The evaluation of juvenile ocular hypertension by optical coherence tomography angiography

Ocular hypertension (OHT) is defined as intraocular pressure (IOP) higher than 21 mmHg with normal optic disc structure/function and no previous history of angle closure. Approximately 5% of OHT patients develop glaucoma after 5 to 7 years in follow-ups [1]. Glaucoma is the leading cause of irreversible blindness, including both adult-onset disease (occurring after 40 years of age) and juvenile-onset disease (occurring between the ages of 3 and 40); the IOP of juvenile-onset patients is often extremely high, with a more aggressive clinical course [2, 3]. Thus, OHT in juveniles (JOHT) should be given more attention when considering long-term follow-up. However, JOHT presents few ocular symptoms and little disturbance of visual acuity, making its diagnosis and management a difficult clinical challenge [4]. In addition, previous literature shows that the normal range of IOP in juveniles may be different from that in adults [5]. Currently, there is no consensus in the literature that offers guidance to the clinician in determining when and whether JOHT should be treated. Evidence provided by previous studies concerning OHT mainly focuses on people more than 40 years old under traditional diagnostic standards [1]. The Ocular Hypertension Treatment Study (OHTS) established that medically treating ocular hypertension is efficacious in delaying or preventing the onset of glaucoma, while the European Glaucoma Prevention Study (EGPS) failed to demonstrate the significance of reducing IOP in preventing the onset of the disease [6, 7]. The effects of lowering IOP in JOHT subjects were ambiguous. Moreover, JOHT individuals are increasingly being identified through the prevalence of optical screening, and some of them even demonstrate a large vertical cup/disc ratio (CDR). This is considered a good qualitative predictor for the onset of glaucoma in JOHT, although this correlation has not been empirically verified [8]. The CDR index has long been used in the assessment of JOHT, though the wide range of CDR values in the normal population from 0.00 to 0.87 limits its utility [9, 10]. Thus, decisions regarding the therapy of JOHT can be difficult, especially for those with large CDR [11, 12]. New modalities should be considered to assess JOHT patients and find how the predictors of CDR and IOP impact the evaluation of the disease.

Previous experimental and clinical investigations have provided evidence showing a strong correlation between vascular dysregulation and glaucoma [13, 14]. Recently, the rapidly evolving technology of optical coherence tomography angiography (OCTA) has been utilized to measure local retinal circulation without the need for dye infusion. The split-spectrum amplitude-decorrelation angiography (SSADA) algorithm can be used to quantify blood flow [15]. Radial peripapillary capillaries (RPC) comprise a network of capillary beds located within the retinal nerve fiber layer (RNFL) that supplies the retinal ganglion cell (RGC) axons; damage to RNFL and RGC are typical manifestations of glaucoma [16]. Thus, OCTA could quantitatively characterize the microvasculature around the nerve head to find the preclinical vessel density (VD) changes of JOHT. The IOP of JOHT typically remains fluctuated, especially in subjects more than 10 years old [4]. No typical clinical symptoms or pathological changes can explain the relationship between IOP and retinal perfusion [4]. Previous studies using OCTA have shown that PRC-VD is significantly reduced in glaucoma and related to the severity of the disease [17, 18]. Additionally, the macular VD of the superficial retinal vascular plexus (SVP) has also been investigated, and a reduction in macula has been found [19]. It is valuable to study the peripapillary and macular VD profile of JOHT and to find the relationship between IOP and perfusion. Besides, variation in CDR requires more accurate and quantatitive assessment in evaluating the peripapillary microvasculature. Therefore, investigating the use of non-invasive and high-resolution OCTA techniques for JOHT subjects is an important research avenue to pursue.

The purpose of this study was to evaluate RPC-VD and SVP-VD in JOHT using OCTA and to examine the relevance of IOP and CDR to the VD profile.

