Abstract
AIM:
We investigate if periodontal disease is prospectively associated with cerebrovascular and neurodegenerative markers of dementia and Alzheimer’s pathology.
METHODS AND MATERIALS:
N=1,306 participants (Visit 5 mean age=76.5[standard deviation=5.4] years) in the Atherosclerosis Risk in Communities study with completed dental exams at Visit 4 underwent brain magnetic resonance imaging scans at Visit 5 while N=248 underwent positron emission tomography scans. Participants were classified as edentulous or, among the dentate, by the modified Periodontal Profile Class. Brain volumes were regressed on periodontal status in linear regressions. Cerebrovascular measures and β-amyloid positivity were regressed on periodontal status in logistic regressions.
RESULTS:
Periodontal disease was not associated with brain volumes, microhemorrhages, or elevated β-amyloid. Compared to periodontally healthy individuals, odds ratios[95% confidence interval] for all-type infarcts were 0.37[0.20, 0.65] for severe tooth loss and 0.56[0.31, 0.99] for edentulous participants.
CONCLUSIONS:
Within the limitations of this study, periodontal disease was not associated with altered brain volumes, microhemorrhages, or β-amyloid positivity. Tooth loss was associated with lower odds of cerebral infarcts.
Keywords: Periodontal Diseases, Dementia, Magnetic Resonance Imaging, Positron-Emission Tomography, Cohort Studies
1. INTRODUCTION
Dementia and mild cognitive impairment (MCI) are significant causes of disability among older adults. Despite the rising prevalence of dementia and cognitive decline, its neuropathological etiology is not fully understood.
Systemic inflammation is a major component in MCI and dementia development.1,2 Periodontal disease (PD), induced by oral microbial dysbiosis, may contribute to immune system dysfunction and chronic inflammatory phenotype.3,4 With inflammation as a mediating factor, PD and oral health have been linked to neurocognitive outcomes,5–9 including MCI10 and incident dementia11 as described in prior publications from the Atherosclerosis Risk in Communities (ARIC) study.
Few studies have investigated the relationship of PD and indicators of brain aging and vascular changes via imaging technology. Three key prior studies are noteworthy. A positron emission tomography (PET) study of brain amyloid in elderly adults identified a link between severe PD and elevated deposition of β-amyloid (βA) proteins.12 In another study, neurodegeneration was observed among those with tooth loss.13 Finally, greater periodontal probing depth was associated with more frequent lacunar infarcts indicated via magnetic resonance imaging (MRI).14 While informative, important limitations of these studies include lack of temporal inference due to cross-sectional designs, small sample sizes, and/or lack of full-mouth examinations.
We explored the association between periodontal disease and imaging measures of neurodegeneration, Alzheimer’s disease pathology, and cerebrovascular changes via brain imaging technology in ARIC. We hypothesized baseline periodontal disease status is associated with greater risk of cerebral small vessel disease, lower total and regional brain volumes, and elevated βA measured 13–17 years later.
2. MATERIALS AND METHODS
2.1. Description of Cohort
ARIC is an ongoing prospective study assessing the etiologic factors and clinical outcomes of atherosclerosis across four communities in the United States. The study mostly recruited White participants from Minneapolis, MN and Washington County, MD; Black participants from Jackson, MS; and Black or White participants from Forsyth County, NC.15,16 15,792 individuals aged 45–64 were present at baseline (Visit 1, 1987–1989).15,16 27% of participants identified as non-White, while 55% were female.16,17 Several follow-up examinations have occurred since Visit 1. Of relevance are periodontal examinations (Visit 4, [1996–1998]) and brain imaging examinations in subsets (Visit 5, [2011–2013]). 41% of the surviving Visit 1 cohort was evaluated in the ARIC Neurocognitive Study (ARIC-NCS) at Visit 5.18
2.2. Study Design
Of the n=11,656 participants at Visit 4, we excluded those who i) did not identify as Black or White (n=31), ii) were missing health history and/or demographic information (n=1,545), and iii) did not undergo dental screening (n=2,768). Of the n=7,312 eligible at Visit 4, n=5,962 dentate individuals underwent periodontal examinations as part of the Dental ARIC study, while n=1,350 with no natural teeth were included as an edentulous comparison group. After excluding n=1,509 who died and n=1,712 lost-to-follow-up before Visit 5, a sample of n=4,091 alive at the time of Visit 5 were invited to ARIC-NCS. From this group, a subset of n=1,306 in ARIC-NCS with brain MRI comprised our first analytical subgroup (n=463 with MCI, n=64 with dementia, n=779 dementia and MCI-free). Separately, a subset of healthy control ARIC-NCS participants with MRI scans were recruited into the ARIC-PET ancillary study for Florbetapir PET amyloid imaging, comprising our second analytical cohort (n=248). Figure 1 depicts the study inclusion criteria.
Figure 1:
Inclusion criteria for participants in periodontal disease-brain imaging study, ARIC
2.3. Standard Protocol Approvals, Registration, and Participant Consents
All participants provided written informed consent. The Institutional Review Board at all ARIC study centers approved the use of health data for the purposes of scientific research.
2.4. Periodontal Disease Assessment
Baseline PD status was ascertained via full mouth examination in the Dental ARIC study, an ancillary study that assessed PD and cardiovascular diseases among dentate Visit 4 participants. Certified dental examiners were calibrated versus a standard examiner, as well as against each other.19 We selected the Periodontal Profile Class (PPC), a robust classification system validated in prior ARIC studies as our periodontal definition.11,20,21 PPC integrates the following dental parameters: clinical attachment loss, periodontal probing depth, bleeding upon probing, gingival inflammation index,22 plaque index,23 and number of teeth present.24,25 PPC consists of seven stages: healthy (no disease), mild disease, high gingival inflammation (GI), tooth loss, posterior disease, severe tooth loss, and severe disease.24 To model complete tooth loss due to periodontal disease, edentulous participants were included to form an additional category in our modified PPC+edentulous classification.
Due to limited sample size in the PET cohort, we collapsed PPC into a four level variable (PPC4). Mild disease, high gingival inflammation index, tooth loss, and posterior disease stages were consolidated into a “mild to posterior disease” category while severe tooth loss and severe disease were combined into a “severe tooth loss to severe disease” group. Healthy and edentulous levels remained distinct.
