The incidenceof parasornniasin child bruxers versus nonbruxers

Cynthia L. Weideman,DDS Debbie L. Bush, DDSFrisca L. Yan-Go, MD Glenn T. Clark, DDS, MS ~effrey A. Gornbein, PhD

Abstract study on genetic predisposition reported that parents who demonstrated tooth grinding in childhood more Bruxismin children has been reported to occur in as- 2 sociation with certain parasomnias(i.e., sleep talking, bed frequently have children who grind their teeth. An wetting). Various dental, medical, neurological, and psy- individual’s general health status mayalso determine chologicalrisk factors also have been correlated with brux- propensity to grind. For instance, children grind more ism. A case-control study was therefore conductedto test frequently whenthey are suffering from allergic rhini- the null hypothesis that there is no difference between tis or asthma2Neurological disabilities such as autism and cerebral palsy also have been strongly indicated as bruxers and nonbruxers in the occurrence rate of other 4 parasomniasand these reported risk factors. A 54-item risk factors for bruxism. The most consistently men- survey questionnaire was developedand mailed to 342 pe- tioned etiology of nocturnal bruxism is psychological diatric patients, half of whomwere avowedto be bruxers stress. Studies suggest that bruxismis positively related to emotional tension, anxiety, and the anticipation or by their parents. These patients were selected randomly 5 from a private pediatric practice in NorthernCalifornia. actual experienceof life stresses. One-hundredfifty-two subjects (77 bruxers and 75 con- The relationship between bruxism and a patient’s trols) returnedthe questionnaire,and steprvise logistic re- sleep stage has been investigated. Generally most jaw gression analysis revealed that five of the 54factors (noc- muscle activity occurs during light phases of sleep and has been observed to take place in connection with turnal muscle cramps, bed wetting, colic, drooling while 6 sleeping, and sleep talking) showedsignificant differences bodily movement. The Association of Sleep and betweenbruxers and controls (odds ratios rangedfrom 3.11 Arousal Disorders classifies bruxism as a parasomnia. to 1.95). Thesefindings strongly suggest the possibility of Parasomnias are defined as undesirable physical events a commonsleep disturbanceunderlying these nonsleep-stage that occur exclusively or predominantly during sleep, specific parasomnias.(Pediatr Dent 18:456-60, 1996) usually taking the form of motor or autonomic phe- nomena often associated with variable degrees of 7 ruxism is defined as the habitual, nonfunctional arousal. Normaltransitions between the awakestate, grinding of the teeth and is characterized by rapid eye movement (REM), and nonrapid eye move- B forceful, rhythmic contact of the occlusal tooth ment (NREM)sleep occur in an orderly manner. The surfaces with mandibular movement.1 Although brux- primary sleep parasomnias are disorders of this sleep/ ism occurs during the awakeand sleep states, the lat- wake cycle and present as unusual and sometimes dra- ter is more common.This behavior in children is of matic behaviors. They are classified as NREM,REM, particular interest to the pediatric dentist. Parents com- and non-sleep stage specific depending on when the monlyreport that while their children sleep, they pro- behavior takes place during the sleep cycle. Bruxismis duce a loud, grinding noise that can be heard by other categorized as nonsleep-stage specific (Table 1). Al- family members. Reported consequences of bruxism though it mayoccur at any stage of sleep, bruxism is include headaches, temporomandibular joint dysfunc- most likely to take place during stage 2 of NREMsleep tion, masticatory soreness, and attrition of the teeth. or8 during REMsleep. The literature suggests various dental, physical, psy- The commonfeature of all of the parasomnias is as- chological, and sleep-related factors associated with sociation with abnormal sleep arousal patterns. Ware bruxism. and Rughreported that 85%of bruxism occurrences are Risk factors implicated in bruxism are heredity, gen- associated with limb muscle activity that results in an eral health, and central nervous system disorders. A arousal from sleep. They subsequently found a group

