Alternative Therapies for Dry Eye Disease

Rhiya Mittal; Sneh Patel; Anat Galor


Curr Opin Ophthalmol. 2021;32(4):348-361. 

In This Article

Addressing Dry Eye Disease With Nerve Abnormalities

Studies have shown that a subset of individuals with DED have underlying nerve abnormalities that may contribute to symptoms. For example, some individuals can have corneal nerve abnormalities causing hypoesthesia along with signs of epitheliopathy, which is categorized as a neurotrophic phenotype, or neurotrophic keratitis (NK). Alternatively, an individual may have a corneal nerve abnormality causing hyperesthesia along with pain out of proportion to ocular surface findings, which is considered a neuropathic phenotype, or neuropathic pain (NP).[75] In some instances, individuals may present with signs of NK and NP simultaneously. Specific treatments have been developed to address nerve abnormalities in both NK and NP.

Neurotrophic Keratitis

NK is a clinical phenotype characterized by decreased corneal sensitivity, with or without pain, and with signs of corneal epithelial disruption. It can occur in the setting of various disorders, including diabetes, postviral infection, anesthetic abuse, neurosurgical procedures, and Sjögren's-associated DED.[75,76] NK is typically graded based on the severity of epithelial irregularities on a 1–3 scale (1 = corneal staining, 2 = persistent epithelial defect [PED], 3 = ulceration or perforation). Autologous serum tears (AST) and amniotic membrane transplant (AMT)[77] are first-line treatments in NK, along with tarsorrhaphy or therapeutic contact lenses. Recombinant human nerve growth factor (rhNGF) (Oxervate, Cenegermin, Dompe),[78,79] platelet rich plasma (PRP), and corneal neurotization[80] are newer approaches that have been explored as treatments for NK.

NGF is a naturally occurring polypeptide that contributes to neuronal growth and differentiation. Released in settings of stress, NGF induces neurite sprouting, eventually resulting in healing and restored function of injured nerves.[81] Numerous tissues express NGF receptors, including those on the ocular surface, and studies have shown that topical application of NGF leads to restoration of the epithelium and prevents corneal melting in NK.[81] The REPARO study randomized 156 subjects with stage 2 or 3 NK to rhNGF 20 μg/ml, 10 μg/ml, or vehicle (6 drops/day for 8 weeks) in a 1:1:1 ratio. At week 4, a higher frequency of treated subjects achieved a PED size <0.5 mm compared to controls (20 μg/ml: 58.0%, 10 μg/ml: 54.9%, controls: 19.6%; P < 0.001). At week 8, outcomes were even better (74.0% vs 74.5% vs 43.1%, respectively). Median time to corneal healing (<0.5-mm lesion staining) was significantly faster in both rhNGF groups compared to the vehicle group (rhNGF 20 μg/ml: 28 days vs rhNGF 10 μg/ml: 29 days vs vehicle: 56 days; P = 0.002 each). Most individuals treated with rhNGF (>96%) did not have a PED recurrence after the initial treatment for up to 56 weeks.[78] A follow-up multicenter, randomized study found similar results when subjects were treated with rhNGF 20 μg/ml (n = 24) or vehicle (n = 24). At 8 weeks, a higher proportion of the rhNGF group achieved a PED size <0.5 mm than controls (69.6% vs 29.2%; P < 0.01).[79] Overall, these findings indicate a benefit of rhNGF as a therapeutic option for NK.

Interestingly, one study noted positive effects of rhNGF in subjects with non-NK DED. In an open-label study of 40 subjects with ATD (symptoms, TBUT≤10 s, Schirmer≤10 mm, staining>3) without history of NK, subjects were treated with rhNGF at different concentrations (20μg/mL [n = 20] and 4μg/mL [n = 20], 2×/day for 28 days). Both groups reported similar improvements in symptoms and signs of DED, including OSDI (20μg/mL: 55.5 ± 21.8 to 32.6 ± 16.2 vs 4μg/mL: 52.4 ± 21.8 to 35.7 ± 22.8; P < 0.001 for both) and corneal staining via lissamine green (20μg/mL: 5.2 ± 2.0 to 1.3 ± 2.2 vs 4μg/mL: 5.9 ± 2.7 to 3.0 ± 3.3; P < 0.001 for both).[82] However, the lack of a control population with which to compare the results limits the strength of the study. Common side effects of rhNGF treatment included hyperemia, pain, and foreign body sensation, but perhaps its most significant limitation is the large cost associated with treatment.[83]

