Impact of botulinum toxin type A on ocular pain with neuropathic features

Abstract

Purpose:

To investigate the impact of botulinum toxin type A (BoNT-A) on neuropathic/nociplastic ocular pain (NOP) and identify predictors of patient outcomes.

Methods:

A retrospective study of individuals with NOP symptoms (light sensitivity, wind/air sensitivity, persistent pain despite dry eye treatment) who received ≥1 BoNT-A injection. Primary outcome measures included if (responders vs. non-responders) and to what degree (none vs. mild vs. moderate vs. marked) individuals experienced pain improvement 4–6 weeks post-injection. Demographics and clinical exam information was compared between the groups.

Results:

27 individuals received BoNT-A for NOP symptoms. 74% (n=20) reported an improvement in pain and were classified as responders. Among responders, the degree of benefit varied, with 25% reporting mild, 45% moderate, and 30% marked pain improvement. Improvements in light sensitivity (37%), wind/air sensitivity (33%), and quality of life (QoL) (59%) were reported by fewer individuals. 80% of responders and 0% of non-responders reported QoL improvements afer BoNT-A. In a multivariable model that examined predictors of response (none to marked, 0–3), the presence of fibromyalgia (FM) (β=0.50; p=0.004) portended a better response, while shooting pain (β=−0.47; p=0.007) portended a worse response to BoNT-A, (full model r²=0.53; p<0.001). Degree of pain improvement significantly correlated with improvements in light sensitivity, wind/air sensitivity, and QoL (σ range: 0.42–0.63; p<0.05).

Conclusion:

After BoNT-A, most individuals reported improved ocular pain and QoL, while fewer noted improved light and wind/air sensitivity. Some systemic and ocular factors predicted treatment response and may thus guide treatment.

1. INTRODUCTION

Ocular surface pain is a common complaint in individuals with dry eye disease (DED), a heterogeneous condition caused by a loss of tear film homeostasis.[1] DED symptoms include dryness, grittiness, and other pain-related complaints and can be driven by various nociceptive abnormalities, such as epithelial irregularities, inflammation, and hyper-osmolarity.[2] Ocular surface pain symptoms, however, can be experienced by individuals with minimal or no ocular surface abnormalities. In those individuals, neuropathic/nociplastic mechanisms are thought to contribute to pain, with dysfunction within peripheral and/or central nerves.[3] Neuropathic/nociplastic contributors may explain why some individuals do not respond favorably to topical therapies that target nociceptive sources of pain (i.e., lubricating, anti-inflammatory, and blood-derived drops).[4–6] Therefore, new treatment options that target nerve dysfunction are needed for individuals with neuropathic/nociplastic ocular pain (NOP). Irrespective of origin, experiencing ocular surface pain can impair one’s ability to perform activities of daily living, diminish quality of life (QoL), and impact psychological health, leading to anxiety and depression.[7, 8] Therefore, a greater understanding of NOP can lead to the development of novel approaches that mitigate the social, psychological, and physical burden experienced by patients.

Currently, the diagnosis of NOP relies on clinical examination findings. First, several pain-specific terms, such as “burning” or “shooting,” pain are more commonly reported by those with neuropathic rather than nociceptive sources of pain.[9] In addition to spontaneous pain, individuals with NOP often report evoked pain by wind/air, light, or temperature changes.[9] Second, altered pain perception (often extending beyond the eyes) can be seen in individuals with NOP, including pain to light touch around the eye (i.e., cutaneous allodynia) and co-morbid pain conditions, such as fibromyalgia (FM) and migraine.[10, 11] Third, a positive proparacaine challenge test (i.e., persistent pain after topical anesthetic placement) suggests a central NOP component.[12]

In patients with a clinically suspected neuropathic/nociplastic component to pain, treatments focus on targeting nerve function. Such strategies can include topical agents, like transient receptor potential vanilloid 1 (TRPV1) antagonists[13], oral agents, like gabapentin and pregabalin[14], or periocular therapies, like transcutaneous electrical nerve stimulation (TENS).[15] Another periocular therapy investigated for the treatment of NOP is botulinum toxin type A (BoNT-A), based on its effectiveness in other pain disorders such as migraine, trigeminal neuralgia (TN), painful diabetic neuropathy, and post-herpetic neuralgia.[16–18] The proposed mechanism through which BoNT-A targets neuropathic pain (NP) is by inhibiting the release of nociceptive (pain-conducting) and inflammatory mediators from afferent receptors within trigeminal nerve pathways.[19, 20] Substance P, which has a role in pain perception and neurogenic inflammation, as well as glutamate and calcitonin gene-related peptide (CGRP), are among the molecules thought to be blocked by BoNT-A.[21]

With respect to ocular pain/discomfort, BoNT-A improved symptoms of ocular dryness, as measured by the 5-Item Dry Eye Questionnaire (DEQ-5) (pre-BoNT-A vs. post-BoNT-A; 15.4±2.47 vs. 13.8±4.02, p=0.03), in 38 individuals with chronic migraine and severe dry eye (DE) symptoms.[19] Similar findings were noted in individuals without a history of migraine. In a case series of 4 individuals without a history of migraine who received BoNT-A for NOP symptoms, all individuals reported less severe DE symptoms post-BoNT-A, as measured by DEQ-5 scores (mean difference=−8).[22] Missing from the literature is an examination of the perceived impact of BoNT-A on ocular pain and an investigation of baseline factors that predict a positive treatment response. To address these knowledge gaps, our study examined subjective changes in pain and QoL in a cohort of individuals who received BoNT-A for a suspected neuropathic/nociplastic component of pain, and assessed which baseline factors impacted clinical response.

