Impact of Medical and Surgical Obesity Treatment on Dental Caries: A 2-Year Prospective Cohort Study

The objective was to study the long-term effects on oral health of bariatric surgery compared with medical treatment of obesity. Swedish females with morbid obesity (n = 66; 18–35 years at baseline) were followed prospectively from before obesity treatment until 2 years after treatment. The main response variable was dental caries registered according to the ICDAS-II system. Possible confounding factors, such as sociodemographic characteristics, general health, oral health habits, and oral hygiene, were controlled for. The statistical methods included χ2 tests, Student’s t tests, one-way ANOVA, Wilcoxon’s nonparametric tests, and linear regression models. In the surgically treated patients (n = 40), a significant increase over time in enamel caries (mean increase 4.13 tooth surfaces ICDAS1–2), dentine caries (mean increase 2.18 tooth surfaces ICDAS3–6), and total caries (mean increase 6.30 tooth surfaces ICDAS1–6) was registered (all p < 0.001), which was not seen in the medically treated patients (n = 26). However, the difference between the treatment groups (surgical or medical) was only statistically significant for enamel caries (crude β 4.89, p = 0.003) and total caries (crude β 6.53, p < 0.001). The relationships were stable and independent of differences in confounders as socioeconomy, general health, and oral health behaviors. In conclusion, 2 years after obesity treatment, a significant increase in dental caries was registered in the surgically treated but not in the medically treated women. The dental service should intensify its preventive efforts in individuals undergoing obesity treatment.

ically treated patients (n = 26).However, the difference between the treatment groups (surgical or medical) was only statistically significant for enamel caries (crude β 4.89, p = 0.003) and total caries (crude β 6.53, p < 0.001).The relationships were stable and independent of differences in confounders as socioeconomy, general health, and oral health behaviors.In conclusion, 2 years after obesity treatment, a significant increase in dental caries was registered in the surgically treated but not in the medically treated women.The dental service should intensify its preventive efforts in individuals undergoing obesity treatment.

Introduction
Obesity is a major public health problem globally with consequences for both individuals and society [WHO, 2021].It is a risk factor for a number of chronic medical diseases, such as cardiovascular disease and type 2 diabetes [Dai et al., 2020], as well as for increased mortality [Sjöström et al., 2007;Carlsson et al., 2020].An impact on oral health with associations with periodontal disease and tooth loss has also been demonstrated [Keller et al., 2015;Nascimento et al., 2016], but divergent results have been presented for dental caries [Kantovitz et al., 2006;Shivakumar et al., 2018].
The treatment methods for obesity include behavioral therapy, dietary treatment and increased physical activity; i.e., methods aiming at lifestyle changes, as well as pharmaceutical treatment and surgery [Bray et al., 2016].Both medical and surgical treatments reduce morbidity and mortality, with bariatric surgery being the most effective treatment for morbid obesity (body mass index, BMI ≥40 kg/m 3 ) [Ma et al., 2017;O'Brien et al., 2019].Today, Roux-en-Y Gastric Bypass (RYGB) and Sleeve Gastrectomy (SG) are the predominant techniques for bariatric surgery [Welbourn et al., 2019;SOReg, 2022].In RYGB, a small pocket is created from the upper part of the original stomach which is then connected directly to the small intestine, bypassing the rest of the stomach.In SG, approximately two-thirds of the stomach is removed, leaving a narrow gastric tube that is, a "sleeve."Surgical treatment of obesity has increased over the last decades [WHO, 2021] and women make up the majority of patients seeking obesity treatment, accounting for about 80% of patients [Guerra et al., 2020;SOReg, 2022].
Although the benefits of bariatric surgery are significant with weight loss and reduction of obesity-related comorbidity, e.g., diabetes, there are risks of complications.Short-term complications are due to the surgical procedure itself.The long-term complications are due to the restructuring of the gastrointestinal tract, including malnutrition, reflux, and osteopenia [Shah et al., 2006].Here, some differences between RYGB and SG have been recognized.For instance, RYGB increases the risk of surgical complications such as bowel obstruction while SG increases the risk of, e.g., gastroesophageal reflux [Zhao and Jiao., 2019].Yet, there are still knowledge gaps regarding the long-term side effects [Höskuldsdottir et al., 2020].The impact of bariatric surgery on oral health has been sparsely investigated, with some reports showing associations with periodontal disease and hypersensitivity while others did not [Netto et al., 2012;Cardozo et al., 2014;Jaiswal et al., 2015;Sales-Peres et al., 2015].A systematic review found indications of a greater risk of caries in bariatric patients [Salgado-Peralvo et al., 2018] and clinical observations by both medical and dental staff have given risen to concern [Hague and Baechle, 2008;Taghat et al., 2021].It should be noted that the studies observing the oral health of bariatric patients over time have short follow-up times, usually no more than 6 months.To our knowledge, there are no published studies that have followed the oral health of this patient group over a longer period of time.The possible effects of medical obesity treatment on oral health are also unknown.
Against the background described above, the hypothesis was that the outcome in dental caries would differ between surgically and medically treated (MT) obese patients.Therefore, the objective of this study was to compare the effects of bariatric surgery with medical treatment of obesity on dental caries estimates from before until 2 years after treatment.

