Incidence of trigeminal neuralgia: A population- based study in Central Sweden

Background: The primary aim of this observational study was to determine the incidence of trigeminal neuralgia (TN) in a county in central Sweden. The secondary aim was to investigate TN characteristics including the affected side and nerve branches. Methods: Patients that received the ICD- 10 diagnostic codes TN (G50.0), atypical facial pain (G50.1) and other/unspecified disorder of the trigeminal nerve (G50.8 and G50.9) in Uppsala County, between 2009 and 2017, were eligible for inclusion. Case ascertainment was conducted by the authors by review of the medical records. Results: The incidence of TN was estimated to be 5.5 (95% confidence interval 4.7– 6.4) per 100,000 person- years. The incidence increased with age, from 0.1 in 0- to 19- year- olds to 23.1 per 100,000 person- years in 80 + - year- olds. Females exhibited a higher incidence at 7.3 than males at 3.7 per 100,000 person- years. Most of the trigeminal neuralgia cases were diagnosed in the Neurology department (47%). Trigeminal neuralgia was most frequently right sided (59%) and limited to one cranial nerve V- branch, of which V2 was the most common. Conclusions: Trigeminal neuralgia incidence was estimated to be 5.5 per 100,000 person- years. The incidence was higher for females and increased with older age. Significance: There is limited knowledge about the true incidence of trigeminal neuralgia. This manuscript provides an estimate of 5.5 cases per 100,000 person-years, by using a thorough case ascertainment methodology.


| INTRODUCTION
Trigeminal neuralgia (TN) is typically characterized by brief, paroxysmal episodes of intense facial pain, located in any of the cranial nerve V (CN V) branches (Bendtsen et al., 2020;Cruccu et al., 2020;ICHD, 2013). Classical TN is attributed to a neurovascular conflict (NVC) between a cerebral vessel and CN V in the root entry zone (ICHD, 2013;Thomas & Vilensky, 2014), although additional mechanisms such as ion-channel disturbances and neuroinflammation with arachnoiditis may also be involved in disease progression (Abu Hamdeh et al., 2020;Bendtsen et al., 2020;Ericson et al., 2019;Gambeta et al., 2020;Mazzucchi et al., 2019;Svedung Wettervik et al., 2022;Thomas & Vilensky, 2014). In addition, secondary TN caused by, e.g., multiple sclerosis (MS) lesions of the CN V pathways as well as idiopathic TN when no specific cause can be found are other subtypes of TN with slightly different pathophysiology (Gambeta et al., 2020;ICHD, 2013).
TN remains a rare neurological disease, but there is only a handful of studies that have tried to determine its epidemiology in detail. These few studies have also reported a rather wide estimate of the TN incidence, from 4 up to 29 cases per 100,000 person-years (Dieleman et al., 2008;Hall et al., 2006Hall et al., , 2008Katusic et al., 1990Katusic et al., , 1991Koopman et al., 2009;Lee et al., 2021;MacDonald et al., 2000). This large variation is explained by significant differences both in study methodology and population cohorts. First, some previous studies are based on primary care registries alone (Dieleman et al., 2008;Hall et al., 2006Hall et al., , 2008Koopman et al., 2009), whereas others have assessed combined primary care and hospital cohorts (MacDonald et al., 2000). This is relevant because TN is often diagnosed in the Neurology department (Lee et al., 2021), but also because of case ascertainment. Since TN is relatively rare and there is a plethora of differential diagnosis for facial pain such as trigeminal neuropathy, dental pain and atypical facial pain (ICHD, 2013), the diagnostic accuracy may be low with both false positives and negatives, especially in the primary care setting. This limitation is particularly pertinent for those studies that have based their incidence rates on case numbers with a TN diagnostic code (Hall et al., 2006(Hall et al., , 2008Lee et al., 2021), which might not be reliable. Some studies have addressed this issue by proor retrospective case validation (Dieleman et al., 2008;Koopman et al., 2009;MacDonald et al., 2000). However, even in these studies, the incidence has varied from 8 to 29 cases per 100,000 person-years (Dieleman et al., 2008;Koopman et al., 2009;MacDonald et al., 2000). In addition to the methodological differences among the studies, regional differences may explain some of the variation in TN incidence. To better understand TN epidemiology, it is therefore important for high-quality epidemiological studies from different parts of the world. Consequently, the primary aim was to determine the incidence of TN and the relation to age and sex in a county in central Sweden. The secondary aim was to further determine TN characteristics (e.g. side, involved CN V branches) in this cohort.

