From DEPARTMENT OF CLINICAL NEUROSCIENCE Karolinska Institutet, Stockholm, Sweden
SELF-MONITORING OF INTRAOCULAR PRESSURE AND ITS CLINICAL
APPLICATION
Laurence Quérat
Stockholm 2022
All previously published papers were reproduced with permission from the publisher.
Published by Karolinska Institutet.
Printed by Universitetsservice US-AB, 2022
© Laurence Quérat, 2022 ISBN 978-91-8016-714-7
Self-monitoring of intraocular pressure and its clinical application
THESIS FOR DOCTORAL DEGREE (Ph.D.)
By
Laurence Quérat
The thesis will be defended in public at Eye Center of Excellence, St. Erik Eye Hospital, Aula vån -1, Eugeniavägen 12, Solna, Friday 7th October 2022, 09.00.
Principal Supervisor:
Docent Enping Chen, MD Karolinska Institutet
Department of Clinical Neuroscience Division of Eye and Vision
Co-supervisor(s):
Professor Maria Kugelberg, MD Karolinska Institutet
Department of Clinical Neuroscience Division of Eye and Vision
Opponent:
Docent Mika Harju Helsinki University
Department of Ophthalmology and Otorhinolaryngology
Division of Ophthalmology
Examination Board:
Docent Gauti Johannesson Umeå University
Department of Clinical Sciences Division of Ophthalmology
Docent Helene Hamberg Nyström Karolinska Institutet
Department of Clinical Neurosciences Division of Eye and Vision
Docent Ville Saarela Oulu University
Department of Ophthalmology
À André et Louise
POPULAR SCIENCE SUMMARY OF THE THESIS
Glaucoma, a disease that damages the optic nerve, is the second main cause of blindness in the world.
The causes of glaucoma are still unknown, but years of research have shown that lowering the eye pressure can slow down the progression of the disease. Traditionally, the first step to lower eye pressure consists of the instillation of eye drops. Sometimes, several agents can be combined. If the eye pressure level is not satisfying, the next step in treatment is laser therapy. The last option to lower the eye pressure consists of various surgical operations. Each method of lowering the eye pressure has side-effects, and it is important to prescribe treatment that is the most efficient and has as few side- effects as possible.
Eye pressure is usually measured by healthcare personnel at an eye clinic. However, access to an eye care provider can be a significant challenge due to a lack of resources. Despite continuous efforts, waiting lists for eye clinics are often long. The Covid-19 pandemic has also shown that restrictions to reduce contagion can limit access to eye specialists. Hence, the idea has been proposed of involving patients in their glaucoma care and having them measure their eye pressure themselves. The self- tonometer aimed to do so must be easy to use and safe, and must provide reliable measurements.
Although research has been conducted on this topic during the last decades, various problems have arisen. For one thing, devices based on the traditional method as that used at the eye clinic necessitate anesthetics. Moreover, because the device used by patients comes into contact with the numbed eye, it carries the risk of injury. Thus, this approach did not appear to be a suitable solution for home use.
Other devices that did not directly touch the eye were developed, but patients had problems holding the device steadily, resulting in unreliable measurements.
With the development of rebound tonometry in the late 1990s, new possibilities emerged. Rebound tonometry consists of a magnetized probe with a plastic tip that bounces off the eye. The probe is lightweight (less than one gram) and, because the tip only touches the eye for a brief moment, anesthetics are not needed. Rebound tonometry has been used with great success during the past 20 years in many eye clinics. Using the same principle, a new device called iCare Home has been developed for patients to use themselves.
The purpose of this thesis was to investigate self-tonometry with iCare tonometers—that is, measurement of eye pressure by the patients themselves. Our research showed that both healthy volunteers and glaucoma patients obtained reliable measurements comparable to those made by healthcare personnel. This means that patients can actually measure their eye pressure without visiting the eye clinic. Instead of sporadic single measurements, patients can monitor their eye pressure over longer periods of time, several times per day or even over many days, and thus provide further information for glaucoma management. We also confirmed that eye pressure varies during the day and even from day to day among glaucoma patients and healthy volunteers. This finding highlights the limited information available for eye specialists when eye pressure is measured only two or three times per year. Although such information can be sufficient for some patients, there are cases in which eye specialists need to know more about the patient’s eye pressure.
The reliability of iCare Home has been questioned and we wanted to investigate the variability of the measurements made with iCare tonometers and those made at the eye clinic with the traditional
method. It is very common in larger eye clinics for a patient’s eye pressure to be measured by different healthcare personnel from visit to visit, and it is known that the results can differ among examiners.
As the level of the eye pressure is used to evaluate the effect of glaucoma treatment, reliable measurements are very important. Part of the clinician’s decision will depend on the intraocular pressure (IOP) level. We found that the measurements made by patients themselves varied less than the measurements made by different healthcare personnel using the traditional method at an eye clinic. These results show that self-monitoring of eye pressure could become a reliable method to evaluate the effect of treatment.
Finally, we investigated whether self-tonometry influenced the clinicians’ decision regarding glaucoma treatment. By monitoring the eye pressure over several days, higher IOP levels than those measured at the eye clinic could sometimes be detected, and these levels motivated an intensification of treatment. Thus, we showed that the information collected by the patients played an important role in the clinicians’ decision.
Self-tonometry is a new method, and only a few eye specialists have started to use it. Nevertheless, it is very promising and, in the future, patients may measure their eye pressure at home at individualized intervals. The results from self-tonometry will be digitally transferred to the eye clinic, and glaucoma patients will need to visit their eye specialist only when specific attention is required.
POPULÄRVETENSKAPLIG SAMMANFATTNING
Glaukom, en sjukdom som skadar synnerven, är den andra huvudorsaken till blindhet i världen.
Orsakerna till glaukom är fortfarande okända, men forskning har visat att sänkning av ögontrycket kan bromsa utvecklingen av sjukdomen. Det första steget för att sänka ögontrycket består traditionellt i instillation av ögondroppar. Ibland kan flera komponenter behöva kombineras. Om ögontrycksnivån inte är tillfredsställande är nästa steg i behandlingen laserterapi. Det sista alternativet för att sänka ögontrycket består av olika kirurgiska operationer. Varje metod för att sänka ögontrycket har biverkningar och det är viktigt att ordinera den mest effektiva behandlingen med så få biverkningar som möjligt.
Ögontrycket mäts vanligtvis av sjukvårdspersonal på ögonkliniken. Tillgången till en ögonspecialist kan dock vara en betydande utmaning på grund av resursbrist. Trots effektivisering av vården är väntelistorna till ögonkliniker ofta långa. Covid-19-pandemin har också visat att restriktioner för att minska smittspridning kan begränsa tillgången till ögonspecialister.
Därför har idén föreslagits att låta patienter mäta sitt ögontryck själva. För att kunna göra det måste en självtonometer vara lätt att använda, säker, och ge tillförlitliga mätningar. Även om forskning på detta ämne har bedrivits under de senaste decennierna, har olika problem uppstått. Dels krävde mätningar baserade på samma metod som den som används på ögonkliniken bedövningsmedel.