2.5. Brain Imaging Measures
ARIC-NCS is an ancillary study on the role of cardiovascular risk factors in predicting dementia and MCI risks among elderly ARIC participants.18,26 Technical procedures of imaging processes are described in prior studies. Briefly, each field center conducted brain imaging in a subset of ARIC-NCS participants using 3-Tesla Siemens MRI scanners.18,26,27 Total brain volume, total intracranial volume, and temporal-parietal meta regions of interest (ROI)—to which the latter consists of measures for parahippocampal, entorhinal, inferior parietal lobules, hippocampus, and pre-cuneus volumes28,29— were quantified with 3D magnetization-prepared rapid gradient-echo (MP-RAGE) sequencing.27
Axial gradient recalled echo T2-weighted (T2*GRE) sequencing was used to determine the presence of microhemorrhages.30,31 Defined as lesions of ≤5 mm in maximum diameter, microhemorrhages were classified as subcortical or lobar.31 Using FLAIR sequencing, cerebral infarcts were categorized as cortical infarcts (includes small [hyperintense lesions 5–10 mm in diameter]; and large [≥10 mm] infarcts), or subcortical infarcts (dark colored hyperintense lesions with ≥3 mm diameter).27
Florbetapir PET images were taken in a subset of healthy control Visit 5 participants to evaluate brain βA deposition at three field centers (ARIC-PET).26 βA deposition was computed using the global cortical, which is the weighted mean Florbetapir uptake in the following brain regions: orbitofrontal, prefrontal, and superior frontal cortices; lateral temporal, parietal, and occipital lobes; the precuneus, and the anterior and posterior cingulates.26,32 As previously done in ARIC, global cortex was dichotomized to form a standardized uptake value ratio (SUVR), where SUVR >1.2 indicated elevated mean cortical βA deposition.26,32
2.6. Risk Factors, Confounders, and Other Variables
Sex, education level, race, income, and health insurance status were derived from self-report questionnaires at ARIC Visit 1 (1987–1989) while self-reported physical activity was collected at Visit 3 (1993–1995). Health interview variables included age at periodontal exam, smoking status, hypertension-lowering medication use, self-reported physician-diagnosed coronary heart disease33 (prevalent CHD), and frequency of dentist visits. Strokes prior to Visit 1 were defined via self-reported physician diagnosis, whereas stroke occurrences after Visit 1 were determined with adapted diagnosis code criteria from the National Survey of Stroke.34 Participants who reported taking medications for heart failure and who achieved a Gothenburg score of 3 at Visit 4 were recorded as having prevalent heart failure. Clinical examinations were used to ascertain measures for body mass index (BMI), LDL cholesterol (via Friedewald formula), and systolic blood pressure (SBP; mean of first and second readings) at Visit 4. Prevalent diabetes at Visit 4 was defined as self-reported physician diagnosis, fasting glucose ≥126 mg/dL, non-fasting glucose ≥200 mg/dL, or self-reported medication use for diabetes or high blood sugar. Dementia and MCI were defined using ARIC-NCS-specific ascertainment techniques.11,18 Briefly, results from Visit 5 neurocognitive examinations were integrated into a computer algorithm used to determine dementia and MCI status.11,18 Determinations from this algorithm were then adjudicated by an expert review panel.11,18 Visit 1 APOE ε4 status (0 vs ≥1 alleles) was assessed from blood samples via TaqMan assay (Applied Biosystems, Foster City, CA).
2.7. Statistical Analysis
We assessed baseline characteristics of both cohorts by PPC levels using one-way ANOVA and chi-square tests. Total brain size and temporal-parietal meta ROI volumes were linearly regressed on PPC in the MRI subgroup. Z-scores of volumetric measures were also computed. Odds ratios for microhemorrhages and infarcts by PPC were computed with logistic regression. In the PET subgroup, we examined odds of elevated SUVR by periodontal status via logistic regression. Since APOE ε4 carrier status is a key genetic factor in Alzheimer’s disease development and βA deposition, we assessed it as a potential effect measure modifier. As done in prior reports including ARIC,35,36 we used inverse probability weighting in both participant subgroups to account for selection for brain MRI, attrition due to death, or failure to attend the follow-up neurocognitive exam (censoring).
Multivariable models adjusted for the following: age, sex, education, race, income, insurance status, APOE ε4 allele presence, total intracranial volume (only adjusted for the outcome of total brain volume27), BMI, physical activity, smoking status, LDL cholesterol, prevalent heart failure, prevalent coronary heart disease, prevalent stroke, anti-hypertensive medication use, systolic blood pressure and dental visit frequency.
Analyses were executed in SAS (version 9.4; SAS Institute Inc., Cary, NC). P-values lower than 0.05 were considered statistically significant.
2.8. Data Availability
ARIC data are available, with appropriate permissions, via the NIH NHLBI-sponsored Biologic Specimen and Data Repository Information Coordinating Center (BioLINCC) https://biolincc.nhlbi.nih.gov/.
3. RESULTS
3.1. Participant Characteristics
Among n=1,306 ARIC participants in the MRI subgroup, 58% were female, 26% were Black, and Visit 5 mean age was 76.5 years (SD=5.4). Approximately 30% of participants carried at least one APOE ε4 allele. Prevalent periodontal disease at Visit 4, defined as any periodontal category other than periodontally healthy, was 73% using PPC. Nearly 5% (n=64) of participants in this subgroup were identified with dementia while approximately 35% (n=463) had mild cognitive impairment.
Several participant characteristics including demographics and general health measures differed by PPC in the MRI cohort (Table 1). Overall, participants who identified as Black, did not possess a high school diploma, earned an annual income less than $50,000, or rarely visited a dentist were more likely to be classified with tooth loss, severe periodontal disease, or be edentulous. Medicare/Medicaid and private health insurance beneficiaries were more likely to be periodontally healthy or categorized with mild disease compared to participants without insurance. Major differences in Visit 5 clinical measurements, medication use, and prevalent chronic diseases across PPC levels were not observed. Among the n=248 in the PET subgroup, 57% were female, 40% were Black, while mean Visit 5 age was 75.9 (SD=5.4). About 32% carried APOE ε4 alleles, while 87% of participants had PPC-defined periodontal disease (Table 2).
