Self-reported periodontitis and C-reactive protein in Parkinson’s disease: a cross-sectional study of two American cohorts

Periodontitis triggers systemic repercussions, such as elevated levels of high-sensitive C-reactive protein (hs-CRP). This has never been studied within Parkinson’s Disease (PD). The aim of this study is to compare hs-CRP levels of self-reported periodontitis cases versus cases without periodontitis in PD patients. Data from the National Health and Nutrition Examination Survey (2015–2016 and 2017–2018 waves) were analyzed. PD cases were identified through medication regimens and periodontitis cases through a validated self-report questionnaire. 51 participants were included (24 females, 27 males, with mean age of 62.96 (14.71)). While the self-reported periodontitis group presented elevated levels of circulating hs-CRP (5.36 vs. 1.99 mg/L, p = 0.031), the self-reported without periodontitis group presented higher lymphocyte levels (29.35 vs. 28.03%, p = 0.007). Blood levels of hs-CRP were significantly higher in PD cases with self-reported periodontitis. Apart from the lymphocyte levels, there were no other significant differences according to the self-reported periodontal status. Future studies shall explore this association using clinical measures.


INTRODUCTION
Parkinson's Disease (PD) is the fastest growing neurodegenerative movement disorder, affecting around 10 million people worldwide 1,2 . This chronic, progressive and degenerative condition of both the peripheral and central nervous systems 3,4 is clinically heterogeneous, with various motor and non-motor clinical features 5 . It has been hypothesized that the onset and progression of PD, unclear thus far, is dependent on the conjugation of different key factors such as neuroinflammation, alpha-synuclein induced neuronal dysfunction (through intracellular aggregation into Lewy bodies, which stand as the pathological hallmark of PD), systemic chronic inflammation (translated in the dysregulation of circulating inflammatory cytokines) and even gut and periodontal dysbiosis 6,7 .
In the advanced stages, beyond the debilitating and interfering impact of motor and non-motor symptoms (NMS) on everyday-life activities, PD also has a major detrimental effect on patients' overall quality of life 5,8 . Oral health is no exception and may be deteriorated in PD resulting from impaired oral hygiene and lack of oral care [9][10][11][12][13][14][15][16][17] .
Among the possible oral conditions that may arise from inadequate oral care is periodontitis, a chronic, infectious, and inflammatory condition characterized by the destruction of the periodontium 18 . The physiopathology of periodontitis involves dental plaque dysbiosis and an uncontrolled immune response attacking the periodontal tissues 19 . Even though a clinical periodontal diagnosis is a gold standard, the self-report of periodontitis is an interesting epidemiological strategy that has been successfully developed and validated [20][21][22] . As an example, a recent prospective cohort study analyzed self-reported periodontitis relationship with female fecundability 23 , showing the potential of this self-reported measure in epidemiological scenarios.
The mutual link between PD and periodontitis has been studied recently. On the one hand, fine motor impairments and cognitive decline in PD patients compromise oral hygiene habits and general oral health status [9][10][11][12][13][14][15][16]24 . On the other hand, evidence has surged on bacterial inflammagens-including major virulence factors of key periodontal pathogens such as Porphyromonas gingivalis, like lipopolysaccharide (LPS) and gingipains-fueling a systemic inflammatory state that might be involved on the development of PD 4,25 . Furthermore, periodontitis was associated with a leukocytosis state in PD patients 26 . Also, higher blood levels of amyloid beta were found in periodontitis, mediated by inflammatory markers such as IL-6 and high-sensitive C-reactive protein (CRP) 27 . In fact, CRP is a widely evaluated non-specific biomarker in the clinical context, not only in the diagnosis and monitoring of acute inflammatory and infectious events but also in the management and prediction of chronic inflammatory conditions, such as cardiovascular and neurodegenerative diseases 28,29 . There are also increased levels of pro-inflammatory cytokines in PD, including CRP 6 . However, hs-CRP levels have never been studied in PD cases according to their periodontal status, and this may provide useful information in the PDperiodontitis link regarding its systemic inflammatory burden.
Hence, we aimed to compare the hs-CRP levels of individuals with PD, according to their self-reported periodontal status. indicative of PD, and 51 adults were included for analysis with the detailed reasons for exclusion presented in the flowchart of Fig. 1.