A total of 86 eyes of 45 Ctrl subjects and 65 eyes of 34 JOHT subjects who initially met the eligibility criteria were included in the final analysis. The demographic and clinical characteristics of the two groups were summarized in Table 1. They were comparable for age, gender, SBP, DBP, MAP, PR, BCVA, CDR, GCC thickness, visual field MD, and PSD. Parameters of AL, CCT, IOP, and RNFL thickness were statistically different between Ctrl and JOHT subjects. Thus, IOP was further compared using a corrected formula as follows: Corrected IOP = Measured IOP - (CCT - 520) * 2.5/50 [25]. Corrected IOP of JOHT subjects was significantly higher than that of Ctrl as designed (P <  0.001). RPC-VD and SVP-VD were adjusted for AL, CCT, and RNFL differences in the post-hoc analysis to compare VD between JOHT and Ctrl. Bare differences were found in NI and T peripapillary regions (P = 0.042 and P = 0.033), while other regions showed no difference (Fig. 2).

All data from Ctrl and JOHT were included to evaluate the effect of IOP and CDR on RPC-VD (Table 2). An eye-gender, CCT, AL, and RNFL-adjusted linear mixed model showed that VD of the inside disc was significantly lower in subjects with CDR larger than 0.6 (r = − 13.304, P <  0.001; Table 2). RPC-VD of the T region was marginally negatively correlated with high IOP (r = − 1.329, P = 0.043; Table 2), while other regions showed no correlation with high IOP (all P > 0.05). Average RPC-VD, and N, IN, IT, ST, SN regions of RPC-VD were strongly correlated positively with CDR > 0.6 (r = 2.621, P <  0.001; r = 4.425, P <  0.001; r = 3.110, P <  0.001; r = 2.153, P = 0.001; r = 1.642, P = 0.019; r = 2.463, P = 0.006, respectively). Of the six peripapillary regions in RPC-VD, only the T region showed no correlation with CDR (r = 0.926, P = 0.152).

Linear mixed models adjusted by CCT, AL, RNFL, and eye gender also showed that perfusion of SVP had no correlation with large CDR, while high IOP did not (Table 2). S and N of SVP-VD were marginally lower in subjects with high IOP (r = − 1.877, P = 0.023; r = − 1.693, P = 0.049; Table 2). In addition, parafovea and hemi-S SVP-VD were found to be slightly statistically positively correlated with IOP more than 21 mmHg (r = − 1.530, P = 0.041; r = − 1.570, P = 0.037; Table 2). Unmentioned parameters were not found to be significant differences between the positive and negative values of IOP or CDR in SVP-VD (all P > 0.05, Table 2).

In this study, we reported the RPC-VD and SVP-VD between Ctrl and JOHT subjects. IN and T of RPC-VD were slightly higher in JOHT than in Ctrl subjects. Risk factors of CDR and IOP were further taken into consideration in the assessment of VD. A strong positive correlation between large CDR and RPC-VD in six regions except for T was established. The T region of RPC-VD had a tendency to decrease with IOP more than 21 mmHg, while its weak impact on S and N of SVP-VD was also observed. All of these data suggest that the results of OCTA scans showed the slight VD change in JOHT subjects. It also demonstrated strong correlation with CDR and exhibited useful advantages in long-term follow-up of JOHT.

The capillary blood flow of glaucoma patients, OHT, and Ctrl subjects has been evaluated using scanning laser Doppler flowmetry images in the past [26]. Subsequent research used OCTA to compare the similar parameters between Ctrl and OHT subjects and agreed with our first findings of no obvious alterations in SVP-VD (Fig. 2) [17, 23]. However, IN and T of RPC-VD showed statistical significance between Ctrl and JOHT, although the p-value was closed to 0.05. The average age of the subjects involved in previous studies was more than 40 years old, substantially different from the JOHT subjects we desired to target. Moreover, the CDR of the Ctrl they used accorded with the inclusion criteria of less than 0.4 by ophthalmoscope, while subjects of high CDR were sometimes considered glaucoma suspects or OHT when the IOP was above 21 mmHg. Regarding the wide range of CDR values in the normal population varying from 0.00 to 0.87, grouping was different from previous studies in our research. This contradiction made further analysis of IOP and CDR necessary; we thus applied a linear mixed model to analyze these two factors separately.

