Sleep quality and risk of recurrent aphthous ulcers: A Spanish cohort study

Oral Diseases. 2021;00:1–9. | 1wileyonlinelibrary.com/journal/odi 1 | INTRODUCTION Recurrent aphthous stomatitis (RAS) is a common condition characterized by recurrent small ulcers with circumscribed margins in the oral mucosa (Jurge et al., 2006). Pain is the main symptom of the disease which may last between 3 and 4 days (Scully et al., 2003) and 10– 14 days (Akintoye & Greenberg, 2014). It causes difficulties in eating, swallowing, and speaking (Scully et al., 2003), and as a result, it negatively affects the quality of life in the patients (Tabolli et al., 2009; Zwiri, 2015). It has been estimated that RAS affects 20% of the general population (Scully & Felix, 2005) and is the most frequent oral mucosal lesion in teenagers (Amadori et al., 2017). RAS is a multifactorial disease that can be affected by several risk factors including trauma of the oral mucosa (Wray et al., 1981), family history of RAS (Koybasi et al., 2006), vitamin B12 deficiency (Koybasi et al., 2006; Sun et al., 2015), nonsmoking status (Koybasi et al., 2006; Ussher et al., 2003), psychiatric disorders (Gavic et al., 2014), hormonal imbalance (Balan et al., 2012), and food sensitivity (Wray et al., 1982). Received: 17 May 2021 | Revised: 18 June 2021 | Accepted: 20 June 2021 DOI: 10.1111/odi.13955 ORIGINAL ARTICLE Sleep quality and risk of recurrent aphthous ulcers: A Spanish cohort study Hamid Reza Tohidinik1,2,3 | Almudena Rodríguez1 | Carlos RegueiraMéndez1,4 | Bahi Takkouche1,4 1Department of Preventive Medicine, University of Santiago de Compostela, Santiago de Compostela, Spain 2HIV/STI Surveillance Research Center, and WHO Collaborating Center for HIV Surveillance, Institute for Futures Studies in Health, Kerman University of Medical Sciences, Kerman, Iran 3Department of Epidemiology and Biostatistics, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran 4Centro de Investigación Biomédica en Red de Epidemiología y Salud Pública (CIBERESP, Madrid, Spain Correspondence Bahi Takkouche, Department of Preventive Medicine, Faculty of Medicine, University of Santiago de Compostela, Santiago de Compostela 15782, Spain. Email: bahi.[email protected] Funding information Regional Ministry of Education, Universities and Vocational Training (Consellería de Educación, Universidades y Formación Profesional), Grant/Award Number: ED431C 2018/20 Abstract Objective: Recurrent aphthous stomatitis (RAS) is a condition that affects 20% of the world population and is characterized by painful ulcers in the oral mucosa. So far, the epidemiology and risk factors of RAS have been infrequently studied. Our objective was to determine whether sleeprelated factors are related to the occurrence of RAS in the first prospective study carried out on this topic. Methods: A cohort of 11210 Spanish students, 13– 17 years old, was followed up for one year. Sleep disorders were assessed at baseline using a standard validated questionnaire. Results: We detected 2655 new cases of RAS with a total of 287,262 personweek of followup. Subjects with high (4th quartile) Insomnia Index showed an incidence rate ratio (IRR) of RAS of 1.29 (95% confidence interval (CI) 1.15– 1.45), while subjects with high Hypersomnia Index presented an IRR of 1.42 (95% CI 1.26– 1.61). A high score of sleeprelated phenomena was also associated with an increased IRR: 1.53 (95% CI 1.37– 1.69). Adolescents with high level of sleep satisfaction were at lower risk of RAS: 0.88 (95% CI 0.77– 1.01). Conclusion: These findings suggest that sleep disorders are moderately associated with RAS in adolescents. KEYWORDS adolescents, aphthous stomatitis, hypersomnia, insomnia, oral mucosa, sleep This is an open access article under the terms of the Creative Commons AttributionNonCommercialNoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is noncommercial and no modifications or adaptations are made. © 2021 The Authors. Oral Diseases published by Wiley Periodicals LLC. 2 | TOHIDINIK eT al. A relation between sleeprelated factors and RAS is biologically plausible as, on the one hand, sleep disturbance was linked in previous studies to psychological disorders such as stress among adolescents (Benca et al., 1992; Breslau et al., 1996) and, on the other hand, psychological stress was implicated in the occurrence of RAS (Akintoye & Greenberg, 2014; Keenan & Spivakovksy, 2013). These two relationships are mediated by immunological factors such as growth hormone (GH) and cortisol (Hatzinger et al., 2008; Vgontzas et al., 1999). Epidemiologic studies on the role of sleep patterns in the occurrence of RAS are scarce. Apart from two recent crosssectional studies that revealed a relation of late bedtime and excessive sleep with RAS (Ma et al., 2015; Webb et al., 2013), no prospective study is available. We, therefore, decided to fill the gap and carry out a study to assess the effect of insomnia, hypersomnia, sleep satisfaction, and presence of sleeprelated phenomena on RAS among adolescents, using a prospective cohort design. 