| Setting
This was a retrospective, population-based epidemiological study on the incidence of TN in Uppsala County, central Sweden, between the years 2009 and 2017. During this period of time, the average population increased from 329,543 in 2009 to 365,172 in 2017. The average population for each year in the study was collected from the "Statistiska centralbyrån" (SCB, 2017; Statistics Sweden, a Swedish government agency responsible for providing data in various societal areas) (SCB). This included average population data for each sex (male/female) and for several age groups (0-19, 20-29, 30-39, 40-49, 50-59, 60-69, 70-79 and 80+).

| Source population and special considerations
The healthcare in Uppsala County was mostly based on public hospitals and primary care practices. There were also 24 private practices providing primary care in the region, of which nine could not provide data and six more practices were only able to supply data for some of the years between 2009 and 2017 due to late adoption of the electronic journal data system. To take into account that the data were unavailable for this subset of the study population, those listed in these practices were excluded from the source population. Uppsala County provided monthly capitation lists for these private practices and the yearly capitation number per practice was calculated as the average of the reported number from January and December of that year (https:// www.regio nupps ala.se/sv/Extra nat/For_vardg ivare/ MOT-PATIE NTEN/Sjukv ard1/Husla karka nslie t1/Kapit ering slist or/; Uppsala, 2019). There were no data on sex and age distribution of these non-participating practices and we then assumed that these numbers were representative of Uppsala County as a whole and subtracted a proportional amount from each sex and age bracket.

| Data extraction
Patient data from medical records in Uppsala County were extracted by IMS Health (formerly Pygargus AB). All International Classification of Diseases (ICD) 10codes were excluded, except for those of interest; G50.0 (TN), G50.1 (atypical facial pain), G50.8 (other disorders of the trigeminal nerve) and G50.9 (unspecified disorder of the trigeminal nerve). The latter three were included to evaluate eventual false negatives, who actually suffered from TN. The patients with these diagnostic codes were controlled against the registries of SCB so that only those who were residents within Uppsala County were included in the further analyses. This was important since Uppsala University Hospital was the main hospital and a referral centre for a much larger catchment area than Uppsala County. In addition, those with a foreign personal ID were also excluded, since they were not included in the SCB population statistics. Consequently, 349 TN (ICD G50.0), 47 with atypical facial pain (G50.1), 28 with other disorders of the trigeminal nerve (G50.8) and 2 patients with unspecified disorder of the trigeminal nerve (G50.9) were eligible for inclusion and further case ascertainment.

| Case ascertainment
Case ascertainment was conducted by a retrospective review of the medical records of the selected patient cohort including 349 patients diagnosed with TN (ICD G50.0), 47 patients with atypical facial pain (G50.1), 28 with other disorders of the trigeminal nerve (G50.8) and two patients with unspecified disorder of the trigeminal nerve (G50.9), by two of the authors (TSW [Neurosurgical resident] and SAH [Senior consultant in Neurosurgery]) and the diagnosis of TN was based on the beta version of the International Classification of Headache Disorders, third edition (ICHD, 2013). The interrater reliability was moderate to strong with a Cohen's Kappa at 0.77. TSW and SAH disagreed on the TN diagnosis in 47 cases, in which a third author (HE [Senior Consultant in Neurosurgery]) made the final decision regarding the TN diagnosis. The incidence rates and the reported age of the patients were based on the time point of the diagnosis (within 2009-2017). In total, 159 (of whom 157 had a G50.0 code and 2 had a G50.1 code) were considered to suffer from TN, 187 (54%) of the G50.0 were considered to suffer from another diagnosis than TN and 52 (15%) were considered to have TN but were diagnosed before 2009. Data on TN characteristics (side, branch, etc.) were also extracted from these medical records.

| Statistical analysis
The incidence was estimated with Poisson distribution (95% confidence interval [CI]). The data were described as numbers (proportions) or median (interquartile range [IQR]). Differences between the sexes were analysed with Mann-Whitney U-test or chi-squared test. Missing data were rare and no imputations were done. The statistical analyses were performed in SPSS version 28 (IBM Corp). A p < 0.05 was considered statistically significant.

| Ethics
The study was approved by the Regional Ethics Committee of Uppsala (Dnr 2015/486) and conducted in accordance with the Helsinki declaration and its later amendments. Informed consent was waived since this was a retrospective study and the medical records were analysed in an anonymized fashion.

