Advice

Appendix

Appendix

Contents

Data tables

Table 2: Overview of the Cook et al. (2012) systematic review

Table 3: Overview of the Avitabile et al. (2010) study

Table 4: Overview of the Dahlmann‑Noor et al. (2013) study

Table 5: Overview of the Marini et al. (2011) study

Table 6: Overview of the Moreno‑Montanes et al. (2015) study

Table 7: Overview of the Rosentreter et al. (2013) study

Table 2 Overview of the Cook et al. (2012) systematic review

Study component

Description

Objectives/hypotheses

To assess the agreement of IOP measurements using tonometers available for clinical practice with GAT as the reference standard.

Study design

Systematic review and meta‑analysis.

Setting

n/a

Inclusion/exclusion criteria

Inclusion:

  • directly comparative studies that assessed the agreement of 1 or more tonometers compared with GAT in the same group of people

  • patients 16 years or older.

Exclusion:

  • tonometers that were not commercially available or judged unsuitable for monitoring ocular hypertension in routine clinical practice (including applanation resonance tonometer, ocular blood flow instrument, Schiotz, SmartLens, pneumatonometer and manometry)

  • studies published in a non‑English language and conference abstracts.

Primary outcomes

Agreement between a tonometer and the reference standard, reliability (inter‑ and intra‑observer variation) associated with measurements, and the proportion of participants with a recorded IOP measurement ('recordability').

Statistical methods

95% limit of agreement interval was calculated for each candidate tonometer from pooled estimates of the mean difference between a tonometer measurement and GAT.

Studies and total sample size included

A total of 102 studies involving 11,582 patients (15,525 eyes) were included in the review; 99 of these studies were included in a meta‑analysis.

In the 102 studies, 8 different tonometers were compared with GAT: 14 studies used Icare (model unknown), 32 studies used dynamic contour tonometers, 4 used handheld applanation tonometers, 26 used non‑contact tonometers, 3 used Ocuton S, 12 used ocular response analysers, 14 used Tono‑Pen, and 20 used transpalpebral tonometers.

Results

Non‑contact tonometers had the smallest mean difference compared with mean GAT value (0.2 mmHg) in contrast with Ocuton S, which had the largest difference (2.7 mmHg). Icare had a mean difference of 0.9 mmHg (95% CI 0.4 to 1.4) compared with GAT, ranking fourth out of 8 for accuracy.

Approximately 50% of all tonometer measurements were estimated to be within 2 mmHg of the GAT measurement, including 52% of all Icare measurements.

For all tonometers (except the non‑contact tonometer), a mean difference of greater than 2 mmHg compared with GAT fell within the 95% prediction interval.

The sensitivity analyses did not have a substantial effect on the results, nor did the subgroup analyses provide informative results. Non‑contact tonometer measurements and handheld applanation tonometers were the most similar to GAT, with 66% and 59% of IOP measurements within 2 mmHg of the GAT value respectively.

Recordability was reported in 4 Icare studies with a median of 100% (range 50–100%). The median study size for Icare was 145 people (range 36–150).

Conclusions

The non‑contact tonometer and handheld applanation tonometers achieve the closest values to those using GAT.

Abbreviations: CI, confidence interval; GAT, Goldmann applanation tonometry; IOP, intraocular pressure; n/a, not applicable.

Table 3 Overview of the Avitabile et al. (2010) study

Study component

Description

Objectives/hypotheses

To evaluate the effect of refractive errors and CCT on the measurement of IOP by Icare (model unknown), and its agreement with measurements using GAT.

Study design

Prospective cross‑sectional.

Setting

Italy, recruited from May 2007 to January 2008

No follow‑up period.

Inclusion/exclusion criteria

Inclusion:

  • healthy volunteers recruited among students, personnel, and relatives of patients referred to the Eye Clinic of the University of Catania.

Exclusion:

  • people unable to provide Icare or GAT measurements.

Primary outcomes

Repeatability and reproducibility of IOP measurements.

Statistical methods

Wilcoxon signed‑rank test was used to compare the IOP values measured by the 2 operators for both Icare and GAT. The repeatability coefficient was calculated for inter‑test differences. Bland–Altman analysis was used to assess the clinical agreement of IOP measurements between the 2 methods.

Patients included

327 total (89 male, 238 female): emmetropic (n=78); hyperopic (n=83); myopic (n=87); and astigmatic (n=79). Age range 18–85 years.

Results

No significant difference was found between the IOP values obtained by the 2 operators with Icare or GAT.

Bland–Altman analysis showed agreement between the 2 methods.

In each refraction group, IOP values detected by Icare were higher than those detected by GAT (paired t‑test p=0.000). The greatest mean difference was in myopic eyes (1.6±1.8 mmHg) with 95% limits of agreement from −1.90 to 5.16. The mean difference was less than 1 mmHg in emmetropic, hyperopic, and astigmatic eye groups. No significant difference in CCT was seen among the 4 groups. In all groups the IOP values correlated with CCT (p<0.05) but the discrepancy between Icare and GAT values correlated with refraction (p<0.001).

The difference between GAT and Icare was greater when Icare detected higher IOPs (p<0.001).