2 | MATERIAL AND METHODS 2.1 | Study population and data collection We started the study with 11,210 students aged 13– 17 years registered in 58 Spanish secondary schools in the region of Galicia (Northwest of Spain). Our study population was a random sample of the students' population of the same age in the Galician region. The number of selected students in each institute was proportional to the total number of students of each school. Upon selection, we obtained students' assent and the written informed consent from their parents to participate in the study. Before starting the followup, each participant responded to a baseline questionnaire about RAS, quality and quantity of sleep and other lifestyle variables that could represent potential confounders of the relation between RAS and sleep. We evaluated four different aspects of sleep including insomnia, hypersomnia, and satisfaction with sleep, in addition to a score of sleeprelated phenomena such as snoring or involuntary leg movements. To increase the response rate, both initial and followup questionnaires were anonymous. To link each questionnaire with its corresponding followup, we asked the students to provide on each questionnaire a 12digit code including the code of school (two digits), sex (one digit), and date of birth in the format of day/month/year (six digits) followed by the last three digits of their phone number. We excluded participants with duplicate codes from the study. We also excluded from the followup students who reported RAS at baseline. Students filled in the questionnaires in the presence of professionals who helped them when needed and checked the completion of the questionnaires. The initial and followup questionnaires took 15 and 3 min on average to be filled in, respectively. We followed the students for 1 year. The study was approved by the Regional Ethics Committee (reference 2004/130). The manuscript is in compliance with the STROBE statement as per its checklist. 2.2 | Exposure assessment In the baseline evaluation, we assessed the quality of sleep during the last month using the Oviedo Sleep Questionnaire (OSQ) (García et al., 2000). This scale is a brief semistructured questionnaire which consists of 15 items, 13 of which are grouped into three dimensions, including subjective satisfaction with sleep, insomnia, and hypersomnia. Two additional questions assess the existence of sleeprelated phenomena including snoring, snoring with suffocation, involuntary leg movements and nightmares, and the intake of sleep pills or other remedies. This questionnaire has the advantage of being developed in Spanish in a neighboring population of the Northwest of Spain, and thus did not need any translation or cultural adaptation. The validity and reliability have been evaluated in healthy people (Jiménez et al., 2017) and patients (García et al., 2000; GarciaPortilla González et al., 2009). The questionnaire showed a Cronbach's alpha value of 0.83 for the total OSQ score in healthy people (Jiménez et al., 2017). It also showed a test– retest reliability of 0.87 and an internal consistency of 0.90, 0.88, and 0.91 for total score, hypersomnia, and insomnia, respectively. It demonstrated an acceptable convergent validity and discriminant validity and was able to differentiate between cases and noncases, as well as distinct degrees of sleep disorder severity in patients (GarciaPortilla González et al., 2009). 2.3 | Disease assessment We determined cases of RAS on the basis of a short questionnaire that we administered to the participants every 12 weeks. We based our questionnaire on symptoms described in the literature (Bruce & Rogers, 2003; Scully et al., 2003). We further asked two questions to exclude traumatic wounds caused by dentistry interventions, or piercing. These questions, answered on a yes/no basis, were as follows: Did you have “aphthous ulcers” in your mouth, which are small and painful white sores? Do you associate their occurrence with the fact that you broke a tooth or bit your cheek? Do you associate their occurrence with the contact with your braces or your piercing? We considered cases with two or more episodes of disease appearing on separate occasions during the followup as severe RAS. 2.4 | Validation substudy To assess the validity and reproducibility of our diagnosis of RAS, based on selfexamination, we compared this diagnosis with that established by nine different dentists on a random sample of 125 patients, 74 with RAS and 51 without RAS who attended to the dental clinics for a different motive than mouth ulcers. The diagnosis made by the dentists was considered as the gold standard. These dentists had previously received specific training with visual aids to enable them to make a differential diagnosis between RAS and | 3 TOHIDINIK eT al. similar mouth lesions such as intraoral herpes ulcers or oral lichen planus. We asked the dentists to consider as a case of RAS the following clinical forms of the disease: (1) Minor or Mikulicz type: oval or round shape wounds smaller than 1 cm in diameter lasting for 1– 2 weeks and which normally do not leave any scar after recovery, (2) major type (or Sutton ulcer): wounds that are larger than 1 cm of diameter and which may last more than one month and may leave a scar after recovery, and (3) Herpetiform type: very small (1– 2 mm) but numerous and painful lesions. 