Clinical Comparison of the Proview Eye Pressure Monitor With the GoldmannApplanation Tonometer and the TonoPen

Objective  To compare intraocular pressure (IOP) values obtained by patients usingthe new Proview eye pressure monitor (Bausch & Lomb, Rochester, NY) withthose measured with the Goldmann tonometer and the TonoPen (Mentor, Norwell,Mass).

Methods  Eighty-six patients (a total of 171 eyes) with a diagnosis of glaucomaor glaucoma suspect successfully completed the study. The IOP was measuredby 3 methods in the following order: Goldmann tonometer, TonoPen, and Provieweye pressure monitor. The central corneal thickness was measured by an ultrasonicpachymeter. Separately for each eye, the differences in mean IOP values betweenmeasurement methods were assessed with paired t testsand also in multivariate models that tested the dependence of IOP differenceon central corneal thickness.

Results  There was a significant difference (P<.001)in the mean IOPs measured by the 3 different methods (Goldmann vs Proview,Goldmann vs TonoPen, and TonoPen vs Proview) for both eyes, and the differencewas independent of the central corneal thickness. The differences betweenIOP measured by Goldmann and Proview were similar in all categories of patient-reportedease of using the Proview.

Conclusions  The IOPs obtained with the Proview eye pressure monitor are significantlylower than those measured with Goldmann tonometer and the TonoPen, and variationsof the central corneal thickness do not contribute to the difference. Intraclasscorrelations of IOP values obtained with the Goldmann and the Proview or TonoPenand Proview are not strong. On the other hand, as expected, measurements withGoldmann and TonoPen agreed fairly well.

Elevated intraocular pressure (IOP) is one of the most important riskfactors for glaucoma, and it is currently the only treatable measure of thedisease. Therefore, it is imperative to have an accurate measurement of theIOP in the care of patients with glaucoma. Since its introduction in the 1950s,the Goldmann applanation tonometer has been the gold standard for measuringIOP. However, this instrument can be used only in a physician's office, wherea "snapshot" of the IOP is taken.

It is well known that there are circadian variations of the IOP, andthese fluctuations are more pronounced in patients with glaucoma.^1,2 Studies have demonstrated that IOPpeaks and, more importantly, IOP fluctuations are associated with progressionof visual field loss in patients with glaucoma even though their office IOPwas in the normal range.^3,4

Over the years, various portable tonometers have been developed; theseinclude the Perkins,⁵ TonoPen (Mentor, Norwell,Mass),⁶ Zeimer and coworkers' self-tonometer,⁷ Ocuton-S (EPSa Elektronik & Praezisionsbau, Saalfeld,Germany),⁸ and ProTon (Tomey, Erlangen, Germany).⁹ These instruments are costly and require a skilledoperator. The Schiøtz tonometer that was developed in the early 20thcentury is also portable and is inexpensive. However, patients have to bein the supine position, and there are multiple sources of error with thistechnique.

The Proview eye pressure monitor (Bausch & Lomb, Rochester, NY)uses a psychophysical test based on the entoptic phenomenon of pressure phosphenesto evaluate IOP, and it was designed for the patient to use at home. It isa pencillike device that has a small flat probe, an internal spring, and areadable pressure scale (Figure 1).The device measures the IOP through the eyelid, and it does not require ananesthetic. It is safe, noninvasive, portable, and affordable, and requiresno electric or battery source. If proven reliable, the Proview eye pressuremonitor could be an important instrument in the detection and management ofglaucoma.

A previous study demonstrated close agreement between the Goldmann tonometerand the pressure phosphene tonometer when the IOP measurements by both methodswere performed by clinicians.¹⁰ However, itis not known whether the results would be the same if patients themselvestook the pressure measurements with the Proview. Therefore, the current studywas designed to compare the IOP readings obtained by patients using the Proviewwith those obtained by physicians using the Goldmann and TonoPen instruments,and to assess the ease and discomfort associated with the use of Proview.

Patients with a diagnosis of glaucoma or glaucoma suspect were recruitedfrom the Glaucoma Service of Duke University Eye Center, Durham, NC. The studyprotocol was approved by the institutional review board of Duke University,and informed consent was obtained from all participants. Inclusion criteriawere age greater than 19 years, diagnosis of glaucoma (with mild visual fielddefect such as early nasal steps) or glaucoma suspect, IOP by Goldmann tonometrygreater than 8 mm Hg, and visual acuity of 20/80 or better. Exclusion criteriawere tremor, previous penetrating keratoplasty or abnormal cornea precludingaccurate measurement of IOP by Goldmann tonometry, or temporal visual fielddefects that prevented perception of the phosphene. The IOP was measured by3 methods in the following order: Goldmann tonometer, TonoPen, and Provieweye pressure monitor. Skilled clinicians performed Goldmann tonometry andTonoPen tonometry, and patients obtained measurements with the Proview eyepressure monitor. Two readings were made for each method and the average wasused in the analysis.

