This paper evaluates both the short-term and long-term variability of intraocular pressure (IOP) patterns in eyes with primary open-angle glaucoma (POAG) based on measurements collected with an intraocular pressure sensor (EyeMate, Implandata GmbH, Germany) implanted in 22 patients during cataract surgery. There also is an external reading device that provides power to the sensor via electromagnetic coupling as well as serving as a data relay to a web-based database.1 Patients were trained on how to measure their own IOP with the device, and were instructed to carry out at least four measurement cycles per day. The 24-hour day was split into seven time periods, in which measurement cycles could take place. For these 22 patients over a mean follow-up duration of (19.2 ± 21.3) months, a total of 93,033 IOP measurements based on 15,811 measurement days were collected and analyzed. Data were grouped by eye and by pressure-lowering drug, with cases in which the same eye has been considered more than once since it had been treated with two or more different drugs. The analysis was carried out for short-term variability, based on measurement cycles collected within three months of each other, and long-term variability, based on measurements over a time period of one year or longer. Variability was assessed using intraclass correlation coefficients (ICCs).
The study showed that there is a moderate short-term and a higher long-term variability and, importantly, that daily IOP variations cannot be characterized with only a single measurement. There are many ways in which the analysis can be affected, as well as the results. First, the measurements cycles are strongly dependent on patients' will and possibility to adhere to the instructions, and their adherence patterns greatly vary: the follow-up range between 1 and 58 months, the measurement cycles range from 1 per day to 277 per day. No actual regular measurement patterns seem to exist among patients and a great variability on medication adherence is observed. Moreover, the ways and the amount by which the possible effects of adherence to protocol and to drugs may affect IOP time variability have not properly assessed yet.2 Other possible effects on patients (e.g., hormonal, seasonal, lifestyle, body mass index, seated posture) on IOP and its time variability should be first assessed on healthy subjects enrolled as controls, although it has been shown this is not an easy task.3 By providing altered data as well as the unavoidable sensor calibration issues, such problems could affect comparisons with the usual GAT measurements. Results are shown in terms of ICCs for groups. As for the timeframe, the ICCs were used to assess variability, with two time duration classes, short-term and long-term, without analysis ot the variability in terms of periodicities based on curve amplitudes and features, or with time series analysis. Higher resolution both in terms of timeframe and scenario could definitely strengthen the analysis.4
The main strength of this study is the availability and the use of implantable IOP sensors and their substantial IOP data. On average, there was more than a 600-fold increase of IOP measurements numerosity compared with those obtained in a typical clinical setting that has just 7 measurements per patient in the average follow-up time duration. The authors themselves outlined the imitations of this study ranging from the small patients' sample, the paucity of measurements obtained during the night, and the neglect of health and lifestyle factors that may disrupt circadian rhythms.
Overall, this is an outstanding manuscript, based on a clever and still very promising approach. As mentioned by the authors future studies of long-term variability are needed. Studies on short-term variability should consider and include different timeframes for both data collection and interpretation. It would be interesting to cover the hours of the day more uniformly, particularly at night when the patient is asleep. The authors stressed the importance of a continual IOP monitoring activity for glaucoma patients, possible only with an implanted sensor in a clinical setting, or in an alternative configuration such as a sleep mask.2 These findings can have important implications both for clinical glaucoma management and clinical trials. This IOP monitoring research seeks to obtain optimal and accurate IOP control via appropriate therapeutic intervention.5 We congratulate Kaweh Mansouri, Harsha L. Rao and Robert N. Weinreb for their valuable contribution.