Long term effect of panretinal photocoagulation on retinal nerve fiber layer parameters in patients with proliferative diabetic retinopathy.

PURPOSE: This study aimed to evaluate the long-term effect of panretinal photocoagulation (PRP) on the retinal nerve fiber layer (RNFL) in patients with proliferative diabetic retinopathy (PDR).
METHODS: This was a prospective longitudinal cohort study examining 42 eyes of 42 patients with PDR undergoing PRP. Peripapillary RNFL thickness (RNFLT) was measured using spectral-domain optical coherence tomography at baseline, 1 year, and 3 years following PRP.
RESULTS: The mean "average RNFLT" was 89.88 ± 14.26 μm at baseline, 85.75 ± 11.36 μm at 1-year follow-up, and 83.33 ± 11.96 μm at 3-year follow-up. There was a statistically significant difference in the average RNFL thickness at baseline and 1 year and 3 years after PRP. At 1-year follow-up, superior, inferior, and nasal RNFL measurements reduced significantly from baseline (P < 0.01). The reduction in RNFL remained statistically significant for superior and inferior quadrants 3 years after PRP.
CONCLUSION: PRP causes a reduction in RNFL thickness until 3 years after the procedure. Caution should be exercised while interpreting peripapillary RNFL thickness scans in patients who have undergone PRP for diabetic retinopathy. Copyright:

Introduction

Diabetic retinopathy (DR) is a microangiopathy affecting up to 53% of patients with diabetes mellitus.[12] Macular edema and proliferative DR (PDR) are two of the most important causes of vision loss in patients with diabetes. Literature describing the association between glaucoma and diabetes has been conflicting, but recent evidence supports the increased prevalence of glaucoma in patients with diabetes mellitus.[34] A number of studies have also shown loss of retinal nerve fiber layer (RNFL) in diabetes mellitus without associated retinopathy.[56]

Panretinal photocoagulation (PRP) is the gold standard of treatment for proliferative DR. PRP has been shown to induce thinning of the RNFL up to 2 years following treatment.[78910111213] Several imaging methods are available to assess RNFL: scanning laser polarimetry (SLP), confocal scanning laser ophthalmoscopy, and optical coherence tomography (OCT). Among all the glaucoma diagnostic technologies, spectral-domain OCT (SD-OCT) has been shown to have the highest diagnostic accuracy for the detection of glaucomatous changes.[1415161718] Initial studies have employed the use of SLP or time-domain OCT (TD-OCT) for the detection of post-PRP RNFL changes;[7891011] however, there are only a few studies employing the use of SD-OCT to detect post-PRP RNFL changes, albeit with a shorter follow-up.[1213]

Through this study, we sought to evaluate the long-term impact of PRP on RNFL changes using SD-OCT in patients with PDR.

Materials and Methods

This prospective longitudinal cohort study involved 48 eyes of 48 patients with newly diagnosed PDR. The study was approved by the institutional ethics board and conformed to the ethical standards stated in the 1964 Declaration of Helsinki. Informed consent was obtained from all individuals before enrollment.

The study finally included 42 patients with newly diagnosed PDR without high-risk characteristics in the form of vitreous hemorrhage, and six patients with vitreous hemorrhage or tractional retinal detachment (TRD) were excluded from the study. Detailed assessment involved recording medical history and family history of systemic diseases, measuring intraocular pressure with Goldman Applanation tonometer and detailed stereoscopic examination at slit lamp using 90D lens and fundus examination using direct and indirect ophthalmoscope. RNFL assessments were performed with SD-OCT and Cirrus HD-OCT (Carl Zeiss Meditec, Dublin, CA, USA) after pupillary dilation. All these patients underwent clinical examination, fundus photography, and SD-OCT at baseline, 1-year follow-up, and 3-year follow-up. The following were the exclusion criteria for our study: media haze precluding disc evaluation and previous PRP; refractive error >5 diopters of hyperopia or myopia; glaucoma; factors predisposing to raised IOP such as family history; pigment dispersion; exfoliation; occludable angles; or glycosylated hemoglobin (HbA1C >7%). Patients who developed a recurrence of PDR, macular edema, vitreous hemorrhage, TRD, neovascularization of iris/angle, or required additional sessions of PRP for control of PDR, were eliminated from the analysis to prevent confounding of results.

