Whitacre et al predicted a mean error of 3.5 mmHg for every 70 m change in CCT. Interestingly, the dependence of IOP on CCT was different for males and females. Males showed a stronger dependence on CCT than did females. The strength of this correlation in females was of borderline statistical significance (partially due to lack of power of the performed tests). This weak association between CCT and measured IOP in females may be because of an effect caused by confounding factors not considered in this study (e.g., the onset of menopause or the stage of menstrual cycle).19,20 The difference in the IOP–CCT correlation between the genders needs to be confirmed and, if reproduced, could have important implications for the adjustment of IOP in the clinical setting. The dependency of current tonometry on CCT is an important issue. For example, many patients with ocular hypertension have structurally thick corneas, and after adjustment for CCT, the IOP may fall into the “normal” range. Hence, patients avoid having to undergo costly testing and follow-up examinations, as well as being labeled as having a disease that in some cases simply is an artifact of measurement. For instance, Argus11 states than in his study of 96 patients, 30% of ocular hypertensives with a measured IOP of 21 mmHg or greater had IOPs of 18 mmHg or lower when adjusted for CCT. Furthermore, corneas may be iatrogenically thinned such as after myopic photorefractive keratectomy. In such cases, IOP readings may be falsely lower and need to be adjusted upward to reflect the true IOP and, hence, avoid the diagnosis of glaucoma from being missed. Factors that affect IOP have been studied extensively.19,21 Because of the significant difference in IOP between the genders found in our study and in others20,22,23 in which females tended to have higher IOPs than males, the population was separated to examine the risk factors for elevated IOP between the genders. Hypertension is known to cause increased IOP,24 –26 although it is unclear why its significance seems to differ between the sexes. Even though persons with a family history of glaucoma have been found to have a higher IOP than is common,20,27 the fact that it is significant in predicting IOP only in females in our population is somewhat of an anomaly. This apparent disparity may be misleading, because it may be that the female population was more aware of its family medical history than its male counterparts. Studies have found a positive correlation between IOP and those 40 years of age and older.20,28 Although no such relationship reached statistical significance in our study, there was a trend for IOP to increase with age in females older than 40 years but not in males. This increase in IOP in females older than 40 years of age may possibly be attributed to the onset of menopause.20 Although it has been shown that CCT seems to be independent of many factors in ocular physiology,29 our
study finds that there is a significant difference in the CCT found among Asians, blacks, and whites. However, the difference between the mean CCT of the races (⬍25 m) suggests that the variance is most likely not clinically significant. Moreover, our study found no statistically significant difference among the mean IOP of whites, blacks, and Asians, although the sample sizes for the black and Asian populations were small. Our findings show that IOP readings by the Tono-Pen, like Goldmann tonometry, are directly correlated with the CCT. Thus, clinicians should be aware that there might be a systematic error in Tono-Pen readings. The CCT measurements should be considered to ensure proper interpretation of Tono-Pen IOP in the diagnosis and management of disorders in which the measured IOP does not correspond to other clinical findings, or in cases in which CCT may have been altered such as in myopic photorefractive keratectomy. In conclusion, there is a need for a follow-up multicenter, prospective study evaluating the effect of CCT on IOP in patients with and without glaucoma to see whether there is any significant difference among these groups. Acknowledgment. The authors thank Mentor Canada for supplying the Tono-Pens used in this study.