456 American Academyof Pediatric Dentistry Pediatric Dentistry - 18:7, 1996 school performance. Children TABLE1. PARASOMNIASIN CHILDREN can injure themselves while engaged in parasomnia behav- L prirnc¢~Sleep P~cr~so~rcfc¢~ A. NREMparasornnias iors. If sleep disturbances are 1. Sleepstarts frequent or persistent, psycho- 2. Disordersof arousal logical trauma may be indi- 3. Sleep drunkenness cated as a causal factor. Lastly, B. REMparasomnias sexual abuse has been linked to 1. Dreamanxiety attacks sleepa2 disorders, 2. Hypnagogichallucinations and/or sleep paralysis Diagnosis of bruxism and 3. REMsleep behavior disorder other parasomnias is obvi- C. Nonsleep-stage specific parasomnias ously difficult since these be- 1. Bruxism haviors all occur during sleep. 2. Enuresis Recognition of these events 3. Rhythmic movementdisorder requires verification by par- 4. Periodic movementsof sleep ents because children have no 5. Posttraumatic stress disorder subjective awareness of their 6. Somniloquy nocturnal behaviors. Based on II. Secondart[Sleep Parasomias parental confirmation, this pa- A. Central nervous system per investigates the factors as- 1. Seizures sociated with bruxism and the 2. Headaches relationship between bruxism B. Cardiopulmonary and other parasomnias. 1. Sleep-related arrythmias 2. Nocturnal asthma Materials and methods 3. Miscellaneous Sampleselection 4. Sleep apnea The original sample se- C. Gastrointestinal lected for the case-control 1. Gastroesophagealreflux study consisted of 346 boys 2. Diffuse esophageal spasm D. Miscellaneous and girls, 173 bruxers and 173 1. Panic attacks nonbruxers, between the ages 2. Nocturnal muscle cramps of 2 and 18. A private pediat- 3. Psychogenicdissociative states ric dental practice located in Northern California pro- From Mahowald & Rosen, 1990. vided 2,750 active patient records from which the sub- of patients whocomplained of both sleep problems and jects were selected. The children in this practice are frequent bruxism and speculated that bruxism was cor- middle or upper middle class and reside in a district related with other sleep disorders such as insomnia or outside of Sacramento. excessive daytime sleepiness. 9 Someparasomnias are Potential subjects were called in alphabetical order thqught to occur simultaneously. For example, children until a memberof the family--usually one of the child’s commonlyhave a history of both sleep walking and parents-~confirmedthat the child was a bruxer. A bruxer night terrors, a° Fisher and McGuirefound sleep talk- was described as a child whoseparents or other family ing to be associated significantly with unrecalled night- members have heard unmistakably loud, repetitive, mares, sleep walking, and sitting up in bed. Through grinding sounds comingfrom the sleeping child’s teeth. factor analysis, they discovered a group of associated The control subjects were recruited by selecting the sleep behaviors that occurred in close association with patient who was closest alphabetically and of similar the parasomnias and concluded that an individual with age to the chosen tooth grinder. A telephone call to the oneaa parasomnia was likely to exhibit more than one. subjects’ parents confirmed they did not brux. A po- It has not been proven whether bruxism occurs in tential control child found to be a bruxer was added to association with other parasomnias, but such knowl- the case dataset and the next two records were used as edge would be useful since bruxism is more easily de- controls. A total of 1,396 parents were called in order tected than some of the other parasomnias. Although to find the desired samplesize of 346 subjects (an equal parasomnias in childhood usually represent a normal number of bruxers and nonbruxers) necessary to have variation in the developmental process of the central sufficient size for statistical analysis. nervous system, unfavorable consequences from these Exclusion criteria for the 1,396 parents telephoned behaviors are possible. 12 Their sleep disturbance may included parents who did not speak English, were un- cause excessive daytime sleepiness and result in poor willing to participate, or failed to understandthe study.

