Correlation of retinal nerve fiber layer thickness with perimetric staging in primary open-angle glaucoma - A cross-sectional study.

BACKGROUND: Glaucoma is a progressive optic neuropathy, characterized by structural optic nerve damage with corresponding field defects. Primary open-angle glaucoma (POAG) is the most common. Although perimetry is the gold standard, retinal nerve fiber layer (RNFL) thickness by spectral-domain optical coherence tomography (SD-OCT) has proved reliable in the detection of pre-perimetric glaucoma. There is preferential involvement of various sectors of the peripapillary RNFL in the different stages of POAG.
PURPOSE: The purpose of this study is to assess RNFL thickness and determine preferential involvement of different sectors of peripapillary RNFL in the various stages of POAG using SD-OCT.
MATERIALS AND METHODS: Forty-nine patients with POAG underwent complete ophthalmic examination including visual field testing and RNFL thickness measurement. Perimetric findings were used to categorize them into mild, moderate, and severe stages of glaucoma. The RNFL thickness values were analyzed and compared with perimetric results.
RESULTS: The average RNFL loss in mild, moderate, and severe POAG was 25.44%, 29.67%, and 44.15%, respectively. A statistically significant correlation (P < 0.05) between RNFL loss and severity of glaucoma was found in all except the superior and temporal sectors. A statistically significant (P < 0.05) negative correlation was noted between visual field index and RNFL loss in all sectors except the nasal-superior in moderate POAG and all sectors in severe POAG. Mean deviation and RNFL loss showed a significant positive correlation in temporal-inferior (TI) sector in mild POAG and all sectors in the severe group.
CONCLUSION: RNFL thickness decreases with increase in glaucoma severity and is a reliable parameter to differentiate mild from severe POAG. The TI followed by nasal-inferior RNFL sector is the most sensitive to glaucomatous damage in all three stages. Copyright:

Introduction

Glaucoma is a group of diseases, characterized by optic neuropathy consistent with remodeling of connective tissue of the optic nerve head and loss of neural tissue associated with the development of visual field defects. Approximately 11.2 million people in India suffer from glaucoma, of which 6.48 million have primary open-angle glaucoma (POAG).[1] The structural changes due to retinal ganglion cell (RGC) loss lead to functional visual field loss. Perimetry is used as a base line at diagnosis as well as for monitoring of the disease. The classification of glaucoma into various stages helps in the assessment and documentation of visual field loss, allowing the detection of progression. It is sensitive and specific, but subjective and shows intertest variability.[2] Visual field defects are found after 25%–30% RGC loss,[3] which may lead to patients with early glaucomatous damage remaining undetected.

Optical coherence tomography (OCT) is more sensitive to changes in visual function than perimetry. Since it specifies the RNFL thickness in various quadrants, the area most susceptible to damage can be determined. It is especially useful in patients whose perimetric results are not reliable.[4]

Materials and Methods

Study design

Descriptive, noninterventional, and hospital-based study.

Inclusion criteria

Confirmed cases of POAG with characteristic optic disc changes and corresponding visual field defects, based on Anderson's criteria,[5] reporting to the Department of Ophthalmology, Justice K. S Hegde Charitable Hospital.

Exclusion criteria

Patients with refractive error >−6.00 D, axial length >24 mm, corneal scarring, significant media opacities, secondary glaucoma, severe nonproliferative diabetic retinopathy (NPDR), PDR, retinal diseases affecting the visual field, nonglaucomatous optic neuropathy, diseases affecting visual pathways, patients with a history of ocular trauma or intraocular surgery (except uncomplicated cataract surgery and/or trabeculectomy in confirmed POAG patients).

All the study participants underwent refraction, slit-lamp examination, intraocular pressure (IOP) measurement by Goldmann's applanation tonometry, pachymetry, fundus examination using +90 D lens, gonioscopy, visual field testing using Humphrey's Field Analyzer, Swedish Interactive Threshold Algorithm (SITA) 24-2 (Carl Zeiss Meditec, Dublin, CA, USA). Acceptable visual field tests included false positive and false negatives of <33% and fixation losses <20%. Eyes were classified as having mild, moderate, or severe glaucoma using Hodapp-Parrish-Anderson criteria.[6]

A scan circle of 3.5 mm diameter around the optic nerve was taken and peripapillary RNFL thickness in global, nasal-superior (NS), temporal-superior (TS), temporal-inferior (TI), nasal-inferior (NI), nasal and temporal sectors was measured using spectral domain OCT (SD-OCT) (Heidelberg Engineering SD-OCT system, Spectralis OCT 14603, software 6.5). Acceptable OCT scans were those with signal-to-noise ratio >25 decibels (dB).