Table 1:
Characteristics of n=1,306 MRI cohort by PPC, ARIC
| Categories of Periodontal Profile Class | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Healthy | Mild Disease | High GI | Tooth Loss | Posterior Disease | Severe Tooth Loss | Severe Disease | Edentulous | p-value | |
| %,(N) | 26.6 (347) | 15.3 (200) | 10.7 (140) | 8.1 (106) | 9.6 (125) | 11.6 (151) | 6.3 (82) | 11.9 (155) | |
| Visit 5 age (SE) | 76.4 (0.3) | 76.6 (0.4) | 75.2 (0.4) | 77.7 (0.5) | 76.8 (0.5) | 77.3 (0.4) | 75.6 (0.6) | 76.7 (0.4) | <0.01* |
| Sex %,(n) | <0.01* | ||||||||
| Male | 34.3 (119) | 52.0 (104) | 40.7 (57) | 39.6 (42) | 60.8 (76) | 38.4 58 | 54.9 (45) | 27.7 (43) | |
| Female | 65.7 (228) | 48.0 (96) | 59.3 (83) | 60.4 (64) | 39.2 (49) | 61.6 (93) | 45.1 (37) | 72.3 (112) | |
| Race %,(n) | <0.01* | ||||||||
| Black | 6.0 (21) | 3.5 (7) | 82.1 (115) | 21.7 (23) | 2.4 (3) | 40.4 (61) | 53.7 (44) | 38.1 (59) | |
| White | 94.0 (356) | 96.5 (193) | 17.9 (25) | 78.3 (83) | 97.6 (122) | 59.6 (90) | 46.3 (38) | 61.9 (96) | |
| Education level %,(n) | <0.01* | ||||||||
| Less than high school | 2.6 (9) | 5.5 (11) | 12.9 (18) | 17.0 (18) | 3.2 (4) | 26.5 (40) | 15.9 (13) | 36.8 (57) | |
| High school graduate | 38.3 (133) | 45.0 (90) | 35.0 (49) | 45.3 (48) | 48.8 (61) | 45.0 (68) | 36.6 (30) | 45.2 (70) | |
| College graduate | 59.1 (205) | 49.5 (99) | 52.1 (73) | 37.7 (40) | 48.0 (60) | 28.5 (43) | 47.6 (39) | 18.1 (28) | |
| Smoking status %,(n) | <0.01* | ||||||||
| Current | 6.1 (21) | 4.0 (8) | 8.6 (12) | 13.2 (14) | 10.4 (13) | 11.3 (17) | 8.5 (7) | 18.1 (28) | |
| Former | 41.2 (143) | 41.5 (83) | 38.6 (54) | 45.3 (48) | 57.6 (72) | 42.4 (64) | 40.2 (33) | 40.7 (63) | |
| Never | 52.7 (183) | 54.5 (109) | 52.9 (74) | 41.5 (44) | 32.0 (40) | 46.4 (70) | 51.2 (42) | 41.3 (64) | |
| V5 BMI* (kg/m2),(SE) | 27.1 (0.2) | 28.0 (0.4) | 28.3 (0.5) | 28.3 (0.5) | 27.9 (0.4) | 29.5 (0.6) | 29.5 (0.7) | 30.1 (0.5) | <0.01* |
| V5 sport index† (SE) | 2.8 (0.0) | 2.6 (0.1) | 2.6 (0.1) | 2.5 (0.1) | 2.7 (0.1) | 2.4 (0.1) | 2.6 (0.1) | 2.3 (0.1) | <0.01* |
| V5 systolic blood pressure† (mmHg),(SE) | 129.3 (1.0) | 129.7 (1.2) | 130.9 (1.6) | 124.8 (1.5) | 128.8 (1.6) | 131.9 (1.4) | 128.3 (2.3) | 131.8 (1.6) | 0.05 |
| V5 diastolic blood pressure† (mmHg),(SE) | 65.9 (0.5) | 63.9 (0.7) | 64.9 (1.1) | 69.5 (0.9) | 65.0 (0.9) | 65.8 (1.0) | 65.4 (1.2) | 66.0 (0.9) | <0.01* |
| V5 hypertension† %,(n) | 56.8 (197) | 61.5 (123) | 76.4 (107) | 69.8 (74) | 59.2 (74) | 75.5 (114) | 74.4 (61) | 72.3 (112) | <0.01* |
| V5 anti-hypertensive use (%) | 65.7 (228) | 70.5 (141) | 81.4 (114) | 82.1 (87) | 67.2 (84) | 86.8 (131) | 78.1 (64) | 77.4 (120) | <0.01* |
| V5 HDL cholesterol† (mg/dL),(SE) | 55.4 (0.8) | 56.2 (1.3) | 51.4 (1.3) | 49.8 (1.6) | 51.6 (1.4) | 51.1 (1.2) | 51.0 (1.7) | 51.2 (1.0) | <0.01* |
| V5 LDL cholesterol† (mg/dL),(SE) | 106.2 (1.7) | 110.7 (3.1) | 103.3 (3.7) | 121.0 (3.3) | 103.5 (2.8) | 101.8 (3.0) | 108.8 (4.0) | 106.4 (2.8) | 0.45 |
| V5 total cholesterol† (mg/dL),(SE) | 185.4 (2.2) | 189.9 (3.7) | 180.5 (4.4) | 199.2 (3.5) | 178.8 (3.6) | 179.6 (3.6) | 184.1 (5.1) | 182.2 (3.1) | 0.37 |
| V5 statin use† %,(n) | 47.0 (163) | 51.5 (103) | 42.1 (59) | 50.0 (53) | 49.6 (62) | 58.3 (88) | 52.4 (43) | 52.3 (81) | 0.33 |
| V5 prevalent coronary heart disease %,(n) | 5.8 (20) | 9.5 (19) | 5.7 (8) | 7.6 (8) | 11.2 (14) | 11.9 (18) | 8.5 (7) | 11.0 (17) | 0.20 |
| V5 prevalent heart failure %,(n) | 4.9 (17) | 6.0 (12) | 11.4 (16) | 11.3 (12) | 11.2 (14) | 15.2 (23) | 12.2 (10) | 11.0 (17) | <0.01* |
| V5 prevalent stroke %,(n) | 3.2 (11) | 3.0 (6) | 5.0 (7) | 0.9 (1) | 1.6 (1) | 2.0 (3) | 4.9 (4) | 1.3 (2) | 0.35 |
| V5 prevalent diabetes† %,(n) | 20.8 (72) | 23.0 (46) | 34.3 (48) | 29.3 (31) | 24.8 (31) | 41.1 (62) | 36.6 (30) | 36.8 (57) | <0.01* |
| Mean number of teeth (SE) | 26 (0) | 26 (0) | 21 (0) | 17 (0) | 26 (0) | 8 (0) | 25 (0) | 0 (0) | <0.01* |
| Dental visit freq. %,(n) | <0.01* | ||||||||
| Regular | 93.4 (324) | 87.0 (174) | 59.3 (83) | 77.4 (82) | 90.4 (113) | 44.4 (67) | 53.7 (44) | 7.1 (11) | |
| Occasionally | 6.3 (22) | 11.5 (23) | 39.3 (55) | 21.7 (23) | 8.0 (10) | 50.3 (76) | 42.7 (35) | 81.9 (127) | |
| Never/other | 0.3 (1) | 1.5 (3) | 1.4 (2) | 0.9 (1) | 1.6 (2) | 5.3 (8) | 3.7 (3) | 11.0 (17) | |