2.5 | Data analysis We calculated persontime for each participant, which started when the initial questionnaire was returned and ended with the onset of the disease, termination of the study, or loss to followup, whichever occurred first. For severe RAS, we considered the last episode of disease as the end of the followup. To calculate the persontime for each positive case of RAS, we summed all previous periods with a negative response and half of the last period (6 weeks), assuming constant incidence of RAS during the period. Continuous variables such as score of hypersomnia, score of insomnia, and number of phenomena during sleep were introduced in the model as quartiles of distribution. Poisson regression model was used to calculate adjusted incidence rate ratios (IRRs) and their corresponding 95% confidence intervals (95% CIs). To explore the shape of the curve relating the insomnia and hypersomnia scores to the onset of RAS, we fitted a cubic splines model adjusted for potential confounders. We set three knots (quartiles 1 to 3) at the boundaries of the exposure categories. Variables that were considered in the analysis as potential confounders were age, sex, smoking, passive smoking, total alcohol intake, total caffeine intake, total physical activity, geographical area (coastal/ inland), body mass index, and two uncorrelated dimensions of psychological stress: Negative Affect and Positive Affect (Watson et al., 1988). Except for age and sex, which were introduced in all models, we introduced potential confounders in the final model if their inclusion modified the IRR of the main exposure variable by at least 10% (Greenland, 1995). Stata 12 (Stata Corp) was used for all statistical analyses. 2.6 | Sensitivity analysis To assess the robustness of our results to misclassification bias, in a secondary analysis we corrected the relative risk estimates of the highest level of each sleep variable using the sensitivity and specificity of our diagnosis tool of mouth ulcers obtained from the concurrent validation substudy. We assumed that all subjects in the cohort were followed until the end of the study (Kleinbaum et al., 1982). 3 | RESULTS A total of 9617 participants were analyzed (Figure 1). Five hundred fortyfive out of 11,210 participants were excluded because FIGURE 1 Flow chart of the cohort study on sleep and risk of recurrent aphthous ulcers 4 | TOHIDINIK eT al. of duplicate codes in questionnaires, and 80 further cases were excluded due to insufficient completion of the baseline questionnaire (5.6%). We also excluded participants who had mouth sores related to braces or piercing (n = 265), sores related to involuntary biting (n = 540), or both (n = 163). Female participants represented 53% of the sample, and 68% of the participants were living in urban areas. The mean followup time was 29.9 ± 19.7 weeks. Students were between 13 and 17 years old [mean (SD): 16.2 (1.46) years]. We detected 2655 incident cases of RAS with a total of 287,262 personweeks of followup. The overall incidence rate of RAS was 0.48 year−1 (0.43 in men and 0.52 in women). The incidence rate of severe RAS was 0.19 year−1 (0.16 in men and 0.21 in women). The sensitivity and specificity of the questionnaire for diagnosis of RAS were 65% and 85%, respectively. Absence or low level of satisfaction with sleep was reported by 13.6% of the study population. Mean age was similar in all subgroups of sleeprelated variables. Table 1 represents the distribution of potential confounders in different categories of sleeprelated variables. Table 2 shows the adjusted IRRs for RAS according to four sleeprelated variables. Both insomnia and hypersomnia scores were related in a monotonic fashion to increased RAS incidence rate. Compared to adolescents in the first quartile of the distribution, those of the 3rd and 4th quartiles of insomnia presented the following IRRs: 1.24 (95%CI: 1.10 to 1.38) and 1.29 (95%CI: 1.15 to 1.45). The trend was similar for hypersomnia with IRRs of 1.22 (95%CI: 1.09 to 1.36) and 1.42 (95%CI: 1.26 to 1.61) for 3rd and 4th quartiles, respectively. A similar trend was observed for the score of phenomena during sleep, the IRRs of which for 3rd and 4th quartiles were, respectively, 1.12 (95%CI: 1.01 to 1.26) and 1.53 (95%CI: 1.37 to 1.69). Sleep satisfaction was inversely associated with incidence of RAS, as subjects with sufficient and high sleep satisfaction showed IRRs of 0.87 (95%CI: 0.77 to 0.98) and 0.88 (95%CI: 0.77 to 1.01), respectively, when compared to subjects with low sleep satisfaction. The exposureeffect curve in Figure 2 shows the relation between scores of hypersomnia and insomnia, and the adjusted incidence rate ratio of RAS. For both variables, there are steady increasing trends of IRRs with increasing scores of hypersomnia and insomnia. Our results for severe aphthous ulcers were similar to those found for nonsevere forms. According to Table 3, the increase in the score of insomnia, hypersomnia, and number of sleeprelated phenomena was related in a doseresponse fashion to the rate of severe aphthous ulcer. Adjusted IRR for 4th quartiles of insomnia and number of phenomena compared with their 1st quartiles were 1.26 (95%CI: 1.03 to 1.53) and 1.63 (95%CI: 1.36 to 1.95), respectively. Adjusted IRRs for 3rd and 4th quartile of