Conclusions

Icare gave higher IOP measurements than GAT, with myopic eyes showing the biggest difference. The measurements were reproducible in healthy volunteers.

Abbreviations: CCT, central corneal thickness; GAT, Goldmann applanation tonometry; IOP, intraocular pressure.

Table 4 Overview of the Dahlmann‑Noor et al. (2013) study

Study component

Description

Objectives/hypotheses

To evaluate the agreement of Icare (model unknown) and GAT in children with glaucoma.

Study design

Prospective cross‑sectional study.

Setting

England, recruited from January 2009 to March 2010.

No follow‑up period.

Inclusion/exclusion criteria

Inclusion:

  • patients attending paediatric glaucoma service at Moorfields Eye Hospital, London.

Exclusion: not stated.

Primary outcomes

Accuracy and reliability of Icare compared with GAT for measuring IOP, reproducibility when used by different observers, child preference for measurement method.

Statistical methods

Summary statistics and Bland–Altman limits of agreement method for intra‑ and inter‑observer agreement.

Patients included

102 children with glaucoma (53 male, 49 female); mean age 11.9±3.2 years (age range 4.9–19 years).

Results

Two different observers measured IOP using Icare (the first observer took 2 measurements while the second took 1 measurement) and a third observer measured IOP using GAT. The mean difference between the 2 Icare readings taken by observer 1 was 0.135±1.45 mmHg (p=0.427). The limits of agreement for intra‑observer readings were −2.71 to 2.98 mmHg. The mean difference between observer 1 and observer 2 Icare readings was 0.11±2.99 mmHg (p=0.8) with limits of agreement from −5.75 to 5.97 mmHg.

Icare frequently gave higher readings than GAT, with a mean difference of 3.3±5.31 mmHg (p<0.001). The degree of disagreement depended on the level of IOP being assessed, with smaller differences being seen with lower measurements (<21 mmHg).

GAT readings were missing for 12 children. For 45 children, only the first Icare reading by observer 1 and the Icare reading by observer 2 were available. Data were available for 74 children who had 2 readings by observer 1.

Icare was the preferred method for 70% of the children.

There was increased disagreement (larger discrepancies) in IOP measures between Icare and GAT with higher CCT values. The normal range is defined as 460–650 micrometres. The median pachymetry reading (n=67) was 581 micrometres and disagreements in children with readings of 581 micrometres could be greater than 10 mmHg.

Conclusions

The authors concluded there was poor agreement between Icare and GAT in children with glaucoma due to Icare overestimating IOP.

Abbreviations: GAT, Goldmann applanation tonometry; IOP, intraocular pressure.

Table 5 Overview of the Marini et al. (2011) study

Study component

Description

Objectives/hypotheses

To compare Icare (model unknown) and GAT measurements of IOP in people with glaucoma and ocular hypertension and evaluate CCT influence on Icare readings.

Study design

Prospective cross‑sectional study.

Setting

Spain, recruited from January to August 2009.

No follow‑up period.

Inclusion/exclusion criteria

Inclusion:

  • people with glaucoma or ocular hypertension.

Exclusion:

  • contact lens wearers

  • patients with previous corneal refractive surgery or corneal diseases associated with oedema or scarring

  • people with astigmatism greater than 2 dioptres.

Primary outcomes

Comparison of GAT and Icare, effect of CCT on IOP readings obtained with Icare.

Statistical methods

Descriptive statistics, Bland–Altman method and limits of agreement.

Patients included

347 eyes in 347 people (194 female, 153 male); mean age 62.3 years (range 43–82 years). n=89 with ocular hypertension, n=258 with glaucoma.

Results

Icare measurement of IOP was taken first (IC1), followed by a GAT measurement, and finally a second Icare measurement was taken (IC2).

Mean IOP measured with IC1 (18.1±4.3 mmHg) was significantly higher (p<0.001) than with GAT (15.6±3.3 mmHg). The mean IOP reading taken with IC2 (16.3±3.9 mmHg) was significantly lower than IC1 values (p<0.001), even though still significantly higher than for GAT (p=0.011).

Mean difference was 2.54±2.47 mmHg with 95% limits of agreement between –2.3 and 7.38 for IC1 compared with GAT. Mean difference for IC2 compared with GAT was 0.71±2.20 mmHg with 95% limits of agreement between –3.6 and 5.02. A significant linear correlation was identified between CCT and both IC1 and IC2, where a 4.6 and 4.1 mmHg increase in IOP was seen for each 100 micrometres increase in CCT respectively.

Conclusions

The authors concluded that when used first, Icare significantly overestimated IOP compared with GAT, however, differences decreased when Icare was used immediately after GAT. Agreement between the instruments was acceptable for low IOP values but worsened with increasing IOP. Icare was significantly influenced by CCT.

Abbreviations: CCT, central corneal thickness; GAT, Goldmann applanation tonometry; IOP, intraocular pressure.

Table 6 Overview of the Moreno‑Montanes et al. (2015) study

Study component

Description

Objectives/hypotheses

To compare patient‑obtained IOPs using Icare ONE (the predecessor to Icare HOME), clinician‑obtained values using Icare PRO and GAT values and analyse the ease of use of Icare ONE.

