Intraday repeatability of macular layers measurements in glaucomatous and non-glaucomatous patients using spectral-domain optical coherence tomography

Glaucoma is a heterogeneous group of optic nerve diseases and leads to characteristic morphological changes of the optic nerve (increasing optic disc excavation), loss of retinal ganglion cells, and thinning of the retinal nerve fiber layer [1‐3]. The specific visual field changes are also the main features of this chronic progressive optic neuropathy. If left untreated, glaucoma leads to severe visual acuity loss even to permanent blindness [1, 4, 5].

Thanks to modern technological advances in the last decades, spectral-domain optical coherence tomography (SD-OCT) has become a significant and established examination technique in ophthalmology and the diagnosis of glaucoma in particular [6‐8]. SD-OCT measurements in glaucoma are based on Bruch's membrane opening and include key quantitative measurements such as an analysis of BMO-minimal rim width, circumpapillary retinal nerve fiber layer thickness, and macular architecture in the posterior pole [7, 9]. Compared with confocal scanning laser ophthalmoscopy (cSLO), which was the gold standard until recently, SD-OCT is a more reliable method for morphometric examination of the optic disc [6, 10‐16].

In diagnosing glaucoma with SD-OCT, the most significant and frequently used parameter is the assessment of the circumpapillary retinal nerve fiber layer thickness of the optic nerve head [15, 17‐20]. However, assessment of macular architecture (especially thickness of retinal nerve fiber layer, ganglion cell layer, and inner plexiform layer) can serve as an important additional diagnostic tool in the early diagnosis of glaucoma [17, 21‐24]. With more advanced and highly reproducible segmentation algorithms of all retinal layers in the macular region, this type of analysis is of interest for clinical use in the detection of glaucomatous lesions [9, 17, 25‐27].

In addition, with an increasing number of follow-up SD-OCT examinations in glaucoma thanks to its ubiquitous use, the logical question arises to what extent intraindividual measurement variability may influence the results of comparative progression analyses [2, 6, 17]. Repeated precise measurements are essential for a correct glaucoma progression analysis of intraindividual disease dynamics [2, 7].

There is a hypothesis that variations in the anatomical structures of the optic nerve are possible due to intraday variations in intraocular pressure (IOP) and blood flow of the optic nerve [19, 28]. Such fluctuations in anatomical structures may also be characteristic of the retina in the macula region and hinder the accurate detection of glaucoma progression depending on the time of day when SD-OCT coherence tomography is performed [17].

Knowing the presence and quantitative values of intraday variations of macular architecture, the true glaucomatous progression and measurement variability of morphometric parameters can be distinguished. However, there is a lack of confirmed data on this to date. Therefore, the purposes of this study are to assess the intraday repeatability of macular architecture measurements in glaucomatous and non-glaucomatous patients using spectral-domain optical coherence tomography and to evaluate the independence from intraindividual intraocular pressure fluctuations.

The present single-center, time-point comparison study was conducted between February 2020 and June 2021 in a German primary ophthalmology center. All procedures performed in this study were following the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Approval for the study was obtained from the Ethics Commission of Cologne University’s Faculty of Medicine (19–1663). Written informed consent was obtained from all participants.

In this study, SD-OCT and Goldmann applanation tonometry were performed at two different time points with a time interval of 5 h on the same day (morning and afternoon). The following inclusion criteria were considered: the ability to give consent or consent by the legal representative, sufficient fluency in the German language to understand the contents of this study, and good cooperation of the patient. The following patients were excluded from this study: underage patients, pregnant patients, patients with a current or past history of systemic diseases with ocular involvement, retinal vascular diseases, visual field defects of other origins than glaucoma, high myopia greater than 6.00 dpt, pathological changes of the macular region or the optic nerve, severe opacities of the refractive media. Patients were classified into the three diagnostic groups according to the 5th edition of the European Glaucoma Society (EGS) guidelines: glaucoma group, ocular hypertension (OHT) group, and healthy patients without glaucoma (control group) [29].

Spectral Domain optical coherence tomography was performed according to the general guidelines using the Glaucoma program (scan size in the macular region 25 × 30°, central position of the pattern, number of scans of the region 61, distance between the regions 120 µm). The SD-OCT Spectralis® Glaucoma Module Premium Edition (Heidelberg Engineering GmbH, Heidelberg, Germany) was used with the device operating software version 6.13. The focus of the study was posterior pole analysis (8 × 8 grid with 64 squares pro layer) and measurement of the following morphometric parameters: entire retinal thickness (RT), the thickness of retinal nerve fiber layer (RNFL), ganglion cell layer (GCL), inner plexiform layer (IPL), and inner nuclear layer (INL). The minimal accepted quality of SD-OCT scans was 15 dB.

