Reliability of Measurements Using Ocular Response Analyzer as a Screening Tonometer and Corneal Hysteresis Values in the Presence or Absence of Glaucomatous Changes in Fundus

Abstract

Précis:

Use of the Ocular Response Analyzer (ORA) as a screening tonometer in clinical practice yielded reliable measurements in over 80% of eyes screened. Including corneal hysteresis (CH) data in screening may improve the accuracy of glaucoma detection.

Purpose:

To examine measurement reliability when the ORA is used as a screening tonometer, and to compare CH measurements in eyes with and those without glaucomatous changes in the fundus.

Patients and Methods:

1488 eyes of 747 patients (mean age: 53.5 ± 20.4 y, range: 6–94 y) underwent intraocular pressure (IOP) measurement using ORA as screening. The percentage of eyes with a waveform score ≥6, the recommended threshold indicating reliability, was calculated. Eyes that had waveform score ≥6 and had undergone fundus photography and optical coherence tomography were assessed for the presence or absence of glaucomatous changes in fundus from optical coherence tomography and fundus images, and CH was compared between the 2 groups.

Results:

Mean ± SD (range) of ORA measurements were: Goldmann-correlated IOP 14.9 ± 4.8 (1.0–63.2) mm Hg, corneal-compensated IOP 16.2 ± 4.7 (3.2–73.6) mm Hg, CH 9.7 ± 1.5 (0.0–20.6) mm Hg, and waveform score 7.3 ± 1.5 (0.1–9.7). Eighty-four percent of eyes had a waveform score ≥6. Among 192 eyes (127 patients, aged 53.5 ± 18.0 y) with waveform score ≥6 and evaluable for glaucomatous changes in the fundus, 53 eyes were determined as positive and 139 eyes as negative. CH was 9.6 ± 1.4 (6.8–13.3) mm Hg in the positive group and 10.2 ± 1.2 (6.9–13.3) mm Hg in the negative group, and was significantly lower in the positive group (P=0.003).

Conclusion:

When using ORA as a screening tonometer, reliable results were obtained in ~80% of the eyes. CH was lower in the glaucomatous change-positive group compared with the glaucomatous change-negative group, but the ranges overlapped between the 2 groups.

Ocular Response Analyzer (ORA) is a tonometer that can measure corneal hysteresis (CH), a parameter that affects the onset^1–3 or progression of glaucoma.^4–10 In recent years, ORA is increasingly being used in clinical care for glaucoma patients. Since ORA is a noncontact tonometer, it is used as a screening device for intraocular pressure (IOP) measurement in routine clinical practice. As a screening tool, CH measured by ORA is also used to assess the severity of glaucoma.¹¹

Many studies to date that investigated the relationship between CH and glaucoma recruited patients who had already been diagnosed with glaucoma, or analyzed measurements obtained after starting treatment with glaucoma eye drops. There is no study on CH measured as a screening test in an unspecified number of individuals without a diagnosis of glaucoma. In addition, studies on the reliability of ORA measurements have been conducted only in healthy subjects,^12,13 and there is no report on the reliability of measurements when ORA is used as a screening tonometer in the clinical setting. Furthermore, although the subjects in almost all the reported studies had glaucoma eyes with visual field abnormalities, research has shown that the pathology of glaucomatous neuropathy is already present before visual field abnormality is detectable, and the pathology can be observed as characteristic changes in the fundus.¹⁴

The purpose of this study was to examine the reliability of measurements when ORA is used as a screening tonometer and to compare the CH measurements in eyes with and those without glaucomatous changes in the fundus.

PATIENTS AND METHODS

We retrospectively reviewed the medical records to identify patients who underwent IOP measurement using an Ocular Response Analyze G3 (Reichert Technologies) as a screening test for abnormal IOP in clinical practice at Yashio Maruyama Eye Clinic between March 1 and May 15, 2021.

In all subjects, ORA measurements were done 3 times while holding the eyelids open if needed, and the average value was used for analysis. The output of the measured results consisted of IOP correlated with the IOP value measured by Goldmann applanation tonometry (IOPg), corneal-compensated IOP (IOPcc), CH, and waveform score which is a coefficient indicating reliability. In this study, the percentage of eyes with a waveform score of 6 or higher was calculated. When the same patient underwent measurement multiple times, the results of the first measurement were used for analysis.

