Background: Children with congenital heart defects (CHD) are reported to have poorer oral health compared with healthy children. The aim of the present study was to evaluate the effect of an intensive oral health care program among children with CHD followed from infancy to the age of five years, by comparing their oral health status at five years with a control group of children with CHD who had not received the program.

Methods: In this longitudinal study, children in western Norway with a need for lifelong follow-up due to congenital heart defects were invited to participate (n=119). Children born in 2008-2011 were offered a promotive oral health intervention program from infancy to the age of five years. The outcome measures for evaluating the intervention were dental caries prevalence, dental erosion, plaque index and gingival bleeding index. The data of the intervention group were compared with cross sectional oral health data of five year old controls with CHD born 2005-2007 (already published).

This was a prospective longitudinal design study, including four cohorts of children with CHD (born 2008-2011) who required lifelong follow-up for their CHD. Parents, who did not respond to the initial letter of invitation for their child, received a reminder by mail two months later. All children in the intervention group were recruited before three years of age. A total of 119 children (66 girls) were invited to participate and receive an oral health intervention; at baseline (below three years), at first follow-up (three years) and at second follow-up (five years). Because parents were responsible for their young children's oral health, we found no reason to exclude children with co-morbidities such as Down's syndrome, Di George syndrome, Noonan syndrome, Cornelia de Lange syndrome, Williams's syndrome and autism. According to their CHD, the children were divided in three groups; i) left-to-right shunts, e.g. atrial septal defect and ventricular septal defect, ii) obstructions, e.g. coarctation of the aorta, aortic stenosis, and pulmonary stenosis, and iii) complex heart defects, e.g. tetralogy of Fallots, transposition of the great arteries and hypoplastic left or right heart syndromes. At five years, data on oral health and oral health related background factors were compared with corresponding published data of 5-year-old children with CHD, also from Western Norway (n=67, born 2005-2007). The latter group served as controls. These children were recruited at the age of five years and had not been offered the intensive oral health intervention program received by children in the intervention group. The control group had been followed in the PDS in line with the general population of children, who are offered the first oral examination from three years of age.

The caries prevalence (at dentine level) of five year olds in the control group had previously been compared with the corresponding prevalence of five year olds in the general population from the national health register "KOSTRA" (Municipality-State-Report of 2010, 2011 and 2012 in three counties in Western Norway, n = 18,974). Likewise the caries prevalence at age five years of the present intervention group were compared with the general population (Municipality-State-Report of 2013, 2014, 2015, 2016, n=26161). Possible time trends in caries epidemiology of five year olds during 2010-2016 were assessed.

Children's clinical oral examinations:

At baseline (below three years of age) a simplified oral examination was performed. Children were examined sitting on their parents' laps in the dental chair or on a regular chair examined by the "lap to lap" method. At the first follow-up examination at three years old, the children sat either on parents' laps or alone in the dental chair. Procedures for the oral health examination of the intervention and of the control group at five years of age were the same. Examinations were performed using a mouth mirror and a dental probe under good lighting conditions. Caries and DDE were registered for all dental surfaces while index teeth were used for dental erosion, dental plaque and gingival bleeding. Bitewing radiographs (BW) were only taken in five year olds, when there were molar contacts. Parents with immigrant background were offered interpreting services if desired. All oral health examinations were conducted by the same two dentists who previously had collected the published inter- and intra-observer reliability data of the control group.

Criteria for the various diagnostic systems were as follows: For caries, a five-graded diagnostic system was used,where enamel lesions were characterised as grades 1 or 2, and dentine lesions as grades 3, 4, or 5. Secondary caries, filled and extracted (due to caries only) teeth and the care index (a fraction with filled teeth as numerator and d3-5mft count as denominator). For dental erosion, a four graded diagnostic system was applied on buccal and palatal surfaces of the primary maxillary anterior teeth (53-63) and the occlusal surfaces of all primary molars. Erosion grade 3 was considered severe and grade 4 very severe. A modified DDE index was selected for DDE registration in which demarcated opacities were coded 1, diffuse opacities 2, demarcated and diffuse opacities 3, hypoplasia 4, hypoplasia and opacities 5, post-eruptive breakdown of enamel 6, and atypical fillings 7. Finally, for dental plaque and gingival bleeding, the PI and GBI were reported around the marker teeth 55, 51, 65, 75, 71, and 85. The presence of plaque in the four amelo-gingival areas was assessed with a dental probe and absence of debris was recorded as score 0 and presence as score 1. Gingival bleeding was recorded by probing cautiously using a WHO periodontal probe with a 0.5-mm ball tip on six surfaces; mesio-buccal, in the middle of the buccal surface, disto-buccal, mesio-lingual, in the middle of the lingual surface and, disto-lingual. If bleeding occurred within 10 seconds in at least one of the places, the tooth was scored 1 (bleeding), otherwise the tooth was scored 0.

