The incidence and etiology of sixth cranial nerve palsy in Koreans: A 10-year nationwide cohort study

We aimed to investigate the incidence, prevalence, and etiology of sixth cranial nerve (CN6) palsy in the general Korean population. The nationally representative dataset of the Korea National Health Insurance Service–National Sample Cohort from 2006 through 2015 was analyzed. The incidence and prevalence of CN6 palsy were estimated in the cohort population, confirming that incident cases of CN6 palsy involved a preceding disease-free period of ≥4 years. The etiologies of CN6 palsy were presumed using comorbidity conditions. Among the 1,108,256 cohort subjects, CN6 palsy developed in 486 patients during the 10-year follow-up. The overall incidence of CN6 palsy was estimated to be 4.66 per 100,000 person-years (95% confidence interval [CI], 4.26–5.08) in the general population. This incidence increased with age, accelerating after 60 years of age and peaking at 70–74 years of age. The mean male-to-female incidence ratio was estimated as 1.41 in the whole population, and the incidence and prevalence of CN6 palsy showed an increasing trend over time in the study period. Surgical incidence for CN6 palsy was only 0.19 per 100,000 person-years (95% CI, 0.12–0.29). The etiologies were presumed to be vascular (56.6%), idiopathic (27.2%), neoplastic (5.6%), and traumatic (4.9%). In conclusion, the incidence of CN6 palsy increases with age, peaking at around 70 years, and shows a mild male predominance in Koreans.

Diseases 10 th Edition (ICD- 10), was adopted in the NHIS-NSC system 11 . Details of the construction and analysis method of the NHIS-NSC are described elsewhere 11 .
We obtained the second version of the NHIS-NSC dataset, which covers the period from 2002 to 2015 and was released in 2017.

Definition of CN6 palsy and surgeries.
The participation flowchart of this study is presented in Fig. 1.
Patients with CN6 palsy were identified using claims with the KCD code for sixth (abducens) nerve palsy (H49.2) assigned by ophthalmologists, neurologists, or neurosurgeons. As in previous studies, subjects who had comorbid dysthyroid exophthalmos (H06.2), thyrotoxicosis (hyperthyroidism, E05), or myasthenia gravis (G70.0) were excluded in consideration of misdiagnosis 4,13,14 . An incident CN6 palsy case was defined based on the earliest CN6 palsy diagnosis for each subject with a preceding 4-year or longer disease-free period. To secure the preceding 4-year disease-free period, patients diagnosed with CN6 palsy between 2002 and 2005 were excluded.
The incidence and prevalence of CN6 palsy were investigated by sex, age-group, and year of diagnosis. A conservative approach (prevalent cases considered those with clinic visits per calendar year) was used to determine prevalence 15 . The annual incidence and prevalence were estimated using the number of patients who qualified for NHIS-in each year 15 . Surgery for CN6 palsy, defined as strabismus surgeries performed after the diagnosis of CN6 palsy during the follow-up period, was identified using the Korean Electronic Data Interchange codes for strabismus surgeries (S5173, 5174, 5175, or 5176).
Etiologies of CN6 palsy. Because medical-record details are not available for the NHIS-NSC database, the causes of CN6 palsy were presumed based on the principal diagnostic codes for comorbidities registered in the NHIS-NSC database within 6 months before and 2 weeks after the first diagnosis of CN6 palsy. Table 1 presents the associated diagnostic codes of possible etiologies of CN6 palsy. The etiologies of CN6 palsy considered in the current study included vascular diseases, cerebral aneurysm, trauma, intracranial neoplasm, intracranial inflammation or infection, and others 5 . Aneurysm was categorized separately from the vascular diseases. Subjects without the diseases mentioned in Table 1 were classified as idiopathic cases. If a subject had multiple co-morbidities, the etiology was determined by the consensus of two examiners (EHJ and B-JC).
Statistical analysis. The incidence and prevalence were estimated with a 95% confidence interval (CI) based on a Poisson distribution. All statistical analyses were performed using SAS Enterprise Guide version 7.1 (SAS Institute, Cary, NC, USA) and R version 3.4.3 (The R Foundation for Statistical Computing, Vienna, Austria).

Results
A total of 1,108,369 subjects were included in the cohort (Fig. 1). Among these, 625 individuals were diagnosed with CN6 palsy at least once from 2002 to 2015. Of these, 26 subjects who had comorbid thyrotoxicosis or myasthenia gravis were excluded. An additional 113 subjects with a history of CN6 palsy from 2002 to 2005 were also excluded. Finally, 486 individuals were identified as having newly developed CN6 palsy during the 10-year follow-up period. Among them, 285 (58.6%) were males and 201 (41.4%) were females.
Incidence of CN6 palsy. Table 2 and Fig. 2 show the incidence of CN6 palsy according to age-group and sex. The estimated overall incidence of CN6 palsy in the entire cohort was 4.66 (95% CI, 4.26-5.08) per 100,000 person-years. The incidence of CN6 palsy was 5.45 (95% CI, 4.84-6.11) per 100,000 person-years for males and 3.86 (95% CI, 3.35-4.42) per 100,000 person-years for females. The overall incidence of CN6 palsy increased with age, with peaks at 65-69 years of age in males and at 70-74 years of age in females. Figure 3 shows the male-to-female ratio of the incidence of CN6 palsy according to age-group. The incidence of CN6 palsy was slightly higher in males than in females, and the average male-to-female incidence ratio was 1.41. Male preponderance was highest in the group of subjects 30-34 years of age and was consistent throughout the cohort, except for the groups of subjects 15-19 and 25-29 years of age. The annual incidence of CN6 palsy grossly increased over time in the study period, as presented in Table 3 and Fig. 4. Of the 486 subjects diagnosed with CN6 palsy from 2006 to 2015, 20 underwent strabismus surgeries. The incidence of surgery for CN6 palsy in the general population was 0.19 per 100,000 person-years (95% CI, 0.12-0.29).
Prevalence of CN6 palsy. Table 3 and Fig. 4 also present the annual prevalence rates of CN6 palsy during the period from 2006 to 2015. The prevalence of CN6 palsy was 3.92 (95% CI, 2.83-5.26) per 100,000 persons in 2006, but gradually increased, reaching 10.08 (95% CI, 8.29-12.12) per 100,000 persons in 2015. There was an overall increasing trend in the prevalence of CN6 palsy over time (Fig. 4). The annual prevalence of CN6 palsy was slightly higher in males than in females, except in 2008 and 2010.