The RNFL imaging was performed using the Optic Disc Cube protocol that images an area of 6 mm × 6 mm with 200 × 200 scans (200 A-scans per B-scan; 200 B-scans) at the optic disc region, analyzes the RNFL thickness at each scan point, and constructs an RNFL map. The in-built algorithm identifies the optic disc center using a graph-based method, and a circle of diameter 3.46 mm is positioned automatically around the disc center to generate average and clock-hour parapapillary RNFL measurements. The average RNFL thickness and the thickness of each quadrant were evaluated. Only scans, of one randomly chosen eye, with signal strengths ≥07 were used for analysis.

PRP was performed using a frequency-doubled Nd: YAG-532 nm laser (532 nm; VISULAS 532; Zeiss, Carl Zeiss Meditec, Jena, Germany) by a single surgeon with the help of wide field Mainster PRP lens (Ocular Instruments, Inc, Bellevue, WA, USA) with a spot size of 300 μm and a total of 1800 spots. The power was set enough to cause grayish–white burns with duration of 0.1 s/spot laser, completed in three sessions, each session 1 week apart. The temporal border of fovea was delineated, with three or four rows of laser burns placed two disc diameters away from fovea. Laser spots were placed approximately one burn width apart.

The data were stored on a computerized database and analyzed using SPSS 20.0 for windows. Repeated measures were tested with the Bonferroni correction. Pearson's Chi-square test was used in the statistical analysis where appropriate, and P < 0.05 was considered statistically significant.

Results

The mean age of the included patients was 56.4 ± 11.3 years; there were 38 males and 4 females. Baseline best-corrected visual acuity was 0.44 ± 0.32 logMAR units. About 4.7% of patients had type 1 DM and 95.3% had type 2 DM. The mean duration of diabetes was 18.09 ± 9.6 (4–36) years. The mean IOP at baseline was 16.26 ± 3.34 (11–32) mmHg. There was no statistical difference noted between baseline and follow-up IOP measurements.

The mean “average RNFLT” was 89.88 ± 14.26 μm at baseline, 85.75 ± 11.36 μm at 1-year follow-up, and 83.33 ± 11.96 μm at 3-year follow-up. There was a statistically significant difference in average RNFL thickness between the baseline and 1-year RNFL measurements. This difference remained significant for average RNFL thickness (RNFLT) at 3-year follow-up [Figure 1]. Average RNFL thickness reduced by 4.61% at 1 year and by 7.28% at 3 years after panretinal photocoagulation [Table 1].

Figure 1

Change in retinal nerve fiber layer thickness following panretinal photocoagulation

Table 1

Longitudinal changes in retinal nerve fiber layer measurements after pan retinal photocoagulation

RNFL	Baseline (0) (n=42)	1 year (1) (n=42)	3 year (3) (n=42)	Difference (%)P 0-1	Difference (%)P 0-3	Difference (%)P 1-3
Average (µm)	89.88±14.26	85.75±11.36	83.33±11.96	4.14 (4.61) <0.01	6.54 (7.28) <0.01	2.40 (2.80) <0.01
Inferior (µm)	112.48±24.44	105.52±20.35	101.60±24.28	6.95 (6.18) <0.01	10.88 (9.67) <0.01	3.93 (3.72) 0.443
Superior (µm)	111.31±20.01	105.38±15.41	103.31±16.74	5.93 (5.33) <0.01	8.00 (7.19) 0.01	2.07 (1.96) 0.87
Nasal (µm)	69.43±12.96	66.24±11.64	66.50±10.60	3.19 (4.59) <0.01	2.92 (4.20)] 0.49	0.263 (0.40) 1
Temporal (µm)	66.64±13.89	64.33±11.97	63.17±14.12	2.31 (3.47) 0.22	3.48 (5.22) 0.38	1.17 (1.82) 1