Among eyes with a waveform score of 6 or higher, those that had undergone fundus photography and optical coherence tomography (OCT) examination were further analyzed. Eyes with a history of intraocular surgery including laser treatment, and eyes treated with glaucoma eye drops were excluded from analysis. The eyes included in the analysis were classified into a group with glaucomatous changes in the fundus (positive group) and a group without glaucomatous fundus changes (negative group), and CH measurements of the 2 groups were compared.

Fundus photographs were taken with a nonmydriatic Auto Fundus Camera AFC-330 (Nidek Inc.). The posterior pole area was photographed at an angle of view of 45 degrees under mydriatic or nonmydriatic conditions. For OCT, an RS-3000 Advance (Nidek Inc.) was used, and the macular map (equivalent to a 9 × 9 mm² area in the Gullstrand model eye) was acquired followed by glaucoma analysis, also under mydriatic or nonmydriatic condition. The signal strength index was not considered in evaluating OCT measurements. A single examiner (K.M.) interpreted the fundus photographs and OCT images in the medical records and determined the presence or absence of glaucomatous changes in the fundus. The presence of glaucomatous changes in the fundus was defined as follows: fundus photograph showing increased optic disc cupping and thinning of disc rim accompanied by retinal nerve fiber layer defect, and OCT inner retinal layer thickness analysis showing thinning along the course of the nerve fibers typically with temporal raphe sign,¹⁵ as well as the exclusion of fundus diseases other than glaucoma that could cause retinal nerve fiber layer defect (including branch retinal vein occlusion, diabetic retinopathy, hypertensive fundus changes, and renal retinopathy). However, even if diseases other than glaucoma were present, eyes that were assessed clearly as complicated with glaucomatous changes were determined to be positive. From the results of the interpretation of fundus findings, only those eyes that were definitively determined as positive or negative for glaucomatous change were extracted, and IOPg, IOPcc, and CH measured by ORA were compared between the positive group and the negative group. F test was used to test the equality of variances, and Student t test was used to compare numerical variables between the 2 groups. All statistical analyses were performed using MedCalc version 12.7.4 (MedCalc Software Ltd.). Statistical significance was defined as P<0.05.

This study was approved by the Ethics Review Committee of the Japan Medical Association (Approval No.: R3-8).

RESULTS

Reliability of ORA Used as a Screening Tonometer in Clinical Practice

A review of the medical records at our clinic identified 1488 eyes of 747 patients who underwent IOP measurement using ORA. The subjects comprised 287 males and 460 females aged 53.5 ± 20.4 (range: 6–94) years, and the eyes comprised 745 right eyes and 743 left eyes.

Mean ± SD (range) of the ORA measurements of the 1488 eyes were: IOPg, 14.9 ± 4.8 (1.0–63.2) mm Hg; IOPcc, 16.2 ± 4.7 (3.2–73.6) mm Hg; CH, 9.7 ± 1.5 (0.0–20.6) mm Hg; and waveform score, 7.3 ± 1.5 (0.1–9.7). The distributions of these ORA measurements are shown in Figures 1–4. Waveform score below 3 was found in 2% of the eyes, scores 3–3.9 in 3%, scores 4–4.9 in 4%, scores 5–5.9 in 8%, scores 6–6.9 in 14%, scores 7–7.9 in 27%, scores 8–8.9 in 32%, and scores 9–9.9 in 10%. Of all the eyes analyzed, 84% (1245 eyes) had waveform scores of 6 or higher.

FIGURE 1.

Frequency distribution of intraocular pressure correlated with the values measured by IOPg in 1448 eyes that underwent IOP measurement using the Ocular Response Analyzer. Mean±SD (range): 14.9 ± 4.8 (1.0–63.2) mm Hg. IOPg indicates Goldmann applanation tonometry.

FIGURE 4.

Frequency distribution of waveform scores in 1448 eyes that underwent IOP measurement using the Ocular Response Analyzer. Mean ± SD (range): 7.3 ± 1.5 (0.1–9.7). IOP indicates intraocular pressure.

FIGURE 2.

Frequency distribution of IOPcc in 1448 eyes that underwent IOP measurement using the Ocular Response Analyzer. Mean ± SD (range): 16.2 ± 4.7 (3.2–73.6) mm Hg. IOP indicates intraocular pressure; IOPcc, corneal-compensated IOP.

FIGURE 3.

Frequency distribution of corneal hysteresis in 1448 eyes that underwent IOP measurement using theOcular Response Analyzer. Mean ± SD (range): 9.7 ± 1.5 (0.0–20.6) mm Hg. IOP indicates intraocular pressure.