| Income %,(n) | <0.01* | ||||||||
| <$5,000-$24,999 | 14.1 (49) | 15.5 (31) | 27.1 (38) | 21.7 (23) | 13.6 (17) | 41.1 (62) | 22.0 (18) | 49.0 (76) | |
| $25,000-$49,999 | 31.4 (109) | 35.5 (71) | 33.6 (47) | 43.4 (46) | 32.8 (41) | 40.4 (61) | 34.2 (28) | 31.0 (48) | |
| $50,000-$74,999 | 24.5 (85) | 21.5 (43) | 19.3 (27) | 20.8 (22) | 29.6 (37) | 11.3 (17) | 18.3 (15) | 9.7 (15) | |
| ≥$75,000 | 28.2 (98) | 24.5 (49) | 13.6 (19) | 11.5 (12) | 21.6 (27) | 5.3 (8) | 18.3 (15) | 4.5 (7) | |
| Missing/not reported | 1.7 (6) | 3.0 (6) | 6.4 (9) | 2.8 (3) | 2.4 (3) | 2.0 (3) | 7.3 (6) | 5.8 (9) | |
| Insurance status %,(n) | |||||||||
| Medicare/Medicaid | 31.4 (109) | 38.0 (76) | 25.0 (35) | 39.6 (42) | 34.4 (43) | 42.4 (64) | 29.3 (24) | 39.4 (61) | <0.01* |
| Private insurance | 66.3 (230) | 58.0 (116) | 66.4 (93) | 55.7 (59) | 64.8 (81) | 48.3 (73) | 65.9 (54) | 48.4 (75) | |
| No insurance | 8 (2.31) | 8 (4.00) | 12 (8.57) | 5 (4.72) | 1 (0.80) | 14 (9.27) | 4 (4.88) | 19 (12.26) | |
| APOE ε4 alleles %,(n) | 0.03* | ||||||||
| 0 | 73.5 (255) | 75.0 (150) | 60.0 (84) | 73.6 (78) | 68.8 (86) | 63.6 (96) | 67.1 (55) | 72.3 (112) | |
| 1+ | 26.5 (92) | 25.0 (50) | 40.0 (56) | 26.4 (28) | 31.2 (39) | 36.4 (55) | 32.93 (27) | 27.7 (43) | |
Abbreviations: PPC=Periodontal Profile Class; GI=gingival inflammation; V5=Visit 5; APOE = Apolipoprotein E.
P-values correspond to statistical differences across PPC;
indicates statistical significance.
Indicates missing observations. Missingness of select V5 variables across PPC is as follows:
BMI: Healthy=2, Posterior disease=1, Severe tooth loss=1, Edentulous=1; Sport Index: Healthy=17, Mild disease=3, High GI=3, Tooth loss=6, Posterior disease=3, Severe tooth loss=2, Severe disease=1, Edentulous=2; Systolic and Diastolic Blood Pressures: Tooth loss=1, Severe tooth loss=1; Prevalent Hypertension: Healthy=3, Mild disease=3, Tooth loss=2, Posterior disease=1, Severe tooth loss=6; HDL and Total Cholesterol: Healthy=2, Mild disease=1, High GI=2, Tooth loss through Edentulous=1; LDL Cholesterol: Healthy=2, Mild disease=2, High GI=3, Tooth loss=2, Posterior Disease=1, Severe tooth loss=4, Severe disease=1, Edentulous=1; Statin Use: Healthy=2, Severe tooth loss=1, Severe tooth loss=1; Prevalent Diabetes: Healthy=7, High GI=2, Posterior disease=1, Severe Tooth loss=2, Severe disease=1, Edentulous=2
Table 2:
Characteristics of n=248 PET cohort by PPC, ARIC
| Categories of Periodontal Profile Class | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Healthy | Mild Disease | High GI | Tooth Loss | Posterior Disease | Severe Tooth Loss | Severe Disease | Edentulous | p-value | |
| N | 42 | 33 | 44 | 28 | 15 | 30 | 23 | 33 | |
| Visit 5 age, (SE) | 74.8 (0.7) | 75.9 (1.1) | 75.1 (0.8) | 78.4 (1.1) | 75.1 (1.2) | 77.3 (1.0) | 74.8 (1.1) | 76.2 (0.9) | 0.11 |
| Sex %,(n) | 0.18 | ||||||||
| Male | 35.71 (15) | 46.48 (16) | 43.18 (19) | 50.00 (14) | 60.00 (9) | 23.33 (7) | 56.52 (13) | 42.42 (14) | |
| Female | 64.29 (27) | 51.52 (17) | 56.82 (25) | 50.00 (14) | 40.00 (6) | 76.67 (23) | 43.48 (10) | 57.58 (19) | |
| Race %,(n) | <0.01* | ||||||||
| Black | 16.67 (7) | 3.03 (1) | 95.45 (42) | 32.14 (9) | 0.00 (0) | 40.00 (12) | 60.87 (14) | 39.39 (13) | |
| White | 83.33 (35) | 96.97 (32) | 4.55 (2) | 67.86 (19) | 100.00 (15) | 60.00 (18) | 39.13 (9) | 60.61 (20) | |
| Mean Number of Teeth (SE) | 26 (2) | 27 (3) | 20 (4) | 15 (5) | 26 (3) | 8 (2) | 25 (4) | 0 (0) | <0.01* |
| Dental Visit Frequency %,(n) | <0.01* | ||||||||
| Regular | 88.10 (37) | 84.85 (28) | 50.00 (22) | 64.29 (18) | 86.67 (13) | 36.67 (11) | 65.22 (15) | 6.06 (2) | |
| Occasionally | 11.90 (5) | 12.12 (4) | 50.00 (22) | 32.14 (9) | 13.33 (2) | 60.00 (18) | 34.78 (8) | 87.88 (29) | |
| Never/other | 0.00 (0) | 3.03 (1) | 0.00 (0) | 3.57 (1) | 0.00 (0) | 3.33 (1) | 0.00 (0) | 6.06 (2) | |
| APOE ε4 alleles %,(n) | 0.30 | ||||||||
| 0 | 80.95 (34) | 69.70 (23) | 56.82 (25) | 71.43 (20) | 80.00 (12) | 66.67 (20) | 56.52 (13) | 66.67 (22) | |
| 1+ | 19.05 (8) | 30.30 (10) | 43.18 (19) | 28.57 (8) | 20.00 (3) | 33.33 (10) | 43.48 (10) | 33.33 (11) | |
Abbreviations: PPC = Periodontal Profile Class; GI = gingival inflammation; APOE = Apolipoprotein E.