Goldmann tonometry was performed in a standard manner. For TonoPen tonometry,the instrument was calibrated each day before use, and readings of the highestreliability (standard error of the mean, ≤5% of the average) were obtained.

Patients were then asked to obtain IOP readings with the Proview device.They were given the diagrammed instruction sheet from the manufacturer, anda clinician went over the steps with them before the procedure. The righthand was used to hold the device and measure IOP in the left eye, and theleft hand to measure IOP in the right eye. In brief, the patient was instructedto keep the head straight and look down and to the side. While keeping theeyelid partially open, the patient gently pressed the Proview probe againstthe upper eyelid just below the edge of the eyebrow at the top of the nose(Figure 2). The pressure was slowlyincreased until a dark spot surrounded with a ring of light, a phosphene,was perceived. The device was then immediately removed and measurement onthe scale was recorded. A 2-minute interval was allowed between the differentmethods. The central corneal thickness (CCT) was measured by an ultrasoundpachymeter (Pachette 2; DGH Technology Inc, Exton, Pa). Six readings weretaken and averaged. In addition, the ease of using the Proview eye pressuremonitor and the discomfort associated with its use were assessed.

Separately for each eye, the differences in mean IOP values betweenmeasurement methods were assessed by paired t testsand also in multivariate models that tested the dependence of IOP differenceon CCT with the Wilks Λ test. Intraclass correlations of IOP valuesfor measurement methods were computed separately for the right and left eyes.An analysis of variance was used to determine whether there was a differenceamong categories of ease for the difference between the IOPs measured by theGoldmann and Proview instruments.

Ninety-one patients were recruited into the study and 86 successfullycompleted the study. Five subjects were unable to successfully perform IOPmeasurement with the Proview instrument because they either were unable tosee the phosphene or could not physically perform the procedure. Of the 86participants, 48 were female and 38 male. The age of the patients ranged from27 to 83 years, and visual acuity from 20/20 to 20/80. No discomfort was associatedwith the use of the Proview eye pressure monitor.

Ease of use of the Proview was assessed by the following grading system:1, very easy; 2, easy; 3, moderate; 4, difficult; and 5, very difficult. Seventypatients (81%) described using the Proview as "easy" or "very easy," whereas6 (7%) described its use as "difficult." The mean CCT was 549 µm (range,449-637 µm) in the right eye and 552 µm (range, 452-629 µm)in the left. In the right eye, the mean IOPs were 17.2 mm Hg (range, 9-44mm Hg), 15.6 mm Hg (10-39.5 mm Hg), and 13.8 mm Hg (8-27.5 mm Hg) by Goldmann,TonoPen, and Proview, respectively, and in the left eye, the values were 16.2mm Hg (8.5-25 mm Hg), 14.7 mm Hg (8-23.5 mm Hg), and 13.4 mm Hg (8-23 mm Hg),respectively.

There was a significant difference (P<.001)between methods for the mean IOP measurements in both eyes (Table 1). For Goldmann and Proview, the difference in IOP measurementswas similar across a wide spectrum of corneal thickness (Figure 3). The Wilks Λ test indicated that the differencewas independent of CCT (P = .70 for the right eyeand P = .26 for the left eye). In the right eye,only 30% of the Proview readings were within ±2 mm Hg of the Goldmannreadings and 51% within 3 mm Hg; these figures were 47% and 61%, respectively,for the left eye (Figure 4). Therelationship between the IOP values obtained with Goldmann and Proview areshown in Figure 5; for Goldmannand TonoPen in Figure 6; and forProview and TonoPen in Figure 7.The intraclass correlations between the IOP values obtained with Goldmannand Proview were 0.07 for the right eye and 0.23 for the left eye. For Goldmannand TonoPen, the coefficients were 0.79 and 0.78 for the right and left eyes,respectively. For TonoPen and Proview, the intraclass correlations were 0.20and 0.39 for the right and left eyes, respectively. The differences betweenIOP measured by Goldmann and Proview were similar in all categories of patient-reportedease of using the Proview, and there was no statistical significance for eithereye (Table 2).

Invented by Fresco,¹⁰ the Proview eyepressure monitor is based on an entoptic phenomenon or phosphene (Greek for"to show a light"), which is a sensation of light elicited by nonphotic stimuli.The likely basis of its operation was thought to follow the Imbert-Fick law:the perception of a phosphene occurs when the retina is deformed, which occurswith the application of a force over a given area, which can then be relatedto pressure.¹⁰ The site of a phosphene in theretina was suggested to be bipolar cells, or the parts of rod and cone cellslying inside the external limiting membrane.¹¹

We separately analyzed IOP measurements obtained with the Proview fromeach eye because measurements were obtained with different hands. Althoughthe IOP difference between the Goldmann and Proview was smaller for the lefteye (a mean of 2.78 mm Hg vs 3.40 mm Hg), it was statistically significantfor both eyes (P<.001 for both eyes).