Blood-derived products have long been used in the treatment of ocular surface disorders, owing to their high concentration of growth factors and regenerative capacity.[84–86] More recently, PRP has been explored for the treatment of DED and NK. Benefits of PRP are that platelets can adhere to damaged epithelium and release growth factors (e.g. platelet-derived growth factor (PDGF)), impact fibroblasts, and provide structural support.[87] PRP has in fact been noted to contain higher concentrations of growth factors (e.g. PDGF, NGF) and anti-inflammatory cytokines (e.g. TGF-β) compared to conventional AST.[88]

PRP have been studied in Sjögren's,[89] graft-vs-host disease,[90] post-LASIK DE,[91] and NK. In a retrospective study of 28 individuals with NK due to postinfectious etiologies, 11 of 11 individuals in the PRP group vs 12 of 17 individuals in the AST group achieved re-epithelization. Healing rate was also faster in the PRP vs AST group (PRP: 1.0 ± 0.3 mm2/day vs AST: 0.5 ± 0.1 mm2/day; P = 0.04).[92] PRP (5×/day along with preservative-free artificial tears and vitamin A ointment at night) also healed (n = 20) or improved (n = 4) NK associated ulcers in 25 subjects with nonhealing PEDs (average size: 3.7 ± 1.1 mm × 2.8 ± 0.9 mm).[91] In randomized study, 83 individuals with ATD (OSDI>13, Schirmer<5.5 mm) were divided into treatment (n = 44, PRP 6 drops/day) and control groups (n = 39; artificial tears 6 drops/day bilaterally). At 30 days, greater changes from baseline were seen in the PRP vs control group in OSDI (ΔPRP: −24.9 ± 15.7 vs Δ control: −5.6 ± 5.7; P = 0.001), Schirmer (Δ PRP: +1.9 ± 2.1 mm vs Δ control: +0.3 ± 1.2 mm; P = 0.002), and staining (Oxford score; Δ PRP: −2.3 ± 1.1 vs Δ control: −0.3 ± 0.5; P < 0.001).[93] Fortunately, no adverse side effects were noted with PRP across studies.

Corneal neurotization is a surgical procedure that involves redirecting sensory innervation from one region of the somatosensory system to another in order to reinnervate a target tissue. This procedure is reserved for the most severe cases of NK and involves grafting a peripheral nerve (e.g. sural nerve) to the trigeminal nerve system in order to restore function to corneal subbasal nerve fibers.[94,95] Corneal nerves secrete growth factors that help maintain the corneal epithelium and by restoring innervation, neurotization is thought to restore these factors and improve the epitheliopathy.[96] One retrospective study examined 23 subjects with postherpetic NK who underwent neurotization. After treatment (6–20 months postop), corneal sensation improved on Cochet Bonnet esthesiometry in all patients (1.6–3.6 cm; P = 0.03), with full sensation returning in 2 subjects. In addition, all 4 subjects with a PED had complete resolution without the need for further treatment.[80] In another study, PEDs healed in 3 patients with NK who underwent neurotization. In addition, corneal sensation (0±0 mm baseline to 11.7 ± 16.5 mm at 3 months to 15.0 ± 21.2 mm at 6 months to 30.0 ± 14.4 mm at 12 months), and nerve density (0±0n/mm2 baseline to 12.5 ± 5.1n/mm2 at 3 months to 18.7 ± 5.1 at 6 months to 25.1 ± 10.2n/mm2 at 12 months) increased over time.[97] Side effects of neurotization include surgical-related risks and hypoesthesia at the donor site.[94,95]