Dessutom, eftersom apparaten som användes av patienter kom i kontakt med det bedövade ögat, medförde den risk för skador. Detta tillvägagångssätt verkade således inte vara en lämplig lösning för hemmabruk. Andra apparater som inte direkt rörde ögat utvecklades, men patienterna hade problem med att hålla apparaten stadigt, vilket resulterade i opålitliga mätningar.
I och med utvecklingen av rebound tonometri i slutet av 1990-talet uppstod nya möjligheter. Rebound tonometri består av en magnetiserad sond med en plastspets som studsar mot ögat. Sonden är mycket lätt (mindre än ett gram) och eftersom spetsen bara vidrör ögat en kort stund behöver patienterna inget bedövningsmedel. Rebound tonometri har använts med stor framgång under de senaste 20 åren på många ögonkliniker. Baserat på samma princip har en ny apparat, iCare Home, utvecklats för patienter att använda själva.
Syftet med denna avhandling var att undersöka självtonometri med iCare-tonometrar, det vill säga mätning av ögontrycket av patienterna själva. Vår forskning visade att både friska frivilliga och glaukompatienter fick tillförlitliga mätningar, jämförbara med de som gjordes av sjukvårdspersonalen.
Det innebär att patienter faktiskt kan mäta sitt ögontryck utan att besöka ögonkliniken. Istället för sporadiska enstaka mätningar kan patienter övervaka sitt ögontryck under längre tidsperioder, flera gånger per dag eller till och med under många dagar, och därmed ge mycket värdefull information till glaukom specialisten. Vi kunde också bekräfta att ögontrycket varierar under dagen och från dag till dag bland glaukompatienter och friska frivilliga. Detta belyser den begränsade information som finns tillgänglig för ögonspecialister när ögontrycket endast mäts två eller tre gånger per år. Även om sådan information kan vara tillräcklig för vissa patienter, finns det fall där ögonspecialister behöver veta mer om patientens ögontryck.
En annan aspekt som vi undersökte var variationen i de mätningar som gjordes med iCare-tonometrar och de som gjordes på ögonkliniken med den traditionella metoden. Det är mycket vanligt på större
ögonkliniker att en patients ögontryck mäts av olika sjukvårdspersonal från besök till besök och det är känt att resultaten kan skilja sig åt mellan två undersökare. Eftersom nivån på ögontrycket används för att utvärdera effekten av glaukombehandlingen är tillförlitliga mätningar mycket viktiga. En del av läkarens beslut beror på trycknivån. Vi fann att de mätningar som gjordes av patienterna själva varierade mindre än de mätningar som gjordes av olika vårdpersonal med den traditionella metoden på ögonkliniken. Dessa resultat visar att självtonometri skulle kunna vara en tillförlitlig metod för att utvärdera effekten av behandlingen.
Slutligen undersökte vi om självtonometri påverkade läkarnas beslut om glaukombehandling. Genom att övervaka ögontrycket under flera dagar kunde ibland högre trycknivåer än de som uppmätts på ögonkliniken upptäckas och dessa nivåer motiverade en intensifiering av behandlingen. Således visade vi att informationen som samlats in av patienterna spelade en viktig roll i läkarnas beslut.
Självtonometri är en ny metod och endast ett fåtal ögonspecialister har börjat använda den. Ändå är det mycket lovande och i framtiden kommer patienterna förhoppningsvis att mäta sitt ögontryck hemma med specifika tidsintervall. Resultaten från självtonometri kommer att föras över digitalt till ögonkliniken och glaukompatienter kommer kanske att besöka sin ögonspecialist huvudsakligen när särskild uppmärksamhet krävs.
ABSTRACT
Purpose Measuring the intraocular pressure (IOP) is one of the most common ophthalmologic examinations, especially within glaucoma management. An elevated IOP is a major risk factor for the development and progression of glaucoma, and is so far the only modifiable parameter. Treatment consists of lowering the IOP to slow down the disease, and the effect of treatment is evaluated by monitoring the IOP level and visual field status. Until recently, IOP measurements were performed by healthcare personnel at the eye clinic or by optometrists and opticians. A decade ago, a new tonometer was launched on the market, aimed at allowing patients to measure their own IOP. The purpose of this thesis was to evaluate the feasibility of self-tonometry by glaucoma patients and healthy volunteers, and the reliability of their measurements compared with the gold-standard method, Goldmann applanation tonometry (GAT). We also wanted to observe IOP variations and IOP patterns over consecutive days. Furthermore, we wanted to investigate the inter-user variability with both methods. Finally, we wanted to examine the impact of self- tonometry on clinicians’ decision-making regarding glaucoma treatment.
Methods In the research reported in Papers I, II, and III, glaucoma patients and healthy volunteers were trained to use a self-tonometer at the eye clinic. Measurements from the participants and trainers were recorded at the eye clinic. In Papers I and II, participants borrowed the self-tonometer over a few days to collect additional IOP measurements. In Paper III, additional measurements made by healthcare personnel at the clinic were collected as well. IOP values obtained by different users and different methods were analyzed, as well as IOP levels over consecutives days. In Paper IV, medical records from glaucoma patients who had performed self-tonometry were retrospectively reviewed. Different parameters were analyzed to evaluate which parameters had the most impact on the clinicians’ decisions.
Results We found that more than 85% of the participants were able to perform self-tonometry (Paper I, II, and III). Overall, approximately 70% of the measurements made with iCare self-tonometers were within 3 mmHg of GAT measurements. IOP means were similar between the different users and methods.
In Papers I and II, more than 60% of glaucoma patients had their highest IOP level in the morning. Between 9%–16% of glaucoma patients and healthy volunteers had an IOP peak outside office hours. There was good agreement between the methods of self-tonometry and GAT, although Bland-Altman analyses showed a bias with a cut-off at 18–20 mmHg in Paper I and 15 mmHg in Papers II and III. There was good repeatability of measurements, although we found a statistically significant difference in Paper III between the trainers’
values obtained with GAT and those obtained with a self-tonometer, as well as between the trainers’ and other healthcare personnel’s values obtained with GAT. The reliability between the users was excellent with iCare Home, at 0.903 (95% CI 0.880–0.959) and good with GAT, at 0.741 (95% CI 0.558–0.849). In Paper IV, we found a statistically significant difference between the clinicians’ decision to keep the existing treatment or escalate therapy based on maximum and mean IOP.
Conclusions Our studies showed that self-tonometry was feasible and that the measurements made by participants were reliable. Different IOP patterns from day to day and the presence of IOP peaks outside office hours support the idea of monitoring IOP over several days. We showed that tonometry with iCare Home used by the patients themselves had similar or less inter-user variation compared with GAT measured by healthcare professionals. Finally, we found that high IOP measurements collected during IOP phasing with self-tonometry could motivate additional treatment. Self-tonometry appeared to be a useful method to confirm that the present treatment was probably adequate when no deviating IOP values were observed.
Thus, self-tonometry provided clinicians with a solid basis on which to make their decisions and avoid a possible under- or over-treatment for the benefit of patients. Self-tonometry with iCare Home appears to be a valuable complement to traditional glaucoma care.