Pediatric Dentistry - 18:7, 1996 American Academy of Pediatric Dentistry 457 If at least two unsuccessful at- TaBLe2. Factors reLatEo to BRUXING(P< 0.15)" tempts were made to contact the family, the next appropriate chart Yes (%) was located. The number of par- Variable BY~I~F$ Controls P-value ents who were called and who fit 1. Lip biting 6 (7.7) 0 0.014 the exclusion criteria was 772, of 2. Drooling while sleeping 45 (57.7) 29 (38.2) 0.015 which 278 were never reached. 3. Colic 18 (23.4) 7 (9.7) 0.026 Each of the 346 potential sub- 4. Jaw soreness upon waking 7 (9) 1(1.3) 0.034 jects agreed to participate. They 5. C-section delivery 11 (14.3) 20 (28.2*) 0.038 were listed separately as bruxers 6. Sleep talking 55(70.5) 42 (55.3) 0.05 and nonbruxers and the parents 7. Gum chewing 8 (10.3) 2 (2.6) 0.055 were sent an anonymous ques- 8. Nocturnal muscle cramps 15 (19.2) 7 (9.2) 0.076 tionnaire with a self-addressed, 9. Poor health during school years 40(55.6) 42 (72.4*) 0.076 stamped envelope. 10. Child opens eyes while sleeping 33 (42.3) 22 (29) 0.084 11. Bed wetting 0.114 This study was exempt from 14 (18) 7(9.2) 12. Age child slept through the night 12.5~ 8.6+ 0.117 the Human Subject Protection 13. Induced labor 20 (26) 11 (15.5) 0.117 Committee and did not require 14. Sleep walking (wake up in a different room) 11(14.1) 18 (23.7~) 0.128 informed consent for the follow- 15. Siblings have similar sleep problems 21 (26.9) 29 (38.2*) 0.137 ing reasons: 1) this research pre- 16. Child awakens irritable 39 (50) 29 (38.2) 0.139 sents no more than minimal risk 17. Quick jerks of arms or legs when sleeping 29 (37.2) 20 (26.3) 0.148 of harm to subjects and involves 18. Child moves around while sleeping 49 (62.8) 39 (51.3) 0.149 no procedures for which consent "Chi-squareanalysis of the percentageof "yes" responsesbetween bruxers and controls. is normally required outside of ~ Meanvalues of the age(in months)at whichchild first slept throughthe night. the research context, and 2) the * Responsehigher in controls--ProtectiveFactor. only record linking the subject Meansfor numericalvalues compared via T-tests. N = 152(77 bruxersand 75 controls). and the research would be the consent13 document. representative of the entire sample, 20 of the 194 Questionnaire development nonrespondents (equal number of bruxers and nonbruxers) were contacted randomly by phone and Specialists in dental research (oral motor disorders), the questionnaire completed by these nonrespondents pediatric dentistry, pediatric medicine, pediatric neu- by phone interview. The responses of the initial 132 rology, and sleep disorders helped develop a question- subjects were compared to the group of 20 using a naire. The following source materials were also used chi-square test. Because there were no significant dif- in the development process: ferences (P < 0.15), the two subsamples were com- 1.14 The Children’s Sleep Behavior Scale 15 bined to form a total sample of 152 (77 bruxers and 2. The Pediatric Questionnaire 75 nonbruxers). 3. Teeth grinding questionnaire and Standardized Analysis of difference in the proportions of yes re- 16 Questionnaire sponses between the bruxer and nonbruxer groups for 4. Mail and Telephone Surveys: The Total Design each yes/no question then was performed using a chi- 17 Method. square methodology. For questions in which numeri- The questionnaire was developed to be completed cal means were available, a t-test for significant differ- easily by parents or others most knowledgeable about ences was performed. All P values given are two-sided. a child’s sleep habits. The term "tooth grinding" re- The yes/no questions that were shown to be placed "bruxism" in the questionnaire to facilitate its strongly related to bruxing via chi-square analysis were completion. Six categories divided the 54 items in- included in a stepwise multiple logistic regression cluded in the survey: general tooth grinding informa- where bruxism (yes or no) was the outcome (Y vari- tion, night-time behavior, morning behavior, daytime able). This model allowed us to determine which of behavior, family history, and general medical history. these factors were still associated after controlling for The first draft of the questionnaire was piloted with all the other factors. The P value to enter for this a small sample of staff members (N = 3) of the pediat- stepwise procedure was set at P = 0.15. A liberal alpha ric dental office who were parents and parents (N = 7) level was chosen since this is an exploratory study, not who were selected at random during an office visit with a confirmatory study. Odds ratios and their corre- their children. After the pilot study, the questionnaire sponding 95% confidence intervals are reported based was modified, approved by the team, and mailed to the onTM this logistic model. selected sample. Data analysis Results Of the 346 subjects available, 132 initially responded A total of 152 questionnaires from 77 tooth bruxers to the survey. To confirm that the respondents were (39 females and 38 males) and 75 control subjects (34