Written informed consent was obtained from all the participants, and Institutional Ethics Committee clearance was obtained.

Statistical analysis was performed using the SPSS software version 25 (IBM Corp., Armonk, N.Y., USA). P < 0.05 was considered statistically significant.

Results

Forty-nine eyes of 27 patients, of which 22 (44.9%) were male and 27 (55.1%) were female, were included in this study. The age of the patients ranged from 41 to 83 years, with a mean age being 56.22 ± 10.1 years [Figure 1]. The difference in age, BCVA, central corneal thickness (CCT), and IOP values between the three groups was not statistically significant [Table 1].

Figure 1

Age and sex distribution

Table 1

Age, best-corrected visual acuity, intraocular pressure, and cup-disc ratios

Parameter	Mean (n=49)	POAG
		Mild (n=14)	Moderate (n=9)	Severe (n=26)
Age (years)	56.22±10.1	52.07±8.61	57.67±11.73	57.96±9.98
BCVA (logMAR)	0.08±0.13	0.05±0.1	0.07±0.2	0.09±0.12
CDRH	0.65±0.19	0.6±0.21	0.58±0.2	0.7±0.18
CDRV	0.67±0.18	0.61±0.18	0.62±0.2	0.72±0.18
CCT (µm)	529±17.32	534.43±13.33	531±14.09	526.81±19.95
IOP (mmHg)	19.24±5.3	19.64±2.98	17.44±3.81	19.65±6.57

BCVA: Best-corrected visual acuity, IOP: Intraocular pressure, CCT: Central corneal thickness, POAG: Primary open-angle glaucoma, CDRH: Cup disc ratio horizontal, CDRV: Cup disc ratio vertical

The average visual field index (VFI), mean deviation (MD), and pattern standard deviation (PSD) were 65.35 ± 29.95%, −11.84 ± 8.88 dB, and 7.77 ± 4.05 dB, respectively. The Chi-square test comparing VFI, MD, and PSD in eyes with mild, moderate, and severe glaucoma showed a statistically significant difference in all the three parameters between the three groups (P < 0.001) [Table 2]. Post hoc Tukey test comparing VFI and MD revealed a statistically significant difference between mild versus severe (P < 0.001) and moderate versus severe groups (P = 0.002, P < 0.001) but not between mild and moderate groups. A similar analysis done for PSD showed a statistically significant difference among all the groups [Table 3]. There was a significant positive correlation between MD and VFI with a correlation coefficient (r) of 0.9218 [Figure 2]. There was a nonlinear correlation between PSD and VFI [Figure 3].

Table 2

Comparison of visual field parameters in the glaucoma sub-groups

Visual field parameter	POAG			Chi square test (P)
	Mild	Moderate	Severe
VFI (%)	94.29±4.14	75.78±25.01	46.16±24.95	<0.001
MD (dB)	−2.85±1.72	−7.03±3.11	−18.37±6.98	<0.001
PSD (dB)	3.72±2.49	7.29±3.77	10.11±2.96	<0.001

VFI: Visual field index, MD: Mean deviation, PSD: Pattern standard deviation, POAG: Primary open-angle glaucoma

Table 3

Post hoc - Tukey test for visual field indices

Visual field parameter	Mild versus moderate difference (P)	Mild versus severe difference (P)	Moderate versus severe difference (P)
VFI (%)	18.51 (0.115)	48.13 (<0.001)	29.62 (0.002)
MD (dB)	4.17 (0.177)	15.51 (<0.001)	11.34 (<0.001)
PSD (dB)	−3.57 (0.021)	−6.39 (<0.001)	−2.82 (0.049)

VFI: Visual field index, MD: Mean deviation, PSD: Pattern standard deviation

Figure 2

Correlation of mean deviation with visual field index (r = 0.9218, P < 0.00001)

Figure 3

Correlation of pattern standard deviation with visual field index (r = 0.4636,P = 0.000808)

On OCT, the mean global RNFL thickness was 62.45 ± 17.75 μm which is significantly lower than normal. The percentage RNFL loss was 36.1%, 34.2%, 35.4%, 33%, 45.1%, 51.3%, and 30% in the global, TS, NS, nasal, NI, TI, and temporal sectors, respectively. Mean RNFL thickness in the three glaucoma groups was calculated [Figure 4 and Table 4]. In all three groups, maximum RNFL loss occurred in the TI sector followed by the NI. There was a statistically significant difference in the amount of RNFL lost with severity of POAG in all except the superior and temporal sectors [Figure 5 and Table 5].

