Frequency of periodontal pathogens and Helicobacter pylori in the mouths and stomachs of obese individuals submitted to bariatric surgery: a cross-sectional study

ABSTRACT Objectives This cross-sectional study compared the frequency of oral periodontopathogens and H. pylori in the mouths and stomachs of obese individuals with or without periodontitis submitted to bariatric surgery. Material and Methods One hundred and fifty-four men and women aged 18-65 were conveniently distributed into four groups. Two groups were composed of individuals who underwent bariatric surgery with (BP) (n=40) and without (BNP) (n=39) periodontitis and two obese control groups with (CP) (n=35) and without (CNP) (n=40) periodontitis. The oral pathogens Porphyromonas gingivalis, Aggregatibacter actinomycetemcomitans, Parvimonas micra, Treponema denticola, Tannerella forsythia, Campylobacter rectus, and Helicobacter pylori were detected by a polymerase chain reaction technique using saliva, tongue and stomach biopsy samples. Results Statistical analysis demonstrated that periodontopathogens were highly frequent in the mouth (up to 91.4%). In the bariatric surgically treated group, orally, P. gingivalis, T. denticola and T. forsythia were more frequent in periodontitis, while C. rectus was more frequent in non-periodontitis subjects. Stomach biopsies also revealed the high frequency of five oral species in both candidates for bariatric surgery (91.6%) and the bariatric (83.3%) groups. H. pylori was frequently detected in the mouth (50.0%) and stomach (83.3%). In the stomach, oral species and H. pylori appeared in lower frequency in the bariatric group. Conclusions Obese individuals showed high frequencies of periodontopathogens and H. pylori in their mouths and stomachs. Bariatric surgery showed an inverse microbial effect on oral and stomach environments by revealing higher oral and lower stomach bacterial frequencies.


INTRODUCTION
High prevalence of obesity and its complications have become a global health concern. Besides different systemic comorbidities 16 , obesity has been risk for the development and severity of periodontal disease 7,8,20,22,23 among the most prevalent oral diseases in different populations. However, a recent systematic review pointed out that evidence on this subject is still limited 21 . Periodontitis triggers local and systemic variations in pathophysiology state, thereby increasing susceptibility to metabolic syndromes. In addition, given that the adipose tissue, particularly the white adipose tissue, acts as a main endocrine organ secreting a number of bioactive substances (such as cytokines), it can also affect the periodontal response and it can also be affected by periodontal infections 20 .
Periodontal disease is a multifactorial infectious disease associated with a microbiota predominantly composed of Gram-negative species. The link between obesity and periodontal disease is based agents could be different, both show high levels of 18,20 . Despite its relevance, the number of surveys that have attempted to analyze the relationship between obesity and periodontal microbiota is still small. However, a few preliminary studies have found different oral obese groups 10,11 , therefore indicating the need for further studies in this area. Although Belstrøm, et al. 3

(2014) failed to establish an association between
Human Microbiome project has shown that shifts in our microbiota are associated with many diseases such as obesity 14 . Recently, a cross-sectional study performed in Japan revealed an association between Porphyromonas gingivalis, Treponema denticola, and Tannellera forsythia and obesity 17 .
Although microbial studies focusing on obesity and periodontal status are limited, when obesity and bacterial frequency are mentioned, Helicobacter pylori undoubtedly plays a key role. Oral sites could represent a reservoir for H. pylori 9,28 and, in theory, the mouth-stomach route used by this pathogen can also be used by periodontal pathogens to access other parts of the body. Bariatric surgery is a commonly recommended treatment option for severely obese subjects 19 . After surgery, several personal changes can be observed; however, the remains controversial 24,30 . Although there have been reports on the possible association between on stomach frequency of periodontal pathogens had not been gathered yet. Moreover, there are no reports in the literature describing whether periodontal pathogens can translocate from the mouth to colonize the gastric mucosa. Therefore, the present study was performed with the aim of evaluating the oral, and especially stomach, presence of certain target periodontal pathogens and H. pylori in bariatric surgically treated obese individuals.