Timely detection of the progression of optical neuropathy is an important aspect of glaucoma management since early recognition of changes in optic nerve morphology and macular architecture can help clinicians to adjust patient treatment and prevent permanent vision loss [30]. Accurate and reliable SD-OCT measurements are crucial for a correct analysis of the glaucoma progression and dynamics of the disease [2, 7]. Difficulties in right diagnosing of the glaucoma progression include not only the complexity of distinguishing true changes associated with the disease from normal aging processes of the retina but also the variability of SD-OCT measurements themselves [17, 31].

In this study, we could show that the macular morphometric parameters (RT, RNFL, GCL, IPL, INL) have a high degree of agreement between the measurements in the morning and the afternoon. According to Bland–Altman plots and intraclass correlation coefficients, it can be concluded that the variations of the morphometric parameters during the day are not clinically significant and that the measurement results in all three groups (glaucoma, ocular hypertension, and healthy groups) speak for the overall excellent repeatability.

Ctori et al. also reported good repeatability and reproducibility of the individual measurements of the retinal layer thickness in young healthy volunteers. In this study, two investigators independently performed two SD-OCT examinations for each participant in a single visit. In the literature, there are also works devoted to the assessment of the variability of morphometric parameters of the optic nerve head. Enders et al. reported high intraday repeatability after assessing the variability of global BMO-MRW, global RNFL thickness, and global BMO-MRA at two different time points with a time gap of more than 5 h [19].

On the other hand, the analysis using random-effects model indicated a statistically significant influence of time points on a measurement of macular architecture. This variability between intraday measurements could be explained by the inability of performing an absolutely identical SD-OCT examination in practice, due to the patient cooperation during the examination and the skills of the staff performing tomography. The fluctuating quality of SD-OCT images can play a crucial role in the variation in the thickness of the retina and its layers since the correct determination of the boundaries of these layers is done automatically by the software of the OCT device. The higher the quality of the images, the more correctly the segmentation algorithms of the software work and, consequently, the more accurately thickness of retinal layers is detected. Each performed examination should also be verified by a physician and manual corrections should be made to the automatic segmentation of the retinal layers, if necessary [7, 32]. As has already been confirmed, manual segmentation correction increases the diagnostic power of another SD-OCT morphometric parameter such as circumpapillary RNFL to detect glaucomatous changes of the retina [9, 33, 34].

The absolute mean variability of the RT, RNFL, IPL, and INL did not correlate in our study with the absolute IOP variations on the two SD-OCT examinations performed. The Spearman correlation of the mean absolute difference of the GCL and the mean absolute difference of the IOP has a value of rho = 0.109 (P = 0.031) and can be described as weak positive correlation, because it is in the range of 0.1 to 0.3 [35]. However, it cannot be claimed why there is a weak positive correlation only with ganglion cell layer. Possibly, this layer of the retina is more susceptible to IOP variations than the other retinal layers or the entire retina itself. Further studies are necessary to gain a comprehensive understanding of the potential impact of IOP changes on macular morphometric parameters, due to the small detected intraday IOP fluctuations. Enders et al. found no significant correlation between intraindividual variability in morphometric parameters of the optic nerve head (global BMO-MRW, global RNFL thickness, global BMO-MRA) and intraindividual IOP changes in two examinations performed on the same day [19].

Limitations of this work include the relatively small number of patients in the glaucoma group and the OHT group. In this study, only the inner retinal layers were examined because of the already proven changes that occur in them by the progression of glaucoma [23, 24]. To obtain a complete picture of the possible intraindividual variations of macular morphometric parameters during the day, it would be useful to study all retinal layers. The good documentation of the morphometric parameters of the retina thanks to high-resolution SD-OCT is an advantage of this work. The variation of retinal vascular characteristics of the macular region and its possible influence on intraindividual variation in macular morphometric measurements should be investigated with the help of optical coherence tomography angiography (OCT-A) in future studies.

In summary, the data of this study showed that the macular morphometric parameters of SD-OCT demonstrated excellent repeatability in two measurements performed in the morning and afternoon (time interval of 5 h 36 min). According to intraclass correlation coefficients, the degree of agreement of each morphometric parameter in three studied groups in two time points was very good. The time points had a significant influence on studied variables. The fluctuation of thickness of the RT RNFL, IPL, and INL does not seem to be affected by the moderate variations of IOP. There was a weak positive correlation between the mean absolute difference of IOP and the mean absolute difference of GCL.