Study design
This cross-sectional study was organised at NIMR in Hungary. Examination and data collection were performed from August 2016 to September 2020. This research has been approved by the Medical Research Council, Scientific and Research Committee of Hungary (No: ETT-TUKEB IV/1433–1/2020/EKU), and was performed according to the Declaration of Helsinki [20]. All patients provided written consent for participation. The study followed the recommendations of STROBE (Strengthening the Reporting of Observational Studies in Epidemiology).
Study size
The sample size was determined based on the results of previous cross-sectional study in this subject area using the G*Power 3.1 software (v.3.1.9.3, 2017, Institut für Experimentelle Psychologie, Heinrich-Heine-Universität, Düsseldorf, Germany). Our calculation was based on the results of Orsós et al. [21] according to which if α (false positive rate) was set at 0.05, to reach a power of 95% with a 1:1 distribution ratio between study groups the minimal sample size should be at least 97 (association between smoking and dental status) and 132 (association between alcohol consumption and dental status) per study group.
Patients
410 post-stroke patients (174 female, 236 male), whose primary diagnosis was subarachnoid hemorrhage, intracerebral hemorrhage or cerebral infarction (International Classification of Diseases, 10 th revision (ICD-10): I60x, I61x, I63x, I67x) were included in this study. In processing the data, the two main types of hemorrhagic stroke (intracerebral and subarachnoid) were examined separately according to the International Classification of Diseases recommended by the World Health Organization (WHO) [22]. All of the stroke patients were recruited while receiving rehabilitative care at NIMR. Inclusion criteria were medical stability and confirmed stroke diagnosis. Inpatients were excluded if they had had nasogastric tube placement or if they indicated communication difficulties (unable to follow a 1-step command) or if they had recurrent stroke.
Personal medical history and functional assessment
On admission, the medical reports of the patients were collected, the history, the initial diagnosis, the type of stroke (subarachnoid or intracerebral and cerebral infarction), time since diagnosis, presence of risk factors for stroke (diabetes, hypertension, hyperlipidemia, atrial fibrillation, ischemic heart disease, heart failure), post-stroke sequelae [hemiplegia/hemiparesis, tetraplegia (paralysis in the upper and lower body), dysphagia (swallowing problems), dysarthria/anarthria (complete loss of speech motor ability), ataxia and depression], Functional Independence Measure (FIM) scores [23] and Barthel Index scores [24]. The FIM and the Barthel index are widely used outcome measures in neurological rehabilitation that assess the level of independence in activities of daily living [25]. The FIM is an 18-item tool that estimates function in 6 areas (self-care, including sphincter control, tranfers, mobility, communication, cognition) and measures functional disability of inpatients in terms of their need for assistance. Each item is scored from 1 to 7 and total FIM score range from 18 to 126 [26]. The Barthel index define the grade of the self-sufficiency inlcuding 10 personal activities. The total score for the Barthel index is a value between 0 and 100 [27].
Socio-demographic factors
Socio-demographic background (age, gender, level of education, occupational status, location type) were obtained from a validated questionnaire and patient documentation. The patients’ age was classified as 17–24, 25–49, 50–64, 65–79, and 80–89 years. Gender was recorded as male or female. Level of education was categorised into seven groups: no formal schooling, less than primary school, primary school completed, secondary school completed, high school completed, college/university completed, postgraduate degree. Occupational status was recorded according to patients’ responses based on local opportunities, which were classified into three main groups according to the Hungarian Central Statistical Office [28], as active (employed, self-employed), inactive (student, housewife, stay at home parent, pensioner, altered work ability) and unemployed. Location type (permanent address) was categorised into three groups: capital, periurban area and rural.
Behavioural factors
Dental health practices (last dental visit, tooth-brushing habits), frequency of smoking and alcohol consumption were assessed using a validated questionnaire that was based on the Oral health surveys: basic methods, 5th edition by WHO [29].