2. METHODS

2.1. Study Design

Retrospective study of 27 individuals who received BoNT-A as a treatment of NOP at the Miami Veterans Affairs (VA) Healthcare System eye clinic or University of Miami’s Bascom Palmer Eye Institute between February 2021 and March 2023. Inclusion criteria for clinically defined NOP included one or more of the following: the presence of ocular surface dysesthesias, allodynia, hyperalgesia, symptoms out of proportion to clinical signs, and/or failure to respond to therapies targeting ocular surface health.[23] Those without features suggestive of NOP were excluded. The study methods adhered to the tenets of the Declaration of Helsinki, and approval for this retrospective study was obtained by the Institutional Review Boards (IRB) of the Miami Veterans Affairs Medical Center (IRB reference ID 3011.01) and the University of Miami (IRB reference ID 20200357).

2.2. Data Extracted

Demographics and medical history, including co-morbidities and medications (including prior BoNT-A injections for NOP), were acquired by patient report at the time of their visit and verified through chart review. Baseline symptom and clinical information prior to BoNT-A injection were gathered based on availability, with specifics outlined below.

Collected baseline questionnaire data:

1. Characteristics of ocular pain such as pain laterality (unilateral or bilateral), location(s) (behind the eye(s), temple(s), and/or above the eyebrow(s)), and frequency (intermittent or constant).

2. Presence or absence of one or more pain descriptors based on a modified McGill questionnaire[24], with choices including throbbing, shooting, stabbing, sharp, aching, heavy, electric shock, or tingling/pins and needles.

3. NP complaints experienced within the past 24 hours using four select questions from the Neuropathic Pain Symptom Inventory-Modified for the Eye (NPSI-Eye) questionnaire. The first question addressed burning pain, rated on a scale of 0 (no burning) to 10 (worst burning imaginable). Subsequent questions addressed evoked eye pain by wind (#2), light (#3), or contact with something hot/cold (#4), with possible responses ranging from 0 (no pain) to 10 (worst pain imaginable). The cumulative score ranged from 0–40.[9]

4. Intensity and frequency of DED symptoms, including ocular discomfort, dryness, and watering during a typical day within the last month, assessed using the DEQ-5 (range: 0–22).[25] Frequency, effect on functioning, and triggers of ocular symptoms (including light sensitivity, grittiness, soreness, blurred and poor vision) within the past week assessed using the Ocular Surface Disease Index (OSDI) (range: 0–100).[26]

5. Pain interference to vocational, social/recreational, and family/marital functioning based on select questions from the West-Haven Yale Multidimensional Pain Inventory (WHYMPI) questionnaire. The first question (#1) assessed how a patient’s ocular pain condition interferes with day-to-day activities along a 7-item scale ranging from 0=no interference to 6=extreme interference. Then, questions assessed specific domains of interference, including the ability to function in vocational activities (#2), social/recreational (#4) situations, household chore completion (#8), marital/family relationships (#6), and other relationships (#9). Finally, individuals indicated how, since the development of the pain condition, their perceived satisfaction or enjoyment from the noted categories changed: social/recreational (#3), vocational (#7), and family relations (#5). Responses were rated on a 7-item scale (0=no change to 6=extreme change), with an option of N/A as not applicable.[27]

Collected baseline clinical information (as available, listed in the order performed):

1. Matrix metalloproteinase (MMP)-9, a marker of inflammation, using InflammaDry (Quidel, San Diego, CA, USA), graded 0–3 based on the presence and intensity of the pink stripe (0=negative, 1=mild, 2=moderate, 3=severely positive)[28];

2. Corneal sensitivity, measured qualitatively, using a cotton tip applied to the central cornea and graded 0–3 (0=none, 1=reduced, 2=normal, 3=increased sensation)[29];

3. Upper and lower eyelid laxity based on the degree of outward rotation (0=0–25%, 1=25–50%, 2=50–100% rotation) and snap back test (0=prompt snap back, 1=slowed return, 2=does not return without blink), respectively;

4. Evoked pain by room light and pen light, each graded 0−10 (numerical rating scale [NRS])[30];

5. Anterior blepharitis, graded 0–3 (0=none, 1=mild, 2=moderate, 3=severe)[31];

6. Lower eyelid vascularity, graded 0–3 (0=none, 1=mild, 2=moderate, 3=severe)[31];

7. Lower eyelid plugging, graded 0–3 (0=none, 1=<1/3, 2=1/3–2/3, 3=>2/3 lid involvement);

8. Palpebral hyperemia, graded 0–2 (0=none, 1=a little, 2=a lot);

9. Papillae located in the palpebral conjunctiva of the lower eyelid (0=absent, 1=present);

10. Conjunctivochalasis based on the folding of epithelium above the lower eyelid margin nasally, medially, and temporally (0=absent or 1=present for each, total from 0–3);

11. Tear film stability assessed by applying 1 drop of fluorescein dye and measuring tear break-up time (TBUT) over 15 seconds. A measurement of >10 seconds indicates healthy tear stability[32];

12. Corneal staining in five quadrants (nasal, temporal, central, inferior and superior), graded as 0=none, 1=mild, 2=moderate, 3=diffuse and summed according to the National Eye Institute scale[32];

13. Proparacaine challenge test comparing pain rating (0–10, NRS) before and after application of one drop of topical anesthesia (proparacaine hydrochloride 0.5%, manufactered by Alcon Laboratories, Inc, Ft Worth, TX, USA). The presence of persistent pain (i.e., pain score≥1) 30 seconds after anesthesia installation suggests a central, non-ocular, or mixed component to pain [12, 14];

14. Basal tear production using Schirmer’s strips placed in the inferior fornix (temporally) after administering one drop of topical anesthetic. The length of wetting estimated tear production over five minutes (range: 0–35 mm/5 minutes) with eyes closed for the duration of the test. A score <5 mm/5 min indicates reduced tear production[32]; and

15. Meibomian gland secretion conducted via contact expression with a cotton swab onto the lower eyelid margin. Quality and presence of secretion observed at gland orifice opening were graded (0=clear/normal viscosity, 1=cloudy, 2=granular, 3=toothpaste-like consistency, 4=no meibum extracted/non-expressible)[33].

