Conflict/competing interest: No stated conflict of interest.
Diaton tonometry: an assessment of validity and preference against Goldmann tonometry
Version of Record online: 2 SEP 2011
© 2011 The Authors. Clinical and Experimental Ophthalmology © 2011 Royal Australian and New Zealand College of Ophthalmologists
Clinical & Experimental Ophthalmology
Volume 40, Issue 4, pages e171–e175, May/June 2012
How to Cite
Doherty, M. D., Carrim, Z. I. and O'Neill, D. P. (2012), Diaton tonometry: an assessment of validity and preference against Goldmann tonometry. Clinical & Experimental Ophthalmology, 40: e171–e175. doi: 10.1111/j.1442-9071.2011.02636.x
Funding sources: No specific funding.
- Issue online: 14 JUN 2012
- Version of Record online: 2 SEP 2011
- Accepted manuscript online: 30 JUN 2011 01:30PM EST
- Received 7 February 2011; accepted 29 May 2011.
- Goldmann applanation tonometry;
- intraocular pressure
Background: To assess agreement between the Diaton, a new transpalpebral tonometer, and Goldmann applanation tonometry, the accepted gold standard.
Design: Comparative study of two devices in a hospital setting.
Participants: Two hundred and fifty-one patients attending the eye casualty and general ophthalmology clinics at St James' University Hospital, Leeds between February and December 2009.
Methods: Intraocular pressure was measured using Goldmann applanation tonometry and Diaton tonometry by one examining ophthalmologist. Patient preference for either technique was also recorded.
Main Outcome Measures: Intraocular pressure measured by Diaton was compared with intraocular pressure measured by Goldmann applanation tonometry. Limits of agreement were determined using the Bland-Altman method.
Results: Two hundred and fifty right eyes underwent both Goldmann applanation tonometry and Diaton tonometry. Mean intraocular pressure was 13.8 ± 3.6 mmHg using Goldmann applanation tonometry and 13.2 ± 4.3 mmHg using Diaton tonometry. Upper and lower limits of agreement were +8.4 mmHg and −9.6 mmHg, respectively. Order of intraocular pressure measurement and positioning did not influence limits of agreement in a clinically significant manner. Overall, more patients expressed preference for Diaton tonometry (40.2%) than Goldmann applanation tonometry (30.3%). Those aged 50 or less were more likely to prefer Diaton tonometry.
Conclusions: The Diaton tonometer is portable, lightweight, user-friendly and well tolerated by patients. However, it shows poor agreement with Goldmann applanation tonometry, thereby precluding it from being regarded as a substitute in routine clinical practice.
The Diaton tonometer (BICOM Inc, Long Beach, NY, USA), is a portable, handheld device which measures intraocular pressure (IOP) through the upper eyelid. It is an updated version of the TGDc-01 tonometer (Rjazan State Instrument Making, Rjazan, Russia). Unlike its predecessor, the Diaton incorporates a vertical alignment sensor and inbuilt software for the provision of an average IOP reading. It is promoted by its manufacturers as an effective alternative to Goldmann applanation tonometry (GAT) in situations where the latter is difficult or impossible to perform: in the presence of conjunctival or corneal disease, following corneal surgery, in immobile patients, in children and in contact lens-wearers. Portability, digital IOP display, measurement of IOP in more than one position and ease of use, have been cited as advantages of Diaton tonometry (DT) over GAT.
Because DT appears to be an attractive alternative to GAT and has been the subject of only one small comparative study,1 we felt the need to undertake a further study to validate its use. This study was accordingly undertaken to assess the agreement between DT and GAT among patients attending our ophthalmology department.
Between February and December 2009, patients attending the eye casualty and general ophthalmology clinics at St James' University Hospital, Leeds, were recruited into this study. A comprehensive information sheet was offered to all adult patients explaining the nature of the study. Ethical approval was obtained from the regional ethics committee. Those consenting to participation underwent further assessment. Inclusion criteria were: (i) absence of lid or adnexal disease precluding DT; (ii) absence of corneal disease precluding GAT; and (iii) ability to be examined using both GAT and DT.
All participants were examined by the same ophthalmologist who was familiar with DT and GAT. Prior to each session, both devices were calibrated in accordance with manufacturer's instructions. The same DT device was used throughout the study. GAT was undertaken using a standard slit-lamp arrangement, after instillation of 1 drop of a proxymetacaine/fluorescein preparation (Minims, Bausch & Lomb, Rochester, NY, USA) in the eye being examined. The rotating drum on the Goldmann tonometer was set to 10 mmHg prior to each measurement of IOP. The mires were always aligned without looking at the drum. A single measurement was taken.