P-values correspond to statistical differences across PPC;
indicates statistical significance.
3.2. Structural Brain Measures
Mean total brain and temporal-parietal meta ROI volumes by PPC are shown in Table 3. Estimated mean total brain volumes ranged from 969.8±8.1 cm3 to 1057.5±9.8 cm3 across PPC levels (p<0.01). After full multivariable adjustments, the association between PPC status and brain volume attenuated as brain volumes were generally similar across PPC, ranging from 1009 to 1017 cm3. In univariate modeling of temporal-parietal meta ROI volumes, mild and moderate disease PPC levels showed mean estimates approximate to 60 cm3 while more severe categories showed slightly lower mean volumes. Adjusting for all confounders, temporal-parietal meta ROI volumes were consistent across exposure levels. Multivariable adjusted Z-scores for brain volumes are displayed in Supplemental Table 1.
Table 3:
Mean volumetric brain measures (SEM) of n=1,306 MRI cohort by PPC at Visit 5, ARIC
| Categories of Periodontal Profile Class | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Healthy | Mild Disease | High GI | Tooth Loss | Posterior Disease | Severe Tooth Loss | Severe Disease | Edentulous | p-value | |
| N | 347 | 200 | 140 | 106 | 125 | 151 | 82 | 155 | N=1306 |
| Total Brain Volume (cm3) | |||||||||
| Crude | 1028.20 (5.92) | 1033.38 (7.64) | 993.83 (9.24) | 1011.08 (10.48) | 1057.54 (9.80) | 991.95 (8.51) | 1009.60 (11.89) | 969.75 (8.07) | <0.01* |
| Model 1 | 1017.11 (2.90) | 1018.11 (3.61) | 1022.31 (4.78) | 1016.66 (4.82) | 1011.18 (4.63) | 1015.02 (4.03) | 1013.16 (5.57) | 1011.47 (3.90) | 0.62 |
| Model 2 | 1014.73 (2.92) | 1016.24 (3.56) | 1021.80 (4.67) | 1017.66 (4.73) | 1009.40 (4.57) | 1016.90 (4.00) | 1014.42 (5.46) | 1017.65 (4.32) | 0.79 |
| Temporal-Parietal Meta ROI Volume (cm3) | |||||||||
| Crude | 60.11 (0.38) | 60.25 (0.49) | 57.34 (0.59) | 58.82 (0.67) | 61.93 (0.63) | 57.07 (0.55) | 58.40 (0.76) | 56.00 (0.52) | <0.01* |
| Model 1 | 59.29 (0.33) | 58.66 (0.40) | 59.12 (0.54) | 59.16 (0.54) | 59.51 (0.519) | 58.50 (0.45) | 58.16 (0.62) | 58.02 (0.44) | 0.24 |
| Model 2 | 59.26 (0.33) | 58.67 (0.40) | 59.16 (0.53) | 59.33 (0.54) | 59.38 (0.52) | 58.49 (0.45) | 58.24 (0.62) | 58.06 (0.49) | 0.42 |
Abbreviations: PPC = Periodontal Profile Class; GI = Gingival Inflammation; ROI = Regions of Interest.
Temporal-parietal meta ROI outcome is comprised of combined volumetric measures of parahippocampal, entorhinal, inferior parietal lobules, hippocampus, and pre-cuneus brain regions.
P-values are from tests for significant difference between PPC levels.
Denotes significant p-values (p<0.05).
Estimates of mean volumetric brain measures were computed via multiple linear regression.
Model 1: crude + age + sex + education + race + income + insurance + APOE ε4 allele presence + total intracranial volume
Model 2: Model 1 + body mass index + physical activity + smoking status + LDL cholesterol + prevalent heart failure + anti-hypertensive medication + prevalent coronary heart disease + prevalent stroke + prevalent diabetes + systolic blood pressure + dental visit frequency.
Note: Total brain volume was additionally adjusted for total intracranial volume starting in Model 1.
3.3. Presence of Cerebrovascular Abnormalities
Odds ratios for the associations between PPC and microhemorrhages and infarcts are presented in Table 4. After full confounder adjustment, PPC was not a significant predictor of microhemorrhages nor were its subcortical and lobar subtypes. Compared to healthy PD status, participants with mild disease, high inflammatory index, moderate and complete tooth loss, and severe disease had non-significantly higher odds of all-type and subcortical microhemorrhages. Odds of lobar microbleeds were non-significantly lower among moderate and severe disease, high gingival inflammation, and moderate and severe tooth loss categories.