2.3. Botulinum toxin type A (BoNT-A) Injection Protocol

Using preservative-free 0.9% sodium chloride, BoNT-A was diluted to reconstitute 5 units (U) per 0.1 milliliter (mL) and administered to select sites via a 31-gauge needle. Based on a modified migraine protocol[22], the patients received 35–100 U of BoNT-A. BoNT-A preparations included incobotulinumtoxinA (50 U vial; Merz Pharmaceuticals GmbH, Frankfurt, Germany) and onabotulinumtoxinA (100 U vial; Allergan, an AbbVie Company, North Chicago, Illinois, USA). All participants initially received 35 U of BoNT-A, administered in the following areas: the glabellar region, targeting the corrugators (10 U total dispersed over two sites), procerus (5 units total in one site), and frontalis (20 U total dispersed over five sites) muscles.[20, 22] This protocol was altered in some individuals on subsequent treatments based on pain location (i.e., over the temporal or periocular regions) and response to BoNT-A. These alterations were not standardized and could include up to 14 additional injection sites (targeting temporalis or orbicularis muscles, 2.5–5 U at each site).[20, 21] All patients were informed that if they had a positive clinical response, they could continue therapy at approximately three-month intervals.[34]

2.4. Post Botulinum toxin type A (post-BoNT-A)

As part of our clinical protocol, all individuals were contacted 4–6 weeks after receiving BoNT-A and asked to respond to a series of questions. The post-BoNT-A questions were uniquely chosen to capture subjective responses to therapy and specific pain-related information based on our current understanding of the potential utility of BoNT-A for the treatment of ocular pain.[22, 35]

Information captured post-BoNT-A included:

1. Subjective changes occurring 4–6 weeks post-BoNT-A injection across four parameters. Individuals were asked to specify whether a change occurred and, if so, the extent of that change for each parameter. The first three questions assessed symptomatic change, each inquiring about a specific symptom: ocular pain (#1), sensitivity to light (#2), and sensitivity to wind/air (#3). Finally, individuals were asked (#4), “After having botulinum toxin injection for ocular pain, do you feel your QoL has changed overall, with respect to normal, day-to-day activities, mood, and social functioning?” Individuals first indicated whether a change was present for each question (#1–4) with a response choice of either “no, not at all,” or “yes.” With each “yes” response, a follow-up question assessed if the experienced change was an improvement or worsening in condition and to what degree (mildly, moderately, or markedly).

2.5. Outcome measures

The main outcome measure focused on subjective pain responses 4–6 weeks after BoNT-A. Response was evaluated through two qualitative variables: 1) Presence or absence of change, and 2) Degree of change, on a scale from 0 to 3 (0=none, 1=mild, 2=moderate, 3=marked change). Individuals were classified into two groups based on their change status: responders included those reporting pain improvement to any degree (mild, moderate, or marked), while non-responders comprised individuals who did not experience any change in pain. Demographics, co-morbidities, and clinical features were examined for associations with subjective pain response measures.

2.6. Statistical analysis

Statistical analyses were performed using the SPSS 28.0 statistical package (SPSS Inc, Chicago, IL). Descriptive variables were first summarized to detail the patient population, with qualitative variables reported as frequencies and quantitative as mean±standard deviation (SD). Next, we divided the population into two groups based on subjective response to therapy: individuals who reported some (mild, moderate, or marked) versus no improvement (termed responders and non-responders, respectively) 4–6 weeks following BoNT-A injection. Group comparisons between responders and non-responders were conducted utilizing Chi-square analysis with Fisher’s exact tests and Mann-Whitney U tests, as appropriate. A linear-by-linear Chi-square analysis was used to assess for associations between degree of pain improvement (range 0 to 3; 0=no improvement, 1=mild, 2=moderate, and 3=marked improvement) and binary variables. The Spearman’s rho (σ) correlation coefficient (95% confidence interval [CI]) was used to examine relationships between degree of pain improvement and continuous variables. A forward, linear regression analysis (95% CI) was used to examine the impact of potential confounders on degree of pain improvement. A p-value of < 0.05 was deemed significant for all measures. In this study, we opted to provide information on all variables being compared, instead of correcting the p-value (e.g., Bonferroni) since the latter methodology has its own limitations.[36]

