Original Article
Sleep disorders increase the risk of burning mouth syndrome: a
retrospective population-based cohort study
Chun-Feng Lee a,b, Kuan-Yu Lin c, Ming-Chia Lin d, Cheng-Li Lin e, Shih-Ni Chang
f,
Chia-Hung Kao g,h,*
1. Introduction
Sleep disorders (SD) have affected approximately 4–6 million
people in Taiwan according to a survey by the Taiwan Society of Sleep Medicine (TSSM) [1]. Sleep is a critical factor formaintaining mental and physical health [2]. Lack of sleep and sleep deprivation threatens a person’s health. The primary effects of sleep deprivation include
physical effects, such as sleepiness, chronic fatigue syndrome, hypertension, cognitive disorder (eg, deteriorated attention and
motivation, diminished concentration and intellectual capacity, and increased risk of accidents during working and driving), and mental health problems [3]. SD impairs the ability to think, manage stress, and maintain a healthy immune system and emotions. Complete sleep deprivation has been fatal in certain animal study models [4].
It is implicated that sleep is a crucial and essential factor of quality of life. Recent studies have shown an association between the immune system and SD [5–7]. The mechanisms by which SD affect health are unclear. However, certain recent studies indicated that SD might also affect or aggravate chronic pain and pain sensation by producing a hyperalgesic state in healthy people, which influences pain perception [8]. The relationship between SD and pain
is very complex and the possible mechanism might be sleep and pain are both processed via the high pathway of the central nervous system (CNS) and through a pain–sleep interaction mechanism [9].
Clinically, pain can directly affect the sleep quality and quantity of patients with pain related to their underlying medical diseases, such as fibromyalgia, rheumatoid arthritis, and cancers [10,11] and psychiatric
disorders such as anxiety and depression [9,12]. Burning mouth syndrome (BMS) is an idiopathic, chronic pain
condition that affects large populations in modern society, characterized by a burning, stinging, and itching sensation of oral mucosa
in the absence of any organic disease [13,14]. BMS is also defined by the International Association for the Study of Pain [15] as a burning pain in the oral mucous membrane and tongue, with normal signs and laboratory data lasting 4–6 months [16,17]. Butlin and
Oppentheim first described it as glossodynia [18]; however, it is currently referred to as glossopyrosis, oral dysesthesia, sore tongue,
stomatodynia, stomatopyrosis, and most commonly as BMS. The BMS mechanism is not fully elucidated at present, but the neuropathic mechanism for BMS is currently acceptable [19]. Studies have revealed trigeminal nerve alterations in both hyper- and hyposensitivity
as well as fiber neuropathy [19–21]. BMS is also associated with a high prevalence of psychiatric symptoms or mental disorders [22,23].
Therefore, the pathogenesis mechanism remains to be determined.
Previous studies have reported sleep dysfunction as a risk
factor of patients with BMS [12,24]. However, no large populationbased study has outlined the relationship between BMS and SD in
Taiwan. Thus, we investigated whether SD, which includes nonapnea SD and apnea SD (or obstructive sleep apnea disorder), increases the risk of BMS. The original database was derived from the Taiwanese National Health Insurance (NHI) system in Taiwan. The
results presented in this paper were derived from a retrospective cohort study to assess the possibility of a lower risk of BMS with clinical management of apnea and nonapnea SD.
2. Materials and methods
2.1. Data sources
In March 1995, the Taiwanese government implemented the National Health Insurance (NHI) program, which provides universal
health insurance coverage to 99% of the population of Taiwan. The National Health Research Institutes (NHRI) compiles all inpatient and outpatient medical-benefit claims in the NHI program and releases the database for research purposes. The National Health
Insurance Research Database (NHIRD) contains medical information, including inpatient and outpatient care facilities, drug
prescriptions, insurant sex, date of birth, date of visit or hospitalization, and diagnoses coded in the International Classification of
Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) format.
Previous studies have described the details of the NHIRD.We analyzed the one million beneficiaries randomly selected from all
insurants from 1996 to 2000, which has been demonstrated to be representative of the entire population. The study conformed to
STROBE Guidelines. The NHRID encrypts the patients’ personal information for privacy protection and provides researchers with
anonymous identification numbers associated with the relevant claim information, which includes the patient’s sex, date of birth, registry of medical services, and medication prescriptions. Patient consent is not required for accessing the NHIRD. This study was approved by the Institutional ReviewBoard of China Medical University
in Central Taiwan (CMU-REC-101-012).
2.2. Study participants
For the case cohorts, we identified patients newly diagnosed with sleep apnea syndrome (ICD-9-CM 780.51, 780.53, and 780.57) between 1998 and 2001 as the apnea-SD cohort, and patients newly diagnosed with nonapnea SD (ICD-9-CM 307.4 and 780.5, except 780.51, 780.53, and 780.57) as the nonapnea-SD cohort. Patientswere excluded if they were younger than 20 years of age or were diagnosed with BMS (ICD-9-CM 781.1, 529.0, and 529.6) before the index date. In total, 47,941 patients with SD comprised the case group.