LIST OF SCIENTIFIC PAPERS
Paper I Chen E, Quérat L, Åkerstedt C. Self-tonometry as a complement in the investigation of glaucoma patients. Acta Ophthalmol. 2016 Dec;94(8):788-792. doi: 10.1111/aos.13129. Epub 2016 May 26. PMID: 27227556.
Paper II Quérat L, Chen E. Monitoring daily intraocular pressure fluctuations with self- tonometry in healthy subjects. Acta Ophthalmol. 2017 Aug;95(5):525-529. doi: 10.1111/aos.13389.
Epub 2017 Mar 14. PMID: 28296082.
Paper III Quérat L, Chen E. iCare Home vs. Goldmann applanation tonometry: Agreement of methods and comparison of inter-observer variation at a tertiary eye centre. Eu Ophthalmol. May 2022. doi:10.1177/11206721221099252.
Paper IV Quérat L, Chen E. Impact of self-tonometry on glaucoma treatment decision. Accepted for publication. Acta Ophthalmol. 2022 Sep.
LIST OF OTHER PUBLICATIONS
Tawfique K, Khademi P, Quérat L, Khadamy J, Chen E. Comparison between 90-degree and 360- degree selective laser trabeculoplasty (SLT): A 2-year follow-up. Acta Ophthalmol. 2019
Jun;97(4):427-429. doi: 10.1111/aos.13949. Epub 2018 Oct 15. PMID: 30318741.
Chen E., Samadi B., Quérat L. (2019) Patient management. In: Sun X., Dai Y. (eds) Medical Treatment of Glaucoma. Springer, Singapore. https://doi.org/10.1007/978-981-13-2733-9_9.
Chen E., Quérat L. Gynnsamt att låta patienten mäta ögontrycket själv. Läkartidningen. 2020,117:
FXUM. Lakartidningen.se 2020-01-23.
Quérat L, Chen E. Clinical use of the iCare Home tonometer. Acta Ophthalmol. 2020 Feb; 98(1):e131- e132. doi: 10.1111/aos.14169. Epub 2019 Aug 20. PMID: 31430031.
CONTENTS
1. Introduction ... 1
1.1 Glaucoma and intraocular pressure ... 1
1.2 Measuring the IOP ... 2
1.3 IOP fluctuations ... 4
1.4 Continuous IOP monitoring ... 5
1.5 Self-tonometry ... 6
2. Research Aims ... 8
3. Materials and Methods ... 9
3.1 Ethical considerations ... 9
3.2 Participants ... 10
3.3 Methods... 11
3.4 Statistical analysis ... 13
4. Results and Discussion ... 14
4.1 Feasibility of self-tonometry ... 14
4.2 IOP levels with GAT and iCare self-tonometers ... 14
4.3 IOP variations ... 14
4.4 Agreement between GAT and iCare self-tonometers ... 16
4.5 Inter-observer variability ... 18
4.6 CCT influence on IOP measurements ... 19
4.7 Influence of self-tonometry on clinicians’ decision ... 20
5. Conclusions ... 22
6. Points of Perspectives ... 23
Acknowledgments ... 24
References ... 26
LIST OF ABBREVIATIONS
ANOVA analysis of variance CC corneal curvature CCT central corneal thickness CH corneal hysteresis CI confidence interval
GAT Goldmann applanation tonometry IOL intraocular lens
IOP intraocular pressure LoA limits of agreement OHT ocular hypertension PEX pseudo-exfoliation
POAG primary open-angle glaucoma RoP rate of progression
RT rebound tonometry
SD standard deviation
SLT selective laser trabeculoplasty VFI visual field index
1
1 Introduction
1.1 Glaucoma and intraocular pressure
Glaucoma is a multifactorial ocular disease that causes progressive damage to the optic nerve and retinal nerve fibers and consequently leads to a progressive reduction of the visual field. A glaucoma diagnosis relies on an estimation of this damage—that is, an assessment of the visual field and an evaluation of the optic nerve head and retinal nerve fiber layer. Although the causes of glaucoma are still unknown, elevated intraocular pressure (IOP) has been identified as the most important risk factor for glaucoma and its progression [1, 2]. To date, IOP is the only modifiable risk factor in glaucoma care, and research has shown that lowering the IOP can reduce the risk of glaucoma progression [1-6]. Therefore, glaucoma patients are treated to lower the IOP in order to prevent further damage.
The IOP refers to the pressure inside the eyeball, which is caused by accumulation of the aqueous humor, a liquid that circulates in the anterior part of the eye (Fig. 1). The aqueous humor is produced in the posterior chamber of the eye by the ciliary body, a structure behind the iris, and flows to the anterior chamber through the pupil. Most of the aqueous humor is drained out of the anterior chamber through the trabecular meshwork and further on through the Schlemm’s canal and the episcleral veins. A minor quantity of aqueous humor is drained by the uveoscleral pathway. The IOP is a product of the balance between the production of the aqueous humor and its drainage. If the drainage of the aqueous humor is lower than its production, the IOP will rise, and vice versa.
Figure 1. Pathway of the aqueous humor from the ciliary body through the pupil to the episcleral vein.
(Printed with permission from the illustrator, J. Alwert and E. Tov)
2 The goal of glaucoma management is to preserve quality of life and quality of vision by slowing down the progression of visual field damage. This is achieved by reducing the IOP to a level (known as the target IOP) below which no further progression is expected. The target IOP is a subjective estimation made by the clinician that varies from patient to patient. It must be adjusted regularly during the care process and is achieved with medication, laser treatment, and surgery.
1.2 Measuring the IOP
As the IOP plays an important role in the management of glaucoma, correct IOP measurement is crucial. However, as in practice the true IOP is unknown, a clinical measurement is an estimation of the IOP level. A clinically useful tonometer should be easy to use and should give a result that is close to the true IOP. This measurement should be accurate and repeatable.
Several methods are available to measure the IOP, the most common of which is Goldmann
applanation tonometry (GAT). GAT is the gold-standard method against which all other instruments are compared for validation. GAT involves a slit-lamp mounted contact tonometer (Fig. 2) and requires the cornea to be anesthetized and colored with staining fluorescein. An area of 3.06 mm in diameter is flattened with a probe that contains a double prism. Illuminated by a cobalt blue filter, two half circles formed by the double prism can be observed (Fig. 2). By adjusting the pressure applied to the cornea, a value in grams is obtained on the dial of the tonometer, corresponding to the force needed to flatten this area. This value is then converted to mmHg.
Figure 2. Goldmann applanation tonometry. (Photos 1 and 2 by E. Tov; photo 3 by O. Hagelberg) GAT is based on the Imbert-Fick law: P = 𝐹
𝐴 . The pressure (P) inside a sphere made of an extremely thin, elastic, and dry membrane is equal to the force (F) necessary to flatten a given area (A).
However, this principle does not fully apply, given the properties of the human eye [7, 8]. First, the eyeball is not spherical, as the corneal curvature (CC) is steeper than the curvature of the rest of the eyeball. Second, the eyeball is not a dry surface, as it is covered by a tear film. The surface tension created by the tear film pulls the probe toward the eyeball, influencing the measuring force. Third, the cornea is not infinitely thin, as required by the Imbert-Fick law. GAT is validated for a central corneal thickness (CCT) of 520 µm, which evens out the surface tension of the tear film. Thus, the IOP might be underestimated or overestimated in cases with thinner or thicker corneas. The influence of corneal properties such as CCT and CC have been described in several studies [7, 9, 10]. The possible
3 influence of biomechanical properties such as corneal hysteresis (CH) has also been investigated [11, 12].