3. RESULTS

3.1. Study Population

27 patients received a BoNT-A injection for clinically defined NOP. The mean age of the individuals was 52.6±10.7 years (55.6% female; 81.5% White; 55.6% non-Hispanic). The collected concomitant therapies did not differ significantly between responders and non-responders. Nearly all individuals utilized eyedrops (96.3%, n=26), with lubricating (85.2%, n=23) and anti-inflammatory (51.9%, n=14) therapies (either cyclosporine 0.05% or 0.09% or lifitegrast 5.0%) being the most common. Corticosteroid eye drops were the least frequently used ophthalmic drop among the population (7%, n=2). Oral neuromodulators were also commonly used by our sample (55.6%, n=15), with the majority of individuals on gabapentin (40%, n=6) or pregabalin (40%, n=7). Among the factors recorded, depression (48%, n=13), sleep apnea (44%, n=12), headaches/migraines (74%, n=20), and immune-mediated contributors (i.e., condition or a laboratory test consistent with autoimmunity) (41%, n=11) were the most prevalent health issues within our population. Our sample had varied and overlapping etiologies of NOP, with many individuals having multiple potential causes. 48% (n=13) had a post-surgical contributor (history of cataract surgery [n=2], refractive surgery [n=7], other ocular surgery [n=2], both cataract and refractive surgery [n=1], or both cataract and other ocular surgery [n=1]); 44% (n=12) had a post-traumatic contributor (history of ocular trauma [n=2] or traumatic brain injury [n=10]); 44% (n=12) had a systemic contributor (history of FM [n=1], auto immune-mediated disease [n=7], or both [n=4]); and 74% (n=20) were classified as migraine-like, with a history of headache/migraine.[14, 37]

Concerning pain characteristics, the most common features included constant pain (55.6%, n=15), bilateral pain (92.6%, n=25), and pain located behind the eye(s) (66.7%, n=18). 4 individuals (14.8%) felt unable to localize their symptoms to a specific location. The most common symptom descriptor was aching (74.1%, n=20), followed by stabbing (59.3%, n=16), throbbing (55.6%, n=15), heavy (48.1%, n=13), sharp (44.4%, n=12), and tingling or pins and needles (37.0%, n=10). Fewer individuals characterized the pain as shooting (25.9%, n=7) and electric shocks (14.8%, n=4).

3.2. Post Botulinum toxin type A (post-BoNT-A) Pain Outcomes

74.1% (n=20) of individuals reported pain improvement approximately 1-month post-BoNT-A injection (responders). Among them, the degree of benefit varied, with 25% (n=5) reporting mild, 45% (n=9) moderate, and 30% (n=6) marked improvement. All 7 non-responders were first-time BoNT-A recipients, and none indicated pain worsening. Within responders, 45% (n=9) had previously received a BoNT-A injection, whereas 55% (n=11) underwent an injection for the first time. Of the responders, the degree of improvement did not significantly differ between those receiving their first injection and those who had received prior injections (mild: 27.3% vs. 22.2%; moderate: 36.4% vs. 55.6%; marked: 36.4% vs. 22.2%, respectively).

3.3. Study Population, Subtyped by Pain Improvement Status

Regarding differences in medical history, responders were more likely to have a co-morbid auto-immune mediated contributor than to non-responders (55.0% vs. 0%; p=0.02). Other demographics, co-morbidities, and medications were similar between groups (Table 1). When examining pain features, responders were less likely than non-responders to describe their ocular pain as shooting (10.0% vs. 71.4%; p=0.005) and sharp (30.0% vs. 85.7%; p=0.02). In terms of pain interference (as measured by WHYMPI), responders reported higher levels of satisfaction/enjoyment (2.27±2.05 vs. 5.00±1.16; p=0.03) during work activities prior to BoNT-A compared to non-responders. There was no difference in baseline ocular signs between responders and non-responders (Table 2).

Table 1.

Baseline population characteristics grouped by pain improvement status and degree of pain improvement.

Variable	Responders (n=20)	Non-responders (n=7)		ξDegree of pain improvement [0–3]
Demographics	% (n)	% (n)	‡ p-value	ρ	p-value
Age, years, mean+SD	52.3±10.7	53.4±12.2	0.78	−0.10	0.63
			¢ p-value	¢ p-value
Gender, Female	65.0% (13)	28.6% (2)	0.19	0.15
Race, White	80.0% (16)	85.7% (6)	1.00	0.79
Ethnicity, Hispanic	45.0% (9)	42.9% (3)	1.00	0.79
Co-morbidities and devices
Smoking, prior or current	40.0% (8)	85.7% (6)	0.08	0.67
Depression	50.0% (10)	42.9% (3)	1.00	0.44
Sleep apnea	35.0% (7)	71.4% (5)	0.19	0.07
CPAP	25.0% (5)	71.4% (5)	0.07	0.03
BPH	15.0% (3)	0% (0)	0.55	0.76
Rosacea	35.0% (7)	28.6% (2)	1.00	0.90
TBI	40.0% (8)	28.6% (2)	0.68	0.95
Headache/Migraines	70.0% (14)	85.7% (6)	0.63	0.35
Hypertension	25.0% (5)	28.6% (2)	1.00	0.16
Hypercholesterolemia/Hyperlipidemia	30.0% (6)	42.9% (3)	0.65	0.54
Fibromyalgia	25.0% (5)	0% (0)	0.28	0.02
Arthritis (non-rheumatoid)	25.0% (5)	14.3% (1)	1.00	0.65
aAuto-immune contributor	55.0% (11)	0% (0)	0.02	0.07
Related medications
Lubricating eyedrops	90.0% (18)	71.4% (5)	0.27	0.32
Anti-inflammatory eyedrops	55.0% (11)	42.9% (3)	0.68	0.93
Anti-allergy eyedrops	10.0% (2)	42.9% (3)	0.09	0.04
Corticosteroid eyedrops	10.0% (2)	0% (0)	1.00	0.53
AST eyedrops	25.0% (5)	14.3% (1)	1.00	0.96
IOP lowering eyedrops	20.0% (4)	0% (0)	0.55	0.66
bOral neuromodulators	50.0% (10)	71.4% (5)	0.41	0.54