CH Measurements With/Without Glaucomatous Changes in Fundus

Among the 1488 eyes, 1245 eyes had a waveform score of 6 or higher. Six hundred and sixty-three eyes that had a history of intraocular surgery were excluded. Of the remaining 582 eyes, 341 eyes had interpretable fundus photographs and OCT images, but the presence or absence of glaucomatous changes could not be determined in 149 eyes. Eventually, 127 patients (mean age: 53.5 ± 18.0, range: 9–87 y) with 192 eyes (53 eyes in the positive group and 139 eyes in the negative group) were included as subjects for analysis (Fig. 5). This study population included 2 patients with unilateral glaucoma, in whom 1 eye was classified in the positive group and 1 eye in the negative group.

FIGURE 5.

Selection of eyes for analysis of glaucomatous changes in the fundus. Eyes with waveform scores below 6, eyes with a history of intraocular surgery, eyes without interpretable fundus photographs and OCT images, and eyes unevaluable for the presence or absence of glaucomatous changes were excluded. Eventually, 192 eyes (53 eyes positive and 139 eyes negative for glaucomatous change) were analyzed. IOP indicates intraocular pressure; OCT, optical coherence tomography; ORA, Ocular Response Analyzer.

Figure 6 shows the frequency distributions of ORA measurements of IOPg, IOPcc, and CH in the positive group and negative group. There were no significant differences in variance between the 2 groups in all 3 parameters (IOPg: P=0.17, IOPcc: P=0.16, CH: P=0.09).

FIGURE 6.

Frequency distributions of measured values of A, Goldmann-correlated intraocular pressure (IOPg); B, IOPcc; and C, CH in the glaucomatous change-positive (+) group and the glaucomatous change-negative (−) group. Open bars: (+) group: 53 eyes with glaucomatous changes in the fundus. Filled bars: (−) group: 139 eyes without glaucomatous changes in the fundus. IOPg: F test, P=0.17. IOPcc: F test, P=0.16. CH: F test, P=0.09. CH indicates corneal hysteresis; IOPcc, corneal-compensated IOP; IOPg, Goldmann applanation tonometry.

Figure 7 shows the boxplots of IOPg, IOPcc, and CH in the positive group and negative group. IOPg was 15.7 ± 3.3 (range: 10.2–24.3) mm Hg in the positive group and 16.2 ± 3.9 (8.1–29.8) mm Hg in the negative group, with no significant difference (P=0.42). IOPcc was 17.0 ± 2.7 (12.8–24.0) mm Hg in the positive group and 16.8 ± 3.2 (10.5–28.3) mm Hg in the negative group, also with no significant difference (P=0.65). On the other hand, CH was 9.6 ± 1.4 (6.8–13.3) mm Hg in the positive group and 10.2 ± 1.2 (6.9–13.3) mm Hg in the negative group, and was significantly lower in the positive group compared with the negative group (P=0.003). However, the ranges of the 2 groups overlapped almost completely.

FIGURE 7.

Boxplots of Goldmann-correlated intraocular pressure (IOPg), IOPcc, and CH in the glaucomatous change-positive (+) group and the glaucomatous change-negative (−) group. Open bars: (+) group: 53 eyes with glaucomatous changes in the fundus. Filled bars: (−) group: 139 eyes without glaucomatous changes in the fundus. Data below are expressed as mean ± SD (range). IOPg: (+) group; 15.7 ± 3.3 (10.2–24.3) mm Hg, (−) group; 16.2 ± 3.9 (8.1–29.8) mm Hg. IOPcc: (+) group; 17.0 ± 2.7 (12.8–24.0) mm Hg, (−) group; 16.8 ± 3.2 (10.5–28.3) mm Hg. CH: (+) group; 9.6 ± 1.4 (6.8–13.3) mm Hg, (−) group; 10.2 ± 1.2 (6.9–13.3) mm Hg. IOPg and IOPcc did not differ between the 2 groups, but CH was significantly lower in the (+) group than in the (−) group (IOPg: P=0.42, IOPcc: P=0.65, CH: P=0.003, Student t test). CH indicates corneal hysteresis; IOPcc, corneal-compensated IOP; IOPg, Goldmann applanation tonometry.