hypersomnia compared with the 1st quartile were 1.39 (95%CI: 1.14 to 1.69) and 1.65 (95%CI: 1.33 to 2.04), respectively. However, we could not find any evidence of a significant relation between sleep satisfaction and severe aphthous ulcers. In addition to sex and age, further adjustment for smoking, passive smoking, total alcohol intake, total caffeine intake, total physical activity, geographical area (coastal/inland), body mass index, and stress measured through Negative Affect and Positive Affect scales did not meaningfully change the results (data not shown). A total of 1254 subjects (13.04%) of the cohort were lost to followup before the end of the study. To assess whether loss to followup was related to exposure and to potential confounders, we compared the distribution of the four sleeprelated exposure variables, in addition to sex, age, smoking, passive smoking, total alcohol intake, total caffeine intake, total physical activity, geographical area, and body mass index among participants who were lost to followup to the distribution of those variables among subjects who completed the followup. The distribution of all variables, except smoking and alcohol consumption, was similar. On average, subjects with incomplete followup consumed more alcohol than those with complete followup [Mean (SD): 65.1 (3.2) vs. 46.8 (1.1) gr]. However, alcohol consumption did not show any relation with the incidence of RAS in our data, and the introduction of this variable in the model did not modify the results. Furthermore, smoking was more frequent among subjects with incomplete followup than among those with complete followup (29% vs. 19%). Separate analyses showed that, in our data, smoking was related to an increase in the rate of RAS but when this variable was introduced in the model, the results were not altered substantially, probably due to the fact that smoking is not related to the sleeprelated exposure variables in our data and, hence, is not a confounder. To assess further the possible effect of loss to followup, we recalculated the incidence rate ratios in two extreme situations. First, we assumed that all participants lost to followup developed RAS. Second, we assumed that none of them developed the disease. In both scenarios, our results were not meaningfully modified from the original one. The highest change was found in the point estimates of “hypersomnia” of 3rd and 4th quartiles which changed from 1.22 and 1.42 to 1.15 and 1.27, respectively, when we assumed that all participants lost to followup developed RAS. Furthermore, the IRR for 4th quartile of “score of sleeprelated phenomena” was modified from 1.53 to 1.41 when we assumed that all censored participants developed RAS. The secondary analysis aimed at assessing the robustness of the results to misclassification bias of the outcome yielded the following corrected crude relative risk estimate for the highest level of each sleeprelated variable: insomnia: 1.44 (95%CI: 1.30 to 1.59), hypersomnia: 1.60 (95%CI: 1.44 to 1.79), satisfaction with sleep: 0.89 (95%CI: 0.79 to 1.00), and score of sleeprelated phenomena: 1.87 (95%CI: 1.71 to 2.04). The observed (uncorrected) crude estimates were as follows: insomnia: 1.18 (95%CI: 1.07 to 1.30), hypersomnia: 1.24 (95%CI: 1.12 to 1.37), satisfaction with sleep: 0.95 (95%CI: 0.85 to 1.06), and score of sleeprelated phenomena: 1.35 (95%CI: 1.24 to 1.47). 4 | DISCUSSION Our findings indicated that high levels of insomnia and hypersomnia, and the existence of phenomena during sleep are moderately associated with a higher rate of RAS. On the other hand, high levels | 5 TOHIDINIK eT al. TABLE 1 Distribution of potential risk factors for Recurrent Aphthous Stomatitis in students of 58 Spanish schools according to different aspects of sleep Sleeprelated variables Agea (years) Femaleb Passive smokerb Current smokerb Living in coastal areab Living in rural areab BMIa (kg/m2) Alcohol intakea (g/week) Caffeine intakea (mg/ day) Total physical activitya (score) Negative Affecta (score) Positive Affecta (score) Mean N (%) N (%) N (%) N (%) N (%) Mean Mean Mean Mean Mean Mean Insomnia score 1st quartile 16.1 969 (41.7) 1624 (69.9) 371(15.7) 1187 (50.2) 776 (32.8) 21.0 40.7 70.5 21.9 21.8 10.7 2nd quartile 16.2 1122 (51.0) 1647 (75.1) 376 (16.8) 1098 (49.1) 716 (31.9) 21.1 44.1 70.1 21.2 24.2 10.4 3rd quartile 16.2 1235 (56.6) 1731 (79.5) 494 (22.3) 1178 (53.2) 716 (32.3) 21.1 50.2 81.8 21.0 26.8 10.3 4th quartile 16.3 1411 (65.0) 1841 (84.1) 632 (28.5) 1147 (51.8) 638 (28.8) 21.1 62.0 115.0 20.6 32.1 9.8 pvalue <0.001 <0.001 <0.001 <0.001 0.03 0.01 0.2 <0.001 <0.001 <0.001 <0.001 <0.001 Hypersomnia score 1st quartile 15.9 904 (43.5) 1398 (67.6) 317 (15.0) 1083 (51.1) 727 (34.3) 20.9 34.4 68.4 22.1 22.0 10.4 2nd quartile 16.2 1473 (55.1) 1990 (74.7) 481 (17.7) 1396 (51.4) 846 (31.8) 21.1 40.9 72.3 21.2 24.5 10.4 3rd quartile 16.2 1468 (54.1) 2119 (80.5) 588 (21.3) 1429 (51.7) 901 (32.6) 21.2 53.3 91.0 20.8 27.5 10.2 4th quartile 16.5 1041 (61.1) 1496 (86.8) 551 (31.7) 875 (50.3) 465 (26.7) 21.0 72.9 110.2 20.6 31.5 10.1 pvalue <0.001 <0.001 <0.001 <0.001 0.8 <0.001 0.003 <0.001 <0.001 <0.001 <0.001 <0.001 Sleep satisfaction None or a little 16.3 727 (56.9) 1063 (82.9) 363 (27.9) 711 (54.6) 366 (28.1) 21.2 59.8 108.0 21.0 31.5 9.6 Regular 16.3 1394 (59.1) 1853 (78.9) 543 (22.7) 1290 (53.8) 742 (31.0) 21.1 50.9 86.1 20.3 28.4 9.8 Enough 16.2 2092 (53.1) 3008 (76.5) 745 (18.6) 1996 (49.8) 1303 (32.5) 21.0 45.7 75.3 21.2 24.5 10.5 High 16.1 795 (43.7) 1303 (71.1) 332 (17.8) 925 (49.6) 614 (32.9) 21.0 47.0 84.0 22.2 23.0 10.9 pvalue <0.001 <0.001 <0.001 <0.001 <0.001 0.01 0.12 <0.001 <0.001 <0.001 <0.001 <0.001 Score of sleeprelated phenomena 1st quartile 16.2 1605 (47.2) 2414 (71.4) 534 (15.4) 1679 (48.5) 1091 (31.5) 20.9 40.8 72.3 21.4 23.5 10.4 2nd quartile 16.3 924 (56.1) 1295 (78.8) 314 (18.8) 795 (47.6) 530 (31.7) 21.1 44.8 77.2 20.8 25.2 10.3 3rd quartile 16.2 921 (59.5) 1263 (80.6) 359 (22.7) 809 (51.0) 492 (31.1) 21.0 50.7 79.0 21.0 27.3 10.3 4th quartile 16.2 994 (57.1) 1456 (83.4) 497 (28.1) 940 (53.2) 558 (31.6) 21.4 66.8 116.5 21.3 30.7 10.3 pvalue 0.08 <0.001 <0.001 <0.001 0.002 0.9 <0.001 <0.001 <0.001 0.001 <0.001 0.2 aOneway ANOVA was used to compare groups. bChisquare test was used to compare groups. 6 | TOHIDINIK eT al. of satisfaction with sleep are inversely related to the rate of RAS. For severe forms of RAS, except for sleep satisfaction for which no effect was noticed, the findings were similar to those concerning nonsevere forms. These findings are biologically consistent. Indeed, sleep disorders may be associated with RAS through increased serum and salivary cortisol and decreased growth hormone levels as a consequence of psychological disorders such as stress and anxiety TABLE 2 Incidence rate ratios of Recurrent Aphthous Stomatitis in students of 58 Spanish schools, according to different aspects of sleep Sleeprelated variables Sample size Crude IRRa 95% CIa Adjusted IRRa ,b 95% CIa No. of personweeks No. of cases Insomnia score 1st quartile (9 to 13) 2366 1.00 Reference 1.00 Reference 75,046 601 2nd quartile (14 to 16) 2238 1.07 0.95, 1.20 1.07 0.95, 1.20 69,1124 592 3rd quartile (17 to 20) 2216 1.25 1.12, 1.40 1.24 1.10, 1.38 64,304 646 4th quartile (21 to 45) 2216 1.33 1.19, 1.49 1.29 1.15, 1.45 62,172 664 Hypersomnia score 1st quartile (2 to 3) 2118 1.00 Reference 1.00 Reference 67,130 527 2nd quartile (4 to 5) 2718 1.08 0.97, 1.21 1.06 0.95, 1.19 84,442 717 3rd quartile (6) 2763 1.26 1.12, 1.40 1.22 1.09, 1.36 80,912 798 4th quartile (7 to 10) 1739 1.47 1.30, 1.66 1.42 1.26, 1.61 46,554 537 Sleep satisfaction None or a little 1303 1.00 Reference 1.00 Reference 36,762 371 Regular 2395 1.01 0.88, 1.14 1.01 0.89, 1.14 68,844 698 Enough 4003 0.86 0.77, 0.97 0.87 0.77, 0.98 122,546 1067 High 1864 0.86 0.75, 0.98 0.88 0.77, 1.01 57,830 503 Score of sleeprelated phenomena 1st quartile (5) 3458 1.00 Reference 1.00 Reference 108,204 874 2nd quartile (6) 1670 1.03 0.92, 1.16 1.02 0.91, 1.15 51,050 457 3rd quartile (7 to 8) 1584 1.16 1.03, 1.30 1.12 1.01, 1.26 47,638 447 4th quartile (9 to 21) 1768 1.57 1.41, 1.74 1.53 1.37, 1.69 47,716 604 aIRR, Incidence Rate Ratio; CI, Confidence Interval. bAdjusted for sex and age (continuous). Further adjustment for the following variables did not modify the results: smoking, passive smoking, total alcohol intake, total caffeine intake, total physical activity, geographical area (coastal/inland), body mass index, Negative Affect and Positive Affect. FIGURE 2 Incidence rate ratios of Recurrent Aphthous Stomatitis (RAS) in students of 58 Spanish schools, adjusted for sex and age (continuous), according to (a) hypersomnia and (b) insomnia; twotail restricted cubic splines model. Solid line represents point estimates; dotted lines represent 95% confidence intervals | 7 TOHIDINIK eT al. (AlbanidouFarmaki et al., 2008; Bali & Jaggi, 2016). Cortisol is released as a response to acute psychological stress by the hypothalamic– pituitary– adrenocortical (HPA) system (Steiger, 2002). Hormones of the HPA system also regulate the sleep– wake cycle, while their dysfunction can disrupt sleep. In turn, sleep loss can hyperactivate the HPA system (Meerlo et al., 2008). It was also described that patients with insomnia present high levels of cortisol (Vgontzas et al., 2001), and that increased secretion of morning cortisol was associated with poor sleep quality in both children (Hatzinger et al., 2008) and adults (Bassett et al., 2015). Furthermore, the associations between stress and anxiety with RAS on the one hand (AlbanidouFarmaki et al., 2008; Gallo Cde et al., 2009), and between stress and sleep problems on the other hand (Hansen et al., 2018; LeBlanc et al., 2009), have been described before. This evidence could plausibly explain the association between sleep disorders and RAS found in our study. The validity of the questionnaire for diagnosis of RAS was acceptable. In addition, previous studies confirmed the validity of the OSQ tool in the assessment of different aspects of sleep (García et al., 2000; GarciaPortilla González et al., 2009). The prospective design of the study ensured that the quality of sleep was assessed before the diagnosis of RAS and, therefore, prevented the differential misclassification of exposures between RAS cases and noncases. Our study could be limited by the fact that the diagnosis of RAS was selfassessed and a certain amount of misclassification of the outcome could have been introduced. However, our