References 1. Johnson M, Kass MA, Moses RA, Grodzki WJ. Increased corneal thickness simulating elevated intraocular pressure. Arch Ophthalmol 1978;96:664 –5. 2. Hansen FK. A clinical study of the normal human central corneal thickness. Acta Ophthalmol (Copenh) 1971;49:82–9. 3. Whitacre MM, Stein RA, Hassanein K. The effect of corneal thickness on applanation tonometry. Am J Ophthalmol 1993; 115:592– 6. 4. Ehlers N, Bramsen T, Sperling S. Applanation tonometry and central corneal thickness. Acta Ophthalmol (Copenh) 1975; 53:34 – 43. 5. Hessemer V, Rossler R, Jacobi KW. Tono-Pen, a new tonometer. Int Ophthalmol 1989;13:51– 6. 6. Khan JA, LaGreca BA. Tono-Pen estimation of intraocular pressure through bandage contact lenses. Am J Ophthalmol 1989;108:422–5. 7. Khan JA, Davis M, Graham CE, et al. Comparison of Oculab Tono-Pen readings obtained from various corneal and scleral locations. Arch Ophthalmol 1991;109:1444 – 6. 8. Rootman DS, Insler MS, Thompson HW, et al. Accuracy and precision of the Tono-Pen in measuring intraocular pressure after keratoplasty and epikeratophakia and in scarred corneas. Arch Ophthalmol 1988;106:1697–700. 9. Ederer F. Shall we count numbers of eyes or numbers of subjects? Arch Ophthalmol 1973;89:1–2. 10. Whitacre MM, Stein R. Sources of error with use of Goldmann-type tonometers [review]. Surv Ophthalmol 1993;38:1– 30. 11. Argus WA. Ocular hypertension and central corneal thickness. Ophthalmology 1995;102:1810 –2. 12. Boothe WA, Lee DA, Panek WC, Pettit TH. The Tono-Pen. A manometric and clinical study. Arch Ophthalmol 1988;106: 1214 –7. 13. Hessemer V, Ro¨ssler R, Jacobi KW. Comparison of intraocular pressure measurements with the Oculab Tono-Pen vs
1853
Ophthalmology Volume 105, Number 10, October 1998
14.
15. 16. 17. 18. 19. 20. 21.
manometry in humans shortly after death. Am J Ophthalmol 1988;105:678 – 82. Hessemer V, Ro¨ssler R, Jacobi KW. Tono-pen versus manometer. Vergleichende Augeninnendruckmessungen an menschlichen Leichenaugen [Eng abstr]. Klin Monatsbl Augenheilkd 1989;195:23–7. Frenkel RE, Hong YJ, Shin DH. Comparison of the Tono-Pen to the Goldmann applanation tonometer. Arch Ophthalmol 1988;106:750 –3. Kao SF, Lichter PR, Bergstrom TJ, et al. Clinical comparison of the Oculab Tono-Pen to the Goldmann applanation tonometer. Ophthalmology 1987;94:1541– 4. Minckler DS, Baerveldt G, Heuer DK, et al. Clinical evaluation of the Oculab Tono-Pen. Am J Ophthalmol 1987;104: 168 –73. Kooner KS, Cooksey JC, Barron JB, et al. Tonometry comparison: Goldmann versus Tono-Pen. Ann Ophthalmol 1992; 24:29 –36. Colton T, Ederer F. The distribution of intraocular pressures in the general population. Surv Ophthalmol 1980;25:123–9. Armaly MF. On the distribution of applanation pressure. I. Statistical features and the effect of age, sex, and family history of glaucoma. Arch Ophthalmol 1965;73:11– 8. Shields MB. Textbook of Glaucoma, 3rd ed. Baltimore: Williams & Wilkins, 1992;53– 83.
1854
22. Qureshi IA, Xi XR, Huang YB, et al. Distribution of intraocular pressure among healthy Pakistani. Chin J Physiol 1996; 39:183– 8. 23. Hirvela H, Tuulonen A, Laatikainen L. Intraocular pressure and prevalence of glaucoma in elderly people in Finland: a population-based study. Int Ophthalmol 1994;18:299 –307. 24. Dielemans I, Vingerling JR, Algra D, et al. Primary openangle glaucoma, intraocular pressure, and systemic blood pressure in the general elderly population. The Rotterdam Study. Ophthalmology 1995;102:54 – 60. 25. Shiose Y. The aging effect on intraocular pressure in an apparently normal population. Arch Ophthalmol 1984;102: 883–7. 26. Tielsch JM, Katz J, Sommer A, et al. Hypertension, perfusion pressure, and primary open-angle glaucoma. A populationbased assessment. Arch Ophthalmol 1995;113:216 –21. 27. Seddon JM, Schwartz B, Flowerdew G. Case– control study of ocular hypertension. Arch Ophthalmol 1983;101:891– 4. 28. Loewen U, Handrup B, Redeker A. Ergebnisse einer Glaukomreihenuntersuchung (Statistische Auswertung) [Eng Abstr]. Klin Monatsbl Augenheilkd 1976;169:754 – 66. 29. Ehlers M, Hansen FK. Further data on biometric correlations of central corneal thickness. Acta Ophthalmol (Copenh) 1976; 54:774 – 8.