Study design

Prospective cross‑sectional study.

Setting

Spain, recruited from October 2011 to March 2012.

No follow‑up period.

Inclusion/exclusion criteria

Inclusion:

  • patients with IOPs in normal range and those with ocular hypertension or glaucoma

  • best‑corrected visual acuity of 10/20 or better

  • normal corneas.

Exclusion:

  • history of ocular trauma.

Primary outcomes

Accuracy of Icare ONE and Icare PRO compared with GAT, and ease of use as measured by patients.

Statistical methods

The differences between the measurements obtained using the 3 instruments were evaluated using the Wilcoxon test. Linear regression and the Bland–Altman plots were drawn to assess the agreement between the 3 methods and the presence of systemic bias.

Patients included

150 eyes of 150 people (mean age 57.0±18.1 years; range 15 to 89 years). n=60 without glaucoma, n=22 with ocular hypertension, n=68 with glaucoma.

Results

For all participants, the mean IOP values were 16.6±4.43 mmHg with GAT, 17.5±5.42 mmHg with Icare ONE (p=0.32 compared with GAT), and 16.6±4.77 mmHg with Icare PRO (p=0.75 compared with GAT).

The mean IOP values obtained for the first, second, and third measurement using the Icare ONE were 16.5±5.04, 16.7±4.95, and 16.6±4.77 mmHg respectively. The mean of these 3 measurements was used in the following comparisons.

The Bland–Altman analysis of IOP measurements showed that the mean difference between GAT and Icare ONE was −0.33±3.28 mmHg (limits of agreement −6.77 to 6.10); the mean difference between Icare PRO and Icare ONE was −0.33±3.51 mmHg (limits of agreement −7.14 to 6.62); and the mean difference between GAT and Icare PRO was −0.01±2.16 mmHg (limits of agreement −4.25 to 4.22).

Regression analysis indicated that GAT results were higher than Icare ONE for lower IOP values (p<0.001); in contrast, GAT results were lower than Icare ONE with higher IOP values. The regression line comparing Icare PRO and GAT showed a normal distribution of all values (p=0.43).

CCT was related to the IOP measurements using Icare ONE (p<0.05) and Icare PRO (p=0.01). However, the CCT was unrelated to the differences between Icare ONE and GAT (p=0.08) or between Icare PRO and GAT (p=0.06).

The ease of use of Icare ONE was classified as very easy by 37 patients (24.7%), easy by 79 patients (52.7%), complicated by 21 patients (14%), and very complicated by 13 patients (8.7%). Perceived difficulty in using Icare ONE was related to increasing age (p=0.003).

Conclusions

The Icare PRO measurements were more similar to GAT than the Icare ONE measurements.

Abbreviations: CCT, central corneal thickness; GAT, Goldmann applanation tonometry; IOP, intraocular pressure.

Table 7 Overview of the Rosentreter et al. (2013) study

Study component

Description

Objectives/hypotheses

To evaluate the agreement between IOP measurements obtained using an Icare tonometer (model unknown) and GAT or DCT in patients with corneal abnormalities.

Study design

Prospective cross‑sectional study.

Setting

Germany (recruitment dates not reported; no follow‑up period).

Inclusion/exclusion criteria

Inclusion:

  • patients with corneal abnormalities.

Exclusion: not reported.

Primary outcomes

Comparison of GAT and DCT with Icare, effect of CCT, corneal diameter, corneal radius, and axial length on IOP measurements.

Statistical methods

Descriptive analysis, Mann–Whitney test, Spearman correlation, Bland–Altman analysis and linear regression.

Patients included

99 patients (171 eyes with different corneal abnormalities and 26 healthy control eyes with normal corneal status).

Results

About 42% of Icare measurements were estimated to be within 2 mmHg of the GAT measurement and 23% were within 1 mmHg of the GAT measurement.

Icare was successfully used for 171 eyes with corneal abnormalities and all healthy eyes (n=26) resulting in a mean IOP measurement of 12.7±4.1 mmHg. GAT was successfully used for 168 eyes with corneal abnormalities (98%) and all healthy eyes resulting in a mean IOP measurement of 15.5±4.4.

The mean difference between Icare and GAT was –2.8 mmHg, with 95% limits of agreement of −10.5 to 4.9 mmHg. The mean difference between Icare and DCT was −3.8 mmHg, with 95% limits of agreement of −12.2 to 4.6 mmHg.

Icare and GAT readings were not significantly influenced by CCT, axial length, corneal diameter or corneal radius.

Conclusions

IOP was difficult to obtain using GAT and DCT in people with corneal abnormalities because of sutures interfering with the tip of the tonometers, corneal surface irregularities and corneal scars, whereas Icare was able to determine IOP in all eyes with corneal abnormalities. The authors concluded that while there is acceptable agreement between the 3 methods, Icare significantly underestimated IOP in all groups compared with GAT and DCT.

Abbreviations: CCT, central corneal thickness; DCT, dynamic contour tonometry; GAT, Goldmann applanation tonometry; IOP, intraocular pressure.