Figure 4

Mean retinal nerve fiber layer thickness: Normative database, mild moderate and severe group

Table 4

Mean retinal nerve fiber layer thickness normative database mild, moderate, and severe primary open-angle glaucoma

Sector	RNFL thickness (µm) (normative database)[7]	POAG			P
		Mild	Moderate	Severe
Global	97.8±8.6	72.93±15.86	68.78±18.61	54.62±14.95	0.002
Temporal superior	128±19.6	96.93±22.61	90.44±10.5	75.46±32.1	0.057
Nasal superior	112±22.4	77.29±24.04	76.67±17.66	68.27±22.42	0.391
Nasal	81.5±12.5	62.57±17.14	60.67±19.56	48.12±13.12	0.013
Nasal inferior	109.5±21.5	72.29±16.12	65.44±19.03	51.65±17.71	0.003
Temporal inferior	148.8±16.9	91.64±26.95	81.89±37.24	59.08±33	0.01
Temporal	70.5±9.8	54.79±13.26	51.56±16.15	45.54±13.43	0.126

POAG: Primary open-angle glaucoma, RNFL: Retinal nerve fiber layer

Figure 5

Mean retinal nerve fiber layer loss (μm) in the three groups

Table 5

Mean retinal nerve fiber layer loss in each glaucoma sub-group

RNFL sector	POAG, µm (%)			P
	Mild	Moderate	Severe
Global	24.89 (25.44)	29.02 (29.67)	43.18 (44.15)	0.002
Temporal superior	31.79 (24.83)	37.56 (29.34)	52.54 (41.05)	0.061
Nasal superior	34.71 (24.83)	35.33 (31.54)	43.73 (39.04)	0.391
Nasal	18.91 (23.20)	20.83 (25.56)	33.39 (40.97)	0.013
Nasal inferior	37.18 (33.95)	44.06 (40.24)	57.85 (52.83)	0.002
Temporal inferior	57.74 (38.88)	66.91 (44.97)	89.72 (60.30)	0.012
Temporal	15.75 (22.34)	18.94 (26.87)	24.96 (35.40)	0.128

POAG: Primary open-angle glaucoma, RNFL: Retinal nerve fiber layer

A negative correlation was found between VFI and RNFL loss which was statistically significant (P < 0.05) in the global, NI and TI sectors in mild POAG and in all quadrants except the NS in moderate POAG. All sectors showed statistically significant negative correlation in severe POAG [Table 6]. In all the three groups, a positive correlation was found between MD and RNFL loss. It was statistically significant in all the sectors only in severe POAG [Table 7]. A poor correlation was observed between PSD and RNFL loss [Table 8].

Table 6

Correlation of visual field index with retinal nerve fiber layer loss

RNFL sector	Correlation r (P)
	Total	Mild	Moderate	Severe
Global	−0.73 (<0.001)	−0.547 (0.043)	−0.836 (0.005)	−0.677 (<0.001)
Temporal-superior	−0.614 (<0.001)	−0.036 (0.903)	−0.813 (0.008)	−0.649 (<0.001)
Nasal-superior	−0.377 (0.008)	−0.194 (0.507)	−0.575 (0.105)	−0.378 (0.057)
Nasal	−0.618 (<0.001)	−0.374 (0.187)	−0.808 (0.008)	−0.533 (0.005)
Nasal-inferior	−0.702 (<0.001)	−0.569 (0.034)	−0.823 (0.006)	−0.58 (0.002)
Temporal-inferior	−0.661 (<0.001)	−0.696 (0.006)	−0.696 (0.037)	−0.581 (0.002)
Temporal	−0.527 (<0.001)	−0.208 (0.476)	−0.799 (0.01)	−0.447 (0.022)

