Photobiomodulation Therapy for Age-Related Macular Degeneration and Diabetic Retinopathy: A Review

Introduction

Patients with diabetic retinopathy (DR) and age-related macular degeneration (AMD) are at increased risk of vision loss. With an ageing population increasingly afflicted by diabetes, the prevalence of both conditions is expected to increase. An estimated 288 million people will have AMD by 2040 and 191 million people will have DR by 2030.^1,2 Up to one third of DR patients develop diabetic macular edema (DME), a manifestation of DR that produces loss of central vision.³

Treatment for DME began with focal photocoagulation and evolved to anti-Vascular Endothelial Growth Factor (anti-VEGF) injections. The advent of the anti-VEGF era led to landmark trials demonstrating efficacy of these injections in eyes with DME, however, entry criteria was visual acuity (VA) ≤ 20/40.⁴ This left the management of patients with DME and VA ≥ 20/25 Snellen (or ≥ 79 letters) unknown. The DRCR Retina Network Protocol V aimed to compare VA and retinal thickness after three different treatments: initiation of monthly aflibercept, focal laser therapy with deferred aflibercept, or initial observation with deferred aflibercept. At 2 years, patients in each treatment group had similar VA outcomes. Although 66.0% did not require further treatment, 12.0% of this group continued to lose ≥ 5 letters of vision.⁵ These results reveal an absence of preventative treatment for this subset of DME patients. While focal laser, micropulse diode laser, intravitreal steroid injections, and anti-VEGF injections are effective treatments, they are invasive and costly. The DRCR Retina Network Protocol I, which followed 346 participants initially treated with focal laser and ranibizumab for DME, disclosed the five-year cumulative probability of worsening was 18% for eyes with non-proliferative diabetic retinopathy and 31% for eyes with proliferative diabetic retinopathy.⁶

The anti-VEGF era also revolutionized treatment for a subset of patients with advanced exudative age-related macular degeneration (exudative AMD). However, no such treatments exist to control the progression of non-exudative AMD. Although the AREDS/AREDS2 vitamin formulation shows promise in individuals with intermediate AMD, its efficacy for all patients is being called into question: some authors now suggest patients should be selected for AREDS therapy based on genetic testing.^7,8

Photobiomodulation (PBT) is an emerging therapy that shows potential as a treatment for patients with DME and non-exudative AMD. PBT consists of serial, brief illuminations with Near Infrared Spectrum light (600–1000 nm) from a laser or, more commonly, a light emitting diode (LED). This review article explores potential mechanisms of action and current clinical activity around PBT treatment in these applications.

Methods

This narrative review was performed with the primary objective of this study to summarize available literature on photobiomodulation therapy employed for any stage of AMD as well as DR including DME. Preclinical studies, case reports, case series, observational studies, and randomized controlled trials were considered for inclusion.

The National Clinical Trial (NCT) database was searched for “Photobiomodulation” and “Age Related Macular Degeneration” as well as “Photobiomodulation” and “Diabetic Retinopathy OR Diabetic Macular Edema.” PubMed was queried with the following search query: (((Age-Related Macular Degeneration) OR (exudative AMD) OR (neovascular AMD) OR (nonexudative AMD) OR (dry AMD) OR (diabetic macular edema) OR (diabetic retinopathy)) AND ((photobiomodulation) OR (PBT) OR (PBM) Or. (photobiomodulation therapy))). Additional articles were identified from a manual search of reference lists within included articles.

Studies were considered and classified dichotomously. Inclusion criteria were: English language, description of PBT for any stage of AMD or any stage of DR including DME. The full text of each article was reviewed by JCM and MWR. The subjects, eyes, device used, wavelength, dose, delivery parameters, and results were extracted. If applicable, NCT number was provided. For trials pending results, the date and outcome measures were reported.

Results

After screening 60 candidate publications, a total of 23 studies were included. 37 Studies were eliminated as they did not report on PBT as an intervention (n=30), combined therapies (n=5), non-English language (n=2). Fourteen studies examining the application of PBT for AMD can be found in Table 1. Nine studies examining the application of PBT for DME can be found in Table 2.