SD: standard deviation, ρ: Spearman’s rho correlation coefficient, CPAP: continuous positive airway pressure, BPH: benign prostatic hyperplasia, TBI: traumatic brain injury, AST: autologous serum tears, IOP: intraocular pressure

^ξDegree of pain improvement where 0=none, 1=mild, 2=moderate, and 3=great

^‡Includes values obtained using Mann Whitney U test

^¢Includes values obtained using Chi-Square analysis

^aIncludes a diagnosis of rheumatoid arthritis, sarcoidosis, Sjögren’s, psoriatic arthritis, ulcerative colitis, scleroderma, graves, and alopecia totalis, or abnormal laboratory testing (late or early markers of Sjögren’s positivity)

^bIncludes gabapentin, pregabalin, duloxetine, amitriptyline, and divalproex

Table 2.

Baseline pain characteristics, pain interference, and clinical signs grouped by pain improvement status and degree of pain improvement.

Variable	Responders (n=20)	Non-responders (n=7)		ξDegree of pain improvement [0–3]
Pain characteristics	% (n)	% (n)	¢ p-value	¢ p-value
Unilateral	0% (0)	28.6% (2)	0.06	0.047
Bilateral	100% (20)	71.4% (5)	0.06	0.047
Behind eye(s)	70.0% (14)	57.1% (4)	0.65	0.33
Temple(s)	50.0% (10)	42.9% (3)	1.00	0.67
Above eyebrow(s)	55.0% (11)	14.3% (1)	0.09	0.046
Intermittent	50.0% (10)	28.6% (2)	0.41	0.34
Constant	50.0% (10)	71.4% (5)	0.41	0.34
Ocular pain descriptors
Throbbing	60.0% (12)	42.9% (3)	0.66	0.67
Shooting	10.0% (2)	71.4% (5)	0.005	0.03
Stabbing	50.0% (10)	85.7% (6)	0.18	0.13
Sharp	30.0% (6)	85.7% (6)	0.02	0.15
Aching	80.0% (16)	57.1% (4)	0.33	0.07
Heavy	45.0% (9)	57.1% (4)	0.68	0.80
Electric Shock	15.0% (3)	14.3% (1)	1.00	0.66
Tingling/pins and needles	35.0% (7)	42.9% (3)	1.00	0.95
Pain symptoms by questionnaire, [range] (n available)	mean±SD	mean±SD	‡ P-value	ρ	p-value
	NPSI-Eye: Neuropathic Pain Symptom Inventory-Modified for the Eye (24-hour recall)
NPSI-Eye Burning [0–10] (23)	4.06±2.88	6.20±1.92	0.11	−0.34	0.12
NPSI- Eye Wind [0–10] (23)	5.17±3.76	4.80± 3.11	0.79	−0.31	0.14
NPSI- Eye Light [0–10] (23)	6.28±3.39	7.20±2.68	0.60	−0.44	0.04
NPSI-Eye Contact Hot/Cold [0–10] (23)	4.39±3.74	5.40±2.70	0.60	−0.27	0.21
NPSI-Eye total [0–40] (23)	19.89±10.37	23.60±5.55	0.79	−0.39	0.06
	DEQ-5: 5-Item Dry Eye Questionnaire (1-month recall)
DEQ-5 [0–22] (23)	15.33±3.33	15.40±1.82	1.00	−0.29	0.18
	OSDI: Ocular Surface Disease Index (1-week recall)
OSDI [0–100] (23)	67.37±21.49	62.00±22.00	0.37	−0.10	0.64
	WHYMPI: West-Haven Yale Multidimensional Pain Inventory
How much has your pain changed your:
Day-to-day activities [0–6]	3.55±1.82	4.86±1.46	0.10	−0.20	0.32
Ability to work [0–6]	3.65±1.90	5.50±1.00	0.05	−0.41	0.07
Ability to work, N/A, retired for reasons other than pain, % (n)	15.0% (3)	42.9% (3)	0.29	¢0.65
Ability to participate in recreational and other social activities [0–6]	3.85±1.84	4.14±2.04	0.67	−0.06	0.76
Ability to do household chores [0–6]	2.70±2.06	3.57±1.99	0.31	−0.37	0.06
Marriage and other family relationships [0–6]	2.15±2.28	3.14±2.55	0.33	−0.23	0.25
Friendships with people other than your family [0–6]	2.0±2.22	2.86±2.80	0.50	−0.07	0.72
How much has your pain changed the amount of satisfaction or enjoyment you get from:
Work [0–6]	2.27±2.05	5.0±1.16	0.03	−0.16	0.52
Work, N/A, due to not presently working, % (n)	25.0% (5)	42.9% (3)	0.63	¢0.67
Participating in social and recreational activities [0–6]	4.0±1.84	4.14±2.19	0.69	−0.07	0.72
Family-related activities [0–6]	2.95±2.21	3.86±2.19	0.34	−0.25	0.22
a Clinical exam finding, [range] (n available)
MMP-9, [0–3] (23)	0.71±0.92	0.50±0.55	0.82	−0.07	0.76
Corneal sensation, [0–3] (21)	1.81±0.66	1.80±0.84	0.81	0.02	0.94
Eyelid laxity, upper, [0–3] (26)	0.16±0.37	0.14±0.38	0.93	0.12	0.57
Eyelid laxity, lower, [0–3] (26)	0.21±0.42	0.14±0.38	0.70	0.25	0.22
Pain with room light, [0–10] (22)	4.0±2.91	5.0±3.69	0.63	−0.22	0.33
Pain with pen light, [0–10] (21)	5.20±3.05	6.50±3.62	0.29	−0.11	0.64
Blepharitis, [0–3] (25)	0.74±0.93	0.83±1.17	0.92	−0.26	0.21
Vascularity, [0–3] (27)	0.95±1.05	0.86±0.69	0.98	0.01	0.96
Plugging, [0–3] (27)	1.45±0.83	1.0±0.82	0.25	0.17	0.40
Hyperemia, [0–2] (25)	0.74±0.65	0.83±0.98	0.92	−0.13	0.55
Papillae, % (n), [0–1] (26)	30.0% (6)	0% (0)	0.28	¢0.82
Conjunctivochalasis, [0–3] (26)	0.53±0.70	1.14±1.21	0.23	−0.25	0.22
TBUT, seconds, [0–15] (24)	6.68±4.0	8.2±4.09	0.33	0.03	0.90
Corneal staining, [0–15] (27)	3.35±4.42	1.29±1.89	0.25	−0.01	0.96
Proparacaine Challenge Status, no improvement, % (n) (24)	58.8% (10)	42.9% (3)	0.66	¢0.24
Proparacaine Challenge, average improvement, [0–10] (24)	1.71±2.62	0.86±0.90	0.84	−0.10	0.64
Schirmer’s, mm wetting/5 minutes, [0–35](24)	8.0±5.66	11.2±7.26	0.29	−0.34	0.10
Meibum quality, [0–4] (24)	1.67±1.94	2.0±1.67	0.45	−0.41	0.047