RNFL: Retinal nerve fiber layer

Quadrant-wise analysis showed that inferior and superior RNFLT thinned progressively after PRP over a period of 3 years. There was a statistically significant reduction in superior and inferior RNFL thickness from the baseline at 1-year follow-up and 3-year follow-up (P < 0.01). For the nasal quadrant, the RNFLT showed a statistically significant decline at 1-year follow-up, but stabilized thereafter at 3 years. For the temporal quadrant, there was no significant change noted in RNFLT at 1 and 3 years of follow-up.

Discussion

Our study showed that average RNFL measurements showed a significant decline for up to 3 years after PRP. Superior and inferior quadrants showed a continued reduction in RNFL thickness 3 years after PRP. Nasal RNFL measurements showed an initial decline at 1 year after PRP but stabilized thereafter. Temporal quadrant RNFL measurements did not show any significant change at 1 or 3 years after PRP.

PRP has been found to cause an increase in the RNFL thickness in the short term, presumably related to laser-induced retinal inflammation and edema.[101213] This early post-PRP increase in RNFL has been shown to last until 1–6 months followed by a gradual decline. Whereas in the study by Yazdani et al.,[13] the RNFL measurements returned to baseline at 3 months post-PRP, the data published by Kim et al. showed that RNFL measurements remained above baseline even at 12 months post-PRP.[12] In the long term, PRP has been found to have a negative effect on RNFL thickness which may be attributed to axonal loss secondary to direct or indirect effects of PRP treatment [Table 2].

Table 2

Summary of studies using optical coherence tomography for evaluation of pan retinal photocoagulation induced changes in retinal nerve fiber layer thickness

Authors/journal	Study, number of patients	Type of OCT (TD vs. SD)	Follow-up duration	Results
				Change in average RNFL (change/decrease in RNFL)	Quadrants most affected
Lim et al. ArchOphthalmol 2009	Cross-sectional, 220 eyes	Stratus OCT 3	None, 3 months post-PRP	No significant change	Inferior (significant reduction in RNFL)
Kim et al. Retina 2012	Retrospective, 68 eyes	Stratus OCT 3	24 months	Reduced (4.9 µm)	Superior and inferior
Lee et al. Retina 2013	Prospective 31 eyes	Stratus OCT 3	24 months	Reduced (3.5 µm)	Superior, nasal, inferior
Park et al. K or JOphthalmol 2014	Retrospective, 105 eyes	Stratus OCT 3	12 months	Conventional PRP: Reduced (5.42 µm) Pascal PRP: No change	Superior and inferior
Kim et al. Br JOphthalmol 2014	Prospective, 35 eyes	Cirrus HD-OCT	12 months	Reduced (2.9 µm)	Superior and inferior
Yazdani et al. Optom Vis Sci 2016	Prospective 42 eyes	SD-OCT Topcon	6 months	Reduced (2.4 µm)	Superior
Present study	Prospective	Cirrus SD-OCT	36 months	Reduced (6.55 µm)	Superior and inferior

OCT: Optical coherence tomography, PRP: Pan retinal photocoagulation, RNFL: Retinal nerve fiber layer, SD-OCT: Spectral domain-OCT

Like most diagnostic technologies in the field of ophthalmology, OCT has evolved from its previous ID system to SD system. The SD-OCT has been shown to produce a better scan quality, preperimetric RNFL defects, and have excellent reproducibility.[1920] Initial studies using TD-OCT showed a significant decline in average RNFL thickness 2 years after PRP.[910] These studies showed an initial increase in RNFL until 6 months after PRP, with measurements returning to baseline at 1 year and then a statistically significant decrease at 2-year follow-up.[910] However, there have been few more recent studies using SD-OCT for evaluation of post-PRP RNFL thickness. Yazdani et al. recently published a study evaluating the effect of PRP on RNFL thickness in 42 eyes with a shorter follow-up of 6 months.[13] Kim et al. studied the effect of PRP on RNFL with a 12-month follow-up period.[12] Both the studies showed a significant reduction in average RNFL measurements following PRP, like ours. Our study shows that the effect of PRP on RNFL is long term and RNFL thickness may continue to decline for up to 3 years following the procedure. The difference in results at 1-year follow-up could possibly be accounted for by the different OCT systems used in the earlier studies.

