Blink index as a response predictor of blepharospasm to botulinum neurotoxin‐A treatment

Abstract Purpose We investigated the blink profiles and blink index using ocular surface interferometer in the patients with blepharospasm (BSP) and identified points to consider predictive factor after BSP treatment. Methods In total, 117 eyelids of 59 elderly patients and 20 eyelids of 10 age‐matched control group were studied. All BSP patients applied botulinum toxin‐A (BoNT‐A) injection for treatment of BSP. An ocular surface interferometer (LipiView; TearScience, Morrisville, NC, USA) was used to measure blink profile and blink index; total and incomplete blinks/20 s, and the partial blink ratio (PBR). Eyelid blink time (including lid closing time, closure time, lid opening time), interblink times (IBT), closing speeds (OS), and opening speeds (OS) were analyzed using 600 blinks recorded over 20 s. Results Total blink rate was significantly higher in BSP patients compared to the age‐matched control group (p = .029) but other time‐related and speed‐related index including interpalpebral fissure, PBR, blink time, closure time (CT), interblink time, CS, and OS were not significantly different. In the responder of BSP patients, the average age was higher, CT was shorter, CS was faster than nonresponder (age; p = .016, CT; p < .001, CS; p = .042). Conclusion The blink index by analyzing the blink profile using ocular surface interferometer, and this blink index may be used as a predictive factor for evaluating the clinical response after BoNT‐A injection in blepharospasm patients.

about 40% of blink rate and 60% of Blepharospasm Disability Index (BSDI) in all BSP and increased blinking (Ferrazzano et al., 2015) and Individualized BoNT-A injection reduces clinical symptoms (Sung et al., 2015).
Apraxia of eyelid opening, as a variant feature of blepharospasm, carries a complex differential diagnosis that is most consistent with late-onset dystonia and BoNT-A injection is considered effective treatment (Forget et al., 2002). Pathogenesis of apraxia of eyelid opening and blepharospasm is poorly understood, but there a few hypotheses based on animal model studies. Nigro-striatal basal ganglia pathways may project to the premotor control of eyelid coordination. Therefore, it is associated with dysfunction in the corticothalamic, basal ganglia, and focal cranial nerve circuitry (Weiss et al., 2010).

However, clinical symptoms do not improve in all BSP patients after
BoNT injection, so we want to investigate the dynamics of blink in the BSP patients using ocular surface interferometer, analyze the blink profile, define the blink index, and evaluate the correlation between blink index and clinical responses of BoNT-A treatment response. can cause abnormal blinking of the periocular muscles, were excluded.

Patients and clinical evaluations
Only patients followed up for at least 1 month after BoNT-A injection were included. Clinical outcomes were evaluated by Jankovic rating scale (JanKovic et al., 1990) or change of injection dose. The frequency and severity score of all patients were higher than 3 points. We defined the responder group improved to 0 or 1 point in the rating scale and maintained the initial injection dose next time injection. Nonresponder group was defined as presenting higher than 2 points in the rating scale at initial injection and next injection with the 50% increased dose of BoNT-A.

Blink dynamics analysis
An analysis of blink dynamics was based on the 20 s videos (600 frames), which recorded in LipiView interferometer. Total blink rate (TBR) and partial blink ratio (PBR) was counted in the internal program of LipiView interferometer. We set the blink cycle as the blink time

Statistical analysis
All statistical analyses employed SPSS for Windows, version 25.0 (IBM Corp., Armonk, NY, USA). The Mann-Whitney test was used to compare the control, responder, and nonresponder in BSP groups. A p value < .05 was regarded as statistically significant.