It has been suggested that applanation tonometry can induce a lowering of the IOP when measurements are repeated, due to both the mechanical applanation itself and the use of
anesthetics [13-15]. Studies have also shown that IOP results may vary even when the same method is used for the same patient on the same occasion by the same operator (intra-observer variation) [8, 16]. Similar results have been obtained when comparing two or more operators (inter-observer variation) [7, 17-19] .
Aside from the influence of corneal properties and different observers, patients’ behavior may affect the IOP [20]. Apprehension during the examination can cause breath holding or lid squeezing. More specific ocular features can affect the IOP, such as the position of the gaze or accommodation [7].
Physical features are also relevant, as the IOP is higher when a patient is lying down rather than sitting, or even depending on how the patient is positioned in front of the tonometer [7]. All these factors must be considered when evaluating the results of IOP measurements with GAT.
Figure 3. Rebound tonometry using the iCare IC100 and disposable metal probes. (Photos by E. Tov)
Rebound tonometry (RT) is another contact tonometry method, but it does not require any anesthetics. All rebound tonometers by iCare are handheld and used with patients in a sitting position. A few models (Pro, IC200, and Home2) can be used with patients lying down. A disposable metal probe, which is 40 mm long and fitted with a 1.8-mm-thick plastic tip, is projected toward the cornea and bounces off from it (Fig. 3). The IOP is calculated from the speed of deceleration after the probe comes into contact with the cornea. The average of four measurements is recorded, after the highest and lowest values have been eliminated [21]. Some devices display a reliability indicator, whereas others request new measurements if there is too much variation, indicating poor quality. RT
4 has been used by healthcare professionals since 2003, and clinical studies have demonstrated high agreement between GAT and iCare results [10, 22-25]. RT measurements may be affected by corneal thickness and biomechanical properties [22, 26-30] . Compared with Perkins, a handheld GAT, RT has shown variable agreement [31, 32]. Since 2013, a new device that patients can use themselves has been available: the iCare Home device [33-37].
One difficulty with iCare tonometers is aligning the probe with the apex of the cornea. Muttuvelu et al. found that peripheral and angular measurements significantly underestimated the IOP values, in comparison with central measurements with up to 4 mmHg when using iCare Pro [38]. Beasley et al.
found that angular deviations of 5° and 10° had no statistical significance, except for 10° nasally;
however, they concluded that the latter should have no clinical significance [39].
1.3 IOP fluctuations
To evaluate the effect of glaucoma treatment at follow-up visits, ophthalmologists use contact or, sometimes, non-contact tonometry, usually limiting IOP measurements to office hours. However, IOP varies over the course of the day in both healthy subjects and glaucoma patients [40-42] . Differences between the highest and lowest IOP values of up to 7 mmHg in healthy subjects and 8 mmHg in glaucoma patients have been reported [42-44] . A single IOP measurement may miss IOP peaks and may lead to an erroneous clinical decision [40, 45]. Indeed, Jonas et al. found that there is a 75% risk of missing an IOP peak when the IOP is measured only once during the daytime (7 am to 9 pm), and Hughes et al. recorded more than 50% of IOP peaks as occurring outside office hours. Therefore, an estimation of the IOP must take into account the variability of the IOP—that is, the diurnal
fluctuations of the IOP—as well as the limitations of the method being used [40, 45].
Several studies have investigated whether or not IOP fluctuations are a risk factor for the progression of visual field damage. The influence of IOP fluctuations on glaucoma progression is controversial, which could be due to different definitions of IOP fluctuations. In the literature, long-term
fluctuations are often defined as the difference between the highest and lowest IOP values (range) measured at occasional routine visits or as the standard deviation (SD) of single IOP measurements over a period of time, such as several weeks, months, or years apart. According to these definitions, long-term fluctuations are based on single daytime measurements obtained during office hours.
Thus, such fluctuations do not reflect IOP variations occurring early in the morning or later during the day, and therefore possibly miss IOP peaks. Two studies have defined long-term fluctuations as the mean of the range of daytime IOP curves obtained every 2 or 3 months over a few years [47, 48].
Regardless of the definition used, results on the influence of IOP fluctuations on glaucoma progression remain contradictory. On the one hand, some studies have found a significant
association between an increased SD and the rate of visual field progression [49-52]. On the other hand, some investigations were unable to find any association [2, 53-56]. Long-term fluctuations based on SD might be a way to evaluate the risk of glaucoma progression when IOP curves are not available. However, this retrospective method requires time and is not appropriate for evaluating the immediate effect of a specific treatment on IOP level or fluctuation.
Short-term fluctuations are commonly defined as the difference between the highest and the lowest IOP values measured during office hours over 1 day (diurnal fluctuations) or during 24-hour
hospitalization (circadian rhythm). Once again, results differ and come to opposite conclusions.
5 Asrani et al., Baskaran et al., Kim et al., and Baek et al., for example, found a correlation between short-term IOP fluctuations and visual field progression, whereas Jonas et al., Fogagnolo et al., and Wang et al. did not [42, 53, 57-61]. All these studies provide important information on IOP
fluctuations. They also question the certainty of single IOP measurements to evaluate the effect of glaucoma treatment.
Despite contradictory conclusions in the literature, most ophthalmologists agree on the usefulness of IOP curves. In-patient measurements are expensive and time consuming for both healthcare systems and patients, which explains why IOP curves are rarely obtained in routine care nowadays. As IOP peaks can occur outside office hours [41, 62, 63], accurate devices that allow patients to measure IOP themselves at home on a 24-hour basis or over a few days could help in clinicians’ decision-making regarding the most suitable glaucoma treatment. Furthermore, as patients must be confined within clinical settings in order to clinically obtain a 24-hour IOP curve, patient inactivity may influence the results by not reflecting IOP variations during the patients’ daily activities. Another aspect of IOP fluctuations that is still relatively unknown is whether the IOP pattern observed over 24 hours is the same in both eyes and on consecutive days [53]. Even though the influence of IOP fluctuations on glaucoma is still controversial, research is needed to develop methods to permit continuous monitoring of the IOP outside healthcare premises.
1.4 Continuous IOP monitoring
Continuous monitoring offers the possibility to obtain IOP measurements throughout the day and while patients are asleep—that is, without disturbing the circadian rhythm of the IOP. Such methods are based on contact lens sensors (extraocular sensors) or implantable sensors (intraocular sensors).
Extraocular sensors consist of contact lenses that are embedded with a sensor that can capture the physiological changes of the cornea. Nano-techniques allow the development of new materials and devices to monitor the IOP [64]. For example, Triggerfish (SENSIMED AG, Lausanne, Switzerland) is a contact lens sensor that patients wear for 24 hours. Changes in the curvature and circumference of the cornea are recorded during a 30-second period every 5 minutes by an external recorder. The measurements are reported in millivolt equivalents (mVeq). After 24 hours, a curve of these changes is obtained, which is assumed to reflect the IOP fluctuations. The sensor allows patients to carry on with their normal daily activities [65]. However, this method has some limitations. Only one eye can be measured at a time, and the involvement of a care provider is required to fit in and remove the contact lens. Patients experience blurred vision and significant corneal edema; conjunctival hyperemia has also been observed after 24 hours [66, 67]. Furthermore, the correlation between mVeq and mmHg is unclear [68].