458 AmericanAcademy of Pediatric Dentistry PediatricDentistry - 18:7, 1996 bias was minimized.Third, this was a case-control study in which subjects Variant OddsRatio LowerLimit UpperLimit were selected as randomlyas possible. 1. Nocturnal muscle cramps 3.11 0.99 9.75 We are confident that the data 2. Bed wetting 2.62 0.57 11.97 collected in this study are reflective 3. Colic 2.57 0.91 7.26 of a true association between these 4. Droolingwhile sleeping 2.0 0.98 4.08 behaviors. The limitations of our 5. Sleep talking 1.95 0.94 4.05 data include restricted generaliza- 6. Agechild slept throughthe night (years) 1.22 1.19 1.26 tion of the sample and reliance on 7. C-section" 0.4 0.17 0.97 audible reports of bruxism sounds to "Thisis a protectivevalue (OR < 1 ). identify bruxers instead of actually measuring bruxism. Bruxers who do females and 41 males) was returned. Of the 52 ques- not makenoise maynot have been identified. tions, 45 required a yes/no response. Seventeenof these Whenthe statistically related predictors of bruxism yes/no questions and a single numerical response were combined, the following composite picture of a question were found to be strongly related to bruxing child bruxer was developed. The child begins bruxing in children (P < 0.15). The numberof "yes" responses, at a meanage of 3.7 years and has a 2.57 times higher the percentage of positive responses, and the mean likelihood of experiencingcolic as an infant. The bruxer value for each of these questions are seen in Table 2. first slept throughthe night at a later age (12.5 months) These 18 questions were analyzed by stepwise logis- than the nonbruxer (8.6 months), suggesting that brux- tic regression whichdetermined that six of the 18 ques- ism was preceded by an earlier disturbed sleep pattern. tions (Table 2) were still associated with bruxismafter The bruxing child also has a 3.11 times higher likeli- controlling for the other factors. Table 3. reports the hood of skeletal muscle cramping and 2.0 times higher odds ratio and the upper and lower limits (based on likelihood of drooling during sleep. Finally, the child 95%confidence interval) for each of these questions. whobruxes has a greater tendency to sleep talk (1.95) For example, bed wetting had an odds ratio of 2.62 and and wet the bed (2.62). Wespeculate that these findings therefore the bruxing child had a 2.62 times higher like- suggest the possibility of a commonsleep disturbance lihood of wetting the bed than the nonbruxer. Accord- underlying these nonsleep-stage specific parasomnias. ing to the reported upper and lower limits, 95%of the The questions that showed no group difference are bruxing children had an odds ratio for bed wetting also of importance. Specifically, only 14%of the par- between 0.57 and 11.97. Caesarean-section showed an ents in this study reported their child’s bruxing began odds ratio less than 1 and is thus a protective factor. or worsenedafter a traumatic life event (i.e., death, Caesarean-section delivery was more prevalent divorce, or physical or psychological abuse). Addition- amongthe nonbruxers than among the bruxers. ally, only 34%of parents responded that their child’s bruxing started or was accentuated after a certain event Discussion in their child’s life (i.e., toilet training, birth of a new Results from the logistic regression analysis suggest sibling, or starting school). The true occurrenceof these that a total of six night-time behaviors and physical stressful events, which children do not rememberand conditions occurred in association with bruxism. Over- parents maynot observe, is difficult to measure. Dis- all, the data reject the null hypothesis and the findings crepancies amongreported rates of parasomnias in cor- support the idea that bruxism and other parasomnias relation with stressful events should not be overlooked. occur together in children predisposed to these sleep- Researchers claim that one’s general health status related events. These data agree with a prior report that mayprompt bruxing. Marks implied that children who demonstrated that children commonlyhave a history suffer from allergic rhinitis or asthma brux more fre- of multiple parasomnias. Our findings disagree with quently. Our data dispute the idea that the poorer an Fisher and McGuire’sfactor structure of the Children’s individual’s health, the morelikely the display of noc- Sleep Behavior Scale, which showedno behavioral fac- turnal tooth grinding. Parents in our study actually tor that corresponded exactly to the parasomnias. They reported that nonbruxing children tended to have a also reported that enuresis (bed wetting) was not re- significantly poorer health status than bruxers during latedu to any of the sleep behaviors assessed, their school years. Another question revealed that This study has several advantages over prior stud- bruxers and nonbruxers were reported to have the ies. First, the questionnaire was carefully designed in same incidence of asthma and ear infections. Bruxers accordance with the standards outlined by Dillman and and controls showed equal proportions of sleepiness the previously described team. The items were easy to during the day, dispelling speculation about the rela- read and organized into a continuous and logical for- tionship between daytime sleepiness and bruxism. mat. They were ordered in degree of difficulty and Stepwiselogistic regression analysis revealed that cae- most of them required a yes or no response. Second, sarean-section delivery had an odds ratio of less than pretesting was conducted prior to the actual study and 1 and is thus consideredto be a protective factor. There-