Participants
(121 females and 33 males, 37.58±11.36 years of age) recruited from two health centers for the treatment of obesity (Trauma One Clinic located in Belo Horizonte, Minas Gerais, Brazil and Dental Medic Clinic located in Lorena, São Paulo, Brazil) composed the population of this study. All of them underwent a complete periodontal clinical examination between December 2009 and December 2013. Inclusion criteria were adults between 18 and 65 years of age, both genders, who had been bariatric surgically treated by the gastric bypass Roux-en-Y -Fobi-Capella technique, and obesity (BMI>30 and <40) for the control groups. In addition, the diagnosis of periodontitis (described a posteriori) was a requirement for two groups while a non-periodontitis diagnosis was an inclusion criterion for the other two groups. The exclusion criteria were: those aged under 18 or over 65; underweight Body Mass Index (BMI<19 for surgically treated patients) or morbid obesity (BMI>40); possession of fewer than 15 natural teeth; pregnancy; antibiotic intake or regular use of chlorhexidine within the three months previous to the examination.
Participants were carefully informed about the objectives of the study and after their approval they signed a consent form. This study was approved by the Federal University of Minas Gerais Research Committee (ETIC 57807) and University of Taubaté Ethics Research Committee (protocol 52210).
The study population was conveniently the bariatric surgically treated groups composed of patients that undergone the surgery at least 24 months later (39.37±15.80 months after surgery). This group was subdivided according to the presence of periodontitis (bariatric with periodontitis -BP) or absence of periodontitis (bariatric with no periodontitis -BNP) and was composed of 40 and 39 individuals, respectively. The second included two obese groups: one composed of obese patients showing signs of periodontitis (control with periodontitis -CP) and the other composed of obese patients without signs of periodontitis (control with no periodontitis -CNP). These groups were composed of 35 and 40 individuals, respectively. The number of 32 individuals per group, as the minimum needed, was determined based on results from a pilot study, which included sampling and laboratorial processing of 10 obese individuals. Statistics revealed that this number would be enough to identify a minimum safety margin was adopted. Data regarding demographic information and oral hygiene habits were collected from each participant while their medical histories were obtained from their medical records. Anthropometric measurements, including weight (Kg) and height (m), were measured while the subjects were wearing light clothing and no shoes by a center trained and calibrated professional nutritionist. These measurements were used to calculate BMI (Kg/m 2 ) and classify obesity 29 . Twenty individuals had their height and weight re-measured to determine BMI reproducibility values. The Kappa test showed values greater than 0.90, thus indicating good reproducibility.

Periodontal clinical examination
For each participant, a full-mouth periodontal examination was performed in a hospital gurney with a photophore. Two periodontists (A.L.P. and S.C.C), trained and calibrated at the beginning of the study, measured probing depth (PD) and clinical attachment level (CAL). After seven days, the periodontal examinations of 10 participants were repeated, showing intra-and inter-examiner reproducibility scores that were higher than 0.85 (Kappa test) for PD and CAL clinical parameters. Intra-class correlation tests showed scores higher than 0.90. The following periodontal parameters were registered with a manual periodontal probe (North Carolina University model -UNC -#15, Hu-Friedy; Chicago, IL, USA) at six sites per tooth: bleeding on probing (BOP), PD and CAL. 15 . Oral hygiene was assessed using plaque index (PI) 26 .