To monitor IOP over longer periods of time, research is currently developing micro-
electromechanical systems (MEMS) that can be implanted inside the eye [64]. The sensors are surrounded by the aqueous humor, and any change in volume will affect the IOP. The sensor is placed in the ciliary sulcus in the anterior chamber, or in the capsular bag. Implants can be placed in the eye together with an intraocular lens (IOL) during cataract surgery. Measurements are usually recorded with an external reader. Clinical studies have shown good results, although further research must be conducted to evaluate the safety and long-term effects of these devices [69, 70]. Other types of sensors are tested on animals to observe the accuracy of the measurements and the long-
6 term tolerance and device safety of permanent implants. Although these new techniques could make continuous IOP monitoring feasible, such methods are invasive and expensive.
1.5 Self-tonometry
Involving patients in their own healthcare is already feasible for a few diseases. Patients suffering from diabetes mellitus have been monitoring their personal blood sugar level for many years with the help of specially designed devices. In the same way, patients with high blood pressure (BP) can monitor their BP themselves. The devices patients can use to monitor blood sugar and BP enable the closer follow-up of patients, a more adapted treatment, and (hopefully) better compliance with medical treatment [71-74].
Similarly, glaucoma management requires new devices to allow patients to measure their own IOP outside of the eye clinic. This would provide information about patients’ IOP fluctuations over time under everyday life conditions. Such devices should be easy for patients to use and should yield repeatable measurements. Several devices have been developed over the last decades for this purpose [75]. Ocuton S is a self-tonometer developed in the 1990s; it is based on applanation tonometry and requires anesthetics. Studies show that, despite training, only half of patients can obtain valid measurements; moreover, these measurements differed significantly from those obtained with GAT [76, 77]. A Proview pressure phosphene tonometer is a portable instrument for self-tonometry that calculates the force needed for a patient to visualize a halo when applying pressure to the eyeball over the eyelid. Although it does not require anesthetics and is easy to use, the measurements obtained do not correlate with GAT measurements [78-80].
Figure 4. iCare Home. (Photos by E. Tov)
iCare self-tonometers are based on RT and have been developed for patients to use themselves. Both iCare One (the first version) and iCare Home (the improved version; see Fig. 4) have been evaluated and show good agreement with GAT [12, 33-37, 81-86]. A major difficulty with iCare self- tonometry
7 is the need for the patient to position the tonometer correctly and make a measurement at the apex of the cornea. A green light circle around the probe helps the patient to align the tonometer with the optical axis, and a green “DONE” light at the back of the tonometer indicates a valid measurement (Fig. 4). Studies have demonstrated that a deviation from the apex has statistical but not clinical significance for the IOP measurements [39, 87, 88].
In today’s practice, IOP curves are rarely obtained, mainly because applanation tonometry implies that patients must visit the eye care center several times during the day to monitor their IOP with GAT at different time points. These IOP curves are limited to IOP fluctuations during office hours over a single day. Self-tonometry with iCare Home tonometers offers the possibility of monitoring the IOP outside office hours and over several days (Fig. 5). This method appears to be a valuable and
accessible way of investigating IOP fluctuation [82, 89, 90] and could open up new ways of following up with glaucoma patients. Glaucoma management could start by obtaining an IOP curve to define the patient’s baseline IOP and evaluate IOP fluctuations and patterns. Clinicians would then have a more reliable basis to choose the appropriate treatment. Reducing the risk of under- and over- treatment is important [91], and self-tonometry could be used to achieve this goal.
Figure 5. Example of IOP phasing of a patient obtained using iCare Home; notably, this patient exhibited
“normal” IOP during office hours and IOP peaks outside office hours.
8
2 RESEARCH AIMS
Paper I To validate IOP self-monitoring by glaucoma patients using iCare self-tonometers Paper II To observe the pattern of diurnal variation of IOP in healthy subjects measured with
self-tonometers
Paper III To estimate the variability of measurements made by different users with GAT and iCare Home
Paper IV To evaluate the impact of self-tonometry on glaucoma treatment decisions by clinicians
9
3 MATERIALS AND METHODS
3.1 Ethical considerations
As patients at our clinic, the participants in Papers I and III had a certain dependency on the eye specialist who recruited them. At all study visits, the research coordinator, who was not involved in the clinical procedures, offered patients the opportunity to discuss any matters more openly. This left room for the patient to receive further information and the possibility to discontinue
participation if the patient so wished. The eye specialist and the study coordinator were available throughout the projects and could easily be contacted by the study patients to discuss any matters related to their participation in the study.
Participation in the projects took place on a free basis and the patients received no compensation.
They knew that their participation would contribute to important information on a new method to measure IOP and that it would hopefully bring better care for glaucoma patients in the near future.
Papers I and II
The Ethical Review Board of Stockholm, Sweden, determined that these prospective non-randomized studies were quality assurance studies and considered no ethical approval to be needed. In regard to the information given to the patients, the Board had no ethical objections to the studies. All
participants gave their written consent after being provided with verbal and written information about the study.
Paper III
Approval was obtained from the Ethical Review Board of Stockholm, Sweden, for this prospective non-randomized study, and all participants gave their written consent after being provided with verbal and written information about the study.
Paper IV
Approval was obtained from the Ethical Review Board of Stockholm, Sweden, for this retrospective study. Access to the patients’ data was strictly limited to the researchers involved. According to the Board, no patient consent was needed.
10
3.2 Participants
Paper I
Consecutive patients diagnosed with primary open-angle glaucoma (POAG) or ocular hypertension (OHT) were enrolled in the study. Of the 87 enrolled patients, six were excluded from the analysis of tonometer accuracy due to early withdrawal (4) or problems in handling the iCare tonometer (2). The mean age was 64 ±13 years, and the female/male ratio was 32/49. Ten patients were one-eyed.
Fourteen patients were excluded from the analysis of IOP variations due to non-compliance with the schedule.
Paper II
Participants were recruited among hospital personnel, volunteers, and relatives of glaucoma patients. Sixty subjects were recruited to the study. The mean age was 48 ± 15 years (range 24–80 years), and the female/male ratio was 36/24. Four subjects dropped out after the first visit, and two patients failed to obtain valid measurements. Eight (13%) subjects were excluded due to missing more than one daily measurement, non-compliance with the time schedule, or obtaining several erroneous values (0 mmHg) over 3 days.
Paper III
Consecutive glaucoma patients diagnosed with POAG, pseudo-exfoliation (PEX) glaucoma, or OHT who were referred to perform an IOP phasing with iCare Home between April and September 2020, as well as volunteers with no known eye disease, were invited to participate in this study. The majority of the 61 participants were first-time users, but seven had tested iCare Home a few years previously. Five participants were excluded due to a difference between the trainer and participant measurements of over 5 mmHg and two participants were excluded due to an iCare Home range of over 7 mmHg at the same session. The mean age was 56 ± 17 years; the female/male ratio was 32/22, and the ratio of glaucoma patients to volunteers was 24/30.
