Effects of Breastfeeding and Sucking Habits on Malocclusion in a Birth Cohort Study

Efeitos da amamentação e dos hábitos de sucção sobre as oclusopatias num estudo de coorte ABSTRACT OBJECTIVE: To estimate the prevalence of malocclusion and to examine the effects of breastfeeding and non-nutritive sucking habits on dentition in six-year-old children. METHODS: A cross-sectional study was carried out nested into a birth cohort conducted in Pelotas, Southern Brazil, in 1999. A sample of 359 children was dentally examined and their mothers interviewed. Anterior open bite and posterior cross bite were recorded using the Foster & Hamilton criteria. Information regarding breastfeeding and non-nutritive sucking habits was collected at birth, in the fi rst, third, sixth and 12 th months of life, and at six years of age. Control variables included maternal schooling and child's birthweight, cephalic perimeter, and sex. Data were analyzed by Poisson regression. RESULTS: Prevalence of anterior open bite was 46.2%, and that of posterior cross bite was 18.2%. Non-nutritive sucking habits between 12 months and four years of age and digital sucking at age six years were the main risk factors for anterior open bite. Breastfeeding for less than nine months and regular use of pacifi er between age 12 months and four years were risk factors for posterior cross bite. Interaction between duration of breastfeeding and the use of pacifi er was identifi ed for posterior cross bite. CONCLUSIONS: Given that breastfeeding is a protective factor for other diseases of infancy, our fi ndings indicate that the common risks approach is the most appropriate for the prevention of posterior cross bite in primary or initial mixed dentition. 344 Malocclusion and sucking habits Peres KG et al. ABSTRACT OBJETIVO: Analisar a prevalência de oclusopatias e o efeito da amamentação e dos hábitos de sucção não nutritivos aos seis anos de idade.

With the decline in dental caries among children in many countries, 15 including Brazil,* other oral health disorders, such as malocclusion, have become relatively more important public health issues.Malocclusion is a developmental disorder of the craniofacial complex that affects jaw, tongue and facial muscles, 17 and is the result of an interaction of genetic and environmental factors. 3uch disorders can appear in primary dentition, where anterior open bite and posterior cross bite are the most prevalent conditions. 7Malocclusion causes functional and aesthetic disturbances in affected individuals, and treatment is often costly. 15ere are a number of controversies regarding the major causes of malocclusion in primary dentition and whether or not these are predictive of malocclusion in the permanent dentition.Anthropological studies that investigated population trends with respect to malocclusion have shown that environmental conditions are the primary explanation for changes observed in the occlusion patterns of populations. 2,24Notable among these environmental conditions are dietary habits, including the tendency in recent decades to adopt foods with a softer texture that require less chewing strength, habits of non-nutritive sucking such as the use of a pacifi er, bottle-feeding, and early weaning. 23though these conditions are potential causal factors for malocclusion, little is known about the effects of time of onset, duration, and cessation of these exposures and their possible combinations.A search of the National Library of Medicine -Medline database covering the period from January 1995 to June 2005, with no restriction on language, for articles on the relationship between breastfeeding and malocclusion in the primary dentition found fi ve studies, all from developed countries, examining the hypothesis that breastfeeding could be a protective factor against the occurrence of malocclusion.In addition, only two of these studies used a longitudinal design.
The aim of the present study was to determine the prevalence of malocclusion and analyze the effect of breastfeeding and non-nutritive sucking habits on the occurrence of malocclusion in six-year-old children participating in a birth cohort study.