RNFL: Retinal nerve fiber layer

Table 7

Correlation of mean deviation with retinal nerve fiber layer loss

RNFL quadrant	Correlation r (P)
	Total	Mild	Moderate	Severe
Global	−0.697 (<0.001)	−0.589 (0.027)	−0.347 (0.36)	−0.759 (<0.001)
Temporal-superior	−0.651 (<0.001)	−0.186 (0.525)	−0.052 (0.895)	−0.777 (<0.001)
Nasal-superior	−0.362 (0.011)	−0.214 (0.462)	−0.177 (0.648)	−0.44 (0.025)
Nasal	−0.541 (<0.001)	−0.396 (0.161)	−0.243 (0.529)	−0.535 (0.005)
Nasal-inferior	−0.667 (<0.001)	−0.446 (0.11)	−0.317 (0.406)	−0.658 (<0.001)
Temporal-inferior	−0.635 (<0.001)	−0.781 (0.001)	−0.495 (0.175)	−0.61 (0.001)
Temporal	−0.494 (<0.001)	−0.447 (0.109)	−0.265 (0.491)	−0.567 (0.003)

RNFL: Retinal nerve fiber layer

Table 8

Correlation of pattern standard deviation with retinal nerve fiber layer loss

RNFL quadrant	Correlation r (P)
	Total	Mild	Moderate	Severe
Global	0.468 (0.001)	0.521 (0.056)	0.457 (0.216)	−0.007 (<0.975)
Temporal-superior	0.158 (0.277)	−0.045 (0.879)	−0.002 (0.997)	−0.147 (0.473)
Nasal-superior	0.303 (0.034)	0.366 (0.199)	0.192 (0.62)	0.224 (0.272)
Nasal	0.369 (0.009)	0.326 (0.256)	0.298 (0.436)	−0.026 (0.901)
Nasal-inferior	0.438 (0.002)	0.435 (0.12)	0.504 (0.166)	−0.062 (0.765)
Temporal-inferior	0.527 (<0.001)	0.529 (0.052)	0.793 (0.011)	0.126 (0.539)
Temporal	0.246 (0.089)	0.307 (0.285)	0.306 (0.423)	−0.15 (0.464)

RNFL: Retinal nerve fiber layer

Discussion

Forty-nine eyes with established glaucoma were included in this study. There was no significant difference in age or sex distribution among the three glaucomatous groups. The CCT values progressively decreased and IOP increased with the severity of glaucoma. Although this is consistent with a study done by Moghimi et al.,[8] the results in our study were not statistically significant. Since several subjects in our study had already been diagnosed with POAG and were on treatment, the IOP values included were not the baseline IOP that the patient presented with. In patients with relatively higher IOP values, the readings may be explained by inadequate treatment or poor compliance with medication. Of the 27 study participants, 22 had established glaucoma in both eyes. In four patients, the second eye had normal optic disc findings; IOP, GHT, and RNFL thickness within normal limits. These eyes were not included in our study but followed up to detect the appearance of glaucomatous changes. One patient had central retinal vein occlusion.

VFI values were significantly different in each group in our study, which was also found in another study that concluded that VFI is a reliable parameter for staging the severity of glaucoma.[9] There was statistically significant difference in VFI and MD between the mild versus severe and moderate versus severe groups, but not between the mild and moderate groups. This may be because the number of patients in moderate group (n = 9) were lower than in other groups. As multiple parameters are considered in perimetric staging, categorization may be difficult and less accurate. A strong positive correlation (r = 0.9218, P < 0.00001) occurred in this study between VFI and MD. Similar results (r = 0.956) were reported by Iutaka et al.[10] A study of 60 patients found a significant correlation between VFI and MD as well as VFI slope and MD slope.[11] Dorairaj et al.[12] found an excellent correlation between MD and VFI (r2 = 0.95) in moderate-to-severe glaucoma.[13] The comparison of VFI and PSD in our study showed a nonlinear correlation. PSD is not useful for detecting progression since it peaks at an MD of about − 12 dB. As the glaucomatous damage increases further, it reverses toward 0 dB such that patients with mild and advanced glaucoma have similar PSD values.[14]

The difference in RNFL thickness between the mild, moderate, and severe glaucoma groups in the TS, NS, nasal, NI, and TI sectors in our study was significant, implying that the susceptibility of each sector to glaucomatous damage differs with the stage of the disease. This corresponds with the study done by Kaw et al.[4] Zivkovic et al.[15] reported a significant difference in RNFL thickness in superior, inferior, nasal, and temporal quadrants in the different stages of glaucoma (χ2= 273.36, DF = 3, P < 0.001). A statistically significant difference in RNFL thickness between normal and glaucomatous patients as well as early, developing and late stages of glaucoma was present in a study by Liu et al.[16]