The same exclusion criteria were also applied to the non-SD control.
The non-SD cohort was 1:2 frequency matched to the case group by sex, age, and index year (n = 95 882).We excluded patients with medication history of angiotensin-converting enzyme (ACE) inhibitors.
Finally, there were a total of 39,349 subjects in the SD cohort and 86,299 subjects in the non-SD cohort.
2.3. Outcome measurement and comorbidities
The index date for each participant was the first SD diagnosis date.We identified the study endpoint as the first diagnosis of BMS from outpatient claims or hospitalization records from 1998 to 2010.
All of the study participants were followed from the index date to endpoint occurrence, withdrawal from the database, or the end of 2010, whichever date came first.
We also incorporated inpatient and outpatient diagnosis records to ascertain the baseline comorbidities, including diabetes (ICD-9- CM 250), hypertension (ICD-9-CM 401–405), hyperlipidemia (ICD- 9-CM 272), dementia (ICD-9-CM 290.0–290.4, and 331.0),
parkinsonism (ICD-9-CM 332), trigeminal neuralgia (ICD-9-CM 350.1), temporomandibular joint disorder (ICD-9-CM 524.6), anxiety (ICD-
9-CM 300.00), and depression (ICD-9-CM 296.2–296.3, 300.4, 311).
2.4. Statistical analysis
We compared the baseline characteristics between apnea SD, nonapnea SD, and non-SD controls by using the chi-square test.
The age- and sex-specific incidence densities (IDs) were determined under the Poisson assumption. Cox proportional hazards
regression models were performed with adjustment for age, sex, and comorbidities of diabetes, hypertension, hyperlipidemia, dementia, parkinsonism, trigeminal neuralgia, temporomandibular
joint disorders, anxiety and depression. For estimating the cumulative incidence of BMS in SD patients and non-SD patients, we
performed a survival analysis by using the Kaplan–Meier method, with significance based on the log-rank test. All statistical analyses were performed using SAS (Version 9.2; SAS Institute, Cary, NC, USA). Statistical significance was determined by a data type I error of 0.05.
3. Results
Table 1 presents the baseline characteristics of the patients in
the three groups. The distribution of age varied somewhat different, but significant. Approximately half of the study participantswere 41–65 years of age. The mean ages of the SD cohort and non-SD cohort were 49.8 (±15.6) and 50.6 (±15.8) years, respectively. In the apnea SD cohort, most patients were men (62.0%), which inverted in the nonapnea SD cohort (36.0%). Patients in the nonapnea SD cohort were likely to have diabetes (5.64%, P < 0.001), dementia (0.39%, P < 0.001), parkinsonism (1.01%, P < 0.001), anxiety (3.99%, P < 0.001), and depression (6.17%, P < 0.001), whereas the apnea SD cohort patients were likely to have hypertension (27.4%, P < 0.0001), hyperlipidemia (21.5%, P < 0.0001), trigeminal neuralgia (0.27%, P < 0.0001), and temporomandibular joint disorders (0.54%, P < 0.0001).
Compared with the non-SD cohort, the apnea SD (IRR = 2.84, 95%
CI = 2.45–3.30; adjusted HR = 2.56, 95% CI = 1.30–5.05) and nonapnea SD (IRR = 3.07, 95% CI = 2.95–3.19; adjusted HR = 2.89, 95% CI = 2.51–
3.34) were associated with a significantly higher risk of BMS
(Table 2). Figure 1 presents the cumulative incidence of BMS compared with the apnea SD cohort and the nonapnea SD cohort. The
risk of BMS was significantly higher for patients both in the apnea
SD cohort (log-rank P = 0.001) and the nonapnea SD cohort (logrank P < 0.001) than for participants without SD.
Female apnea SD patients (IRR = 4.63, 95% CI = 3.82–5.61) had
a higher risk of developing BMS than did male patients (IRR = 1.76, 95% CI = 1.39–2.24) (Table 2). By contrast, the risk of BMS was higher
in male nonapnea SD patients (IRR = 3.50, 95% CI = 3.28–3.73) than in female patients. For age stratification, the risk of BMS was highest in apnea SD patients aged >65 years (IRR = 3.67, 95% CI = 2.64–
5.10), and in nonapnea SD patients aged ≤40 years (IRR = 3.34, 95%
CI = 3.10–3.61). Table 2 lists the adjusted HRs of BMS according to various age groups. Compared with patients younger than 40 years of age, the HR increased with increased age in the apnea SD cohort and in the nonapnea SD cohort.