| Incidence of trigeminal neuralgia and relation to age and sex
The incidence of TN was estimated to be 5.5 (95% CI 4.7-6.4) per 100,000 person-years (Table 1, Figure 1). The incidence gradually increased with age, as it was estimated to be 0.1 (95% CI 0.0-0.8) in 0-to 19-year-olds age group and 23.1 (95% CI 16.1-32.5) per 100,000 person-years in the 80+-year-olds age group.

| Trigeminal neuralgia characteristics
The age of the TN patients was in the median of 61 (IQR 48-74) years, the majority were females (66%) and most cases were diagnosed at the Neurology department (47%) rather than any other hospital department (28%) or the primary care (25%), as demonstrated in Table 2

T A B L E 2 TN characteristics in relation to biological sex.
F I G U R E 1 Cumulative incidence of trigeminal neuralgia in male and females. The figure demonstrates that the percentage of TN cases below 40 years of age was low (less steep slope), but thereafter the incidence was higher (steeper slope), particularly in the female population. The incidence of TN in males also increased from 40 years and higher, but the males exhibited a more prominent incidence peak after 80 years of age. TN = Trigeminal neuralgia.
CN V-branch, of which V2 (34%) was the most commonly affected branch. The median age of the female TN patients was lower than the male patients (58 [IQR 45-71] vs. 73 [IQR 58-83], p = 0.001), whereas the level of concomitant continuous pain was slightly lower (11% vs 24%, p = 0.04). There were otherwise no differences between the sexes in TN type or affected side or CN V-branch.

| DISCUSSION
In this population-based study of a county in central Sweden, the incidence of TN was estimated to be 5.5 per 100,000 person-years. The incidence increased with age F I G U R E 2 (a-c) TN in relation to age, sex and place of diagnosis. The figure demonstrates the distribution of the patients diagnosed with TN in relation to age groups (a), biological sex (b) and where the diagnosis was made (c). TN, trigeminal neuralgia. and was higher for females, whereas males generally exhibited a lower incidence in all age groups except for those aged 80 years and older. Most patients were affected in only one CN V branch, of which V2 was the most common.