Table 4:
Adjusted odds ratios (95% CIs) for cerebrovascular outcomes of n=1,306 MRI cohort by PPC at Visit 5, ARIC
| Categories of Periodontal Profile Class | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Healthy | Mild Disease | High GI | Tooth Loss | Posterior Disease | Severe Tooth Loss | Severe Disease | Edentulous | p-value | |
| N | 347 | 200 | 140 | 106 | 125 | 151 | 82 | 155 | N=1306 |
| Microhemorrhages | |||||||||
| All type | ref. | 1.43 (0.93, 2.21) | 1.15 (0.66, 2.02) | 1.11 (0.64, 1.94) | 0.67 (0.37, 1.19) | 0.76 (0.44, 1.31) | 1.14 (0.61, 2.13) | 1.55 (0.89, 2.69) | 0.07 |
| Subcortical | ref. | 1.45 (0.91, 2.31) | 1.38 (0.77, 2.50) | 1.37 (0.77, 2.43) | 0.67 (0.35, 1.28) | 0.84 (0.47, 1.50) | 1.37 (0.71, 2.64) | 1.29 (0.71, 2.33) | 0.19 |
| Lobar | ref. | 1.31 (0.67, 2.57) | 0.82 (0.33, 2.01) | 0.73 (0.27, 1.98) | 0.65 (0.26, 1.64) | 0.92 (0.40, 2.08) | 0.75 (0.27, 2.05) | 1.10 (0.46, 2.63) | 0.83 |
| Infarctions | |||||||||
| All type | ref. | 0.67 (0.42, 1.08) | 1.00 (0.56, 1.79) | 1.42 (0.84, 2.41) | 0.83 (0.49, 1.40) | 0.37 (0.20, 0.65) | 0.93 (0.50, 1.73) | 0.56 (0.31, 0.99) | <0.01* |
| Cortical | ref. | 1.14 (0.64, 2.03) | 0.68 (0.30, 1.51) | 1.52 (0.78, 2.97) | 0.62 (0.29, 1.33) | 0.37 (0.16, 0.87) | 0.54 (0.21, 1.36) | 0.76 (0.34, 1.71) | 0.06 |
| Subcortical | ref. | 0.52 (0.30, 0.90) | 0.98 (0.51, 1.87) | 1.02 (0.56, 1.85) | 0.74 (0.41, 1.35) | 0.36 (0.19, 0.67) | 1.00 (0.51, 1.97) | 0.36 (0.19, 0.68) | <0.01* |
| Lacunar | ref. | 0.50 (0.29, 0.87) | 0.99 (0.52, 1.89) | 1.01 (0.56, 1.85) | 0.66 (0.36, 1.22) | 0.34 (0.18, 0.64) | 1.03 (0.53, 2.02) | 0.34 (0.18, 0.66) | <0.01* |
Abbreviations: PPC = Periodontal Profile Class; CI = confidence intervals; GI = gingival inflammation.
P-values are from tests for significant difference of odds between PPC levels.
Denotes significant p-values (p<0.05). Bold typeface indicates significant 95% confidence intervals.
Odds ratio estimates shown are from full multivariable logistic regression models. Covariables of age, sex, education, race, income, insurance status, APOE ε4 allele presence, body mass index, physical activity, smoking status, LDL cholesterol, prevalent heart failure, anti-hypertensive medication use, prevalent coronary heart disease, prevalent stroke, prevalent diabetes, systolic blood pressure, and dental visit frequency were adjusted.
PPC was a strong predictor of all-type infarcts as well as subcortical and lacunar infarcts after covariable adjustment (p<0.01) as shown in Table 4. The odds ratio (OR) [95% confidence interval] for all-type infarcts was 0.37 [0.20, 0.65] in the severe tooth loss category and 0.56 [0.31, 0.99] for edentulous participants, indicating lower odds of infarcts compared to healthy PPC. Similarly, decreased odds were consistent for subcortical and lacunar infarcts between severe tooth loss and edentulism PPC levels. Additionally, those with mild disease had at least 50% lower odds of subcortical and lacunar subtype infarcts compared to healthy individuals. PPC was associated with cortical infarcts overall, as severe tooth loss was protective compared to healthy, although imprecise (OR=0.37 [0.16, 0.87]). Tooth loss was non-significantly associated with elevated odds of all-type infarcts and each subtype. MRI findings without inverse probability weighting are presented in Supplemental Table 2.
3.4. Deposition of β-Amyloid
Table 5 displays the odds of elevated βA deposition by PPC4 among the PET cohort. Overall, PPC was not associated with βA in sequentially adjusted logistic modeling. Mild to posterior disease and severe tooth loss-severe disease categories had comparably non-significant lower risks relative to healthy participants across multivariable models. After adjustment for all confounders, odds of elevated βA were non-significantly highest among the edentulous (OR=2.74 [0.74, 10.17]) versus the periodontally healthy, although this estimate was imprecise. No significant interaction was observed between PPC and APOE ε4 after full covariable adjustment (p=0.13). When stratified by presence of APOE status, ε4 allele carriers tended to have higher odds of elevated βA deposition compared to non-carriers, though precision was poor for the stratified analyses. PET findings without inverse probability weighting are presented in Supplemental Table 3, while global cortical SUVR across both periodontal classifications are found in Supplemental Figures 1 and 2.
Table 5:
Odds ratios (95% CIs) of β-amyloid positivity (SUVR>1.2) of n=248 PET cohort by PPC4 at Visit 5, ARIC
| Main Effects | |||||
|---|---|---|---|---|---|
| Healthy | Mild-Posterior Disease | Severe Tooth Loss- Severe Disease | Edentulous | Total | |
| Sample size | 42 | 120 | 53 | 33 | N=248 |
| N events | 19 | 61 | 29 | 22 | 131 |
| p-value | |||||
| Crude | ref. | 1.19 (0.58, 2.46) | 1.67 (0.74, 3.79) | 2.43 (0.95, 6.19) | 0.20 |
| Model 1 | ref. | 0.82 (0.37, 1.82) | 0.93 (0.37, 2.33) | 1.82 (0.62, 5.34) | 0.38 |
| Model 2 | ref. | 0.87 (0.38, 2.02) | 0.87 (0.31, 2.39) | 2.74 (0.74, 10.17) | 0.21 |
| PPC4 x APOE Interaction | |||||
|
APOE ε4 alleles <1
(n=169) |
|||||
| N (%) elevated βA | 16 (21.05) | 33 (43.42) | 15 (19.74) | 12 (15.79) | |
| Crude OR | ref. | 0.78 (0.34, 1.77) | 1.28 (0.50, 3.28) | 1.38 (0.48, 3.95) | |
| Adjusted OR | ref. | 0.63 (0.25, 1.58) | 0.69 (0.22, 2.21) | 1.67 (0.41, 6.86) | |
|
APOE ε4 alleles ≥1
(n=79) |
|||||
| N (%) elevated βA | 3 (5.45) | 28 (50.91) | 14 (25.45) | 10 (18.18) | |
| Crude OR | ref. | 3.67 (0.68, 19.76) | 3.43 (0.56, 20.93) | 18.88 (1.34, 266.70) | |
| Adjusted OR | ref. | 4.38 (0.60, 32.09) | 3.44 (0.42, 28.05) | 38.29 (1.83, 801.58) | |
| Interaction p-value | 0.13 | ||||
Abbreviations: PPC4 = Condensed 4-level Periodontal Profile Class; CI = Confidence Intervals; APOE = Apolipoprotein E. Bold typeface indicates significant 95% confidence intervals (p<0.05). P-values from testing whether odds of outcome are significantly different between PPC4 levels.