Paper IV
In this paper, the medical files of 133 patients who performed IOP phasing using an iCare Home self- tonometer between January and December 2019 were retrospectively reviewed. Forty-three patients were excluded, some due to concomitant reasons. The reasons for exclusion were: missing data (8), no follow-up due to external referral (11), IOP phasing for less than 2 days (7), missing one or more daily measurements (8), problems handling the device (1), a difference between GAT and iCare Home greater than 7 mmHg (8), a difference between iCare Home measurements by the patient and healthcare personnel greater than 5 mmHg (2), and patients’ iCare Home measurements being lower than 5 mmHg (12). Patients were diagnosed with POAG (55), PEX glaucoma (22), OHT (1),
11 uveitic glaucoma (7), glaucoma suspect (2), and angle-closure glaucoma (3). At the referral visit, seven patients were without treatment, 69 instilled eye drops only, eight had received selective laser trabeculoplasty (SLT) and were on eye drops, four had received surgery and were on eye drops, and two had received surgery with no other treatment. The mean age was 72 ± 12 years, and the female/male ratio was 40/50.
3.3 Methods
Protocol for the iCare self-tonometer
All study participants (i.e., glaucoma patients and volunteers) using iCare self-tonometers were trained by a qualified ophthalmic nurse or an optometrist. During the training session, the trainer measured the participant’s IOP with the self-tonometer. Afterward, the participants practiced until they obtained acceptable measurements, which were recorded. The training session ended with the nurse or optometrist measuring the participant’s IOP with GAT (a single measurement). The iCare measurements were blinded to both the trainer and the participant, as no values are displayed on the device. All IOP measurements were made in this sequence. The self-tonometry results were downloaded at the eye clinic by connecting the iCare tonometer to the iCare Link software, recording the date, time of day, eye, and IOP value for each measurement.
Paper I
Each patient visited the clinic twice. During the initial visit, the patients learned how to use the iCare tonometer (see protocol above). The iCare One and iCare Home devices apply the same IOP
measurement principle. iCare One is the first version of the handheld tonometer and cannot detect which eye is being measured. The patients used iCare One on the right eye for 3 days and then on the left eye for the following 3 days. As an improved version, iCare Home can identify and record which eye is being measured. With this device, IOP was measured on both eyes at the same time points on three consecutive days. The iCare One device indicates a valid IOP measurement on a scale ranging from 5 to 50, whereas iCare Home indicates only whether a valid reading has been obtained, by illuminating a green indicator light.
Between the two visits to the clinic, the patients measured their IOP at home on three or six
consecutive days according to the following schedule: 4 am, 8 am, 12 pm, 4 pm, and 8 pm. They were instructed to take their medication at the prescribed times. The IOP pattern was defined as the highest IOP value in the morning (i.e., at 4 am, 8 am, or 12 pm) and in the afternoon/evening (i.e., at 4 pm or 8 pm). The results gave an ascending, descending, or flat pattern. An outside-office-hour peak was defined as a value measured outside office hours that was 4 mmHg higher than the highest value measured during office hours.
Paper II
All participants were examined for the following on visit 1: best-corrected visual acuity, CCT, and slit- lamp bio-microscopy. During this first visit, the participants learned how to use the iCare Home
12 tonometer (see protocol above). The participants then borrowed the tonometer to measure their IOP for three consecutive days according to the following schedule: 6 am, and 10 am, 2 pm, 6 pm, and 10 pm. On visit 2, the participants measured their IOP with the iCare Home tonometer, and the optometrist measured the IOP with the same device and GAT. An IOP peak was defined as a
difference of 4 mmHg or more between the highest IOP value recording outside office hours and the highest IOP value during office hours.
Daily patterns comprised a comparison between the highest IOP measurement made in the morning and the highest in the afternoon/evening. Pattern 1 described a pattern in which the highest IOP was measured at 6 or 10 am. Pattern 2 described a pattern in which the highest IOP was measured in the afternoon/evening (i.e., at 2, 6, or 10 pm).
Paper III
The participants were first introduced to the self-tonometer, and a single measurement with iCare Home (measurement 1) was obtained by one of three eligible training personnel. Following a training session, the participants measured their IOP with the iCare Home tonometer to obtain three valid measurements (measurements 2–4). Afterward, the training personnel measured the participants’
IOP once with GAT (measurement 5). On a routine basis, 4–10 personnel (physicians,
ophthalmological nurses, or optometrists) worked at the clinic. Within 15 min from the first GAT measurement, one of the available personnel, who was blinded to the previous results, was
randomly asked to measure the IOP once with a different GAT instrument (measurement 6). All IOP measurements were made in this sequence. For each participant, a total of six measurements were obtained on a single visit. Finally, the iCare Home results were downloaded by connecting the tonometer to the iCare Link software.
One eye was randomly selected for each participant, by alternating between right eye for patient 001, left eye for patient 002, and so on. If only one eye was available, it was eligible without randomization. At least two iCare Home measurements obtained by the participants were needed.
Participants who could not perform self-tonometry due to reduced hand and arm mobility (e.g., due to rheumatism or tremor) were excluded. According to the manufacturer’s recommendations, participants were excluded if the difference between the trainer’s GAT measurement and the participants’ first measurement was greater than 5 mmHg and/or the range of the participants’
measurements (i.e., the difference between the highest and lowest values) was greater than 7 mmHg.
Paper IV
IOP phasings performed between January and December 2019 within routine care by the patients themselves using an iCare Home tonometer were reviewed. IOP phasing consisted of four to five daily measurements recorded on two to three consecutive days. Patients’ medical records were reviewed to compile data such as age, gender, diagnosis, visual field index (VFI), rate of progression (RoP), IOP measurements from the referral visit and self-tonometry, and type of treatment pre- and post-IOP phasing. IOP data obtained by the patients were used to calculate the mean and maximum
13 IOP as well as the range (fluctuations) and peaks over consecutive days. Range was defined as the difference between the maximum and minimum IOP over 1 day or over a 3-day period. Level of treatment was defined as low (0–2 hypotensive agents), medium (3–4 hypotensive agents), or high (5 hypotensive agents or a combination of hypotensive agents with selective laser
trabeculoplasty/surgery or surgery only).
Only one eye was included in the analysis. If both eyes fulfilled the inclusion criteria, then the eye with the fastest progression of visual field defect according to the clinicians’ record was included. If the progression was the same for both eyes, the right eye was included.
3.4 Statistical analysis
The data mainly consisted of IOP measurements obtained with two different methods—iCare self- tonometers and GAT—and different users. Results were presented as the mean and SD. Sample size was calculated with the G*Power program (version 3.1.9.2, University Dusseldorf, Germany).
Windows Excel was used for all statistical analyses and SPSS version 23 (SPSS Inc., Chicago, IL, USA) was used for statistical analysis in Papers II, III, and IV.
In Papers I and II, both eyes were included. In Papers III and IV, only one eye was included.