Parental questionnaire and anamnestic factors:

Parents of children in the intervention group were asked to respond to a questionnaire at baseline and at the second follow-up, while parents in the control group were asked at examination at age five years. The questionnaire contained 30 structured items. The topics covered were socio-demographic and anamnestic items, items about parental oral health attitudes, oral health related behaviours, and oral health promotion.The anamnestic background variables were coded cyanotic (1), not cyanotic (0), heart failure (1) and heart failure not present (0), taking heart medication (1) and does not take heart medication (0), low birth weight <2500 g (1) and normal birth weight >2500 g (0).

Intervention At baseline (before three years) all parents in the intervention group received a standardised "knowledge package", included "Lift the lip program". Both general and local interventions were repeated at first follow-up at three years and second follow-up at five years. Risk assessment was made by the following background factors: 1) dental factors such as visible caries, visible plaque, gingivitis, erosion and DDE; 2) parental oral health related behaviours such as oral hygiene habits, dietary habits and use of fluoride; 3) social factors such as immigrant background where parents showed poor language skills; and 4) medical factors such as comorbidity and use of medication with drugs sweetened with sucrose.

The recommendation from the caries risk assessment was the most intense preventive program for children with the highest caries risk (those children with the highest number of caries risk determinants). The children were evaluated based on data from the oral examination, anamnestic information and information obtained from the questionnaire and accordingly assigned to different promotive oral health intervention programs, denoted as Program I, II and III. As all the participants had CHD, stated previously as a risk for reduced oral Health, the most modest program applied was Program I. The children offered this program had no other known risk factors for poor oral health than CHD. Program II was assigned to children who had association with other known risk groups in terms of poor oral health or had signs of oral disease such as a) enamel caries grade1-2, erosion or DDE; b) findings from the questionnaire corresponding to additional risk for oral disease beyond what is normal for the group; c) immigrant background where parents show poor language skills and simultaneously caries lesions; and d) co-morbidities with additional risk for oral disease. Program III was assigned to children who also had other known risk determinants like very poor oral health, signs of severe oral disease or where findings from the questionnaire corresponded to serious risk of oral disease. After categorising the children to the various programs, each child's responsible dentist or dental hygienist at the local PDS clinic was contacted by telephone with information about the project and the findings of the examination. A written report with a treatment plan and recommendation for follow-up was sent after each oral examination: at baseline, first follow-up and second follow-up.

The initial intervention plan for recommendations for the PDS dentists for follow-up of the highest risk children (Program III) included an option of recalls every month during a three month period. Compliance by the PDS staff with the recommendation for follow-up was evaluated when the second follow-up for all children was completed. The child's local PDS-clinic was contacted and feedback in the form of a written list of the child's appointments in the PDS in the period from baseline to the second follow-up was received. There was no information about clinical procedures undertaken, only dates of appointments.

Statistical methods Descriptive statistics were used to characterise the intervention and the control group. The effectiveness of an early oral health intervention was estimated by comparing intervention and control groups on clinical oral health outcomes at age five years using a multiple logistic regression model for each outcome variable. The model was fitted unadjusted and adjusted for one of the following independent variables at a time: brushing habit, start age of tooth brushing, diet habit, parent's origin, parent's education, bottle-feeding, night meals, sugar water, sex, heart problems, cyanosis, birth weight, heart medication and syndrome.The various dependent dichotomised outcome variables for evaluating the effect of the intervention program were caries experience, dental erosion, PI and, GBI. The independent variables used in the logistic regression were selected in relation to known caries risk determinants and demographic background factors, which could influence the oral outcomes. In the final model, maximally adjustments were included to guarantee sufficient power. The adjustments with the largest deviation from odds ratio (OR) were selected in the unadjusted model if the deviation was at least 10% of the unadjusted OR. For other comparisons chi-square and independent sample t-test were used. The general significance level was set to 0.05. The Bonferroni correction was used to address multiple comparisons leading to the marginal significance levels of 0.0028 when comparing characteristics between the intervention and control group, and 0.01 for the logistic models. All analyses were performed using SPSS version 23 (Inc. Chicago, IL, USA) and Matlab 9.0 (Mathworks Inc.).

Inclusion Criteria:

Children with congenital heart defects (CHD) in western Norway, born 2008-2011 who require lifelong follow-up for their CHD.

Exclusion Criteria:

-