Microbiological assessment
Samples of saliva 6 and from the dorsum of the tongue 5 were collected from the total population (n=154). Saliva samples were collected in the morning, between 8:00 and 11:00. The patients were instructed not to eat or drink prior to sampling. Immediately before sampling, individuals rinsed their mouths with water. During collection, they remained seated with their heads tilted forward (approximately 45°) and 2.0 mL of unstimulated, whole saliva were collected into sterile Falcon tubes. Samples were centrifuged for 10 minutes at 15,000×g at 4°C, and the supernatants were immediately stored at -80°C. Scrapings from the tongue dorsum were taken from areas of approximately 1 cm 2 using a cotton swab dipped in reduced Ringer's solution, rotated six times. Each swab was transferred into a microtube also containing reduced Ringer's solution (1 mL).
According to systemic medical needs, additional stomach biopsies were collected from a representative subgroup of 49 subjects requiring endoscopy. Following the protocols of the medical centers, a gastroenterologist physician collected the stomach biopsies using an endoscope. Tissue samples were inserted into Eppendorf microtubes containing 1 mL of reduced Ringer's solution and were immediately placed in a polystyrene box with ice and then stored at low temperatures (-20°C) until processing.
Microbial processing procedures were performed as previously described 5 . Each collected sample was immediately conditioned in a Styrofoam box with ice and then stored at low temperatures (-80°C) until processing. The genomic DNA of each sample was extracted using PureLink™ CA, USA) according to the manufacturer's a polymerase chain reaction (PCR) was carried out using unspecific "Universal primers" (16S rRNA) (5'-GATTAGATACCCTGGTAGTCCAC-3' and 5'-CCCGGGAACGTATTCACCG-3') to detect bacterial DNA in the samples. After this procedure, the presence of P. gingivalis (5'-AGGCAGCTTGCCATACTGCG-3' a n d 5

Statistical analysis
The frequency of each bacterium was separately evaluated in the samples taken from the saliva and dorsum of the tongue. The occurrence of each bacterium in a given subject in at least one of these sites was also checked. This last evaluation was referred to as oral representation. An additional analysis considering the red complex oral bacterial species was also performed, thereby indicating the simultaneous presence of P. gingivalis, T. denticola and T. forsythia.

The comparison of the frequencies of bacteria
Chi-square and Fisher's test. To determine the significance between groups, the Bonferroni correction was applied. The characteristics of the participants' variables were described using frequency distribution for categorical variables and median, mean and standard deviation for continuous variables. Normality data distribution was verified by the Kolmogorov-Smirnov test. The ANOVA test was used for variables of normal distribution and the Kruskal-Wallis test was used for those of non-normal distribution, followed by the Tukey's test and the Mann-Whitney test for comparisons between groups. Tests of accuracy tested the relationships between the stomach and oral frequencies of each target bacterial species.
All statistical tests were performed using SSPS (Statistical Package for Social Sciences, version 16 for Windows) software (SPSS Inc., Chicago, IL,

RESULTS
As expected, BMI was greater among the controls when compared with the bariatric groups. In addition, within the periodontal groups (with or without periodontitis), BMI values were similar ( Table 1).
analysis of oral samples. In periodontitis subjects, P. gingivalis and T. denticola higher in the saliva samples taken from the bariatric group compared with the control group, suggesting that bariatric surgery was accompanied by higher salivary frequency of these two species in this showed higher frequency of a third periodontopathic species, T. forsythia, but in the tongue dorsum instead (Table 2).
Interestingly, even in the non-periodontitis groups, the BNP group exhibited a higher frequency of C. rectus than that which usually precedes the colonization of the mouth by the red complex species. This bacterium was again observed in higher frequency in the saliva and tongue samples taken from the bariatric group. Individuals without reduced frequency when bariatric and control groups were compared: P. micra appeared in lower frequency in the saliva taken from bariatric subjects.
Furthermore, the analysis of the oral representation brought a new perspective to the   difference.
Along with the oral analysis, a second set of results that was derived from the stomach biopsies processed in the laboratory is further presented. frequency of well-known periodontal pathogens in both the bariatric (up to 83.3%) and control (up to 91.7%) groups. In the stomach, despite the health/disease status of the periodontium, nearly frequencies in the bariatric group. Thus, in the long term (at least 24 months after surgery), bariatric surgery was accompanied by a reduced frequency of periodontal pathogens, as well as H. pylori in the stomach. Hence, bariatric surgery showed an Rather unexpectedly, the accuracy analysis (Table 3) demonstrated that the frequency observed in the stomach was similar to that observed in the mouth. However, the presence of H. pylori was higher in the stomach than in the mouth, as it was expected.
H. pylori on the detection of the species periodontally sought. Some associations between H. pylori and oral species were found; for example, P. gingivalis, T. denticola, T. forsythia, and C. rectus were involved in these associations. Interestingly, these oral results observed in the previously conducted analysis. Among the subjects colonized by H. pylori, differences of combined H. pylori -P. gingivalis (Hp-Pg) frequency in saliva were found among the groups (p=0.033). This association showed a tendency for increased bacterial occurrence in bariatric subjects with periodontitis (BP group; p=0.021). In periodontitis subjects, the association between H. pylori and T. forsythia (Hp-Tf) was higher in the tongue samples, especially in the bariatric group (BP group; p=0.006). The association between H. pylori and C. rectus (Hp-Cr) was statistically higher in the bariatric groups with periodontitis (BP) than in the non-periodontitis bariatric (BNP) groups (p=0.006). In the previous analysis, C. rectus did not show any with periodontitis. However, C. rectus appeared again as a key pathogen in the non-periodontitis individuals. The association between H. pylori and C. rectus (Hp-Cr) was statistically higher in the bariatric group (BNP; p<0.001). Considering the oral representation, the data previously showed was frequency of H. pylori and periodontal pathogens in the stomach were not observed (data not shown).  When evaluating the influence of age and tobacco use in the groups, there were no statistically significant differences, either in relation to bariatric surgery or to the presence or absence of periodontitis.