Study Design and Setting
A prospective cohort study was designed and the date of registering the trial was March 03, 2015; NCT03152617.The data collection was carried out in the Swedish Region of Västra Götaland.

Participants
The participants constituted a subgroup of women from the BAriatric surgery SUbstitution and Nutrition (BASUN) study, which included individuals (n = 971) who were referred for obesity treatment to the Regional Obesity Center [Höskuldsdottir et al., 2020].The criteria for inclusion were age >18 years and BMI ≥40 kg/m 2 or, if obesity-related comorbidities were present, BMI ≥35 kg/m 2 .Sufficient knowledge of the Swedish language to be able to understand information was required.Baseline data were collected from May 2015 to November 2017.The method of obesity treatment (i.e., medical or surgical treatment) was determined based on medical status and taking into account the patient's wishes.The study protocol includes follow-ups after 2, 5, and 10 years.All females 18-35 years old were further asked to participate in a substudy on oral health and bone health (the latter will be reported elsewhere).The follow-ups in this subgroup were planned to be performed before the onset of the menopause.Reports on baseline findings have been published earlier, in the main BASUN study [Höskuldsdottir et al., 2020] and in the odontological substudy [Taghat et al., 2022].The current publication reports from the 2-year follow-up in the substudy.
The Regional Ethical Review Board of Gothenburg (reg.no.673-14) and the Swedish Radiation Protection Committee (reg.no.14-39) reviewed and approved the protocol.All participants provided written informed consent.

Intervention
The intervention was obesity treatment, either medical or surgical.The medical treatment included a period on a very low energy diet (VLED) followed by food reintroduction during a 12week period, further dietary advice and general counseling.The patient received written instructions about a diet where the food intake is distributed in energy percentage (15-20 E% protein, 30 E % fat, 50-55 E% carbohydrates), and with three to five intakes per day [Parrett et al., 2016;Höskuldsdottir et al., 2020].It was completely advised against drinking sugary-sweetened beverages and strictly limits the intake of ordinary sugar.The surgical method was either Roux-en-Y gastric bypass, RYGB or sleeve gastrectomy, SG [Lönroth et al., 1996;Gagner, 2016;Höskuldsdottir et al., 2020].Both surgically and MT patients were prescribed dietary supplementation of vitamins and minerals according to Nordic Nutrition Recommendation [Laurenius et al., 2018] and received scheduled follow-ups over a 12-month period.