Finally, preclinical studies have explored the role of omega-3 polyunsaturated fatty acids (PUFA) on nerve regeneration in the setting of DED. Docosahexaenoic (DHA), a PUFA, promotes the release of docosanoids (e.g. neuroprotectin-1), which work in a cascade along with other growth factors to increase synthesis of NGF.[98] One study applied a collagen shield soaked in DHA+PEDF or vehicle (changed every 72 h over 6 weeks) to rabbit eyes after lamellar keratectomy. Six weeks after initiating treatment, corneal nerve density was higher in the PEDF-DHA vs vehicle group (treatment: 26.7 ± 2.6% vs vehicle: 11.7 ± 1.7%; P = 0.01). Likewise, corneal sensation was greater in the PEDF-DHA vs vehicle group.[99] In humans, one study examined the effect of PUFA in diabetes (a frequent cause of NK). In an open label study, 36 individuals with type 2 diabetes and clinically diagnosed DED were treated with 3 months of oral PUFA (170 mg EPA+115 mg DHA). Marginal improvements in symptoms and signs of DED were noted including OSDI (26.1 ± 8.2 to 22.0 ± 7.5; P < 0.001), TBUT (3.4 ± 0.9 s to 4.4 ± 1.4 s; P < 0.001), and Schirmer (5.2 ± 2.7 mm to 6.6 ± 2.3 mm; P < 0.05).[100] However, the lack of a control group limits the strength of this study. Furthermore, a much larger double-blind, randomized controlled study of 535 individuals with ATD (OSDI>25, TBUT<7, Schirmer<7) failed to show a significant difference in symptoms (OSDI) and signs (conjunctival/corneal staining, TBUT, Schirmer) of DED following treatment with either 3000 mg oral omega-3 fatty acids (n = 349) or placebo (n = 186) for 12 months.[101]

Overall, strong evidence is available for the efficacy of rhNGF and PRP as treatments of NK, whereas weaker data are available regarding the benefits of corneal neurotization and PUFA.

Neuropathic Pain

NP is defined as 'pain caused by a lesion or disease of the somatosensory nervous system. '[102] NP can occur due to an abnormality in corneal sensory nerves (e.g. peripheral NP) or central nerves (e.g. central NP), or, in some cases, both simultaneously. Certain symptoms are more common in individuals with neuropathic ocular pain (NOP), such as the presence of burning, and sensitivity to light and wind.[103] Peripheral NOP can be treated with AST, whereas oral medications (α2γ ligands e.g. gabapentin, selective serotonin-norepinephrine reuptake inhibitors e.g. duloxetine, or tricyclic antidepressants e.g. nortriptyline) are often used when a central component is suspected.[75] For individuals who have persistent pain despite first-line modalities, secondary therapies may be considered.

Cutaneous nerve blocks may be utilized in individuals with NP. This modality entails injection of a nerve-blocking anesthetic that causes reversible blockade of sodium channels that propagate pain, combined with a long-acting steroid to potentiate the pain-blockade.[104] Demonstrating this use, a case series on 11 subjects with clinically diagnosed NOP were treated with periocular (supraorbital, supratrochlear, infratrochlear, and infraorbital) nerve blocks (4 mL of 0.5% bupivacaine + 1 mL of 80 mg/mL methylprednisolone acetate). In total, 7 subjects experienced pain relief lasting from hours to months, whereas no change was reported in the remaining 4.[105] Limitations of this study include its retrospective nature, small study population, and lack of long-term follow-up. Common side effects of this therapy include discomfort and tenderness at the injection site and adverse reactions to the steroid (swelling, skin discoloration) or anesthetic (temporary headache, nausea).[104]

Transcutaneous electrical stimulation (TENS) has also been studied in subjects with DED and suspected NP. Purported analgesic mechanisms of TENS include segmental 'gate control' and supra-spinal descending modulatory mechanisms.[106] A retrospective study found that an in-office 30-min TENS session improved ocular pain in 14 individuals with ocular surface pain and suspected NP (e.g. burning, photophobia). Mean pain intensity was reduced 5 min posttreatment (0–10 numerical rating scale (NRS): OD: 4.5 ± 3.2 baseline to 1.9 ± 2.5 posttreatment; P = 0.01; OS: 4.5 ± 3.4 baseline to 2.0 ± 2.4; P = 0.01).[107] TENS can further be integrated as an in-home treatment and used with other topical and systemic therapies. For example, in a retrospective study of 10 subjects who used TENS at home for a period of 6.6 ± 3.6 months, pain scores were significantly improved posttreatment, with a mean overall decrease of 27.4% (0–10 NRS: 7.7 baseline vs 5.6 posttreatment; P = 0.02), without any adverse events reported during the study period.[108] However, these studies are limited by their retrospective nature, limited number of patients, and lack of control group.