secondary analysis, carried out to assess this effect of outcome misclassification due to an imperfect diagnosis tool, showed that our results were conservative and that the unknown true value of the risk increase was probably more extreme. We also adjusted the associations for potential confounders and performed sensitivity analyses to assess the effect of loss to followup, but the results were not meaningfully modified. Therefore, the associations we found are unlikely to be due to confounding by measured variables, misclassification, or loss to followup. However, we cannot rule out a possible effect modification by a genetic factor. Indeed, previous studies showed that the HLAB51 gene as well as genes that control heat shock proteins or cytokines may be implicated in the occurrence of RAS (Mizuki et al., 1995; ShohatZabarski et al., 1992). In our study, we did not measure those genetic factors. The effect of sleep disturbance on RAS may then be different depending on whether subjects harbor a polymorphism in those genes or not. Furthermore, we cannot exclude potential confounding by unmeasured factors either. The relation between psychological stress, sleep disorders, and RAS is not straightforward. Although we did not find sufficient evidence in our study of a potential role of TABLE 3 Incidence rate ratios of severe Recurrent Aphthous Stomatitis in students of 58 Spanish schools, according to different aspects of sleep Sleeprelated variables Sample size Crude IRRa 95% CIa Adjusted IRRa ,b 95% CIa No. of personweeks No. of cases Insomnia score 1st quartile (9 to 13) 1967 1.00 Reference 1.00 Reference 63,218 202 2nd quartile (14 to 16) 1869 1.18 0.97, 1.42 1.15 0.95, 1.39 59,211 223 3rd quartile (17 to 20) 1783 1.22 1.01, 1.49 1.18 0.97, 1.44 54,341 213 4th quartile (21 to 45) 1771 1.32 1.09, 1.60 1.26 1.03, 1.53 51,894 219 Hypersomnia score 1st quartile (3) 1752 1.00 Reference 1.00 Reference 55,847 161 2nd quartile (4 to 5) 2246 1.19 0.98, 1.46 1.18 0.97, 1.45 71,109 245 3rd quartile (6) 2247 1.42 1.17, 1.72 1.39 1.14, 1.69 68,908 282 4th quartile (7 to 10) 1393 1.65 1.34, 2.04 1.65 1.33, 2.04 40,067 191 Sleep satisfaction None or a little 1052 1.00 Reference 1.00 Reference 30,693 120 Regular 1928 1.02 0.82, 1.27 1.04 0.83, 1.30 57,977 231 Enough 3299 0.89 0.73, 1.10 0.91 0.73, 1.12 103,811 363 High 1539 0.93 0.74, 1.17 0.97 0.77, 1.23 49,016 178 Score of sleeprelated phenomena 1st quartile (5) 2872 1.00 Reference 1.00 Reference 90,588 288 2nd quartile (6) 1391 1.08 0.89, 1.32 1.07 0.88, 1.31 42,979 148 3rd quartile (7 to 8) 1279 1.12 0.92, 1.37 1.10 0.90, 1.35 39,866 142 4th quartile (9 to 21) 1383 1.67 1.40, 1.99 1.63 1.36, 1.95 41,217 219 aIRR, Incidence Rate Ratio; CI, Confidence Interval. bAdjusted for sex and age (continuous). Further adjustment for the following variables did not modify the results: smoking, passive smoking, total alcohol intake, total caffeine intake, total physical activity, geographical area (coastal/inland), body mass index, Negative Affect and Positive Affect. 8 | TOHIDINIK eT al. confounders of stressrelated variables such as Negative Affect and Positive Affect, we cannot exclude that other dimensions of stress should be taken into account. Our results were consistent with the study of Ma et al. which found a relation between late bedtime and RAS (Ma et al., 2015). In another crosssectional study, Webb et al. revealed a possible association between RAS and increased sleep (Webb et al., 2013). However, both studies used a crosssectional design, which makes it difficult to draw any causal inference. In summary, our cohort study showed a moderate association between lack of satisfaction with sleep, insomnia, hypersomnia, and a high score of sleeprelated phenomena; and the occurrence of RAS in adolescents. These findings are biologically plausible and unlikely to be due to confounding by measured variables or selection bias. However, residual confounding due to unmeasured variables, including stressrelated variables, as well as some misclassification of the outcome due to the selfassessed diagnosis of RAS cannot be dismissed. Sleep hygiene measures and observance of the amount of sleep recommended by experts to reach optimal health in adolescents should be promoted (Paruthi et al., 2016). ACKNOWLEDGMENTS This work was funded by the Regional Ministry of Education, Universities and Vocational Training (Consellería de Educación, Universidades y Formación Profesional) [Grant number ED431C 2018/20], Santiago de Compostela, Spain. CONFLICT OF INTEREST The authors declare no potential conflicts of interest with respect to the authorship and/or publication of this article. AUTHOR CONTRIBUTIONS Hamid Reza Tohidinik: Conceptualization; Formal analysis; Writingoriginal draft; Writingreview & editing. Almudena Rodríguez: Data curation; Formal analysis; Investigation; Writingreview & editing. Carlos RegueiraMéndez: Conceptualization; Investigation; Methodology; Writingreview & editing. Bahi Takkouche: Conceptualization; Methodology; Project administration; Supervision; Writingreview & editing. ETHICAL APPROVAL The authors certify that informed consent was obtained from patients and that the study was performed in accordance with the Declaration of