F I G U R E 3 (a-c) TN characteristics-type
In our study, the incidence was in the lower range in comparison with previous studies, which have estimated it to be between 4 and 29 cases per 100,000 person-years (Dieleman et al., 2008;Hall et al., 2006Hall et al., , 2008Katusic et al., 1990Katusic et al., , 1991Koopman et al., 2009;Lee et al., 2021;MacDonald et al., 2000). Our study was based on a relatively large number of positive cases (n = 159) and personyears (n = 24,155,993) over a longer period of time (9 years), which increase the reliability of our findings. It was also clear in our cohort that case ascertainment was important to exclude false positives. This methodological step significantly reduced the false positives, as 69% of the cases with a G50.0 code were excluded because they either had another diagnosis or had been diagnosed with TN prior to the study period. We also tried to detect false negatives by evaluating the nearby diagnoses G50.1, G50.8 and G50.9, but only 2 (1%) more TN cases were found in these cohorts. We cannot exclude that there were more patients in the county that remained undetected with an incorrect diagnosis or who did not seek medical care. However, we anticipate that this number was relatively low since the natural history of TN typically leads to severe symptoms that require medical care with appropriate medications and/or surgery. In previous studies with similar case ascertainment methods, the estimated incidence has been from 8 to 29 cases per 100,000 person-years (Dieleman et al., 2008;Koopman et al., 2009;MacDonald et al., 2000). The previous studies were conducted in the United Kingdom (MacDonald et al., 2000) and the Netherlands (Dieleman et al., 2008;Koopman et al., 2009), whereas our study was based on a county in central Sweden and represented approximately 4% of the Swedish population. Although all of these studies were done in Europe, there may be a significant geographical variation in TN incidence. There is clearly a need for similar studies in other parts of the world to better understand the global TN incidence.
In Sweden, it has been suggested that there should be a maximum capitation list of 1100 patients per general practitioner in primary care (SLF, 2022). Consequently, each general practitioner would only encounter one new case of TN every 17 years in Uppsala County. Altogether, taking into account the low incidence of TN, the low frequency of these patients in primary care and the low diagnostic accuracy with 54% of the TN diagnoses that were incorrect in our study, we conclude that it is reasonable with a referral to the Neurology Department for better diagnostic accuracy when TN is suspected in routine clinical care.
Furthermore, the incidence of TN was two times higher for females than males, which was largely consistent with previous epidemiological TN studies (Hall et al., 2006;Koopman et al., 2009;Lee et al., 2021). We also found a significant rise in TN with higher age, particularly for those aged 80 years and older, which was also consistent with previous studies (Hall et al., 2006;Koopman et al., 2009). Interestingly, it seemed that females exhibited a prominent increase from 40 years of age, whereas males exhibited a much later and prominent incidence peak at 80 years and older. The association between old age and TN is in line with the idea that ageing goes with increased brain atrophy and elongation of the cerebral arteries, which increase the risk to develop an NVC. After the NVC has occurred, it may also take some time for further disease progression including demyelination and nerve atrophy before TN is elicited, which altogether predispose to a late onset of the disease in life (Jannetta, 1980). However, it is clear that NVC is not the only mechanism for the development of TN, since an NVC could not be detected in many cases (107 idiopathic TN cases, 47 with an MRI which was negative). The role of biological sex may then have a certain role in TN pathophysiology which is not fully elucidated, but it has been demonstrated that females exhibit a smaller posterior fossa, which predisposes to more cases of the idiopathic subtype of TN without an NVC. This TN subtype often has an earlier onset than TN with NVC (Hardaway et al., 2019;Ko et al., 2015), which would be in line with that the female patients showed an earlier increase in incidence in our study. In addition, females generally exhibit an increased propensity to develop various autoimmune diseases and inflammatory conditions (Quintero et al., 2012) and this may also involve the risk to develop TN since inflammatory pathophysiological mechanisms seem to be relevant both in primary Svedung Wettervik et al., 2022) and secondary TN (e.g. MS) (Cruccu et al., 2009). In addition, males appeared to more often experience concomitant facial pain than females. Concomitant continuous pain has been suggested as a specific subtype of TN (Maarbjerg et al., 2014a), characterized by more pronounced nerve atrophy (Di . There was a non-significant trend towards a longer time from symptoms to diagnosis for males in our cohort and it could be speculated that the increased burden of concomitant continuous pain reflected a later stage in disease pathophysiology for the males. Furthermore, differences in genetic factors in ion channels, (e.g. transient receptor potential channels) and certain signalling proteins in pain regulation could also contribute to the difference in TN epidemiology between the sexes (Smith et al., 2021). Altogether, more studies on TN pathophysiology are needed to clarify these epidemiological findings.
Lastly, the right CN V was more often affected than the left (1.5:1), which corroborates the findings in previous studies (Maarbjerg et al., 2014b). We also found a predominance for the affection of the V2 branch (34%) followed by V2-3 (25%) and V3 (20%), which was similar with some variations to previous TN case series (Haviv et al., 2016;Maarbjerg et al., 2014b).

| Methodological considerations
First, this was a retrospective study and although the medical records typically reported the pain condition in a clear and thorough way, we cannot exclude that certain important clinical details were left out. For example, the accuracy in anamnesis and reporting might have been lower from the primary care who encountered this patient group less frequently than, e.g., at the Neurology department. Hence, this might have particularly led to some degree of false negatives. Second, we cannot exclude that some residents of Uppsala sought medical care for their TN outside the county. However, since Uppsala provides a full thirdlevel service for TN including Neurology, Neurosurgery (including MVD and ablative procedures) and pain medicine, this limitation was most likely negligible. Third, classical, secondary and idiopathic TN epidemiology were studied as one cohort of TN patients. However, excluding, e.g., the secondary TN cases would likely not have yielded any different results in patient characteristics since they did not differ in age or sex compared to the other two groups. Fourth, the definition of classical TN was based on descriptions of an NVC in medical records, but it was not always explicitly stated if there was concurrent nerve atrophy, which decreases the reliability to some extent. At the same time, only 47 of the 107 cases had conducted an MRI and it is likely that more classical TN cases would have been detected if all patients had done an MRI.

| CONCLUSIONS
The incidence of TN was 5.5 cases per 100,000 and year in this population-based cohort of central Sweden. The incidence was higher in females where it plateaued at 40 years of age and forward, whereas males exhibited a generally lower incidence and a later but more pronounced incidence peak after 80 years of age. Most patients were affected in only one CN V branch, of which V2 was the most common.