Population
The sample consisted of 24 females (47.06%) and 27 males (52.94%), with a mean group age of~63 years (Table 1). Most participants were non-Hispanic whites (60.78%), reported an educational level higher than high school (58.82%), and were non-smokers (54.90%). However, only one statistically significant association was found between the self-report of periodontitis and sociodemographic data, namely the marital status (p = 0.025). The "no periodontitis" group presented a higher number of singles (28.57%), and the "periodontitis" group presented the majority of married/living with partner status (65.22%).
Regarding general health status, this group of PD patients presented a mean value of 2.55 total chronic medical conditions, 25.49% suffered from diabetes mellitus, and 66.67% from hypertension, although cases were evenly distributed (Table 1). Also, the sample presented approximately a mean value of 7.47 missing teeth. All in all, no statistically significant differences were found on general health covariates according to the self-report of periodontitis.

Blood and biochemical parameters
The levels of biochemical parameters and complete blood count with 5-part differential were analyzed in order to assess the systemic status of these participants according to the self-reported periodontal status (Table 2).
Overall, statistically significant differences were found for hs-CRP levels (p = 0.031) and lymphocyte percentage (p = 0.007). The "periodontitis" group presented higher mean levels of hs-CRP (5.36 vs. 1.99 mg/L) when compared to the "without periodontitis" group, while the "without periodontitis" group presented a higher mean percentage of lymphocytes when compared to the "periodontitis" group (29.35 vs. 28.03%).
In order to explore potential confounding variables on the hs-CRP values, we observed that PD patients with diabetes mellitus (p = 0.130), hypertension (p = 0.844), coronary heart disease (p = 0.405), emphysema (p = 0.365), asthma (p = 0.184), hepatic conditions (p = 0.888) or cancer (p = 0.354) had non-significant differences in serum levels of this marker. Similarly, active smokers did not present significant differences in serum levels of hs-CRP (p = 0.343).