Table 1 Summary of Available Articles Describing the Application of Photobiomodulation for Age-Related Macular Degeneration. Subjects, Eyes, Wavelength Treated, Device Used, Dose and Delivery Parameters as Well as Results of the Studies are Presented

Table 2 Summary of Available Articles Describing the Application of Photobiomodulation for Diabetic Retinopathy. Subjects, Eyes, Wavelength Treated, Device Used, Dose and Delivery Parameters as Well as Results of the Studies are Presented

Discussion

Mechanism of Action

Proposed mechanisms of action for PBT include enhanced photoreceptor mitochondrial function, counteracting inflammation, and enhanced supporting cell function.

In PBT, near-infrared spectral light (660–1000 nm) induces a photochemical reaction on the cellular level, starting with complex IV in the electron transport chain.^9–14 Complex IV colocalizes to retinal layers rich in mitochondria such as the nerve fiber layer, around retinal ganglion cell nuclei, the inner/outer plexiform layers, the photoreceptor inner segment and basal surface of retinal pigment epithelium (RPE) cells.¹⁵ Complex IV is directly activated by NIR light, increasing respiratory chain function and mitochondrial activity.^9,16 Photic stimulation uncouples nitric oxide, which would otherwise inhibit complex IV, thereby indirectly increasing mitochondrial activity.^12,14,17 The freed nitric oxide also triggers downstream cascades to increase anti-antioxidant production, anti-apoptotic pathways, and cellular metabolism.¹⁸ Within photoreceptors, these changes amount to globally improved photoreceptor performance as measured by ATP production and electroretinogram (ERG).^{11,12,14,19,20} Although enhanced photoreceptor function shown in ERG is reported in diabetic mouse models, the same models suggest the complex IV direct and indirect pathways may not mediate the effect.^21–23 Confounders, such as severity of hyperglycemia, might explain this result. It is possible that the PBT’s mechanism of action is specific to disease state, triggering pathways in a patient DME which may not be shared with those triggered in a patient with AMD. Further basic science studies might clarify if disease specific pathways are present.

The anti-inflammatory effect of PBT might be mediated across multiple pathways. Murine models for AMD and DME agree that PBT reduces C3 and C2 complement expression within the outer retina.^20,23–25 Other studies upheld and expanded this observation, showing C3 presence shifted from the photoreceptors to RPE/Bruch’s membrane.^20,26 This could be seen as the retina is not only expressing less early inflammatory triggers, but also clearing them. Studies have also demonstrated inhibition of superoxide as well as increased expression of the antioxidant mitochondrial antioxidant manganese superoxide dismutase (mnSOD).^23,25 Others have shown decreased signals of cellular inflammation such as Glial Fibrillary Acidic Protein (GFAP), Vimentin, Cyclooxygenase (COX), calcitonin, Tumor Necrosis Factor (TNF), 4 hydroxyneal, and Intracellular Adhesion Molecules (ICAMs).^20,23,24,26 These studies suggest a net anti-inflammatory effect of PBT which may be mediated by complement downregulation.

Supporting cells in the retina also react to PBT in both diabetic and AMD murine models. Müller cells show a decrease in Vimentin and GFAP, which are cell-specific markers for stress and inflammation, and possibly a prolonged lifetime.^22,24,27 Tang et al demonstrated rescue of retinal ganglion cells, while Fuma et al demonstrated PBT increased phagocytic activity of RPE cell lines.^25,28 Supporting cells are crucial for a healthy retina and evidence of their rescue appears positive.

Finally, PBT therapy can alter retinal gene expression. Natoli et al analyzed the change in expression of 175 neuroprotective retinal genes and ncRNAs after five days of exposure to PBT therapy. 126 of these entities increased expression as a result.²⁹ This effect is potentiated when combined with dietary saffron intake as suggested confirmed in other studies.^29,30 The effect of isolated PBT was further confirmed by Heinig et al who utilize gene expression analysis to identify upregulation of α-crystallins; a result the authors interpreted as indicating rescued mitochondrial function.¹³

Murine models suggest several mechanisms of action across multiple cell types, pathways, and modalities. However, these models are limited for several reasons: the disease states discussed are induced, murine and human physiology differ, and targeting light efficiently into the murine eye is an imperfect science. Due to the difficulty in targeting rodent eyes specifically, there is the possibility that PBT’s effect is may be mediated by the skin and propagated systemically through an unspecified pathway.²¹ Although these results must be considered with the limitations discussed above, they form the basis for clinical trials.