METHODS
In 1993, a cohort study was begun which included all live births occurred in Pelotas, Southern Brazil, a city of approximately 300 thousand inhabitants.The main objective of this cohort study was to investigate health characteristics in the perinatal period and infancy, and included fi ve major sub-studies: perinatal, follow-up, infant mortality, hospital admissions, and psychological development of the child. 22 the perinatal sub-study (N=5,249), the fi ve hospitals in Pelotas were visited twice a day by a team of medical residents between January 1 st and December 31 st , 1993.The children identifi ed accounted for 99% of the babies born to mothers living in the urban area of the city.During the course of these visits an interview was carried out with the mothers which included questions on breastfeeding and pacifi er use.In addition, information was collected regarding socioeconomic and demographic characteristics, gestation, access to health care services, morbidity, and maternal occupation.Mothers were weighed and measured and children were weighed, measured and examined.In the follow-up substudy, a systematic sample of children was selected and visited during their fi rst (N=649), third (N=644), sixth (N=1,414), and 12 th month (N=1,383) of life.In 1998, a sample consisting of all low birthweight children plus 20% of the remainder was revisited.Of 1,460 eligible children, 87% (1,270 children) were located.In each of these cohort follow-ups, breastfeeding and non-nutritive sucking habits were investigated in detail in order to determine whether the mother had begun breastfeeding, and if so, if she was still breastfeeding or when she had stopped.Use of pacifi er was also investigated in these follow-up visits.Details on methodology and the major results of the fi ve sub-studies can be found elsewhere. 22e oral health study began in 1999 and adopted a cross-sectional design nested into the aforementioned cohort study.A sub-sample of the 1998 follow-up visits was examined in order to estimate the prevalence of anterior open bite and posterior cross bite.This was done to test, among other things, the hypothesis that breastfeeding acts as a protective factor against the development of malocclusion at age six years.The sample size needed to estimate the prevalence of these two conditions was calculated using the prevalence of anterior open bite described in the literature -approximately 45% -with an error of 5 percentage points.The calculation for a fi nite population of 5,249 children resulted in a sample size of 355.The sample size required to test the association between breastfeeding and malocclusion was estimated for an exposure defi ned as duration of breastfeeding < 9 months.Considering the detection of relative risks of at least 1.9 for anterior open bite and 2.5 for posterior cross bite, with respective prevalences of 54% and 20% in children breastfed for less than nine months (exposed), 8 a sample of 342 children was needed to provide 80% power at a signifi cance level of 5%.
The sample was infl ated by 10% to allow for losses or refusals, resulting in a rounded value of 400 children.Since low birthweight children (9.7% in the perinatal study) were overrepresented in the sample (29.7%), all analyses were weighted using a factor of 0.34 for low birthweight children and 1.27 for the remainder. 14piInfo 6.04 software was used for sample size calculations.
Dental examination included evaluations of malocclusion, dental caries, and soft tissue lesions.Fieldwork was carried out by three dentists, responsible for oral examination, and three interviewers, who administered the questionnaires.The dentists underwent a training and calibration exercise with 20 children not included in the sample and of the same age as the children studied.One of the examiners was considered as representing the "gold standard".Diagnostic reliability was determined for each of the clinical conditions by means of the Kappa coeffi cient, following methodology described in detail elsewhere. 13Diagnosis of malocclusion was based on the classifi cation by Foster & Hamilton. 5nterior open bite was defi ned as lack of vertical contact between the upper and lower teeth in the anterior region and unilateral or bilateral posterior cross bite was defi ned as transverse and reverse inter-relationship of one or more posterior teeth in one or both hemi-arches.Children were examined in their homes, seated, under artifi cial light.When necessary the mandible was gently guided towards centric occlusion by the examiner.Examiners wore suitable clothing and all recommended biosafety measures were observed.The outcomes studied (dependent variables) were anterior open bite and posterior cross bite, dichotomised (absence/presence), at six years of age.
Following examination, a structured interview was carried out with the mother, which included questions on the habit of digital sucking since the last cohort follow-up in 1998.Confounders and other risk factors of interest were obtained from the data collected in the previous follow-ups, and included parental schooling (>8 or ≤8 years of study), birthweight (adequate or low), cephalic perimeter at birth (>10 th percentile or The parents or guardians of the children included in the study were informed of the study's objectives, and written informed consent was obtained prior to the interviews and to oral examination.The study protocol was approved by the Committee for Ethics in Research of the Faculdade de Medicina -Universidade Federal de Pelotas.