Data Sources and Variables
The data sources were records from clinical dental examinations and specified information collected in the main study BA-SUN.The procedure for the dental examination followed exactly the same protocol, including both clinical and radiographic registrations, at baseline and follow-up and has previously been described in detail [Taghat et al., 2022].
The main predictor variable was the treatment method for obesity.For the analyses, both three groups (very low energy diet, RYGB, SG) and two groups (medical or surgical) were used.
The response variable was dental caries registered according to the International Caries Detection and Assessment System, IC-DAS-II [Ismail et al., 2007], measured at two time points: 3-4 weeks before the start of treatment (corresponding to baseline) and 2 years after treatment.Two dentists (N.T., main examiner; A.L.Ö.) performed all the clinical registrations after an extensive, both clinical and radiological, calibration program, prior to the baseline examinations, with good intra-examiner and interexaminer reliability [Taghat et al., 2022].Prior to the 2-year examination, the two examiners repeated the introduction program for ICDAS-II recording (www.iccms-web.com) and examined the two same patients separately (128 tooth surfaces per patient).The clinical findings were then compared and discussed chairside.Great agreement was noted and consensus was reached.The records were categorized as enamel caries (ICDAS codes 1-2), dentine caries (ICDAS codes 3-6), and total caries (ICDAS codes 1-6) at the tooth surface level.
Complementary variables were used for the description and control of confounding.The choice of these variables was based partly on the fact that some are known to influence similar associations and partly on previous own investigations [Taghat et al., 2022].From the main BASUN study, data were retrieved about BMI and pharmaceutical treatment (glucose-lowering treatment, blood pressure treatment, lipid-lowering treatment, treatment for anxiety/depression, treatment with antipsychotics, pain medication, hypothyroidism treatment, treatment for attention-deficit hyperactivity disorder).Oral hygiene was represented by dental plaque (0/1) registered according to Ramfjord, 1967 and bleeding on probing, BoP, at two sites per tooth (0/1).Self-report forms were filled out by the participants immediately before the dental examination and provided information about age (in whole years), education (postsecondary schooling >12 years vs. maximum secondary school ≤12 years), and marital status (cohabiting vs. not cohabiting).Furthermore, the form contained items on smoking (never smoked/stopped smoking ≥1 year ago vs. smoke occasionally/smoke daily), toothbrushing (twice daily vs. once a day/a few times a week/seldom or never), interdental cleaning (daily/3-5 times a week vs. 1-2 times a week/never), visits to the dentist the last 5 years (once a year/3-4 appointments vs. 1-2 appointments/never), and reason for the most recent dental appointment (routine visit vs. pain/other problems).
Information about demography, socioeconomics, and habits was collected both at baseline and at follow-up.Baseline data were used for the control of possible confounding factors, as these circumstances were assumed to have prevailed during most of the 2-year follow-up period.

Data Management and Statistical Methods
The control and management of data were carried out in Microsoft Excel ® (Microsoft Corporation, WA, USA).The SPSS ® statistical software (Statistical Package for Social Sciences), version 26 (IBM Corporation, NY, USA), was used for the statistical analyses.
Categorized and dichotomized variables as described above were used.The descriptive statistics comprised numbers, percentages, means, standard deviations (SDs) and median values, when indicated.Distribution of data for response and main outcome variables was checked graphically; all graphs were bellshaped and symmetrical with a tendency toward positive skewness for ICDAS 3-6 .χ 2 tests, Student's t test, and one-way ANOVA were applied to analyze differences between treatment groups versus categorical and continuous variables, respectively.Wilcoxon's signed-rank test was used to examine and compare the scores for dental caries, dental fillings, and BMI from baseline to follow-up.Graphic presentations describe mean values for dental caries in the different treatment groups from baseline to follow-up.To explore the impact of possible confounders, the difference in caries experience for each ICDAS category (dependent variable, main outcome) by obesity treatment groups (independent variable) between baseline and follow-up was analyzed using linear regression models.Crude analyses were followed by limited and extended models.The limited models included baseline information about age, socioeconomic status (cohabiting, educational level), and general health variables (medication, smoking, BMI) together with baseline information about oral health habits (toothbrushing, interdental cleaning, regular dental visits, reason for last dental visit) and oral hygiene (plaque, bleeding on probing).The extended full models included all variables from the limited models, together with baseline information about number of filled tooth surfaces and baseline total caries (ICDAS 1-6 ).The level of statistical significance was set to 0.05.