Diabetic Macular Edema: Clinical Applications

In addition to the aforementioned mechanisms, separate experimenters demonstrated a significant reduction in retinal microvascular leakage observed in the inner plexiform layer, inner nuclear layer and outer plexiform layer.^22,31 These would suggest a unique application for PBT in diabetic retinopathy.

Two clinical studies for treatment of DME followed murine studies. Eells et al conducted a randomized prospective study in 10 patients to compare anti-VEGF alone (n=4) or anti-VEGF plus PBT (n=6). 670 nm-light treatment was administered at a transpupillary dose of 4.5 J/cm² three consecutive days per week for 8 weeks.³² At the end of 24 weeks, the authors reported a reduction in central retinal thickness (CRT) of −24±5 microns and improvement of 6±5 letters in the PBT group as compared to a +120 ± 97 microns gain and −3 ± 4 letter VA decrease in the standard treated group (P values not reported). Tang et al reported four consecutive cases with bilateral non-center involving DME. One eye was selected for PBT, while the fellow eye was an untreated control.³³ 670 nm light treatment was administered at a dose of 9.0 J/cm² for 2 to 9 months. After treatment, thickened areas on spectral domain OCT were reduced by a mean of 20.0% (±11.7%) in the treated group as compared to −3.0% (± 8.0%) in the control group.³³

These results offer a new application for PBT in center involving DME (CI-DME) in the absence of a truly preventative treatment for this subset of DME patients. Shen et al delivered 670 nm light at varying energy levels (25, 100, 200 mW/cm²) to 42 eyes for 5 weeks. A maximal reduction of −129±51 μm in central macular thickness (CMT) for patients dosed 100 mW/cm² was observed.²² Although CMT reductions were noted across all dosages, no significant VA gains were noted.

These results prompted a randomized, multi-center, sham-controlled clinical trial, Protocol AE, through the Diabetic Retinopathy Clinical Research Network on patients with CI-DME and good VA.³⁴ Other ongoing trials investigate combinations of ketogenic diet and PBT to attenuate DME as well as treatment refractory DME, however details on these trials are limited.^35,36 Trials such as Shen et al and Protocol AE show a focus on patients with VA ≥ 20/25 and Central Subfield Thickness (CST) ≥ 300 μm as inclusion criteria, which limits generalizability of available results and conclusions.^22,34

At the time of this writing, evidence would suggest CI-DME patients with VA ≥ 20/25 could benefit from PBT. However, the role of PBT in DR treatment remains to be well defined in future trials standardizing devices, treatment patterns, and inclusion criteria must determine this pending the results of current work.

Age-Related Macular Degeneration: Clinical Applications

The mechanisms suggested for PBT function - an increase in mitochondrial function, reduced oxidative damage and decline in complement expression - directly foils the declining mitochondrial function, declining photoreceptor and supporting cell performance, and complement activation that characterize AMD.^37–40

Ivandic treated 348 eyes with VA ≤ 20/20 at various stages of AMD with 670 nm light aimed through the conjunctiva and sclera at the macula. Patients with concomitant vision impairing disease or prior treatment affecting vision and VA >20/20 were excluded. VA by the Snellen chart improved (P<0.00001) as did color vision.⁴⁰ This preliminary work did not provide details on VA measures, CST changes or fundus photography. In the following TORPA 1 trial, ETDRS acuity, contrast sensitivity and fixation stability were tested in 18 AMD eyes after 18 PBT treatments.⁴¹ Eyes that were AREDS stages 2–4 from patients 50 or older with VA ranging 20/200 - 20/20 were included. Eyes with previous or active exudative AMD, other retinal diseases, previous retinal surgery, significant media opacity or contraindications to dilation drops were excluded. At 12 months, VA improved from 0.25 logMar units (20/36) to 0.13 logMar units (20/27). Contrast sensitivity improved while fixation stability remained unchanged.