RESULTS
The response rate in the present study was 89.8%, with 359 children being investigated.Losses were due mostly to families moving out of Pelotas.Diagnostic reliability, as assessed by the Kappa coeffi cient, was at least 0.85 (the value obtained for posterior cross bite).The sample comprised 193 (53.8%) boys and 166 (46.2%) girls.Prevalence of anterior open bite was 46.2% (95% CI: 41.0;51.4) and that of posterior cross bite was 18.2% (95% CI: 14.2;22.2).Prevalence of breastfeeding for less than one month, 1-3.9 months, 4-8.9 months, and nine months or longer were 15.7%, ≤10 th percentile), and regular use of a pacifi er before the age of four (yes/no).Information regarding the onset and duration of breastfeeding was also obtained from previous follow-up visits. 22ude associations between outcomes and predictors were estimated and signifi cance was determined using the Chi-square test.Since prevalence of the fi rst outcome was approximately 40%, Poisson regression was used in multivariable analyses to allow the estimation of correct prevalence ratios. 1 Initially, all risk factors of interest (time of breastfeeding, use of pacifi er, and habit of digital sucking) were included in the model.Then, regardless of their statistical signifi cance, variables sex of the child, birth weight, cephalic perimeter, and maternal schooling were used for adjustment in the fi nal model.This last variable was considered as an indicator of family socioeconomic status. 21The interaction between breastfeeding and pacifi er use was explored for the outcomes of interest, using the standard 5% signifi cance level.All analyses were performed using Stata 9.0 software.The interaction between duration of breastfeeding and use of pacifi er was tested but did not show statistical signifi cance (p=0.146).
Table 3 presents the result of Poisson regression for posterior cross bite.In contrast to anterior cross bite, here, the interaction between breastfeeding and use of pacifi er was statistically signifi cant (p=0.03).These risk factors showed similar prevalence ratios (7.4 and 7.9 respectively), and the interaction term was 0.1.This interaction works basically by zeroing the effect of one of the exposures in the presence of the other.This is shown in the Figure , where the prevalence of posterior cross bite is presented for the four exposure groups.
Children who either were breastfed for less than nine months or used pacifi er regularly between 12 months and four years are under a risk that was about 7.5 times greater than those who breastfed longer and did not use a pacifi er.These results show that the concomitant presence of breastfeeding for at least nine months and the non-regular use of pacifi er between 12 months and four years of age is necessary to ensure a protective effect (Table 3).