Oral examination
The oral examination was carried out by two calibrated dentists using a plane mouth mirror, a dental probe and an artificial light. The examiners were trained and calibrated by an experienced dental epidemiologist. The calibrator examined 25 subjects who were also examined by two dentists. The diagnosis found by reference examiner was compared to those of the calibrated dentists. To assess the consistency of each individual examiner and the variations between examiners, each examiner first practised the examination on a group of 10 subjects. Every examiner then independently was examined the same group of 20 subjects and compared the findings with the other examiner. The kappa value of intra-examiner reproducibility was 0.85, while for inter-examiner reproducibility was 0.87. Diagnostical criteria and codes of dentition status were recorded according to the criteria of WHO [29]. Accordingly, a tooth was coded as sound if it shows no evidence of treated or untreated clinical caries. A tooth was diagnosed as decayed (code:1), if there was a lesion in a pit or fissure, or if on a smooth tooth surface had an unmistakable cavity, undermined enamel, or a detectably softened floor or wall. A tooth was recorded as filled with decay (code:2) when the crown had one or more restorations and one or more areas that were decayed. A crown was filled with no decay (code:3) when one or more restorations were present without caries. A missing tooth as a result of decay (code:4) or any other reasons (code:5) was used for teeth that have been extracted. The prevalence of caries was assessed based on Decayed, Missing and Filled Teeth Index (DMFT), which could be derived directly from the codes before. The D component (D-T) includes all teeth with codes 1 or 2. The M component (M-T) contains teeth coded 4 and 5. The F component (F-T) comprises teeth only with code 3. The wisdom teeth were included in the calculation, so the highest score was 32.
The type of prosthetic appliances (fixed prosthetic appliances, removable prosthetic appliances) and the number of prosthetically replaced teeth were also recorded. The number of prosthetically replaced teeth per person was evaluated using the Prosthetic Index, which was expressed in a percentage [30]. The level of dental care (restorative index) was expressed as [(F/D + F)*100] at tooth level, as in a previous study [18, 19]. The restorative index described the proportion of the decayed and filled teeth that had been treated restoratively.
Data collection
All data was recorded on tablets based on an Android operating system. The data collection and analysis software were created in a collaboration with scientists from the Wigner Research Centre for Physics, located in Budapest. All the collected information was uploaded directly to the Hungarian Academy of Science’s cloud-based data storage. For statistical analysis, data could be retrieved from the cloud and converted to Microsoft Office Excel, version 2019.
Statistical analysis
The analysis of data was performed using the statistical software R, version 4.0.2 (R Core Team, 2020, Vienna, Austria) using the coin (Hothorn, Hornik, van de Wiel and Zeileis, 2008), tidyverse (Wickham et al., 2019), knitr (Xie, 2021), RcmdrMisc (Fox, 2020), summarytools (Comtois, 2021), ggpubr (Kassambara, 2020) packages. For the descriptive analysis of categorical variables, case number and percentage were computed, while in the case of continuous variables, patient number, mean, Standard Deviation (SD), median, 25% and 75% quartiles (IQR) were calculated. The socio-demographic factors, stroke-related factors, functional assessment and oral health-related behaviour were examined for their association with the number of decayed, missing and filled teeth. The correlation between FIM and frequency of tooth cleaning was also investigated. To identify the association between dental status and various factors, we used the Chi-square test and Fisher’s exact test in case of categorical data, the Welch test for normally distributed variables, and the Mann–Whitney U test and Kruskal–Wallis test for non-normally distributed variables. Analysis of Variance technique by Blocks (ANOVA) was used to detect the statistical differences between three or more independent groups. Correlation analysis was used to measure linear relationships between FIM, Barthel index and D-T, M-T, F-T, DMFT score. The level of significance was set at p < 0.05. The total sample size for the study was 410, however, data was missing in 8 cases in FIM scores and 13 cases in Barthel Index scores. Patients with these missing data (8 patients) were excluded from the statistical calculations where association between functional status and dental status or frequency of tooth cleaning was calculated, however, they were included in all other calculations because patients weren’t treated as individual data but as groups.