Paired samples t-tests were used to analyze the difference between parametric groups (eyes, visits, and methods) in Paper I.
Independent Student’s t-tests were used to analyze the difference in clinical decisions between parametric groups in Paper IV.
The Mann-Whitney U-test was used to compare the difference in clinical decisions between non- parametric groups in Paper IV.
Agreement between the methods and/or users was assessed by means of the Bland-Altman analysis in Papers I, II, and III. The 95% of the limits of agreement of the compared groups were calculated as mean difference (= bias) ± 1.96 × SD.
One-way analysis of variance (ANOVA) was used to assess the repeatability of the measurements in Papers II and III. It was also used to assess the effect of age, VFI, and RoP on the clinicians’ decision in Paper IV.
The repeatability coefficient was calculated as 2.77 × (SD (examiner 1 – examiner 2)/√2).
The reliability (i.e., agreement and correlation) between measurements was assessed with the interclass correlation coefficient (ICC) in Paper III. The ICC agreement scale was: <0.50: poor
reliability; 0.50–0.75: moderate reliability; 0.75–0.90: good reliability; and >0.90: excellent reliability.
The level of significance was p = 0.05 for all papers.
14
4 RESULTS AND DISCUSSION
4.1 Feasibility of self-tonometry
Our results showed that self-tonometry could be performed by a majority of users, despite a short training period. We found no statistically significant difference between measurements made by patients older or younger than 70 years (p = 0.182) (Paper I). This is very important, as the majority of glaucoma patients tend to be old. Even glaucoma patients with severe visual impairment were successful, with only two out of 10 one-eyed patients being excluded because they could not position the tonometer properly to obtain valid measurements (Paper I). In Paper II, two patients (4%) failed to obtain valid measurements. In Paper III, we followed the manufacturer’s recommendations and excluded participants with a difference between the trainer’s and participant’s measurements of over 5 mmHg or with an iCare Home range of over 7 mmHg for the participant at the training session.
Consequently, nine glaucoma patients or healthy volunteers (15%) were excluded. In Paper IV, 10 patients (8%) were excluded for that same reason.
Overall, nearly 70% of the measurements performed with self-tonometers were within 3 mmHg of the GAT measurements, with 72% in Papers I and III and 68%–73% in Paper II. These findings are in accordance with those of Huang et al., who found that 71.5% of participants’ measurements were within this interval [92]. However, this level of agreement is lower than studies by Dabasia et al. and Cvenkel et al., whose results showed 83.3% and 84%, respectively [34, 85]. In the latter studies, the majority of the participants were glaucoma or OHT patients, and thus were probably more familiar with IOP measurements. In Papers II and III, the majority of the participants were healthy subjects and were not accustomed to IOP measurements. Nervousness might lead to contraction of the eyelids, breath holding, or even looking away from the tonometer, resulting in misalignment. To obtain the most accurate measurements, perhaps users should be given additional training time in achieving optimal alignment at the center of the cornea. The latest version of self-tonometer, iCare Home 2, has sensors to help the users position the probe at the apex of the cornea, hopefully leading to better precision.
4.2 IOP levels with GAT and iCare self-tonometers
In all our studies, the IOP was measured with iCare self-tonometers and GAT. The mean IOP values obtained by the different users are listed in Table 1.
4.3 IOP variations
In Papers I and II, we evaluated IOP variations. The results are summarized in Table 2. The IOP range is similar in our two studies with glaucoma patients and healthy volunteers. As other studies had shown, our results confirmed that a large majority of glaucoma patients have higher IOP in the morning than in the afternoon [93-95]. We found a high incidence of IOP peaks (9%–16%), which is in accordance with earlier studies [45, 53, 55]. To the best of our knowledge, no previous study had
15 investigated IOP fluctuations over consecutive days for glaucoma patients. Our results indicate that IOP fluctuations showed different patterns. Almost half of the eyes in glaucoma patients had IOP patterns that differed on consecutive days. For healthy volunteers, 63% of the study eyes showed different patterns on three consecutive days. This finding agrees with the results obtained from groups of healthy subjects, glaucoma and OHT patients, where the IOP was measured with GAT [96]
and a non-contact tonometry [97].
Table 1. Mean IOP measurements made with GAT and the iCare self-tonometer.
Participants Eye Method Visit 1
(mean IOP ± SD in mmHg)
Visit 2 (mean IOP ± SD
in mmHg)
Paper I Glaucoma patients
Right GAT
RT patients
15 ± 5 15 ± 5
15 ± 5 14 ± 6
Left GAT
RT patients
14 ± 5 15 ± 6
15 ± 5 15 ± 6
Paper II Healthy volunteers
Right
GAT RT volunteers
RT trainers
15 ± 3 15 ± 4 14 ± 4
14 ± 3 15 ± 5 14 ± 4
Left
GAT RT volunteers
RT trainers
15 ± 3 15 ± 4 15 ± 4
14 ± 3 15 ± 5 14 ± 4 Paper III
Glaucoma patients and healthy
volunteers
-
RT participants RT trainer GAT, trainer GAT, extra personnel
14 ± 5 13 ± 5 15 ± 5 13 ± 5
-
Paper IV
Glaucoma patients -
GAT RT patients
RT trainer
15 ± 4 14 ± 5 14 ± 5
- RT: rebound tonometry; GAT: Goldmann applanation tonometry
Table 2. IOP range, maximum, peaks, and patterns.
Paper I Glaucoma patients
N = 126 eyes
Paper II Healthy volunteers
N = 92 eyes
Range (mmHg) 11- 19 10 - 17
Higher IOP in the morning (%) 64 (day 2) 65 (day 3)
51 (day 1) 57 (day 2) 58 (day 3) Peaks outside office hours (%) * 9 (day 2)
16 (day 3)
10 (day 1) 16 (days 2 and 3)
Different patterns (%) ** 47 63
Range: mean minimum IOP and mean maximum IOP over 3 days
* if the difference between the highest value outside office hours (4, 6 am or 6, 8, 10 pm) and the highest value during office hours (8, 10 am or 12, 2, 4 pm) equals or exceeds 4 mmHg
** day 2 vs. day 3
16 Our results raise questions about the evaluation of glaucoma treatment effects. Indeed, this
evaluation is often based on a single IOP measurement with GAT obtained at the eye clinic during opening hours, and it may not be repeatable the next day due to IOP pattern variation. This implies that, in some glaucoma investigations, it would be useful to measure the IOP several times per day to estimate the daily variation and the effect of treatment. Self-tonometry is becoming increasingly popular as a way to evaluate and compare the effects of different treatments [98, 99].
4.4 Agreement between GAT and iCare self-tonometers
Bland-Altman plots of the correlation between the methods indicated good agreement between the GAT and the iCare measurements made by the different users. The mean difference between the measurements made with GAT by the trainer and with iCare devices by the glaucoma patients and the participants varied between 0.06 and 0.93 mmHg (Figure 6). However, the iCare self-tonometers tended to underestimate values below and overestimate values above 18–20 mmHg (Paper I) and 15 mmHg (Papers II and III) compared with the GAT measurements. This finding should be considered when clinically applying IOP values measured using iCare devices.