Helsinki. IRB: Comité Ético de Investigación Clínica de Galicia. Nº of approval: 2004/130 Date of approval: 7/7/2004. PEER REVIEW The peer review history for this article is available at https://publo ns.com/publo n/10.1111/odi.13955. DATA AVAILABILITY STATEMENT Anonymous data are available from the corresponding author upon reasonable request. ORCID Bahi Takkouche https://orcid.org/0000-0002-0739-2241 REFERENCES Akintoye, S. O., & Greenberg, M. S. (2014). Recurrent aphthous stomatitis. Dental Clinics of North America, 58, 281– 297. https://doi. org/10.1016/j.cden.2013.12.002 AlbanidouFarmaki, E., Poulopoulos, A. K., Epivatianos, A., Farmakis, K., Karamouzis, M., & Antoniades, D. (2008). Increased anxiety level and high salivary and serum cortisol concentrations in patients with recurrent aphthous stomatitis. The Tohoku Journal of Experimental Medicine, 214, 291– 296. https://doi.org/10.1620/ tjem.214.291 Amadori, F., Bardellini, E., Conti, G., & Majorana, A. (2017). Oral mucosal lesions in teenagers: a crosssectional study. Italian Journal of Pediatrics, 43, 50. https://doi.org/10.1186/s1305 201703677 Balan, U., Gonsalves, N., Jose, M., & Girish, K. L. (2012). Symptomatic changes of oral mucosa during normal hormonal turnover in healthy young menstruating women. Journal of Contemporary Dental Practice, 13, 178– 181. https://doi.org/10.5005/jpjourn als10024 - 1117 Bali, A., & Jaggi, A. S. (2016). An integrative review on role and mechanisms of ghrelin in stress, anxiety and depression. Current Drug Targets, 17, 495– 507. https://doi.org/10.2174/13894 50116 66615 05180 95650 Bassett, S. M., Lupis, S. B., Gianferante, D., Rohleder, N., & Wolf, J. M. (2015). Sleep quality but not sleep quantity effects on cortisol responses to acute psychosocial stress. Stress, 18, 638– 644. https:// doi.org/10.3109/10253 890.2015.1087503 Benca, R. M., Obermeyer, W. H., Thisted, R. A., & Gillin, J. C. (1992). Sleep and psychiatric disorders. Archives of General Psychiatry, 49(8), 669– 670. https://doi.org/10.1001/archp syc.1992.01820 08005 9010 Breslau, N., Roth, T., Rosenthal, L., & Andreski, P. (1996). Sleep disturbance and psychiatric disorders: A longitudinal epidemiological study of young adults. Biological Psychiatry, 39, 411– 418. https:// doi.org/10.1016/00063223(95)00188 - 3 Bruce, A. J., & Rogers, R. S. (2003). Acute oral ulcers. Dermatologic Clinics, 21, 1– 15. https://doi.org/10.1016/s0733 - 8635(02)00064 - 5 Gallo Cde, B., Mimura, M. A., & Sugaya, N. N. (2009). Psychological stress and recurrent aphthous stomatitis. Clinics (Sao Paulo), 64, 645– 648. https://doi.org/10.1590/s1807 - 59322 00900 0700007 García, J. B., GarciaPortilla González, M. P., Martínez, P. A. S., Fernández, M. T. B., Alvarez, C. I., & Domínguez, J. M. F. (2000). Propiedades psicométricas del cuestionario Oviedo de sueño. Psicothema, 12, 107– 112. GarciaPortilla González, M. P., Saiz, P. A., DiazMesa, E. M., Fonseca, E., Arrojo, M., Sierra, P., & Bobes, J. (2009). Psychometric performance of the Oviedo Sleep Questionnaire in patients with severe mental disorder. Revista de Psiquiatría Y Salud Mental, 2, 169– 177. https:// doi.org/10.1016/s1888 - 9891(09)73235 - 5 Gavic, L., Cigic, L., Biocina Lukenda, D., Gruden, V., & Gruden Pokupec, J. S. (2014). The role of anxiety, depression, and psychological stress on the clinical status of recurrent aphthous stomatitis and oral lichen planus. Journal of Oral Pathology & Medicine, 43, 410– 417. https://doi.org/10.1111/jop.12148 Greenland, S. (1995). Doseresponse and trend analysis in epidemiology: alternatives to categorical analysis. Epidemiology, 6, 356– 365. https://doi.org/10.1097/00001 64819950 700000005 Hansen, Å. M., Grynderup, M. B., Rugulies, R., Conway, P. M., Garde, A. H., Török, E., Mikkelsen, E. G., Persson, R., & Hogh, A. (2018). A cohort study on selfreported role stressors at work and poor sleep: does sense of coherence moderate or mediate the associations? | 9 TOHIDINIK eT al. International Archives of Occupational and Environmental Health, 9, 445– 456. https://doi.org/10.1007/s0042 001812947 Hatzinger, M., Brand, S., Perren, S., Stadelmann, S., von Wyl, A., von Klitzing, K., & HolsboerTrachsler, E. (2008). Electroencephalographic sleep profiles and hypothalamicpituitaryadrenocortical (HPA)- activity in kindergarten children: early indication of poor sleep quality associated with increased cortisol secretion. Journal of Psychiatric Research, 42, 532– 543. https://doi.org/10.1016/j.jpsyc hires.2007.05.010 Jiménez, K. M., PereiraMorales, A. J., & Forero, D. A. (2017). Val158Met polymorphism in the COMT gene is associated with hypersomnia and mental healthrelated quality of life in a Colombian sample. Neuroscience Letters, 644, 43– 47. https://doi.org/10.1016/j. neulet.2017.02.050 Jurge, S., Kuffer, R., Scully, C., & Porter, S. R. (2006). Number VI recurrent aphthous stomatitis. Oral Diseases, 12, 1– 21. https://doi. org/10.1111/j.16010825.2005.01143.x Keenan, A. V., & Spivakovksy, S. (2013). Stress associated with onset of recurrent aphthous stomatitis. Evidencebased Dentistry, 14, 25. https://doi.org/10.1038/sj.ebd.6400919 Kleinbaum, D. G., Kupper, L. L., & Morgenstern, H. (1982). Epidemiologic research: Principles and quantitative methods. John Wiley & Sons. Koybasi, S., Parlak, A. H., Serin, E., Yilmaz, F., & Serin, D. (2006). Recurrent aphthous stomatitis: Investigation of possible etiologic factors. American Journal of Otolaryngology, 27, 229– 232. https:// doi.org/10.1016/j.amjoto.2005.09.022 LeBlanc, M., Merette, C., Savard, J., Ivers, H., Baillargeon, L., & Morin, C. M. (2009). Incidence and risk factors of insomnia in a populationbased sample. Sleep, 32, 1027– 1037. https://doi.org/10.1093/ sleep/ 32.8.1027 Ma, R., Chen, H., Zhou, T., Chen, X., Wang, C., Chen, Y., Rao, S., Ge, L., & Lin, M. (2015). Effect of bedtime on recurrent aphthous stomatitis in college students. Oral Surgery, Oral Medicine, Oral Pathology and Oral Radiology, 119, 196– 201. https://doi.org/10.1016/j.oooo.2014.10.014 Meerlo, P., Sgoifo, A., & Suchecki, D. (2008). Restricted and disrupted sleep: Effects on autonomic function, neuroendocrine stress systems and stress responsivity. Sleep Medicine Reviews, 12, 197– 210. https://doi.org/10.1016/j.smrv.2007.07.007 Mizuki, N., Ohno, S., Sato, T., Ishihara, M., Miyata, S., Nakamura, S., & Inoko, H. (1995). Microsatellite polymorphism between the tumor necrosis factor and HLAB genes in Behcet's disease. Human Immunology, 43, 129– 135. https://doi. org/10.1016/01988859(94)00159 - n Paruthi, S., Brooks, L. J., D'Ambrosio, C., Hall, W. A., Kotagal, S., Lloyd, R. M., Malow, B. A., Maski, K., Nichols, C., Quan, S. F., Rosen, C. L., Troester, M. M., & Wise, M. S. (2016). Recommended amount of sleep for pediatric populations: A consensus statement of the American Academy of Sleep Medicine. Journal of Clinical Sleep Medicine, 12, 785– 786. https://doi.org/10.5664/jcsm.5866 Scully, C., & Felix, D. H. (2005). Oral medicine– update for the dental practitioner. Aphthous and other common ulcers. British Dental Journal, 199, 259– 264. https://doi.org/10.1038/sj.bdj.4812649 Scully, C., Gorsky, M., & LozadaNur, F. (2003). The diagnosis and management of recurrent aphthous stomatitis: A consensus approach. Journal of the American Dental Association, 134, 200– 207. https:// doi.org/10.14219/ jada.archi ve.2003.0134 ShohatZabarski, R., Kalderon, S., Klein, T., & Weinberger, A. (1992). Close association of HLAB51 in persons with recurrent aphthous stomatitis. Oral Surgery, Oral Medicine, and Oral Pathology, 74, 455– 458. https://doi.org/10.1016/00304220(92)90294 - z Steiger, A. (2002). Sleep and the hypothalamopituitaryadrenocortical system. Sleep Medicine Reviews, 6, 125– 138. https://doi. org/10.1053/smrv.2001.0159 Sun, A., Chen, H. M., Cheng, S. J., Wang, Y. P., Chang, J. Y., Wu, Y. C., & Chiang, C. P. (2015). Significant association of deficiencies of hemoglobin, iron, vitamin B12, and folic acid and high homocysteine level with recurrent aphthous stomatitis. Journal of Oral Pathology & Medicine, 44, 300– 305. https://doi.org/10.1111/jop.12241 Tabolli, S., Bergamo, F., Alessandroni, L., Di Pietro, C., Sampogna, F., & Abeni, D. (2009). Quality of life and psychological problems of patients with oral mucosal disease in dermatological practice. Dermatology, 218, 314– 320. https://doi.org/10.1159/00019 6973 Ussher, M., West, R., Steptoe, A., & McEwen, A. (2003). Increase in common cold symptoms and mouth ulcers following smoking cessation. Tobacco Control, 12, 86– 88. https://doi.org/10.1136/tc.12.1.86 Vgontzas, A. N., Bixler, E. O., Lin, H.- M., Prolo, P., Mastorakos, G., VelaBueno, A., Kales, A., & Chrousos, G. P. (2001). Chronic insomnia is associated with nyctohemeral activation of the hypothalamicpituitaryadrenal axis: Clinical implications. Journal of Clinical Endocrinology and Metabolism, 86, 3787– 3794. https://doi. org/10.1210/jcem.86.8.7778 Vgontzas, A. N., Mastorakos, G., Bixler, E. O., Kales, A., Gold, P. W., & Chrousos, G. P. (1999). Sleep deprivation effects on the activity of the hypothalamicpituitaryadrenal and growth axes: Potential clinical implications. Clinical Endocrinology, 51, 205– 215. https://doi. org/10.1046/j.13652265.1999.00763.x Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: the PANAS scales. Journal of Personality and Social Psychology, 54, 1063– 1070. https://doi.org/10.1037//00223514.54.6.1063 Webb, J. R., Webb, B. F., Schroeder, M. C., & North, C. S. (2013). Association of aphthous ulcers with selfreported symptoms of depression in a sample of smartphone users. Annals of Clinical Psychiatry, 25, 266– 270. Wray, D., Graykowski, E. A., & Notkins, A. L. (1981). Role of mucosal injury in initiating recurrent aphthous stomatitis. British Medical Journal, 283, 1569– 1570. https://doi.org/10.1136/bmj.283.6306.1569 Wray, D., Vlagopoulos, T. P., & Siraganian, R. P. (1982). Food allergens and basophil histamine release in recurrent aphthous stomatitis. Oral Surgery, Oral Medicine, and Oral Pathology, 54, 388– 395. https://doi. org/10.1016/00304220(82)90384 - x Zwiri, A. M. (2015). Anxiety, depression and quality of life among patients with recurrent aphthous ulcers. Journal of Contemporary Dental Practice, 16, 112– 117. https://doi.org/10.5005/jpjourn als10024 - 1646 How to cite this article: Tohidinik, H. R., Rodríguez, A., RegueiraMéndez, C., & Takkouche, B. (2021). Sleep quality and risk of recurrent aphthous ulcers: A Spanish cohort study. Oral Diseases, 00, 1– 9. https://doi.org/10.1111/ odi.13955