Estimates for odds ratios were computed via multivariable logistic regression.
Model 1: crude + age + sex + education + race + income + insurance + APOE ε4 allele presence
Model 2: Model 1 + body mass index + physical activity + smoking status + LDL cholesterol + prevalent heart failure + anti-hypertensive medication + prevalent coronary heart disease + prevalent stroke + prevalent diabetes + systolic blood pressure + dental visit frequency.
Interaction analyses were conducted with full covariable adjustments (Model 2).
4. DISCUSSION
In this population-based study, we explored the relationship between PD status and brain imaging outcomes assessed across a median of 14.7 years of prospective follow-up among a cohort of older participants in ARIC. Our findings show that (i) baseline PD was not associated with altered total brain volume, nor were there major differences in brain region volumes by PD status; (ii) PD status was not linked to MRI-detected microhemorrhages; (iii) periodontal categories of mild disease, severe tooth loss, and edentulism were associated with lower risk of MRI-indicated subcortical and lacunar infarcts; and (iv) elevated βA accumulation assessed by PET imaging was not associated with baseline PD status, although there was a trend for an association among APOE-ε4 allele carriers.
The present results from ARIC stand in contrast to some prior reports. A 2018 cross-sectional MRI-based study among elderly Swedish adults (n=2,715) showed lower total brain volume among participants with advanced tooth loss versus dentate individuals,13 whereas such relationship was not observed in our findings. Excessive secretion of pro-inflammatory molecules from continual periodontal infection may circulate into the brain and elicit downstream neuroinflammatory processes contributing to brain atrophy.13,37 Although there is biological plausibility to expect more serious PPC categories to present brain volumes indicative of neurodegeneration and Alzheimer’s pathology, our null findings suggest that PD status may not change brain structure.
The inverse relationship between PD and risk for subclinical strokes along with null findings for PPC and microhemorrhages were surprising and did not support our a priori hypotheses. Furthermore, our current findings did not align with the broader literature that has consistently shown PD to be associated with greater risk of subclinical markers of cerebral infarcts14 as well as incident ischemic stroke.38–41 A cohort study among elderly Japanese participants found increasing periodontal pocket depth to be modestly associated with increased odds of MRI-detected lacunar infarcts (OR=1.916 [0.979,3.751]; p=0.058).14 A case-control study among Korean adults found clinical attachment loss ≥6 mm to be strongly linked to elevated odds of hemorrhagic stroke post confounder adjustment (OR=2.53 [1.14,5.61]).42 And in a longitudinal analysis among the same ARIC population with similar study constructs, advanced PPC levels were associated with greater risk of incident ischemic stroke, as well as for lacunar, cardioembolic, and thrombotic subtypes after adjusting for health and lifestyle factors.20
An important factor possibly contributing to the inverse associations is the intentional design of ARIC-NCS. Only a subset of cognitively healthy Visit 5 participants were recruited to undergo brain MRI.27 Similarly, an even smaller subset of participants with completed MRI scans were selected for PET,26 thus resulting in limited sample sizes of our imaging cohorts. Selection bias and competing risks may have also contributed to the reduced risk of cerebrovascular outcomes by select PPC levels. Individuals with advanced PD and tooth loss were less likely to attend Visit 5 compared to individuals with milder forms (Supplemental Table 4). Likewise, participants who were censored or died prior to brain imaging were more likely to experience CVD-related comorbidities compared to brain imaging participants (Supplemental Table 5). Alternatively, partial to complete tooth loss restricts sufficient biofilm generation between the teeth and gingiva, reducing the risk of dysbiotic inflammation. This may limit systemic inflammatory burden linked to pathways associated with cerebrovascular disease, ultimately lowering the risk of infarcts. Lastly, information bias in the PPC might contribute to these inverse findings. We assume severe tooth loss and edentulism reflect periodontal infections. However, it is possible that such states were due to dental caries, trauma, or surgical procedures, all of which have no or very modest evidence in the literature linking them to increase risk for cerebrovascular disease.
While we did not observe an association between PD and βA, we found a non-statistically significant strong association between PD and elevated βA deposition among APOE ε4 allele carriers, which is in line with prior research and is biologically plausible. It is hypothesized that βA may serve as an antimicrobial agent, as it could potentially be beneficial in early periodontal pathogenic exposure.43 However, these antimicrobial mechanisms under prolonged chronic periodontal infection may lead to the overproduction of βA and the aggregation of toxic gingipains (toxic Porphyromonas gingivalis-derived proteases) that proteolyze βA, resulting in neuroinflammation.
Our investigation has several strengths. First, the use of brain imaging technology provides insight into the relationship between periodontal status and preclinical markers of dementia and Alzheimer’s pathology. Second, PD severity was objectively assessed via full mouth dental examinations prior to brain imaging, thus establishing temporality. Third, inverse probability weights in all our analytical models allowed us to account for attrition and/or censoring. Finally, the addition of an edentulous category, as a possible reflection of historical periodontal disease, informs the potential value of improved oral hygiene throughout the life course.
Important limitations of our study should be noted. First, our underpowered analyses for MRI outcomes and βA positivity due to the design of brain imaging in ARIC-NCS limits precision of our findings. Second, selection bias might have biased our results to the null. While possible, it is important to note that the low sample size included imaging studies arose largely through intentional design by the investigators.18,26 Additionally, a known characteristic of the odds ratio is that it remains unaffected by selection bias even if it occurs differentially by either exposure or outcome; therefore, substantial differential loss-to-follow-up would have needed to occur synergistically by both exposure and disease status to produce large bias. Third, given the lack of repeated dental exams between ARIC Visits 4 and 5, as well as the absence of an updated dental exam concurrent with neuroimaging, we were unable to capture longitudinal changes in PD status across follow-up. Moreover, while PD status provides insight into current and/or historical exposure, it is a surrogate for the underlying pathogens hypothesized as etiologic agents. Assessing the presence and abundance of microbial organisms could be more helpful in identifying specific microbial signatures of oral dysbiosis that are related to neurodegeneration and associated cerebrovascular outcomes. Fourth, the PPC construct does not completely agree with other definitions such as the Centers for Disease Control/American Academy of Periodontology classification (CDC/AAP).44 The lack of a dose-response association between PPC levels and imaging outcomes in our study may in part be due to the elaborate nature of PPC, potentially causing information bias in our findings. The inconsistency between the PPC and CDC/AAP definitions is shown in Supplemental Table 6, and also described in a prior ARIC study.11 Fifth, lack of information on reasons for tooth loss precludes inference about whether tooth loss reflects chronic PD inflammation, poor oral hygiene, or differential access to care. Sixth, although inverse probability weights incorporated to reduce bias due to extensive loss-to-follow-up, it appears they did not have much impact on our findings. Seventh, sampling weight for PET attendance was not available, restricting us to use the MRI attendance sampling weight when employing inverse probability weights in the PET cohort. Finally, unmeasured confounders and competing risks may not have been accounted for, thus potentially biasing our results.