It is not known whether the position of the eye on inferior temporalrotation when the Proview was used affects the IOP and therefore the measurementreadings. However, this is unlikely a significant factor. Axial length, refractivestatus, condition of the vitreous, and previous eye surgery may play a rolein the IOP measurements with the Proview, and these were not analyzed in thepresent study.

Two recent abstracts published by the Association for Research in Visionand Ophthalmology reported a mean difference in IOPs between the Proview andGoldmann instruments of 0.5 mm Hg (with a range from −5 to +16 mm Hg)¹² and 3.2 mm Hg.¹³ TheIOPs obtained with the Proview eye pressure monitor in our study were significantlylower than those measured with Goldmann tonometer and the TonoPen. It is aremote possibility that the order in which the IOP was taken contributed significantlyto the difference because at least 2 minutes was allowed to elapse betweenmeasurement methods. Recep et al¹⁴ noted thatthe time interval between successive IOP measurements should be 2 or 10 minutesfor accurate tonometry. The mean IOPs measured by TonoPen were lower thanthose by Goldmann tonometry. Although the TonoPen tonometry was always performedafter Goldmann tonometry, we do not believe that drying and therefore thinningof the cornea is a factor responsible for the IOP difference, because patientswere frequently reminded to blink their eyes during the measurements. It iswell known that the CCT affects the measurement of IOP by Goldmann tonometer^15-17 and, to a lesserdegree, by TonoPen,^18,19 althoughcontradictory reports exist in the literature.^20,21 Ameta-analysis of literature found that the mean corneal thickness of eyesreported to be normal was 544 µm by ultrasonic pachymetry.¹⁶ Themean CCT in our group of patients was 549 µm (right eye) and 552 µm(left eye). Race, age, sex, and a history of diabetes have all been reportedto influence corneal thickness,^15,17,22 andour group of patients was not homogeneous.

Theoretically, the IOP measurements with the Proview should not be influencedby corneal thickness, and this was found to be the case in a group of 22 patientsundergoing laser in situ keratoplasty.¹³ Therefore,we analyzed the dependence of IOP difference between the Goldmann and Proviewon corneal thickness, in an effort to determine whether the underestimationor overestimation of IOPs by Goldmann tonometry was based on differences inCCT. The CCT in our patients ranged from 449 to 637 µm, and these widevariations did not contribute to the IOP difference obtained with the 2 methods.Regardless of the CCT, a similar IOP difference was present. Scleral thicknessdifferences may play a role in the accuracy of the Proview, but its influencehas not yet been determined.

We also determined whether the difficulty with use of the Proview couldexplain the discrepancy in IOP readings compared with the Goldmann instrument,and we found no difference in all categories of patient-reported ease of usingthe device. Further analysis showed no strong correlation between the IOPsmeasured with the Goldmann and Proview, and no correction factor could bederived and used to accurately estimate the IOP with this new device. In fact,the Proview tonometer was found to act like a spring moving inside a cylinderwith friction, leading to nonlinearity and irreproducibility.²³ Furtherstudies are needed in larger samples to assess whether the Proview will allowus to gauge a range of IOP fluctuation in the same patient.

In summary, we found that the IOPs obtained with the Proview eye pressuremonitor were significantly lower than those measured with the Goldmann tonometerand the TonoPen. Furthermore, the intraclass correlation between the IOP valuesobtained with the Goldmann and the Proview instruments was low, indicatingvery little agreement between these 2 methods of measuring IOP. Therefore,these data suggest that the IOPs measured by the Proview eye pressure monitordo not correlate well with those obtained by Goldmann tonometer. Whether theIOP measurements by the Proview are reproducible for a specific individualpatient and, if so, whether the difference compared with Goldmann tonometryis constant over a range of IOPs in the same patient need to be determined.Nevertheless, our results demonstrated that the Proview eye pressure monitorfailed to measure the IOP accurately in our study sample of patients.

Correspondence: Leon W. Herndon, MD, Department of Ophthalmology,Duke Eye Center, PO Box 3802, Erwin Road, Durham, NC 27710 (hernd012@mc.duke.edu).

Submitted for publication June 16, 2003; final revision received December9, 2003; accepted January 9, 2004.

This study was supported in part by Clinical Vision Research DevelopmentAward EY 11725 from the National Eye Institute, National Institutes of Health,Bethesda, Md (Dr Stinnett).