Laser burns may expand, and often fuse, even until the 4th year of posttreatment follow-up, and the annual rate for laser burn expansion has reported to be as high as 16.5%.[9] Progressive destruction in RNFL and the consequent decrease in its thickness, thus occurs as demonstrated also in our results.

We found that in both superior and inferior quadrants, the RNFL thickness was decreased. There was statistically significant RNFL reduction noted in these quadrants 1-year post-PRP. This trend continued until 3 years post-PRP; however, the difference between the 1-year and 3-year measurements was not statistically significant. These results are in line with previously published data.[1213] Researchers have suggested that superior retina has more microaneurysms and acellular capillaries and greater alterations in the retinal blood flow possibly accounting for the said changes. On the contrary, a few studies have found the inferior quadrant as the most affected one; however, there is no satisfactory explanation for this result.[8] Aging has also been related to decay in RNFL, particularly after 50 years, with a described negative slope of 0.16 μm/year in average RNFL and 0.23 μm/year in the superior quadrant.[21] However, the significant RNFL change noticed in our study is certainly in excess of “permitted” age-related loss.

The effect of diabetes on RNFL has been evaluated in some studies. It has been suggested that diabetes may induce thinning of RNFL in the absence of PRP.[2223] However, this has not been consistently supported by studies using OCT as an RNFL measurement tool.[2425] Lim et al. did not find a significant difference in RNFL thickness between diabetic and control eyes.[8] Similarly, Park and Jee did not find a significant change in RNFL in patients with DM through 1-year follow-up.[11]

Laser photocoagulation treats DR by a process of “gentle destruction.”[26] High-grade retinal laser burns cause a pyramid of retinal damage with the base at the RPE and apex in the retina, eliminating large areas of oxygen-demanding photoreceptors. Correspondingly, a stronger laser can damage the entire retinal layer and ganglion cells eventually leading to axonal atrophy and RNFL thinning.[27] Park and Jee compared the effect of conventional PRP and pattern scanning laser (PASCAL) photocoagulation on RNFL. The PASCAL system reduces the total pulse energy effectively by lessening the laser shot duration. They found that conventional PRP produced significant RNFL thinning, whereas the PASCAL system was protective against RNFL loss.[11]

One of the most important applications of objective RNFL measurements has been in monitoring the progression of glaucomatous changes.[28] OCT has been shown to be more likely to detect progression in preperimetric and early glaucoma cases, as compared to stereo photographs and visual field index.[29] Furthermore, RNFL thickness parameters perform significantly better than ONH and macular thickness parameters for detection of change.[30]

A major design flaw of our study is that there is no concurrent control group of diabetics with PDR who have not had a PRP, especially since Takahashi et al. have reported that RNFL thinning in diabetes increases with worsening disease severity.[6] It would be ideal to have a longitudinal control group with similar DR changes without PRP; however, ethical issues limit that.

Despite that, the clinical relevance of this study lies in delineating that we must exercise caution in interpreting RNFL measurements with lasered PDR, over time. Further studies evaluating visual field changes would be necessary to understand the functional relevance of RNFL loss in both, glaucomatous and nonglaucomatous, eyes. The strengths of our study lie in it being a prospective and longitudinal study, with a 3-year follow-up showing the effect of PRP on RNFL thickness using SD-OCT.

We conclude that PRP may induce significant RNFL thinning until 3-year postprocedure in eyes without preexisting glaucoma. Caution should be exercised while interpreting peripapillary RNFL thickness scans in patients who have undergone PRP for PDR.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.