DISCUSSION
Evidence suggests a two-way relationship between periodontal disease and obesity 20,22 . In addition, a previous study that analyzed 345 subjects undergoing bariatric surgery showed a high prevalence of periodontitis, ranging from 70.69% to 91.66% of individuals before and after bariatric surgery 23   T. forsythia in obese individuals with gingivitis. Moreover, Goodson, et al. 10 (2009) found changes in the salivary bacterial composition in overweight women while Matsushita, et al. 17 (2015) reported that the red complex bacterial species, i.e., P. gingivalis, T. denticola, and T. forsythia, are associated with obesity. In our study, the presence of T. forsythia (up to 72.2%), P. gingivalis (up to 30.0%), T. denticola (up to 35.0%), C. rectus (up to 91.4%), and P. micra (up to 82.9%) were bariatric and control groups. The only periodontal species that was always observed in low frequency (up to 5.0%) was A. actinomycetemcomitans. study is able to corroborate the aforementioned response, which, therefore, influences human microbiota. According to Nagpal, et al. 20 (2015), the penetration of periodontal pathogens or their products in lamina propria may lead to endotoxemia This state may further affect the expression and functioning of important immunoinflammatory molecules, thereby contributing to altered lipid and glucose metabolisms. In our study, the bariatric group showed an increased frequency of the red Haffajee and Socransky 11 (2009), suggest the need for longitudinal monitoring of obese patients, since these bacteria play a key role in the development and progression of periodontal disease. It is known that weight reduction may benefit overweight and obese people in particular, mainly due to responses 4 , but also because it can positively impact their periodontium 20,25 . Furthermore, the control of comorbidities, such as diabetes, could also lead to a reduction of periodontal pathogens. In the present study, most of the target species appeared in greater numbers in the bariatric group, and they were even more evident among individuals with periodontitis. However, it is important to emphasize that the intense and progressive changes that could delay the establishment of the oral microbial of bariatric surgery, oral disorders, such as the loss of periodontal tonus, bleeding, hypersensitivity and xerostomia, have been related to this form of treatment for obesity.
One surprising observation of this study was the high frequency of periodontal pathogens present in the biopsies of stomach tissue, with some bacterial frequency reaching over 90%. Even species such as A. actinomycetemcomitans (58.3%), which is sometimes only found in low numbers in the mouth, were commonly detected in the stomach. In addition, although many H. pylori infections are treated before surgery using systemic antibiotics, in our study this microorganism was frequently found in stomachs and in the mouths of the bariatric and control groups. Overall, longitudinal studies are required to further elucidate oral and stomach In addition, future longitudinal studies could investigate periodontitis in obese patients before and after bariatric surgery. Currently, long-term studies on the topic are scarce. In an unselected population based on periodontal status, Jaiswal, et al. 12 (2015) failed to report pocket depth and clinical attachment level improvements in a sixmonth monitoring period after bariatric surgery. Similarly, Sales-Peres, et al. 24 (2015) observed worsened measurements of pocket depth and clinical attachment levels six months after bariatric surgery, alongside an increased amount of P. gingivalis.