DISCUSSION

This report is the first to examine the reliability of measurements when ORA is used as a screening tonometer in the clinical setting, and compare the CH measurements between eyes with and those without glaucomatous fundus changes. In this study, when ORA was used as a screening tonometer, reliable measurements were obtained in ~80% of the eyes. To rule out the effects of treatments on ORA measurements, only eyes that had not undergone intraocular surgery or treated with glaucoma eye drops were included in the comparison of CH measurements between the positive and negative groups. This analysis showed that eyes in the positive group had lower CH measurements than in the negative group, but the distribution of CH measurements overlapped markedly between the 2 groups.

There are not many reports on the reliability of measurements obtained using ORA. Ayala et al¹² investigated 266 healthy subjects and reported a mean waveform score of 7.39 ± 1.32, with a range of 2.8–9.7. In another study of 64 healthy adults, Lam et al¹³ analyzed a total of 512 measurements obtained by measuring both eyes of the subjects 4 times, and they reported a mean waveform score of 5.49, ranging from 1.58 to 9.06. Since the ORA used in these 2 previous studies was model 2.04, the reliability of the more recent G3 model remains unclear. The present study evaluated ORA G3, the model currently used in clinical practice. By conducting measurements not only in normal eyes but also in eyes with different backgrounds in terms of the status of eye diseases and history of surgery, we obtained a mean waveform score of 7.3 ± 1.5 with a range of 0.1–9.7, which is comparable to or better than previous reports. Our results demonstrate that reliable measurements can be obtained even when ORA is used for IOP measurement as a screening in clinical practice.

Many of the studies to date that discussed the relationship between CH and the status of glaucoma were conducted on selected patients who had already been diagnosed with glaucoma. Abitbol et al¹ compared the CH of 58 eyes with glaucoma (open angle glaucoma 88%, angle closure glaucoma 12%) and 75 healthy eyes, and reported that CH was 8.77 ± 1.4 (range: 5.0–11.3) mm Hg in glaucomatous eyes and 10.46 ± 1.6 (7.1–14.9) mm Hg in healthy eyes, and was significantly lower in glaucomatous eyes. Hirneiß et al² compared CH in the glaucomatous eye and fellow eye of 18 patients with unilateral primary open angle glaucoma treated with eye drops. CH was 7.73 ± 1.46 mm Hg in glaucomatous eyes and 9.28 ± 1.42 mm Hg (range not stated) in fellow eyes, and was significantly lower in glaucomatous eyes. Furthermore, Kaushik et al³ compared the IOP and corneal features including CH between 323 eyes already diagnosed with glaucoma (comprising 101 eyes with the glaucoma-like disc, 38 eyes with ocular hypertension, 59 eyes with primary angle closure disease, 36 eyes with primary open angle glaucoma, and 18 eyes with normal tension glaucoma) and 71 normal eyes (all subjects received no surgery or eye drop treatment). Mean CH measurements for primary open angle glaucoma and normal tension glaucoma were 7.9 mm Hg (range not given, 95% CI, 6.9–8.8 mm Hg) and 8.0 mm Hg (95% CI, 7.2–8.8 mm Hg), respectively, and were significantly lower than that for normal eyes (9.5 mm Hg; 95% CI, 9.2–9.8 mm Hg). In the present study also, CH was lower in the glaucomatous change positive group compared with the negative group. From the data of previous studies, a close examination of the ranges and 95% CIs of the measured values shows that the glaucomatous eyes in most of the study subjects measured lower in CH compared with the normal eyes. On the contrary, the ranges of CH measurements in the present study overlapped remarkably between the positive group and the negative group. This may be attributed to the clinical background of the positive group. In this study, the presence or absence of visual field abnormality was not a criterion of subject selection. Hence, preperimetric glaucoma is predicted to be included in the positive group. In addition, since eyes with a history of intraocular surgery and eyes receiving eye drop treatment were excluded, many undetected or untreated glaucoma cases were probably included in the analysis. Because these factors contribute to the exclusion of late-stage cases and the inclusion of many early-stage cases, it is possible that the eyes in the positive group of this study had higher CH measurements than the eyes diagnosed with glaucoma in previous reports. This may imply that CH determination is insufficient to detect very early-stage glaucoma and preperimetric glaucoma.

The correlation between CH and various parameters measured by OCT has been reported by many previous studies.^7,16–19 In this study, we also examined the correlation between CH and OCT parameters in the positive group but detected no significant correlation (data not shown). A possible reason is that many of the glaucomatous fundus changes in this study were localized nerve fiber layer defects, and the variability in defect location probably resulted in little impact on the OCT parameters. Further study of a larger number of eyes is needed to clarify this issue.