ρ: Spearman’s rho correlation coefficient, SD: standard deviation, N/A: Not applicable, MMP-9: Matrix metalloproteinase 9, TBUT: tear breakup time.

^ξDegree of pain improvement where 0=none, 1=mild, 2=moderate, and 3=great

^¢Includes values obtained using Chi-Square analysis

When examining degree of pain improvement (no vs. mild vs. moderate vs. marked improvement) post-BoNT-A, associations were noted between CPAP usage (71% vs. 40% vs. 22% vs. 17%; p=0.03), the presence of FM (0% vs. 0% vs. 22% vs. 50%; p=0.02), anti-allergy eyedrop usage (43% vs. 20% vs. 11% vs. 0%; p=0.04), unilateral pain (29% vs. 0% vs. 0% vs. 0%; p=0.047), bilateral pain (71% vs. 100% vs. 100% vs. 100%; p=0.047), pain above the eyebrow(s) (14% vs. 40% vs. 56% vs. 67%; p=0.046), and shooting pain (71% vs. 0% vs. 11% vs. 17%; p=0.03) and treatment response. Additionally, degree of pain improvement exhibited an inverse correlation with severity of light sensitivity (as assessed by NPSI-Eye) and meibum quality (ρ range: −0.44 - −0.41; p<0.05) (Tables 1 and 2).

When all significant factors were considered in a step-wise forward linear regression model to examine which ones most closely related to degree of treatment response (none to marked), the presence of FM (β 0.50; p=0.004) remained associated with more favorable response while the presence of shooting pain (β −0.47; p=0.007) remained associated with a less favorable response to therapy, (full model r²=0.53; p<0.001). CPAP use, anti-allergy eyedrops, pain laterality, pain located above the eyebrow, NPSI-Eye assessed light sensitivity, and meibum quality did not remain in the final model.

3.4. Post Botulinum Toxin type A (post-BoNT-A) Quality of Life (QoL), Sensitivity to Light, and Sensitivity to Wind/Air

A total of 59% (n=16) of individuals reported an improvement in QoL after BoNT-A, whereas only a minority reported improvements in light (37%, n=10) and wind/air (33%, n=9) sensitivity (Figure 1). Overall, most individuals (16 of the 20, 80%) who reported a subjective improvement in pain 1-month post-BoNT-A (responders) also reported QoL improvement, whereas none of the BoNT-A non-responders reported an improvement in QoL (p<0.001). Of the 4 individuals who experienced pain improvement but did not report QoL improvements, two reported mild pain improvement, and two reported marked pain improvement post-BoNT-A. In addition, improvements in other symptoms were more frequent in BoNT-A responders versus non-responders, including improved tolerance to light (50% vs. 0%; p=0.03) and wind/air (45% vs. 0%; p=0.06). Overall, the degree of improvement (scale 0–3) in QoL (r=0.63; p<0.001), light sensitivity (r=0.56; p=0.002), and wind/air (r=0.42; p=0.03) sensitivity significantly correlated with the degree of pain improvement (range 0–3). No individuals in our study reported worsening QoL or sensitivity to light and wind/air post-BoNT-A.

Figure 1.

Subjective Responses 1-month Post Botulinum toxin type A (post-BoNT-A)

4. DISCUSSION

To conclude, a majority of individuals treated with BoNT-A for ocular pain with suspected neuropathic/nociplastic contributors reported at least some subjective improvement in pain following injection. Fortunately, most, but not all, individuals who reported a positive response to BoNT-A also reported an improvement in overall QoL. In a multivariable analysis examining degree of pain improvement, individuals with FM reported greater improvements in pain, while individuals with shooting pain reported less improvements in pain after BoNT-A.