DISCUSSION
The results of the present study showed that self-reported periodontitis is associated with higher circulating levels of hs-CRP in PD patients. No differences were found on white cells, red cells, and platelets according to the self-reported periodontal status of PD individuals, except for the lymphocyte percentage.
With this study, we ultimately aimed to strengthen the existing hypothesis that PD patients with periodontitis carry a systemic inflammatory burden, which can be translated into higher hs-CRP levels 27 . In fact, hs-CRP is one of the circulatory inflammatory markers previously known to be aggravated in patients with the periodontal disease when compared to healthy control groups [30][31][32] . Therefore, our results align and corroborate the previous lines of evidence in this regard 32 . Furthermore, to explore the hypothesis that the concomitant presence of other systemic conditions may interfere with the high hs-CRP levels in PD patients, we compared hs-CRP levels in PD patients with and without diabetes mellitus (p = 0.130), hypertension (p = 0.844), coronary heart disease (p = 0.405), emphysema (p = 0.365), asthma (p = 0.184), hepatic conditions (p = 0.888) or cancer (p = 0.354). No statistically significant differences were found, which indicates that the presence of other systemic diseases was not a confounding factor to the elevated hs-CRP levels in PD patients. Likewise, an active-smoker status was also not a confounding factor to high serum levels of hs-CRP (p = 0.343).
The clinical relevance of elevated CRP is worth discussing. This serum biomarker is mostly produced hepatically, triggered by acute and/or chronic inflammatory events 29 . In the past years, CRP has been shown to play a key role in the management of inflammatory diseases such as cardiovascular diseases 33 , neurodegenerative diseases (such as Alzheimer's Disease (AD) and PD) 28,34 or even periodontitis 27 . In addition to activating the complement classical pathway, CRP also binds to several tissues and membranes propelling the inflammatory reaction through cytokines and nuclear antigens. In what PD concerns, this may be of importance because systemic and cerebral inflammation is increasingly cited in its pathophysiological basis regarded as a syndrome by many 35 . Besides, neuroinflammation has been shown to be an important contributor to the pathogenesis of the Parkinsonian process and may aggravate the process of nigral neurodegeneration in animal models of PD 36 . Furthermore, in the periodontitis-PD link, and besides the established effects PD impairments cause in oral health that may ultimately lead to the development of periodontitis [9][10][11][12][13][14][15][16]24 , the infectious nature of periodontitis may have implications on gut microbiota 37 which is known to be abnormal in PD 36 .
Self-reported periodontitis is a validated, efficient, and accepted measure of periodontitis cases, with higher validity upon a combination of several self-report questions [20][21][22] . In fact, the selfreport strategy has been previously validated in other contexts, such as to identify cases of hypertension, diabetes mellitus, hypercholesterolemia 38 , risk factors for cardiovascular disease 39 , and even bruxism in PD patients 40 . All in all, self-report enables larger scale epidemiologic studies and low-cost surveillance of symptoms, risk factors, and diseases of interest 20 .
However, even though the periodontitis-leukocytosis link is well established-especially given the infectious nature of periodontitis whose effects summons WBC to the lesioned site 26,41,42the self-reported "periodontitis" group presented slightly lower lymphocyte percentage when compared to the "without periodontitis" group. This can be explained through the fact that elevated levels of WBC would be more probable upon clinical diagnosis of periodontitis cases. Therefore, as the clinical periodontal diagnosis is far more preferable and reliable, the use of selfreported measures of periodontitis stands as a limitation in this study, even though this method has been previously vali dated [20][21][22] and provides a cost-effective means of great-scale monitoring of oral health 20 . Furthermore, disease severity and activity could not be appraised through a self-report method, which is also fully reliable on the patient's knowledge of the disease and full awareness of a previous clinical diagnosis. Thus, the possibility of unmeasured confounding through this method of identification of periodontitis cases cannot be discarded 23 .
Additionally, the sample size may be considered limited and thus a shortcoming of this study. Perhaps, this might be because we are trying to signpost an underappreciated condition of difficult diagnosis 5 . Despite the small sample number, the collected data is of clinical significance, and we hope to pave the way for future larger studies on this condition worldwide (for instance using the MDS non-motor study group network). Also, the secondary study design based on the available NHANES data has been previously applied and accepted in several recent studies, including the PD-case selection method employed 26,43 . Nonetheless, due to the used PD-case selection method, data on disease duration was not available, which would have been relevant to evaluate disease staging. Furthermore, the observational nature of the study impairs the conclusion of causality, thus robust evidence has been reported regarding periodontitis increasing circulating levels of CRP and hs-CRP 42 .
All in all, the present study evaluates the association between self-reported periodontitis and hs-CRP levels in PD patients.
Hence, future research should continue to focus on the systemic repercussions of the periodontitis infection in PD patients, in the hopes of potentially clarifying the causality of the PD-periodontitis link. Furthermore, future research including in-depth clinical measures of periodontitis (such as periodontal pocket depth and clinical attachment loss), will provide further confirmation on the association with circulating systemic inflammatory surrogates.

Study design
In this secondary study, data was extracted and further analyzed from the National Health and Nutrition Examination Survey (NHANES), a representative and stratified multistage health-related survey conducted on noninstitutionalized U.S. citizens. The STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guideline was followed (Table S1, Supplementary Materials) 44 .