Although the primary site of H. pylori colonization is the stomach, the mouth also harbors this pathogen, even if temporarily, mainly in individuals with chronic gingivitis or periodontitis 2,9,28 . Thus, the oral cavity could represent an extragastric reservoir of H. pylori 9,28 . Interestingly, the professional gastric reinfection of H. pylori. In a study involving 110 individuals, only 19.6% of patients who H. pylori compared with 84.3% of patients without 13 . In our study, there was a high frequency of H. pylori in the subjects' stomachs (from 15.3 to 83.3%), saliva, and tongues (from 30.0 to 50.0%). In the control group, individuals who harbored this bacterium in the stomach were more likely to carry it in their mouths as well. These patients were probably not successfully treated for H. pylori infection and they had higher BMIs than individuals from the bariatric higher prevalence of H. pylori in obese patients than in non-obese patients 1 . Erim, et al. 8 (2008) also found that bariatric surgically treated patients were 1.7 times more likely (95% CI, 1.3 -2.2) to demonstrate H. pylori infection. In general terms, our study showed a simultaneous frequency of H. pylori and periodontal pathogens in both the periodontitis groups and in the bariatric groups.
Tests of accuracy between the presence of periodontal pathogens in the stomach and mouth primary site of H. pylori, which can help explain its statistically higher frequency in the stomach than in the mouth. Similarly, periodontal pathogens demonstrated higher frequency in their primary oral sites when compared with the stomach.
The relationship between microorganisms and obesity is not yet well understood. This complex and intriguing relationship reveals several possibilities in is strongly recommended. Finally, the possible mouth-stomach route highlighted in this study draws attention to other ways for periodontal pathogens to migrate from oral to systemic sites. It is important to consider that this route is not necessarily dependent on the diseased epithelium pockets. Once more, the maintenance of local health and control of oral microbiota appear to impact a person's systemic health. Considering the possible authors also suggest the use of a sample analysis using different microbial techniques, especially quantitative ones. However, under the correct conditions, a small bacterial fragment or even bacterial products could initiate and/or sustain an Some limitations of our study, such as the use of a convenience sample, should also be mentioned. The most obvious criticism about convenience sampling is that the sample is not representative of the entire population, which limits data generalization. In addition, because of their overall compromised systemic condition, the exclusion of morbidly obese participants could have impacted the results drawn from the present study, although we cannot presume to know what type of microbial interactions among oral bacteria and H. pylori would be expected in this group. Moreover, a greater number of stomach biopsies should be microbiologically analyzed by using quantitative techniques to determine any infective patterns. Some of the limitations observed in the study could be corrected in future research by randomly selecting obese patients to more accurately represent the entire population and by including other degrees of obesity and quantifying bacterial levels.
Our study suggests that the stomach, although a different environment from the oral cavity, can know whether oral bacteria in the stomach are able to migrate to other parts of the body and if periodontal pathogens can contribute to stomach disease in obese or bariatric surgically treated individuals. The results of this study showed that obese individuals had higher levels of periodontal pathogens and H. pylori in both their mouths and stomachs. It is clear that bariatric surgery has influenced bacterial frequency in these environments, but the changes that occur after the surgery seem to trigger distinctive effects in the mouth and in the stomach.

CONCLUSIONS
Bariatric surgery showed an inverse microbial effect on oral and stomach environments and was accompanied by higher oral and lower stomach bacterial frequencies.