17
The repeatability of the measurements with both methods and different users is summarized in Table 3.
Table 3. Pairwise comparisons with one-way repeated-measure analysis of variance (ANOVA).
Methods Mean
difference* Std. error Sig.**
Paper II
GAT vs. RT healthy volunteers RT healthy volunteers vs. RT trainer RT trainers vs. GAT
-0.266 0.231 0.035
0.226 0.183 0.212
0.240 0.206 0.869 Paper III
RT participants vs. trainers GAT trainers vs. extra personnel RT participants vs. GAT trainers RT trainers vs. GAT trainers
RT participants vs. GAT extra personnel
0.519 1.315 -0.926 -1.144 0.389
0.255 0.421 0.339 0.311 0.587
0.280 0.017 0.052 0.000 1.000 GAT: Goldmann applanation tonometry; RT: rebound tonometry with iCare Home
*Difference between the mean of all GAT and iCare measurements, respectively
** Adjustment for multiple comparisons: Bonferroni test
No statistically significant difference was found between the methods in Paper II with healthy volunteers. In Paper III, the iCare Home tonometers showed a better repeatability than GAT, which might be caused by several factors, mainly due to the way in which the mean IOP was calculated. On the one hand, participants’ mean IOP with iCare Home was calculated from three measurements that were already the mean of four measurements each. The mean of these multiple measurements probably gives less variation. For the trainer, only one measurement (the mean of four
measurements) was recorded. On the other hand, the mean difference in the inter-user
measurements made with GAT was calculated from single measurements and resulted in a greater variation. Similar to our study, Dielemans et al. compared the first measurements of two examiners with different GAT instruments in different rooms, whereas Tonnu et al. compared the means of three measurements of two examiners [18, 100]. Dielemans et al. and Tonnu et al. had only two examiners, who were trained and certified for their study. In contrast, our study reflects daily clinical practice in a tertiary eye center, where the participating personnel (three trainers and eight different personnel) were randomly selected and were not study certified. This may explain the larger
Figure 6. Bland-Altman plot of agreement between IOP measured using Goldmann applanation
tonometry (GAT) by the trainers and iCare self- tonometry by glaucoma patients (Paper I), healthy volunteers (Paper II), and both glaucoma patients and healthy volunteers (Paper III).
Solid lines indicate mean differences (bias); dashed lines indicate the upper and lower 95% limits of agreement; and dotted lines indicate regressions between the mean and the difference.
18 variations of GAT measurements compared with iCare Home, which are also reflected in the
repeatability coefficients (3.88 and 2.38, respectively). Interestingly, a study by Mihailovic et al.
concluded that, despite education and training to improve agreement between physicians’ and technicians’ IOP measurements, differences remain, indicating that single GAT measurements might need to be regarded with caution [19].
Another factor that might affect repeatability is the calibration of the instruments. iCare Home devices were calibrated at the beginning of each measurement session, while GAT instruments were calibrated once a month at our clinic. Furthermore, the same iCare Home device was used by the participant and the trainer, whereas two different GAT instruments were used on the same occasion by two different healthcare staff (Paper III), possibly adding to the variation of the measurements.
4.5 Inter-observer variability
In Paper III, we compared the inter-observer variation between GAT and iCare Home. In order to reflect tonometry performance at a tertiary eye center, all measurements were performed in routine clinical settings where none of the personal was selected and certified for the study purpose. We found good agreement between the GAT and iCare Home measurements. The repeatability of the three consecutive measurements made by the participants using iCare Home was excellent (ICC = 0.975, 95% CI: 0.960–0.985, p = 0.001). The degree of reliability between the measurements made by different users is summarized in Table 4.
Table 4. Agreement between inter-users’ measurements of IOP made with iCare Home and GAT.
Users ICC
(95% CI) p
RT participant vs. RT trainer 0.93
(0.880–0.959) 0.001 GAT trainer vs. GAT personnel 0.741
(0.558–0.849) 0.001 RT participant vs. GAT trainer 0.852
(0.742–0.915) 0.001 RT participant vs. GAT personnel 0.593
(0.388–0.742) 0.001
RT trainer vs. GAT trainer 0.892
(0.821–0.936) 0.001 RT: rebound tonometry with iCare Home; GAT: Goldmann applanation tonometry ICC: interclass correlation coefficient
The level of reliability with iCare Home measurements was excellent, with an agreement between the participants and trainers of 0.93 (95% CI: 0.880–0.959); this result was similar to the result obtained by Pronin et al. of 0.903 (95% CI: 0.867–0.928) and that obtained by Termühlen et al. of 0.894 [37, 84]. The level of reliability with the GAT measurements was good, with an agreement between the trainers and personnel of 0.741 (95% CI: 0.558–0.849), in accordance with the result
19 obtained by Dielemans et al. (0.81) but much lower than that obtained by Salim et al. (0.989), where two examiners measured three times the IOP with the same instrument [100, 102].
Our studies (Papers I, II, and III) presented several limitations, as they were not randomized. All consecutive patients who agreed to participate were included, which might induce a selection bias.
Our sample with healthy participants was not representative of glaucoma patients, as there was a majority of middle-aged women and not familiar with ocular measurements on themselves.
Furthermore, in Paper III, three different personnel trained the participants, so the latter might have received slightly different instructions that could influence the quality of their measurements. In all studies, the measurements were not made randomly but were always performed in the same order.
In Paper III, as the GAT measurement by extra personnel was always measured last and delayed by up to 15 minutes, we cannot exclude the possibility that the IOP differed because the participants moved to another room or were anxious about meeting different personnel. We also know that repeated measurements with GAT tend to decrease the IOP [15, 17, 100]. However, in our study, only two measurements were made with GAT for each participant, so this effect may be negligible.
In summary, the analysis of agreement, repeatability, and reliability between iCare Home and GAT supports the feasibility of self-tonometry.
4.6 CCT influence on IOP measurements
The influence of CCT on the measurements was analyzed in Paper II. The mean CCT was 560 ± 37 μm for the right eye and 553 ± 48 μm for the left eye. Table 5 illustrates the mean difference between GAT and iCare Home in relation to CCT. As previously observed in other studies, we found that, for CCT values of 500–600 µm, the mean difference between the methods was minimal in our study with healthy volunteers [22, 26, 33, 34, 37]. However, for CCT values over 600 µm, iCare Home
overestimated the IOP compared with GAT measurements.
Table 5. Mean difference between GAT and iCare Home IOP in relation to CCT.
CCT (µm) N (%) Mean difference
GAT – RT (mmHg)
<500 500–600
>600
6 (5) 98 (82) 16 (13)
1 ± 1.17 0 ± 3.29 -2 ± 4.55 CCT: central corneal thickness,
RT: rebound tonometry with iCare Home; GAT: Goldmann applanation tonometry
A correlation between CCT and IOP measured with GAT and RT has been described in many studies [7, 9, 10, 12, 22, 23, 26, 30]. In Paper III, we chose not to consider corneal parameters such as CC, CCT, or corneal hysteresis (CH). We wanted to compare the differences between inter-users within each method measured on the same patients, and corneal parameters would affect these results equally, irrespective of the users; therefore, the influence of these factors was not further studied.