Table 3 presents the IRRs of BMS classified according to
comorbidities. In the non-comorbid subgroup without any
comorbidities, the risk of BMS was significantly higher in the apnea SD patients (adjusted HR = 2.98, adjusted for age and gender, 95%
CI = 1.11–8.04) and in the nonapnea SD patients (adjusted HR = 3.05, adjusted for age and gender, 95% CI = 2.57–3.63) than the non-SD cohort. Apnea SD patients without comorbidity generally had a higher IRR of BMS development than non-SD patients. This association was similar for the risk of BMS among the nonapnea SD
cohort. Table 3 presents the relationships among apnea SD patients with comorbidity, such as hypertension (IRR = 3.56, 95%
CI = 2.78–4.57), hyperlipidemia (IRR = 3.13, 95% CI = 2.30–4.26), anxiety (IRR = 5.04, 95% CI = 2.48–10.3) or depression (IRR = 17.5, 95%
CI = 10.2–29.9), compared with non-SD cohort with comorbidities.
In the nonapnea SD cohort, patients with diabetes, hypertension, hyperlipidemia, trigeminal neuralgia, anxiety or depression was also associated with higher risk of BMS than non-SD cohort with those
comorbidities. Anxiety (adjusted HR = 3.29; 95% CI = 1.53–7.09) remained significant factors of an increased risk of BMS development
in the apnea SD cohort and non-SD cohort after adjusting for age, sex, and other comorbidities. After adjusting for age, sex, and other comorbidities by multivariable Cox model, the risk factors of trigeminal neuralgia (adjusted HR = 5.29, 95% CI = 2.35–11.9), anxiety
(adjusted HR = 1.76, 95% CI = 1.26–2.45), and depression (adjusted HR = 1.41, 95% CI = 1.03–1.95) were increased risk of BMS in the
nonapnea SD and non-SD cohorts.
4. Discussion
BMS has multiple etiological causes and is associated with chronic pain, and psychiatric disorders, such as depression and anxiety. Our study revealed that nonapnea SD and apnea are risk factors of BMS and increased HR in the anxiety and depression comorbidity group in both apnea and nonapnea SD group.
In this study, SD was classified into apnea and nonapnea SD group because apnea SD (obstructive sleep apnea syndrome, OSAS)
is an organic disorder with a hypoxemia mechanism affecting respiratory ventilation [25]. However, nonapnea SD might alter the
immune system and induce a systemic inflammation condition. Previous studies have reported that sleep deprivation can elevate the
levels of proinflammation cytokines including CRP, IL-6, and TNF-a [26,27]. Studies have also shown that SD can alter the immune
system which, mediated by the augmented activity of the
hypothalamic-pituitary-adrenal axis (HPA) and stress, is considered one of the major triggers of insomnia [28,29]. It has been
suggested that SDs exert mutual effects with BMS through pain and the neuropathic interaction pathway which might be related to the HPA stress pathway or immune mediators related to the pain pathway mechanism. Our findings also suggest that female patients with an apnea disorder have a higher risk than do male
patients, which is consistent with the etiology of BMS [13]. The age incidence of our findings being dominant above 41 years, neither in apnea SD nor nonapnea disorder groups, infers that BMS typically develops in middle-age patients with SD. In this study, the
comorbidity-specific incidence was higher in nonapnea SD with trigeminal neuralgia in BMS patients, implying that BMS is linked with
neuropathic disorder and pain [19,30,31]. Our findings are also consistent with these studies.
The strength of our study includes its use of population-based data that are highly representative of the general population.
However, certain limitations to our findings should be considered.
First, the NHIRD does not contain detailed information regarding smoking habits, alcohol consumption, socioeconomic status, and family history of systemic diseases, all of which might be risk factors for SD or BMS. Second, the evidence derived from a retrospective
cohort study is generally lower in statistical quality than that from randomized trials because of potential biases related to adjustments for confounding variables. Despite our meticulous study design and control measures for confounding factors, bias resulting from unknown confounders might have affected our results. Third, all data in the NHIRD are anonymous. Thus, relevant clinical variables, such as blood pressure, imaging results, pathology findings, and serum laboratory data were unavailable regarding our study-patient cases.
However, the data regarding SD or BMS diagnoses were nonetheless reliable.
5. Conclusion
This population-based retrospective cohort study determined that patients with SD have an increased risk of BMS. This implies that management of SD can improve BMS symptoms by eliminating systemic inflammatory cytokines, reduce stress process, and/or
modulating pain sensation. Good quality of sleep might improve the progress of BMS. The underlying mechanism remains unclear, and adequate, high quality sleep might eliminate the risk factors of BMS.
This paper could also provide therapeutic hints for clinicians.
However, additional large-scale studies are necessary to confirm these findings.