Using the linear mixed model to study the risk factor of IOP higher than 21 mmHg, T RPC-VD exhibited a significant reduction in JOHT subjects. Although no statistical correlation was found in the IT region, a consistent trend was observed throughout the comparison. Kerr and associates reported reduced blood flow in the temporal neuroretinal rim of the optic nerve; thus, the decrease of RPC-VD in our study respects the patterns of early glaucomatous changes [27]. Meanwhile, sectors of superior and nasal SVP-VD also exhibited decreases matching the patterns of peripapillary regions. It should be noted that we did not actually measure the blood flow of the retina using OCTA, but the imaging technique could be a useful tool to show the vascular changes by the index of VD. Hence, our results of reduction in both RPC and SVP-VD could be explained by a decrease in MOPP according to the formula which leads to a drop in ocular perfusion, as microvascular networks are of primary interest in glaucoma [20]. On the other hand, high-level evidence has shown that treatments decreasing IOP improve retinal circulation and reduce the risk of developing glaucoma in OHT [28]. This validates the effects of high IOP on ocular perfusion. Besides, individuals with high IOP are more likely to suffer from venous collapse, which then exacerbates impaired blood flow. OCTA proves the difference of capillary VD caused by high IOP in JOHT subjects and could serve as a precise method to monitor disease development, although longitudinal data are still needed to fully evaluate this possibility.

The RPC-VD of all six regions, except for the T sector, showed an increase with CDR > 0.6. No statistical difference was found in the perifoveal areas. CDR has been shown to be positively related to optic disk size, and the number of nerve fibers revealed consistency with the optic disk size, with the macula unaffected [29, 30]. RPC comprises a network of capillary beds located within the RNFL carrying metabolites to nerve fibers; thus, it should be correlated with the total number of nerve fibers. The positive relationship between CDR and RPC-VD in our study further clarifies this vascular structure of blood supply. However, the T sector showed no correlation with CDR. One reason for this could be that the T peripapillary region shares the thinnest RNFL according to ISNT rules. Taking application into consideration, CDR is used as an individual risk factor for OHT development with a wide normal range, which makes it confusing to distinguish large optic disc subjects with glaucoma suspects. OCTA quantifies these changes more precisely, which could aid in the evaluation of early perfusion alterations in JOHT subjects.

The main limitation of our study is the small sample size, especially JOHT subjects exhibiting both high IOP and normal CDR. Basically, our research is an exploratory study instead of a confirmative one. We wish to find some clues in OCTA scans of JOHT to provide any support for clinicians to pursuing further important results. The sample size applied in this study follows the previous publication of OCTA in Ctrl, OHT or glaucoma patients. The majority of the researchers conducted the research with 20–50 participants in each individual group. In future, we will evolve more participants in our study and pursue the long-time follow up to acquire more powerful evidence to support the findings. Additionally, IOP has always been normalized in the majority of juveniles over adolescence with long-term follow-up, which is why the medical terminology of ‘adolescence IOP fluctuation’ or ‘adolescence ocular hypertension’ is used. We only used IOP as a factor directly, and in the future we could test whether IOP alone or corrected IOP can exhibit more powerful correlations with the VD of JOHT subjects [31]. Besides, the software of OCTA applied in this study was not capable of removing the signals of large vessels. As explained above, more researches concerning disc size, CDR, and VD using the new upgraded device of OCTA are necessary. There is no long-term follow-up done to support the value of the clinical implications in OCTA of CDR or IOP [4]. In addition, we are still unclear about how IOP fluctuations accompany, for example, seasonal changes; this is a further target for the group to study.

The current study demonstrated slightly lower nasal-inferior and temporal RPC-VD in JOHT subjects. A marginal negative relationship between T RPC-VD and high IOP was found, while CDR had the strong opposite outcomes. SVP-VD of perifoveal regions had a trend to reduce with high IOP but not by CDR. The results suggest that both the risk factor especially CDR should be considered and analyzed individually when clinicians assess the test results of OCTA scans in JOHT. RPC-VD, especially in the temporal region, has the potential value to assess the development of JOHT.