Regarding the usefulness of CH measurement as a screening tool for glaucoma, Schweitzer et al¹¹ investigated 126 consecutive eyes presenting at a glaucoma subspecialty clinic for the first time and screened the eyes for severity of glaucoma. Eyes were classified into CH ≥10 and CH <10, and the severity of glaucoma assessed by 24-2 Humphrey automated perimeter was compared between the 2 groups. The analysis showed that the percentage of eyes with moderate or severe glaucoma was 2.9 times higher in eyes with CH <10 than in eyes with CH ≥10. Among eyes with CH <10, those with moderate or severe visual field impairment had significantly higher IOP than those with suspected or mild visual impairment. These findings suggest that measuring CH at the initial presentation may be useful to predict the severity of glaucoma.

The results of this study that investigated an unspecified number of undiagnosed cases are significant for the evaluation of CH as a screening for glaucomatous fundus changes. Low CH is known to be one of the risk factors for glaucoma. However, when ophthalmologists consider the detection of glaucoma in clinical practice, the conventional procedure is first to conduct various screening tests such as visual acuity, IOP, anterior segment slit lamp, and fundus examination; if glaucoma is suspected, appropriate tests are added to proceed with diagnosis. When ORA is used as a screening tonometer, a low CH measurement provides evidence for suspecting the presence of glaucoma. However, because the range of CH values in eyes with glaucomatous changes overlaps with that of normal eyes, CH alone is not sufficient and other test results are needed to comprehensively support a suspicion of glaucoma. In this regard, because the number of eyes in the positive group is higher than that in the negative group when CH is below 9, one may recommend that for patients with CH below 9 in particular, glaucomatous changes in fundus should be actively suspected and further workup should be conducted.

Since this study was a single-center, retrospective research, interpretation of results must take into account the effects of various biases. First, there are some limitations in the interpretation of fundus findings due to several features of this study. For example, when interpretable imaging data from other facilities are available, fundus photographs and OCT images are usually not acquired again at our hospital. Hence, we did not perform fundus photography and OCT in all the patients with suspected glaucoma. In addition, since image interpretation was performed by a single examiner, overlooking of some findings or bias in judgment cannot be ruled out. Furthermore, because stereoscopic observation of the optic nerve head was not performed in all the cases, there is a possibility of missing a very early increase in cupping that is difficult to determine by fundus photography and OCT. In this study, the accuracy of OCT measurements (signal strength index) was not considered. However, the OCT data were used only to distinguish between the presence and absence of glaucomatous changes, which does not depend on accuracy, and the impact is likely to be small. The reason why we did not use the OCT circumpapillary retinal nerve fiber layer thickness analysis in this study was that in order to exclude macular disease, the macular region was often imaged even though the papillary region was missing in many subjects. To exclude cases without peripapillary OCT images from analysis would greatly reduce the number of normal eyes in particular. Not including the results of circumpapillary retinal nerve fiber layer thickness analysis may possibly reduce the accuracy of diagnosis. Furthermore, more than 40% of the eyes that were interpreted could not be evaluated and were excluded from analysis, which may also affect the results.

Apart from the interpretation of fundus findings, other factors may also affect the results of the present study. The eyes analyzed in this study were limited to those with an ORA waveform score of 6 or higher and no history of intraocular surgery or glaucoma treatment. Moreover, the fact that the conditions of ORA measurement were not uniform may also impact the results. For example, there is no standard procedure for holding eyelids open for patients with severe eyelid closure, narrow palpebral aperture, or long eyelashes. In the case of eyelid opening, while using topical anesthesia to reduce the blinking reflex and the anticipatory eyelid closure and excluding cases needing help to keep eyelids open are expected to increase the accuracy of ORA measurement, using these methods would not reflect routine clinical practice and would overestimate the reliability of ORA when used as a screening tool in routine practice. Furthermore, because unilateral glaucoma is encountered in routine clinical practice, we included 2 cases of glaucoma, in which 1 eye was classified in the positive group and 1 eye in the negative group, for analysis in the present study. The inclusion of these cases may have some effect on the results. Lastly, in this study, 3 examiners performed the measurements, but the results for each examiner were not analyzed.

Notwithstanding the above limitations, including CH data in screening is expected to improve the accuracy of glaucoma detection. In the future, this may contribute to early detection of glaucoma and early identification of patients with risk factors for disease progression, with the result that severe disease may be prevented, which could have medical economic implications. Further validation in large-scale multicenter prospective studies including analysis of peripapillary retinal nerve fiber layer by OCT for evaluating glaucomatous fundus changes is warranted.