Discussion
In this population-based cohort study, the overall incidence of CN6 palsy in Korea was 4.66 per 100,000 person-years; the incidence rate increased with age, accelerating after the age of 60 years and peaking between 70-74 years of age. An increasing trend in the overall incidence and prevalence of CN6 palsy was observed during the study period. The most common presumed etiology was vascular disease, but the etiologies did vary by age-group.
In the present study, we found that the incidence of CN6 palsy in Koreans is slightly lower than that in Americans reported by Patel et al. (4.66 vs. 11.3 per 100,000 person-years), although the age-distribution and peak incidence of CN6 palsy were similar 5 . Of note, Repka et al. reported that strabismus is less common in Asian races than in the white race 16 ; however, the differences in paralytic strabismus among different ethnicities have not been widely reported. The difference of the epidemiology of CN6 palsy among different races needs to be studied further. In addition, we also observed overall increasing trends in both the annual prevalences and incidences of CN6 palsy over time.
Results of the current study revealed a mild male preponderance of CN6 palsy in Koreans and an overall male-to-female ratio of 1.41. This phenomenon has been observed in several previous studies, with  Table 3. Annual prevalence and incidence of sixth cranial nerve palsy in the general Korean population from 2006 to 2015.  www.nature.com/scientificreports www.nature.com/scientificreports/ male-to-female ratios ranging from 1.2 to 1.8 3,6,17 . Although some studies reported no sex predilection for CN6 palsy 18 , one of the suggested explanations for male preponderance of CN6 palsy is the higher frequency of head trauma in males than in females 19 . In the present study, male preponderance was prominent in the group of subjects aged 30-34 years, and the age group with the highest frequency of head trauma was also young adults aged 30-39 years. Another possible explanation could be that male gender is a risk factor for diabetes or hypertension in Korea 20,21 . The higher risk of microvascular disease in male might be associated with the preponderance of CN6 palsy, especially in older age groups.
Studies on the etiology of CN6 palsy have reported high frequencies of microvascular disease (28-46%) 3,5,18 or unknown origin (24-31%) [3][4][5]10,13,18 . It is widely reported that microvascular diseases are a common cause of isolated unilateral CN6 palsy in patients over 50 years of age 1 . The association between age and incidence in the present study might be related to the age-dependent increase of vasculopathic diseases, such as hypertension, diabetes, or other cerebrovascular diseases. We also observed that vascular disease was the most common cause of CN6 palsy, and the proportion was higher than in previous results of up to 39%. This variety in the etiology distribution may be attributed to variable inclusion criteria for vascular causes. We considered not only diabetes and hypertension but also other cerebro-or cardiovascular diseases as vascular factors. Tamhankar et al. also reported that CN6 palsy in 80.6% of patients over 50 years of age was due to microvascular causes 2 .
Although neuroimaging techniques have improved, the rate of idiopathic CN6 palsy still remains high; recent rates are not lower than rates reported from the 1950s through the 1980s 1,3,5 . On the other hand, the proportion of aneurysmal origin in CN6 palsy is low (0-6%) in the past and recent studies 4,5,8,13,14,18,22,23 . Studies on CN6 palsy in a series of children reported a high frequency of neoplastic etiology (39-45%) 8,23 . Similarly, a high proportion of neoplasm was also observed in the group of subjects aged 10-19 years in the present study.
Recovery rates of CN6 palsy have been reported as high as 60.0 to 87.3% 3,10,24 . Vascular and idiopathic etiologies were associated with higher natural recovery rates than other etiologies of ocular motor nerve palsies 3,4,9,22 25 . In our study, the incidence of surgery for CN6 palsy was also rare. The main causes of CN6 palsy requiring surgery was vascular or idiopathic as in the whole patients with CN6 palsy, and this was due to the high number of vascular and idiopathic groups.
There were a few potential sources of bias and limitations in this study. First, the incidence and prevalence may have been underestimated, as subjects who did not visit the hospitals were missed. In addition, the diagnosis of CN6 palsy was confirmed using only the national health insurance database, not by detailed ophthalmological medical records, because full medical records are not available in the claim database. Nevertheless, the diagnosis of CN6 palsy is usually straightforward, and the chance of misdiagnosis is relatively low. As such, many studies on paralytic strabismus have used the NHIS-NSC dataset 26,27 . Lastly, the causal relationships remain unclear, because the etiologies were attributed to comorbid diseases associated with the CN6 palsy diagnosis.
In conclusion, we found the incidence of CN6 palsy in Korea to be 4.66 per 100,000 person-years. It increased with age, especially after the age of 60 years. The most common cause of CN6 palsy was presumed to be vascular disease in the general population, but the etiologies varied according to age group. Clarification of the epidemiology of CN6 palsy may help us to understand the pathophysiology of the disorder and to build public policies for its prevention and treatment.