Results
The participation rate at follow-up was 58% (66 of 113 individuals): 57% of the MT and 60% of the surgically treated patients (RYGB 57%, SG 67%).The results are presented based on the two main treatment groups, i.e., medical and surgical treatment.Results based on the two different surgical methods (RYGB and SG) are given in supplementary material (online suppl.Table 1-3; for all online suppl.material, see https://doi.org/10.1159/000533609).A flowchart is shown in Figure 1.The mean follow-up time was 27.5 months, and the mode value in delay from the planned 2 years was 2 months.
Description of the cohort by type of treatment at baseline and at follow-up is shown in Table 1.At baseline, only the level of education differed between the treatment groups (medical, surgical; p < 0.001), with a greater proportion with low education among the surgically treated.This difference remained at follow-up.Cohabitation status differed at follow-up between the medically (53.8%) and the surgically (25%) treated (p = 0.017) patients, as did smoking habits (p = 0.011); however, few were smokers.
There was also a significant difference in oral hygiene between the groups at follow-up with a smaller proportion of tooth surfaces with plaque in the MT than in the surgically treated group (p = 0.024).There were no other differences between the treatment groups regarding basic characteristics.Within the treatment groups (p values not in table), both presence of plaque (MT p = 0.016, surgically treated p < 0.001) and BoP (MT p = 0.002, surgically treated p < 0.001) were more common at follow-up than at baseline.Smoking was more common at follow-up (p = 0.034) in the surgically treated group; otherwise, no statistically significant differences from baseline to follow-up were noted within the two treatment groups, respectively.

Outcome: Main Results
At baseline, there were no statistically significant differences in dental caries status between the treatment groups (all p values >0.05).A significant increase in dental caries was seen from baseline to follow-up in the surgical treatment group, in enamel as well as in dentine caries, and in total caries, all p < 0.001 (Table 2).In the medical treatment group, there were no significant changes in any measure of dental caries.The number of fillings had also increased significantly in the surgically treated patients during the study period (p = 0.013), while there was no such difference in the MT group (p = 0.421).Mean BMI had decreased significantly at follow-up among the surgically treated patients (p < 0.001), but not among those receiving medical treatment (p = 0.170).
In Table 3, the differences in decayed tooth surfaces in the treatment groups from baseline to follow-up are further clarified and compared.The mean increase in enamel caries (ICDAS 1-2 ) among the surgically treated was 4.13 (SD 6.7); however, a mean decrease was observed (-0.77,SD 5.5) in the MT patients, and the difference between the treatment groups was statistically significant (p = 0.002).The pattern was similar regarding total caries (ICDAS 1-6 ).For dentine caries (ICDAS 3-6 ), the difference between the treatment groups was not statistically significant: MT patients had a mean increase of 0.54 tooth surfaces and surgically treated patients a mean increase of 2.18 tooth surfaces (p = 0.067).A graphic presentation is given in Figures 2-4 for medical treatment and respective surgical method (i.e., RYGB and SG).
The impact of possible confounders on the relationship between the obesity treatment predictor and the outcome/responsethe difference in dental caries from baseline to follow-upis presented in Table 4.The crude The p value refers to tests between medical and surgical treatment groups for each point in time.*One or more of the following: glucose-lowering treatment, blood pressure treatment, lipid-lowering treatment, treatment for anxiety/depression, treatment with antipsychotics, pain medication, hypothyroidism treatment, ADHD (attention deficit hyperactivity disorder) treatment.

Dental Caries after Obesity Treatment
Caries Res 2023;57:231-242 DOI: 10.1159/000533609 disparity in the difference (difference-of-difference, β) from baseline to follow-up between surgical and medical treatment regarding decayed tooth surfaces was statistically significant for enamel caries (β 4.89, p = 0.003) and total caries (β 6.53, p < 0.001), but not for dentine caries (β 1.64, p = 0.069).All other models generated statistically significant β coefficients.The p values for the adjusted models exhibited a more mixed pattern.The full models were statistically significant for dentine caries (p = 0.020) but not for enamel and total caries.
Of the other variables that is, possible confounders included in the models, a few were statistically significant per se (not in tables).Educational level and reason for last dental visit were statistically significant for differences in dentine caries (full extended model p = 0.008 and p = 0.011, respectively).No other variables included in the models were statistically significant.