The average RNFL loss in our study was 25.44%, 29.67%, and 44.15% in the mild, moderate, and severe groups, respectively. Similar values of 25% RNFL loss in early glaucoma and 48% in severe glaucoma were noted by Sihota et al.[17] El-Naby et al.[18] reported a 14.9%, 25.1%, and 37.2% RNFL loss in the three groups while another study showed a loss of 22.21%, 22.06%, and 42.72% in the three stages.[4] The TI followed by the NI sector was most susceptible to glaucomatous damage in all three groups in our study. Wu et al.[19] reported that the highest correlation between RNFL thickness and visual field defect values was in the TI sector. Zivkovic´et al.[15] found that the temporal quadrant was significantly thinner than all the other quadrants followed by the nasal quadrant while others[202122] found that the inferior quadrant was the thinnest. Sehi et al.[21] found additional average and superior quadrant RNFL thinning in the later stages of glaucoma. The temporal sectors in our study were found to show least thinning in all three glaucoma groups similar to other studies which compared RNFL thickness in normal and glaucomatous eyes.[18] The TI and NI sectors can be monitored for progression in glaucoma patients as these sectors show a significant difference in RNFL loss, especially between the mild and severe perimetric stages. Leung et al.[23] reported that the TI quadrant was also most frequently associated with glaucoma progression.

In the present study, as the VFI decreased, RNFL thickness decreased, that is, the RNFL loss increased. This finding is in agreement with those of Liu et al.[16] and Kimura et al.[11] A nonlinear relationship was found between VFI and estimated RGC counts, with VFI underestimating the amount of RGC loss in early glaucoma by Marvasti et al.[22] Rao[24] reported that VFI showed maximum correlation with average and superior RNFL thicknesses in mild and moderate groups contrary to our study, where it correlated best with RNFL in the TI sector in mild POAG, all the sectors in moderate except NS and all except the TS and temporal sectors in severe POAG.

In our study, there was a positive correlation between MD and RNFL loss as seen in studies by Kaw et al. and Taliantzis et al.[420] who found strong (r = 0.73) and moderate (r = 0.58) correlations of RNFL thickness with MD and loss variance (r = 0.51; r = 0.53), respectively. MD was found to have a significant positive correlation with RNFL thickness by other studies.[182526] In a study by Choi et al.,[27] most of the RNFL sectors had a significant positive correlation with MD while in advanced glaucoma, the superior RNFL quadrant showed the greatest positive correlation (r = 0.452). In our study, MD in the mild group showed a good positive correlation with RNFL loss in the global (r = 0.589, P = 0.027) and TI sectors (r = 0.781, P = 0.001) while in severe glaucoma, all the sectors showed a statistically significant correlation.

An increase in PSD value in our study was associated with a corresponding increase in mean RNFL loss but was not statistically significant. In studies done by Kaw et al.[4] and Arun et al.,[25] a significant association was found between PSD and RNFL thickness. Parisi et al.[26] found a highly significant correlation (r = 0.663, P < 0.001) between average RNFL thickness and corrected PSD. The poor statistical correlation in our study may due to the relatively higher number of severe POAG patients in whom PSD trend reverses when MD worsens beyond −12 dB. Taliantzis et al.[20] reported strong correlations between average RNFL thickness and visual field indices only in those with progressive structural changes (P < 1%) on OCT, although, sectoral RNFL thickness was considered a reliable parameter. In contrast, our study showed that all three groups had a moderate to excellent correlation of global RNFL thickness with visual field indices which corresponds with the findings of El-Naby et al.[18] and Zangwill et al.[28]

Conclusion

Peripapillary RNFL thickness is a useful measure to discriminate between mild and severe stages of POAG. The TI sector followed by NI is most susceptible to glaucomatous damage. RNFL thickness in these sectors can be used to diagnose early POAG.

The limitations of our study are a small sample size (49 eyes) and the difference in number of patients in each glaucoma sub-group which may limit the extent to which the results of this study can be extrapolated to the general population. Further studies to include more data from Asian populations would enable the creation of a normative database and a reference range for normal RNFL thickness values in each sector that is applicable to the Indian population.

Another limitation of this study is that although RNFL thickness measurements by different instruments follow the anatomic distribution, the machines may generate significantly different total and sectoral thickness values.[29] This must be taken into consideration while comparing OCT scans from different instruments.

RNFL thickness values in this study along with more longitudinal studies and creation of a new algorithm and scoring system for the detection of progression can be used to categorize patients into the various stages of POAG based on RNFL thickness.

This will allow the early detection of disease progression so that further loss of visual function can be prevented.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.