RESULTS
A total 117 eyelids of 59 BSP patients (37 of male, 80 of female) who met the inclusion and exclusion criteria were enrolled this study. The average age of patients receiving BoNT-A therapy was 65.3 ± 11.3 years. After BoNT-A therapy, 82.9% of the patients were responders, but the others of 17.1% were nonresponders. Average injection dose of BoNT-A was higher in nonresponder group (responder; 0.7 ± 0.2 cc, nonresponder; 0.9 ± 0.3 cc, p = .004) and number of injection was also higher in nonresponder (responder; 2.7 ± 2.1 times, nonresponder; 4.3 ± 3.9 times, p = .028) (  Figure 2). Normalized TBR (6.8 ± 4.4 times/20 s) was observed 1 month after the BoNT-A injection in responder, but other indexes were not changed significantly (Table 3).

DISCUSSION
The purpose of present study is to analyze predisposing and treatment prognostic factors of patients with BSP using ocular surface

Responder (n = 38 eyes) Nonresponder (n = 6 eyes)
Before  BSP patients in this study had a tendency to increased total blink rate compared to age-matched control group as previous study (Conte et al., 2013). However, other time-and speed-related indexes of the ocular surface parameters were not different from those the control group.
Conte et al. reported that the blink reflex recovery cycle using EMG response and suggested increased R2 recovery cycle in the patients with BSP. The R2 recovery index and blink rate was correlated, suggesting that the changes of R2 recovery cycle in the BSP patients due to contraction of the orbicularis oculi (Conte et al., 2013). Just as previous EMG studies have helped us understand BSP pathogenesis, it suggest that analyze the blink dynamics also helps us understand BSP pathogenesis. Although the previous study did not show a response of BSP treatment, this study analyzed not only blink dynamics of BSP but also the treatment response. Comparing the responder and nonresponder groups after BoNT-A injection, it was confirmed that the increased CT and decreased CS in the nonresponder groups. It seems that shorter CT and longer IBT in the responder groups may appear to be related to an increased blink rate, which is irregular, and orbicularis dystonia in general clinical features of spasm compared to the control group. But on the contrary, longer CT, shorter IBT, and shorter CS in nonresponders may explain that the levator function is also affected.
After 1 month treated with BoNT-A injection, normalized TBR was observed in in responder groups, but other indexes were not changed in both treated groups. CT was also tended close to the control group in responder groups but statistically not significant. It indicates the blink pattern does not change after BoNT-A treatment except TBR. This suggests that the blink character in BSP patients is unique compared to the control group, and that the etiology of muscle dystonia between responders and nonresponders may not be same. From these results, it can be considered that the etiology of BSP is not only dystonia of orbicularis oculi contraction but the decreased function of the eyelid levator muscles in the nonresponder group. As the results in the nonresponder group, we suggest that the CT and CS analyzed blink index can be used as prognostic factors for BoNT-A injection treatment.
The other finding between responder and nonresponder groups was age, and it was found that the nonresponders was slightly older.
This may be interpreted as changes in neurotransmitter (NT) production, NT transmission, and NT receptor functions as age increases. In previous animal model, having both depleting the dopamine NT and eye irritation condition produce the BSP, and cerebellum blocked blink amplitude and duration (Evinger, 2013;Hall et al., 2006 (Chen et al., 2012). Therefore, genetic analysis in BSP patients can also be used as a prognostic factor and remains a question for future study.
There are limitations to this study. It was a retrospective study, the follow-up duration is not the same in all patients, so changes in blink dynamics before and after BoNT-A treatment are not definitely identified. Because not all patients could come to follow-up and test interferometer until the next BoNT-A injection on account of medical insurance issues, we could analyze the blink indexes of patients who were able to take the LipiView interferometer test one month after BoNT-A injection. The other limitation is the injection dose of BoNT-A was variable depending on the degree and extent of BSP. We should determine the long-term results and identify possible correlation factor of treatment response of BSP patients with blink dynamics in future studies.
Despite these limitations, this study suggest that the blink dynamics can be used for analyzing etiology and treatment response of BSP.

CONCLUSIONS
In conclusion, we found the blink index by analyzing the blink profile using ocular surface interferometer which can be easily tested in clinic, and this blink index may be used as a predictive factor for evaluating the clinical response after BoNT-A injection in BSP patients.