Discussion
The key results in this study of young women who were followed from before obesity treatment until 2 years after treatment was a significant increase in dental caries in the surgically treated but not in the MT patients.The relationships were stable and independent of differences in socioeconomic status, general health, oral health habits, oral hygiene, and the caries situation at baseline.

Strengths and Weaknesses of the Study
A strength of the study was the considerably longer follow-up period than in previous similar studies [Marsicano et al., 2012;de Moura-Grec et al., 2014;Marzequin et al., 2022], which strengthens the significance of the findings.In general, most studies regarding bariatric surgery are retrospective, while BASUN has a prospective design.Also, to our knowledge, there are hitherto no other studies comparing the outcome regarding oral health between surgical treatment for obesity versus structured medical treatment.There is a sparsity of such studies also with regard to medical outcomes and complications [Höskuldsdottir et al., 2020].

Dental Caries after Obesity Treatment
Caries Res 2023;57:231-242 DOI: 10.1159/000533609 higher limit in similar studies [WHO, 2021].Another strength is that the dental examinations followed a strict protocol, including radiographs, and were performed by calibrated dentists in a clinical setting.The same standardized conditions were at hand at baseline and at follow-up [Taghat et al., 2022].The detailed scoring system in ICDAS-II [Ismail et al., 2007] enabled registration of different degrees of dental caries, which proved to be crucial in detecting differences between the outcome in enamel caries and dentin caries, respectively, in the study.Earlier studies generally used the DMFT system with only dichotomous outcomes; manifest caries or no caries [Marsicano et al., 2012;de Moura-Grec et al., 2014;Marzequin et al., 2022].
The study has limitations.A randomized design would have been ideal; however, the medical status and wishes of the patients differed and were taken in account, mainly for therapeutic but also for ethical reasons.Nevertheless, the fact that BASUN is designed as "usual care"; that is, how it works in reality, can be considered a strength.It could be seen as a limitation that the sample only included women; however, they make up the great majority of those seeking treatment for obesity [SOReg, 2022], and the follow-up was planned to be completed before menopausal onset for the study of bone health [Lindsay, 2004].Also, loss to follow-up is a common problem in longitudinal studies and in the current study, four out of ten did not participate in the examinations after 2 years.Fortunately, the distribution of individuals in the groups corresponded to the baseline figures [Taghat et al., 2022].The limited delay of dental examinations at follow-up made the observation period somewhat longer, which could, in fact, be considered an advantage, and the delay was evenly distributed in the group.However, it was difficult to check for other dental visits during the intervening period; this information was self-reported, and no journals from the regular dental service were checked.Other aspects that were not investigated and that should be explored in future studies are for instance dietary habits and information about fluoride supplements.