DISCUSSION
The present paper reports a cross-sectional study nested into a birth cohort investigating the prevalence of the main occlusal changes present at aged six years and the infl uence of breastfeeding and of non-nutritive sucking habits on the development of these conditions.In our sample, only 5% of children had permanent molars and/or permanent incisors, which means that dentitions were predominantly composed of primary teeth.Major strengths in the present investigation include high response rates during follow-up studies, high level of diagnostic reliability, and knowledge of the prospective factors investigated, in addition to its population base.A further aspect to note is that, although the outcomes investigated were ascertained by a cross-sectional study, these outcomes are unlikely to have developed before the main exposures analyzed, since breastfeeding and pacifi er use precede dental eruption.The study design makes the occurrence of recall bias unlikely since the information used was collected during or shortly after exposure, leading to short recall periods.Observation bias is are unlikely to have occurred, since observers, when performing oral examination, were unaware of the children's exposure status.Taken together, these methodological aspects ensured high internal validity.In spite of the fact that the dental outcomes analyzed in this study can be easily assessed in the epidemiological context, i.e., during fi eld work, relying on natural light and on a household chair, it is likely that the accuracy of examinations would have been greater in the clinical context, where X-rays, artifi cial lighting, and dental chairs are available.Moreover, the present results should be extrapolated with caution, particularly when considering that Pelotas shows more favorable socioeconomic conditions when compared to Brazil as a whole.
The positive effects of extended breastfeeding on the normal development of occlusion may be explained by the effect of sucking on the skeletal and muscular development of the child's face.The growth and development of facial structures takes place during the fi rst four years of the child's life, and 90% of this process is complete by age 12 years.The mechanics of sucking differs between children who are breastfed and those who are fed from a bottle.Breastfeeding promotes the correct development of the jaws, strengthening the muscles involved in the process of sucking to obtain breast milk.The mother's nipple adapts itself to the internal shape of the oral cavity, enabling a perfect oral seal. 4On the other hand, the artifi cial nipple of a baby bottle is made from more rigid material, which can "force" the interior of the oral cavity, causing inappropriate alignment of teeth and the transverse growth of the palate, conditions Breastfeeding* No pacifi er use between 12 months and 4 years # 0.1(0.2;0.9)0.1 (0.0;0.9) 0.1 (0.0;0. which can lead to the development of posterior cross bite. 4It is for this reason that national standards exist for regulating such products, the restriction of which is recommended, especially during the fi rst months of life.In addition, the position of lower lip and tongue during breastfeeding aids in the development of the habit of physiological swallowing -thus promoting this swallowing mechanism in adult life -, which does not occur with the use of a bottle. 4e habit of regular pacifi er sucking between age 12 months and four years was the main risk factor for the occurrence of anterior open bite at age six years, even after adjustment for breastfeeding duration.Breastfeeding for nine months or more, without regular use of pacifi er between age 12 months and four years, decreased the risk of posterior cross bite.
There are few multidimensional studies 8,12,19 on the etiology of malocclusion, and the majority of these 8,19 analyze the risk factors separately, without measuring concurrent impact or assessing potential interactions.Indeed, there are no previous studies in Brazil that follow a procedure similar to that employed in the present study.
Comparison with international studies showed that the prevalence of anterior open bite found in the current study (46.2%) was higher than that found in Germany (16%) 18 and Italy (13%). 23On the other hand, a study 8 carried out in Recife, Northeastern Brazil, also detected higher prevalence (36.4%).Prevalence of posterior cross bite was 18.2%, which is higher than that found in the Italian study (7%), 23 and in the other Brazilian study (12.1%). 8These differences may be attributed, in part, to the use of different diagnostic criteria, as well as to the differences as to the age ranges studied, which can infl uence the prevalence these outcomes.For example, samples with different ages may bias results, since prevalence of anterior open bite and posterior cross bite tend to be higher in younger children than in six-year-olds, whose primary dentition is complete. 19onsequently, older children have larger exposure time to orthodontic treatment, as well as a tendency to abandon non-nutritive sucking habits.In addition, it is well known that the prevalence and severity of dental caries are risk factors for malocclusion.Since dental caries are more prevalent in developing countries, it is reasonable to suppose that prevalence of malocclusion should be higher in these regions as well.Moreover, prevalence of malocclusion is expected to vary with the prevalence of non-nutritive sucking habits.There are considerable differences in feeding, as well as in artifi cial sucking habits, between different areas of the world and during different periods. 11Finally, substantial variation in epidemiological criteria may also infl uence the prevalence found in these different studies.
In the present study, the nature of feeding during the fi rst years of the child's life, as well as the presence of non-nutritive sucking habits, showed different effects on the different occlusal changes investigated.The protective effect of breastfeeding for nine months or more against the development of anterior open bite was offset by the regular pacifi er use between 12 months and four years of age.Previous studies analyzing this same relationship found confl icting results, showing either an effect 20 or no effect of breastfeeding on the occurrence of malocclusion. 23Such confl ict in results may be related to cultural differences between the populations studied, to different non-nutritive sucking habits, and also to the range of study designs and statistical methodologies employed.
Our fi ndings agree with those of Viggiano et al 23 with regard to the etiological role of the pacifi er and to the lack of -or, at most, small infl uence of -breastfeeding on anterior open bite.The occurrence of anterior open bite appears to be more related to local deformation (related to the use of a pacifi er) than to a defi ciency in muscleskeletal development due to lack of breastfeeding.
An inverse relationship between duration of breastfeeding and presence of posterior cross bite has already been described in a previous study. 23However, in the present study the protective effect of breastfeeding was only detected in the absence of regular pacifi er use.Either breastfeeding for less than nine months or regular pacifi er use implied in a similarly increased risk of malocclusion.
The effect of non-nutritive sucking habits on the development of occlusion has been under investigation for several decades, while studies of the effects of breastfeeding on occlusion are a more recent phenomenon. 23he main question arising from these studies is whether such conditions disappear once the risk factors for malocclusion have been identifi ed and removed.The limited amount of data in the literature show that anterior open bite tends to disappear with the disappearance of the habit, while the same does not occur in the case of posterior cross bite. 9,10Further studies specifi cally designed to elucidate this relationship are required; such studies could also help determine whether or not the presence of malocclusion in primary dentition is predictive of malocclusion in permanent dentition.
Breastfeeding protects against various childhood health problems.The advantages of maternal milk are related to maturation of the baby's immune system, less intense allergic reactions to the nutrients of maternal milk, and protection against malnutrition, diarrhea, and respiratory diseases.Other positive aspects include the nutritional benefi t resulting from the milk's composition, improvements in the child's psychological well-being, and the substantial economic advantages. 6alth care strategies can include intersectorial approaches to health promotion, based on population strategies and on the common risks approach. 16Public policies that include an extension of current maternity leave allowances, as a stimulus for longer duration

Figure .
Figure.Prevalence of posterior crossbite* according to breastfeeding (9 months or more) and pacifi er use (between age 12 months and 4 years) among 6-year-old children.Pelotas, Southern Brazil, 1999.

Table 1 .
Sample distribution according to anterior open bite, posterior cross bite, and associated factors (χ 2 ).

Table 2
presents the results of Poisson regression for anterior open bite.After adjustment for use of pacifi er, breastfeeding duration lost statistical signifi cance, whereas digital sucking became statistically significant.Children who used pacifi er regularly between 12 months and four years of age showed a risk of anterior

Table 2 .
Association between anterior open bite, breastfeeding, and non nutritive sucking habits in six-year-old children.Pelotas, * Breastfeeding and pacifi er use adjusted for each other ** Adjusted for breastfeeding and pacifi er use *** Breastfeeding and pacifi er use adjusted for each other and for birthweight, cephalic perimeter, sex, and maternal schooling **** Adjusted for breastfeeding, pacifi er use, birth weight, cephalic perimeter, sex, and maternal schooling