Diaton tonometry was performed either in the sitting position with the patient's head thrown back horizontally on the head support; or lying, with the patient's head placed horizontally on the couch's cushion. The patient's gaze was directed at 45° to the horizontal in both positions such that the upper eyelid margin coincided with the superior limbus. The device was then held vertically above the upper lid and depressed gently to obtain a measurement of IOP. This process was performed up to six times until an audible bleep indicated a satisfactory average. The IOP measurement displayed on the device was recorded.
The order of IOP measurement with GAT or DT was determined using a random number generator. There was a variable interval between DT and GAT measurements. At the end of the examination, patients were asked if they had a subjective preference for either method of IOP assessment. Their responses were recorded.
The Bland-Altman method was used to assess 95% limits of agreement (mean difference ± 1.96 standard deviation) between DT and GAT. Subgroup statistical analysis was carried out to evaluate the impact of order of IOP assessment (GAT/DT first) and positioning for DT (lying/sitting) on agreement between the two methods. Validity of DT in detecting an IOP difference between two eyes of the same patient was evaluated using Pearson coefficient of correlation. All calculations were performed using MINITAB 15 (Minitab Inc, State College, PA, USA).
Over 11 sessions, 251 patients were recruited into this study. There were 134 (53%) females and 117 (47%) males with a mean age of 60.5 ± 18.7 years (range 17–93). DT and GAT readings were both available for 250 right eyes and 248 left eyes. Statistical analysis was carried out using data from right eyes. Mean IOP in this group was 13.8 ± 3.6 mmHg (range 6–30) using GAT and 13.2 ± 4.3 mmHg (range 4–27) using DT. Figure 1 illustrates data on the number of eyes with IOP in the ranges <16, 17–22 and >22 mmHg for each instrument. Using the Bland-Altman method, upper and lower limits of agreement between DT and GAT were +8.4 mmHg and −9.6 mmHg, respectively (Fig. 2). Results for left eyes were similar (Table 1).
|Subgroup||n||MD ± SD||Significance||Limits of agreement|
|Right eyes||250||−0.61 ± 4.60||+8.4||−9.6|
|Left eyes||248||−0.70 ± 4.30||+7.7||−9.1|
|DT then GAT||122||−0.48 ± 4.50||} P = 0.660||+8.4||−9.3|
|GAT then DT||128||−0.73 ± 4.70||+8.5||−10.0|
|Lying||63||0.49 ± 4.70||} P = 0.003||+9.7||−8.7|
|Sitting||65||−1.90 ± 4.40||+6.8||−10.6|
Diaton tonometry was undertaken before GAT in 122 (48.8%) right eyes. Whether DT or GAT was first used did not significantly influence the mean difference between readings (−0.48 vs. −0.73 mmHg, P = 0.66). The influence of positioning was analysed in 128 (51.2%) patients who underwent DT after GAT. Of these, 63 (49%) had DT in the lying position and 65 (51%) while sitting. Although mean difference was influenced by positioning in a statistically significant manner (−1.9 vs. +0.49 mmHg, P = 0.003), there was no clinically significant difference in limits of agreement. Table 1 summarizes the findings.
The DT and GAT readings were available for both eyes of 247 patients. Figure 3 shows a scatter plot of IOP differences between right and left eyes measured with DT and GAT. Mean difference between right and left eyes using GAT was 0.3 ± 2.8 mmHg. The corresponding value for DT was 0.34 ± 3.8 mmHg. Pearson's coefficient was 0.35 suggesting weak linear correlation between IOP differences in the same participant observed with GAT and DT.