Setting, participants, and study size
Data from the NHANES 2015-2016 and 2017-2018 databases were used for the present study. Our analysis deemed the following inclusion criteria: 18 years of age or older; and undertaking secure PD medication regimens. Edentulous patients, missing data (on sociodemographic and/or systemic health information), and unsecure PD medication regimens as previously defined (Cabergoline, Orphenadrine, and Pramipexole) 26 were excluded.  Variables and data measurement PD definition. PD cases were identified in the NHANES database through the report of specific PD medications, according to a previous study 26 . Hence, the reported use of Benztropine, Carbidopa, Levodopa, Ropinirole, Methyldopa, Entacapone, and Amantadine were considered PD medications indicative of PD, therefore validating a PD case 46,47 . Cabergoline, Orphenadrine, and Pramipexole all present other known clinical applications apart from PD-Cabergoline is used to treat high levels of prolactin hormone 48 , Orphenadrine is used to treat muscle spasms in musculoskeletal conditions 49 and Pramipexole is also used to treat restless legs syndrome (RLS) 50 -and therefore were considered unsecure medications for the selection of PD cases.
Periodontitis definition. Periodontitis cases were pinpointed through a positive self-report on either one of the following oral health-related (OHR) questions, all regarding the moment when the survey was applied: "Do you think you might have gum disease?", "Ever had treatment for gum disease?" and "Ever been told of bone loss around teeth?". This method of self-reporting periodontitis has been previously validated and is indicative of a periodontitis case [20][21][22] .
Demographic characteristics. Age, gender, ethnicity, level of education, marital status, family income to poverty ratio, and smoking status were the self-reported sociodemographic variables collected and analyzed from NHANES datasets.
The level of education in individuals aged over 20 was categorized as follows: "<high school" (including <9th grade, 9-11th grade, and 12th grade with no diploma), "high school" (including high school grad/GED or equivalent) and ">high school" (including some college or AA degree and college graduate or above) 51 .
With regards to the reported family income to poverty ratio, a continuous score from 0 to 5 was given: "0" corresponding to no income, "5" corresponding to an income 5 or more times above the federal poverty threshold 53 .
At last, smoking status was defined as "active smokers" (reporting a consumption of ≥100 cigarettes during their lifetime and still currently smoking), "former smokers" (reporting smoking ≥100 cigarettes during their lifetime and presently ceased smoking) and non-smokers (reporting having smoked <100 cigarettes during their lifetimes) 26 .
Health characteristics. The systemic health status of the included participants was overall characterized through a sum of chronic medical conditions-asthma, psoriasis, gout, congestive heart failure, coronary heart disease, angina, heart attack, stroke, emphysema, thyroid, bronchitis, liver, and cancer-which was statistically considered a continuous variable. Furthermore, Diabetes Mellitus (DM) was separately defined through selfreport information and confirmed with glycated hemoglobin levels (Hba1c) 54 . Levels of Hba1c > 8% were considered uncontrolled DM cases 55 . Also, high blood pressure cases were defined from previous self-reports of medical-informed hypertension and were further confirmed with systolic and diastolic blood pressure levels (>140 mmHg and >90 mmHg, respectively) 56 .
Blood and biochemical parameters. Serum fractions of hs-CRP (mg/L), HDL-cholesterol (mg/dL), and total cholesterol were analyzed from blood specimens of the NHANES database 57,58 . Also, complete blood count with 5-part differential data was gathered, and information on white Blood Cell (WBC) count (10 9  Blood collection occurred~3 weeks following interviews, as detailed in https://wwwn.cdc.gov/nchs/nhanes/continuousnhanes/manuals.aspx? BeginYear=2015 (accessed in April 2021). reported through mean ± standard deviation (SD), while the number of cases (n) and percentage (%) represent categorical variables distribution among group categories. Upon assessment of data non-normality and homoscedasticity, Mann-Whitney test was applied for comparison of continuous variables. Chi-square test was used to evaluate association between the categorical variables. A 5% significance level was used in all inferential analyses.

Reporting Summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.

DATA AVAILABILITY
The analyzed data that supports the findings of this study are available in the NHANES database, a publicly accessible repository that does not issue DOIs, www. cdc.gov/nchs/nhanes.htm. The used datasets can be located under the "Survey Data and Documentation" tab, followed by the "NHANES 2015-2016" and "NHANES 2017-2018" databases tabs. "Demographics Data", "Examination Data", "Laboratory Data" and "Questionnaire Data" were consulted.