Interpretation
The significant increase in caries in the surgically treated patients in the study was in stark contrast to the outcome in the MT group, where a reduction in caries lesions was seen.Regardless of medical or surgical treatment, existing caries lesions could both improve and deteriorate.Some enamel lesions may have proceeded into the dentine, thereby reducing the number of surfaces with enamel caries, but the caries activity may also have slowed down so that the clinical and/or radiological picture had changed for the better [Nyvad and Fejerskov, 1997].At the same time, of course, new lesions could occur.The poor oral hygiene seen at baseline [Taghat et al., 2022] even deteriorated over time and this confirms the continued need for preventive measures regardless of treatment group.One would perhaps expect that individuals who achieved their weight loss goal would be more motivated in terms of self-care and possibly oral hygiene measures.In this study, the opposite was observed.Those who improved their body condition by significantly losing weight were those who had more caries and more plaque accumulation.The individuals might have other problems caused by the restructuring of the gastrointestinal tract [Shah et al., 2006], for instance complications as reflux and that oral hygiene might then be neglected.
Furthermore, physiological and nutritional shifts that a bariatric procedure entail can influence our microbiome and immune system increasing the likelihood of disruption to the natural disease-free homeostasis [Zaura and Cate., 2015].Nutritional deficiencies are for instance common after bariatric procedures [Nuzzo et al., 2021].Meanwhile, recent studies on saliva and fecal excrements have pointed to considerable alterations of both gut-and oral microbiota following bariatric surgery, however, to what extent these changes impact oral health has yet to be elucidated [Adawi et al., 2022;Stefura et al., 2022].There is an urgent need in the next step to investigate the mechanisms behind the differences in caries found between medical and surgical treatment, for instance, diet and saliva properties [Flink et al., 2008;Takahashi and Nyvad, 2011;Mika et al., 2022].
The BMI of the participants in the current study was reduced after 2 years only in the surgically treated group.In the main study BASUN, a decrease in BMI was seen both in the medically and the surgically treated patients, even if the weight reduction was greater in the surgically treated group [Höskuldsdottir et al., 2022].Both medical and surgical obesity treatment have been shown to be effective regarding weight loss [Ma et al., 2017;Lean et al., 2018], although most studies do not compare those treatments but have "care as usual" as a control group.It might be hypothesized that the intestinal differences in anatomy after RYGB and SG could have an impact, for example, on nutritional and hormonal status with subsequent consequences [Sumithran et al., 2011;Höskuldsdottir et al., 2022].At the start of BASUN, RYGB was the predominant obesity surgery method in Sweden while SG surgeries increased during the inclusion period.Today, the two methods are used equally often in Sweden [SOReg, 2022], and both methods reduce weight and comorbidities [Höskuldsdottir et al., 2022;Thaher et al., 2022].The possible significance of the surgery method for oral health needs to be further investigated.
The association between the obesity treatment methodmedical or surgicaland caries outcome revealed stable relationships.However, a few potential confounders were statistically significant per se in the regression models.About seven out of ten in the surgery group had an education ≤12 years, compared with only approximately three out of ten in the medical treatment group, a significant difference that was constant over time.Lower education may make it more difficult to evaluate and use information [Sherlock and Brownie, 2014] and surgery may be seen as an easier and faster way to weight loss.In this study, however, it did not affect the caries outcome.Previous caries experience has earlier been shown to be an indicator to assess the caries risk [Mejare et al., 2014;Schwendicke, 2013], and although the associations were attenuated in our study, it was not a significant factor.
Another salient finding, although not significant in the regression models, was the abundant use of medicines, which is known to have a negative effect, for example, on saliva flow [Närhi et al., 1992].The usage was high at baseline [Taghat et al., 2022] and had further increased at follow-up in both the surgically and MT patients.This should be investigated further but falls outside the scope of the current study.Other factors to examine are for instance: other saliva properties, nutritional status, and dietary habits, which could be hypothesized to have an impact [Lingström and Moynihan, 2003;Nauntofte et al., 2003;Shah et al., 2006].
The external validity; i.e., the generalization of the findings, can probably be considered equal to at least similar Western contexts, even if the circumstances in different populations may differ to some extent.Further research should focus on explanatory factors and mechanisms behind the oral health effect of bariatric treatment demonstrated in the study.
Finally, it can be concluded that at follow-up 2 years after obesity treatment, a significant increase in dental caries was registered in the surgically treated but not in the MT women.The dental service should intensify its preventive efforts in individuals undergoing obesity treatment.

Table 1 .
Sociodemographic characteristics, oral health habits, and hygiene of the cohort in total and by type of treatment at baseline and follow-up after 2 years

Table 2 .
Clinical characteristics by treatment group (medical or surgical) at baseline and at 2-year follow-up Wilcoxon tests for differences between baseline and follow-up in each treatment group.ICDAS, International Caries Detection and Assessment System.

Table 4 .
Difference-of-difference in decayed tooth surfaces between surgical and medical treatment from baseline to follow-up using linear regression models Body Mass Index), oral health habits (toothbrushing, interdental cleaning, regular dental visits, reason for last dental visit), oral hygiene (plaque, bleeding on probing), dental fillings, and baseline scores of ICDAS 1-6 .ICDAS, International Caries Detection and Assessment System.