Overall, 101 (40.2%) participants preferred DT, 74 (29.5%) preferred GAT and 76 (30.3%) were impartial to either method. There was no statistically significant association between gender and preference for either method (χ2 = 5.4, P = 0.068). However, when age was considered, patients aged 50 or less were more likely to prefer DT (χ2 = 18.6, P < 0.05). Table 2 summarizes these findings.
|Overall||101 (40.2%)||74 (29.5%)||76 (30.3%)|
|Males||46 (39.3%)||28 (23.9%)||43 (36.8%)||} P = 0.068|
|Females||55 (41.0%)||46 (34.3%)||33 (24.6%)|
|Age < 50 years||41 (62.1%)||15 (22.7%)||10 (15.2%)||} P = 0.000|
|Age > 50 years||60 (32.6%)||58 (31.5%)||66 (35.9%)|
Sole reliance on mean IOP values obtained with DT and GAT in this study may convey the misleading impression that DT is a reliable alternative to GAT. Bland and Altman have argued that an assessment of agreement between two devices measuring the same physical quantity should reflect the degree of scatter.2 Using their technique, the upper and lower limits of agreement between DT and GAT were found to be +8.4 mmHg and −9.6 mmHg, respectively. These limits are unacceptable in clinical practice and do not compare favourably with the repeatability coefficient of GAT which is generally accepted to be ±2 mmHg.3 Our findings are consistent with those of Toker et al. who compared GAT and DT in smaller study consisting of 162 eyes of 81 patients.1 Even when pachymetry was taken into account, upper and lower limits of agreement were +6.2 mmHg and −7.6 mmHg, respectively.
The Diaton's predecessor, the TGDc-01 tonometer, has been compared with GAT in several studies. Lösch et al. in a study of 218 eyes, found upper and lower limits of agreement of +4.4 mmHg and −11.8 mmHg between GAT and the TGDc-01.4 Sandner et al. in a similarly powered study, compared TGDc-01 readings with GAT within 5 min of each other and in random order.5 They reported a mean difference of 0.71 ± 2.5 mmHg and observed that 81% of TGDc-01 IOP readings were within 3 mmHg of GAT. Although they concluded that the TGDc-01 was reliable enough to be used as a screening tool, we would argue that their limits of agreement (upper: +5.2 mm, lower: −4.3 mmHg) suggest otherwise. Other smaller comparison studies, by Troost et al. and García Resúa et al. have also not shown encouraging results for the TGDc-01.6,7 It is interesting that, despite being considered an improved version of the TGDc-01, the Diaton does not appear to offer better agreement with GAT.
Several secondary findings from our study deserve comment. Although positioning for DT did not have a clinically significant effect on limits of agreement, we noticed a statistically significant variation in mean difference between GAT and DT readings (Table 1). In the lying position, DT readings were on average overestimates. The converse was true in the sitting position. To eliminate postural influences on GAT, we undertook this analysis in patients who had GAT before DT. Our observation suggests that, contrary to manufacturer's advice, positioning does influence DT. We are, however, unable to recommend an ideal position. As part of this study, we also assessed the reliability of DT at identifying IOP differences between both eyes of the same patient. To this effect, IOP differences measured with GAT were plotted against those measured with DT for each patient. Arguably, IOP differences can be as useful as absolute IOP measurements in some clinical situations. The bedside assessment of a patient with suspected unilateral ocular pathology associated with raised IOP is one such example. Figure 3 is a graphical representation of the scatter. The weak correlation between observed differences in IOP using the two instruments (r = 0.35) does not favour DT as a screening tool.
Patients in our cohort appeared to generally favour DT over GAT. We found no gender differences in preference for either technique. However, when age was considered, those aged less than 50 years showed a statistically significant preference for DT (Table 2). This trend is multifactorial. Younger patients are more sensitive to corneal contact and may therefore prefer the transpalpebral approach. A beeping electronic device, with a digital display, may also be more naturally appealing to this group. Likewise, the requirements of adequate positioning for DT may have been more physically challenging in comparison to GAT in the older group. No adverse events were encountered during the course of this study in relation to DT. The technique is well tolerated and safe.
Findings from this study are subject to limitations. Despite a large cohort, our comparison of GAT and DT is restricted to IOP readings up to 30 mmHg. We are therefore unable to comment on agreement between the two techniques over the entire spectrum of possible IOP measurements in clinical practice. Furthermore, although inter-observer variability is absent in this study design, the influence on agreement of the ‘learning curve’ associated with DT is not evaluated. It can also be argued that our reliance on a sole investigator does not control for observer bias. This was, however, minimized by ensuring that GAT measurements were only read after the mires had been aligned. Finally, it is reassuring that our findings concur with previous studies.
In summary, the Diaton is an updated version of the TGDc-01, with additional features to improve accuracy and ease of use. The device is portable, lightweight, user-friendly and well tolerated by patients. However, DT shows poor agreement with GAT, and therefore cannot substitute GAT in clinical practice. Its use as a screening tool requires further evaluation.
- 2Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 8476: 307–10., .