PediatricDentistry - 18:7,1996 AmericanAcademy of PediatricDentistry 459 fore, caesarean-section delivery was more prevalent g. Has 1.95 times higher likelihood of sleep among the control subjects than the bruxers. The only talking. physical finding that showed a positive correlation Dr. Weidemanis a pediatric dentist in Citrus Heights, California; with bruxism was infantile colic. Colic is sudden benign Dr. Bushis a pediatric dental resident, at BaylorUniversity; Dr. abdominal pain during the first 3 months of life, a medi- Yan-Gois a pediatric neurologist at the neuropsychiaticinstitute cal condition that has been associated with insomnia. at UCLA;Dr. Clark is associate dean of UCLAdental research, and Dr. Gornbeinis a biostatistician, at UCLAMedical Center. Colic19 is a commoncause of major sleep disruptions. In conclusion, these results may serve as an infor- 1. Leung AK,Robson WL: Bruxism: howto stop tooth grind- mational tool for the pediatric dentist. Concerned par- ing and clenching. Postgrad Med89:167-68, 1991. 2. Abe K, ShimakawaM: Genetic and developmental aspects ents of a bruxing child in the dental setting need to be of sleeptalking and teethgrinding. Acta Odontologica reassured and educated by the clinician. Bruxism, bed Scandinavica 32:291-356, 1974. wetting, sleep talking, and nocturnal muscle cramping 3. MarksMB: Bruxism in allergic children. AmJ Orthod77:48- may occur simultaneously. Current evidence suggests 59, 1980. that these related sleep disturbances take place in nor- 4. Rugh JD, Harlan J: Nocturnal bruxism and temporoman- dibular disorders. AdvNeurol 49:329-41, 1988. mal children with no underlying physical or psycho- 5. FunchDP, Gale EN:Factors associated with nocturnal brux- logical disorders. If these behaviors do occur simulta- ism and its treatment. ] BehavMed 3:385-97, 1980. neously and become disruptive to the family, referral 6. AhmadR: Bruxismin children. J Pedodontics10:105-26,1986. to a pediatrician, pediatric psychologist or pediatric 7. Thorpy MJ, Govinsky PB: Parasomnias. Psychiatr Clin North Am10:623-39, 1987. neurologist for further evaluation and treatment may 8. MahowaldMW, Rosen GM:Parasomnias in children. Pe- be indicated. diatrician 17:21-31,1990. 9. WareJC, RughJD: Destructive bruxism:sleep stage relation- Conclusions ship. Sleep 11:172-81,1988. 10. Clore ER, Hibel J: The parasomniasof childhood. J Pediatr 1. Chi-square analysis revealed that 18 questions Health Care 7:12-16, 1993. were strongly associated with bruxism in chil- 11. Fisher BE, McGuireK: Do diagnostic patterns exist in sleep dren (Table 1). behaviors of normal children? J AbnormChild Psychol 2. Stepwise logistic analysis determined that six of 18:179-86,1990. these 18 factors (nocturnal muscle cramps, bed 12. Kales A, Soldatos CR, Kales JD: Sleep disorders: insomnia, wetting, colic, drooling while sleeping, sleep sleep walking, night terrors, nightmares and enuresis. Ann Int Med106:582-92, 1987. talking, and the age at which the child first slept 13. HumanSubject Protection Committee,University of Cali- through the night) were still associated with fornia, Los Angeles: HumanSubject Protection Committee bruxism after controlling for the other factors. Application. Los Angeles, 1990. Caesarean-section delivery was found to be 14. Fisher BE, Pauley C, McGuireK: Children’s Sleep Behavior Scale: normativedata on 870 children in grades I to 6. Per- more prevalent among the nonbruxers. cept MotSkills 68:227-36,1989. 3. The child bruxer: 15. Minnesota Regional Sleep Disorders Center, Hennepin a. Begins bruxing at a mean age of 3.7 years CountyMedical Center: Pediatric Questionnare. Hennepin, b. Has 2.57 times higher likelihood of experi- MN,1991. encing colic as an infant 16. Lindquist B: Bruxismand emotional disturbance. Odontol Revy23:231-42, 1972. c. First slept through the night later (12.5 17. Dillman D: Mail and Telephone Surveys: The Total Design months) than the nongrinder (8.6 months) Method.Washington: John Wiley & Sons, Inc, 1977. d. Has 3.11 times higher likelihood of experi- 18. Fisher L, VanBelleG: Biostatistics: A Methodologyfor the encing nocturnal muscle cramps Health Sciences. Washington:John Wiley &Sons, Inc, 1993. 19. Ferber R: Dependent causes of insomnia and sleepiness e. Has 2 times higher likelihood of drooling in middle childhood and adolescence. Pediatrician 17:13- during sleep 20, 1990. f. Has 2.62 times higher likelihood of wetting the bed

460 AmericanAcademy of Pediatric Dentistry PediatricDentistry - 18:7, 1996