Our study shares similarities and differences from prior work that examined BoNT-A as a treatment for head and facial pain. For instance, in a prospective U.S. study, 71 individuals with chronic migraine were treated with ≥2 BoNT-A injections at intervals of 12–15 weeks. Similar to our findings, a majority (76.1%, n=54) were categorized as BoNT-A responders based on ≥ 50% reduction in both headache frequency and Migraine Disability Assessment Scale (MIDAS) scores from baseline. Of the responders, the majority (85%, n=46) responded with some improvement in pain 3 months after the first set of injections.[38] On the other hand, unlike our study, prior studies have reported a lower frequency of response to BoNT-A in FM. In a Spanish study of 22 individuals with FM and cervical neck pain who received BoNT-A as part of a prospective, randomized study, only a minority (31.8%, n=7) reported pain reduction 2 months after injection (8.29±1.67 to 7.84±1.83; p≤0.005 on a 0–10 visual analogue scale (VAS)).[39] It is not clear why FM portended a better response to BoNT-A in our study, but differences in study design, pain location, and population composition may all have contributed to the disparate findings.

Beyond NOP, BoNT-A has been found to improve pain in other pain conditions of the head and neck. An Argentinean study examined the efficacy of BoNT-A versus saline in 36 individuals with TN (both administered over areas of involved trigeminal branches). Interestingly, at 1 and 2 months post-BoNT-A, pain ratings did not significantly differ between the groups (VAS 1 month: 5.05 vs. 6.06; p=0.29; 2 months: 4.9 vs. 6.63; p=0.07). However, by 3 months, pain report was lower in BoNT-A versus saline-treated individuals (4.75 vs. 6.94; p=0.01).[40] Considering the findings from the TN study, we acknowledge that our assessment at an interval of 4–6 weeks post-BoNT-A may have been premature, and our outcomes might have differed if evaluated at a different time point following injection.

Our study identified certain patient-related factors that predicted BoNT-A’s effect on ocular pain, with some enhancing and others diminishing its response. Prior investigations have similarly examined factors that predict BoNT-A’s efficacy in neuropathic pain, with some findings relevant to our study. One randomized, placebo-controlled, multicenter trial (conducted in France and Brazil) performed injections of either BoNT-A or placebo for peripheral NP (defined by Douleur Neuropathique en 4 Questions (DN4) questionnaire score ≥ 4) and collected questionnaire data and skin punch biopsy specimens (of the affected areas) with the goal of identifying factors predictive of treatment response.[41] After 24 weeks, there was a significant difference between the rate of responders in the BoNT-A treated group versus the control group when response was considered as a 30% reduction in subjective pain scores (65% (n=22) vs. 25% (n=8); p=0.001), but not when the threshold for response was raised to a 50% reduction in pain scores (29% (n=10) vs. 16% (n=5); p=0.2). BoNT-A response (≥ 50% pain improvement 24 weeks after ≥1 injection) was predicted by several factors, including: the severity of brush induced allodynia (odds ratio (OR) 4.6 95% confidence interveal (CI) 1.5–13.7, p=0.007), greater intra-epidermal nerve fiber density (IENFD) (OR 6.2, 95% CI 1.5–25.2; p=0.01), and lower thermal deficit (OR not given) at baseline. As such, it was hypothesized that BoNT-A might offer greater benefits for specific sensory phenotypes of peripheral NP associated with a central component to pain, suggested by the presence of allodynia and hyperalgesia.[41] In relation to the eye, allodynia and hyperalgesia often manifest as sensitivity to light and wind/air. [35] As such, the presence of ocular allodynia and hyperalgesia in many of our patients lends further credibility to the use of BoNT-A for NOP.[4]

Other studies have focused on the impact of pain descriptors on BoNT-A responses. In a retrospective analysis of data from two studies conducted in the U.S., 63 individuals received BoNT-A for migraine (injections targeting pericranial and neck muscles).[42] Prior to injection, participants were asked to classify whether their headache pain felt as if it was “exploding” (pressure build-up from inside the head), “imploding” (skull assaulted by forces such as crushing, clamping, or stabbing from outside of the head), or “ocular” (pain in the eyes). Interestingly, 3 months post-BoNT-A, the number of migraine days per month was significantly decreased by 90.0±0.05% (p<0.0001) in the “imploding” group and by 96.5±0.03% (p<0.05) in the “ocular” group but was unchanged in the “exploding” group (values not provided, p>0.9).[42] We hypothesize that some of our individuals with NOP may have shared disease characteristics similar to those placed into the “ocular” migraine category in prior studies. Pain descriptors have also been examined in relationship to BoNT-A in occipital neuralgia.[43] In a prospective U.S. study of 6 subjects with occipital neuralgia, BoNT-A (targeting areas of the greater and lesser branches of the occipital nerve) significantly improved sharp/shooting pain at 12 weeks compared to baseline (5.75±2.37 vs. 3.42±1.40; p=0.04, VAS 0–10 scale), but did not significantly impact dull/aching or pins and needles pain descriptors.[43] While our study design did not assess change in symptom report by descriptors, we found that individuals who reported pain improvement with BoNT-A were less likely to describe their pain as shooting at baseline. Together, these findings suggest that further investigation is needed to elucidate which pain phenotypes will have the most beneficial response to BoNT-A.