Botulinum toxin type A (BoNT-A) has also been investigated as a treatment for NP.[109] BoNT-A exerts analgesic effects by inhibiting the release of neuroinflammatory substrates (e.g. calcitonin gene-related peptide) and inhibiting unmyelinated C-fiber nociceptors in the meninges, that normally relay signals of photophobia and dryness.[110] A retrospective study of 76 subjects with migraine (≥15 headache days/month for ≥3 months with ≥8 days/month having migranous features) found that ocular surface pain improved after BoNT-A treatment (100U–150U). At 4–6 weeks posttreatment, subjects reported decreased migraine severity and interictal photophobia (0–10 NRS: 4.9 ± 3.0 baseline to 3.4 ± 2.5 posttreatment; P < 0.001) as well as decreased light sensitivity (Visual Light Symptom Questionnaire, range 8–40: 29.8 ± 5.1 baseline to 27.7 ± 6.5 posttreatment; P = 0.002). In those with DED symptoms (DEQ5≥6), symptoms also decreased (DEQ5: 15.4 ± 2.5 baseline to 13.8 ± 4.0 posttreatment; P = 0.03).[111] Similar effects were noted in an case series of 4 subjects with ocular pain and photophobia without migraine, where 35U of BoNT-A were injected in 7 areas of the forehead. At one month, all patients noted decreased light sensitivity, with notable improvements in frequency of light sensitivity to outdoor daylight. DED symptoms also improved with notable improvements in frequency and severity of discomfort and dryness.[112] Finally, one study treated 60 subjects with post-LASIK ATD (symptoms, TBUT<10 s, Schirmer<5 mm) with BoNT-A (n = 20) or preservative-free tear substitutes (n = 20). After treatment, DED symptoms and signs were less severe in the BoNT-A vs control group, including in OSDI (BoNT-A: 12.3 ± 8.6 vs control: 25.3 ± 11.8; P < 0.05), TBUT (BoNT-A: 7.1 ± 1.3 s vs control: 4.9 ± 1.5 s; P < 0.05), and Schirmer (BoNT-A: 8.4 ± 4.4 mm vs control: 4.6 ± 2.1 mm; P < 0.05). These differences lasted up to 3 months posttreatment.[113] Common side effects of BoNT-A injection include ptosis, strabismus, diplopia, pain, and tearing.[109]

Finally, acupuncture, or the use of fine needles placed at acupoints, has been investigated as an adjunct treatment for DED.[114–118] Acupuncture is hypothesized to provide analgesia by dampening sympathetic responses to pain stimuli and/or stimulating the release of analgesic factors like endorphins.[119] In a randomized control trial, 49 subjects with DED symptoms were grouped into acupuncture (n = 24; one 45-min therapy of 12 placed needles at acupoints LI1 and LI2) or sham groups (n = 25; 8 needles on the upper shoulders, which do not serve as acupoints). The acupuncture group demonstrated improvements in symptoms at 1 week, 1 month, 3 months, and 6 months posttreatment via OSDI (34.0 ± 17.0 baseline vs 19.0 ± 17.0, 21.0 ± 17.0, 20.0 ± 21.0, and 16.0 ± 12.0 respectively; P < 0.05 for all), however improvements were also noted in the sham group, but to a smaller degree (36.0 ± 20.0 baseline vs 24.0 ± 22.0, 24.0 ± 21.0, 21.0 ± 20.0, and 25.0 ± 18.0 respectively; P = 0.1).[120] Other studies have shown that the BL1,[121] TE23, LI4, and ST1,[122] acupoints also provide benefit in subjects with ocular pain. Of note, besides conventional acupuncture, laser acupuncture has also shown benefits in improving ocular pain.[123] Although these studies did not specifically examine individuals with suspected NP, the studies suggest that acupuncture may improve symptoms by impacting nerve function.

In conclusion, a number of alternative modalities exist for DED with NP, although the number and strength of the studies is lower than for the other DED subtypes. Most robust are the data regarding the use of BoNT-A.