TORPA 2 expanded the TORPA 1 study to 42 eyes. Inclusion and exclusion criteria were identical with the exception that VA >20/200 was considered. At 3 months, +5.14 letters were recovered compared to baseline.⁴² Similarly, LIGHTSITE I evaluated PBT efficacy in 46 eyes AREDS stages 2–4 with VA 20/40-20/200.⁴³ Major exclusion criteria were shared with the TORPA studies. 670 nm light was delivered in 9 treatment installments at baseline and again at 6 months. VA was noted to increase by up to 4 letters, then decline to baseline 6 months after each treatment, suggesting some level of consistent therapy would be required. VA variation included overlap between groups, and a significant change was noted only when analysis generated a high responder group (≥ 5 letters gained at one month).

In both TORPA 2 and LIGHTSITE I, OCT monitoring showed significant reduction in drusen volume and thickness in the setting of an unchanged CRT and central retinal volume. These observations were made relative to a control group, accounting for changes which may result from the natural course of AMD. Both studies reported improvement in contrast sensitivity and fixation stability. These functional and objective metrics are substantiated in individual case series.^44,45 In both trials and case series, these improvements are noted to be last up to 6 months after PBT.

The ongoing LIGHTSITE II and LIGHTSITE III were launched in response to these findings.^46,47 Eligible subjects for these studies are ≥ 50 years of age, with VA of 20/32-20/100, with eyes containing drusen ≥ 63 μm confirmed by a reading center. The ELECTROLIGHT study, which is also ongoing, seeks to assess the effect of PBT by recording electroretinograms from non-exudative AMD patients.⁴⁸ Notable inclusion criteria are documented as non-exudative AMD, VA of 20/32-20/100, and ≥ 50 years of age. Salient exclusion criteria across these studies include GA, media opacity, ocular surgery, other visually significant disease. Results of the above studies are pending at the time of this writing.

As in studies for DME, the methodology of human trials for AMD show variation in treatment delivery, measurement, and outcomes. Additionally, the absence of control arms and sponsorship by device manufacturers raises further concern for bias. For this reason, Grewal et al exposed 42 eyes (12 controls, 30 intermediate AMD) to a simple easily reproducible Light Emitting Diode based light source for 120 seconds daily for 12 months.⁴⁹ At the end of the study period, no significant anatomic or functional changes were reported.

Though the results are contradictory to other studies, a true comparison is difficult. Paradoxically, Grewal’s methodology – using a different device, control group, and much higher cumulative dose and exposure period – differs significantly from prior work. Nonetheless, future, well-appointed trials standardizing devices, treatment patterns, and inclusion criteria are needed to determine the true effect of PBT in AMD. Moreover, these trials have focused largely on AREDS stages 2–4. Although patients with advanced stages (geographic atrophy or exudative AMD) have been included as a subset, future work may investigate the potential for PBT to serve as an adjunctive treatment in exudative AMD and clarify the potential role it may play in a dedicated GA subset without center involvement.^40,42,43,50

Available studies suggest PBT is best suited to AREDS stages 2–4 AMD, where it might alter complement deposition, attenuate oxidative damage, and enhance mitochondrial function. As with DME, the patient subset is narrow, shared in study design limitations, and high-level evidence remains wanting.

Conclusion

As a non-invasive therapy with no adverse effects reported from human subjects as of this writing, PBT may fill a role in treating niches of DR and AMD patients, specifically CI-DME with VA ≥ 20/25 or AREDS stages 2–4 with VA ≥ 20/200. The promising reports surrounding PBT might justify prospective, randomized controlled trials. If possible, a consensus of optimal timing, energy, wavelength(s) to be tested in these trials ought to be employed. Additionally, to prevent bias the trials ought to be initiated and monitored by a neutral party without vested interest.