There is biological plausibility that BoNT-A may impact ocular pain with neuropathic or nociplastic contributors. NOP, stemming from peripheral or central contributors, can occur due to various modifications in ion channels, receptors, and nociceptive/inflammatory mediators, with a subsequent change in neuronal and/or glial cell activity. These changes contribute to aberrant signaling, sensitivity, and a change in the perception of pain.[19, 44, 45] In the context of ocular pain, NOP stems from dysfunction or alteration in the trigeminal pain pathway, with potential contributions from the autonomic nervous system. Interestingly, structural corneal nerve abnormalities have been noted in individuals experiencing ocular pain co-morbid with ocular conditions, including Meibomian gland dysfunction (MGD), and systemic conditions, including Sjögren’s and FM.[46, 47] For example, in a Mexican case-control study, individuals with FM (n=17) had significantly thinner corneal stromal nerves (5.0±1.0 vs. 6.1±1.3 μm; p=0.01) and decreased sub-basal plexus nerve density (85±29 vs. 107±26 density/mm²; p=0.02) compared to controls (n=17).[48] Similar findings have been reported in patients with migraine. Specifically, a case-control study in the U.S. found that individuals with migraine had significantly lower corneal nerve fiber density compared to controls (48.4±23.5 vs. 71.0±15.0 fibers/mm²; p<0.001).[49] It is possible that overlapping neuronal contributors to pain in various pain phenotypes, such as FM and migraine, may provide insight into why shared treatments, such as BoNT-A injections, can be effective across these groups.[29, 46]

There are several proposed mechanisms by which BoNT-A may impact pain. Within the peripheral sensory nerves, proposed mechanisms of BoNT-A include reductions in the pain-evoked protein expression of nociception-related ion channels (i.e., TRPV1), reduced mRNA expression of pronociceptive peptides (i.e., preprodynorphin), a regenerative effect on injured nerves, and the blockage of both nociceptive mediators (i.e., substance P, glutamate) and inflammatory mediators (i.e., CGRP, IL-1β) from release at nerve terminals.[50] One Spanish study of 83 patients with chronic migraine investigated CGRP levels in blood before and after BoNT-A treatment.[51] In the overall population, CGRP levels decreased approximately 1 month (range 3–5 weeks) post-BoNT-A compared to pre-BoNT-A (pre- vs. post-median: 74.09 vs. 51.89 pg/mL; p=0.001). The cohort was then divided into responders (headache episodes longer than 4 hours were reduced by at least 50% and a subjective improvement greater than 50% based on a visual scale of 0 to 100) and non-responders. CGRP levels significantly decreased 1 month after injection in responders (pre- vs. post-median: 76.85 vs. 52.48 pg/mL; p=0.003) but not in non-responders (pre- vs. post-median: 50.45 vs. 51.89 pg/mL; p>0.05). Additionally, baseline CGRP levels (prior to BoNT-A) were higher in responders than non-responders (median 76.85 vs. 50.45 pg/mL; p=0.001). Interestingly, this study also found that when CGRP levels were >72 pg/mL, there was a 28-fold increase in the probability of response.[51] CGRP has been implicated in pain mediation, exerting peripheral pain effects and peripheral sensitization via nociceptor sensitization and the dorsal root reflex (DRR), and has been linked to photophobia and corneal inflammation[52–54]. Our study did not measure CGRP levels. However, a portion of our sample included individuals with comorbidities commonly associated with photophobia (e.g., TBI, migraine, DE), suggesting that CGRP might be relevant as a contributing factor to pain in some of our patients.[35]

Within the central nervous system, BoNT-A has been hypothesized to alter nociceptive neurotransmitter release, attenuate microglia and astrocyte activation and neuroinflammation, modulate ascending pain processing pathways, and enhance opioidergic and GABA neurotransmission.[50] An Italian study used an NP model of mice (induced through chronic constriction injury) to explore the central antinociceptive action of BoNT-A. Specifically, they investigated the effects of BoNT-A on microglia and astrocyte activity, which have a role in NP chronicity, and on opioid receptors (which dampens sensory input to the spinal dorsal horn).[50, 55] Compared to control mice, BoNT-A treated mice had a lower number of cells in the dorsal horn of the spinal cord that were positively immunoreactive for glial fibrillary acidic protein (GFAP), a marker of astrocyte expression (12±1 vs. 25±0.5 cells; p<0.001). Additionally, BoNT-A mice had a higher percentage of astrocytes expressing the μ-opioid receptor (85±1 vs. 63±4 %; p<0.01) compared to control mice after both groups were injected with morphine for 9 days.[55] Therefore, it is hypothesized that BoNT-A is involved in reducing neuroinflammation and counteracting the development of morphine-induced opioid tolerance, thereby increasing its analgesic effect.[55] These various mechanistic explanations may underlie the pain improvement noted after BoNT-A in the majority of individuals in our study.

As with all studies, it is important to consider our findings in light of the study limitations, which include small sample size, retrospective design, reliance on subjective symptom reports, and the potential for uncaptured confounders. Furthermore, heterogeneity in the number and units of injections received by participants and underlying causes of pain (i.e., post-surgical, migraine) may have impacted our findings. Thus, larger studies are warranted in diverse populations. Additionally, while we did see positive relationships between pain relief and QoL metrics, not all BoNT-A responders reported QoL improvements. The lack of QoL improvement may be due to QoL domains not captured in this study (i.e., physical and socioeconomic). As such, future studies are needed to more robustly examine which dimensions of the patient experience are altered by BoNT-A. Despite these limitations, our study builds on prior work that suggests BoNT-A as a potential treatment for ocular pain with neuropathic/nociplastic contributions and identifies several potential markers that may help predict which patients will benefit from treatment. Therefore, our study serves as an important initial stride toward improving the prediction accuracy of BoNT-A in the management of ocular pain.

Highlights.

• A majority of patients with neuropathic ocular pain report improvement with BoNT-A

• Patients with fibromyalgia reported a greater improvement in pain after BoNT-A

• Patients with “shooting” pain reported less improvement in pain after BoNT-A

• Improvements in ocular pain correlated with improvements in quality of life