5. CONCLUSION
Within the limitations of this study, periodontal disease was not associated with altered brain structure, nor does it predict microhemorrhages or βA positivity. Counter to our hypotheses, advanced and complete tooth loss was linked to a reduced risk of ischemic stroke, while severe tooth loss, edentulism, and mild periodontal disease were protective against subcortical and lacunar infarct subtypes. In light of several previous investigation, including in ARIC linking PD to cognitive decline and dementia,11 future studies which integrate repeated longitudinal oral microbiome assessment and brain imaging are necessary to advance our understanding of the temporality and underlying biological mechanisms which link periodontal status and neurocognitive outcomes.
Supplementary Material
Clinical Relevance.
Study Rationale:
Subgingival microbial dysbiosis, common in periodontal disease, is posited to be linked with mild cognitive impairment.
Principal Findings:
We found no clear association between periodontal disease and most brain imaging outcomes. An inverse association was observed between periodontal status and cerebral infarcts; this could be either a chance finding, due to extensive loss-to-follow-up, selective sampling for brain imaging, and/or selection bias, or the result of reduced infectious burden secondary to tooth loss.
Practical Implications:
Future research should incorporate direct measurement of the subgingival microbiome as well as longitudinal microbiome, periodontal and brain imaging measurements.
ACKNOWLEDGEMENTS
The authors thank the staff and participants of the ARIC study for their important contributions.
Study Funding:
The Atherosclerosis Risk in Communities study has been funded in whole or in part with Federal funds from the National Heart, Lung, and Blood Institute, National Institutes of Health, Department of Health and Human Services, under Contract nos. (HHSN268201700001I, HHSN268201700002I, HHSN268201700003I, HHSN268201700005I, HHSN268201700004I).
Neurocognitive data were collected by U01 2U01HL096812, 2U01HL096814, 2U01HL096899, 2U01HL096902, 2U01HL096917 from the NIH (NHLBI, NINDS, NIA and NIDCD), and with previous brain MRI examinations funded by R01-HL70825 from the NHLBI.
The ARIC-PET study was funded by the National Institute on Aging (R01AG040282). The ARIC Dental Study was funded by NIH/NIDCR R01-DE021418, and R01-DE021986, and NIH/NCRR UL1-TR001111.
This study was also supported by the Atherosclerosis Risk in Communities Study Pre-Doctoral Diversity Supplement NIH NHLBI-CON000000080742 and Pre-Doctoral T32 Cardiovascular Disease Training Supplement (PHS Grant number: 5 T32 HL 7779-27).
Conflicts of Interest Statement
H. Adam, K. Lakshminarayan, F. Norby, T. Mosley, K. Walker, R. Gottesman, W. Wang report none. K. Meyer has a research grant from Balchem Corporation. T. Hughes, J. Pankow report none. D. Wong reports non-monetary help from AVID/Lilly on NIH grant study including Tau comparisons, some with AV-45 (Florbetapir) anticipated for screening. C. Jack Jr serves on an independent data monitoring board for Roche, has consulted for and served as a speaker for Eisai, and consulted for Biogen, but he receives no personal compensation from any commercial entity. He receives research support from NIH and the Alexander Family Alzheimer’s Disease Research Professorship of the Mayo Clinic. J. Beck reports none. The PPC-Stages and computational algorithms, as well as the application to Stages and Grades is protected under Copyright 2017–2018 with the University of North Carolina. P. Lutsey, R. Demmer report none.
Appendix 1.
Authors
| Name | Location | Role | Contribution |
|---|---|---|---|
| Hamdi Adam | University of Minnesota, Minneapolis | Author | Data analysis, drafting, and revision of manuscript; |
| Kamakshi Lakshminarayan | University of Minnesota, Minneapolis | Author | Drafting and revision of the manuscript; |
| Faye Norby | University of Minnesota, Minneapolis | Author | Data analysis; revision of the manuscript; |
| Thomas Mosley | University of Mississippi, Jackson | Author | Conception and design of the study; drafting and revision of the manuscript; |
| Keenan Walker | Johns Hopkins University, Baltimore | Author | Drafting and revision of the manuscript; |
| Rebecca Gottesman | Johns Hopkins University, Baltimore | Author | Drafting and revision of the manuscript; acquired funding; |
| Wendy Wang | University of Minnesota, Minneapolis | Author | Data analysis; revision of the manuscript; |
| Katie Meyer | University of North Carolina, Chapel Hill | Author | Drafting and revision of the manuscript; |
| Timothy Hughes | Wake Forest University, Winston-Salem | Author | Drafting and revision of the manuscript; |
| Jim Pankow | University of Minnesota, Minneapolis | Author | Drafting and revision of the manuscript; |
| Dean Wong | Washington University in St. Louis, St. Louis | Author | Drafting and revision of the manuscript; |
| Clifford R. Jack Jr. | Mayo Clinic, Rochester | Author | Drafting and revision of the manuscript; |
| Jim Beck | University of North Carolina, Chapel Hill | Author | Conception and design of the study; drafting and revision of the manuscript; |
| Pamela Lutsey | University of Minnesota, Minneapolis | Author | Conception and design of the study; drafting and revision of the manuscript; |
| Souvik Sen | University of South Carolina | Author | Drafting and revision of the manuscript; |
| Ryan Demmer | University of Minnesota, Minneapolis | Author | Conceptualized manuscript; supervised data analysis; drafting and revision of the manuscript; |
References
- 1.Noble JM, Scarmeas N, Papapanou PN. Poor oral health as a chronic, potentially modifiable dementia risk factor: review of the literature. Curr Neurol Neurosci Rep. 2013;13(10):384. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Daly B, Thompsell A, Sharpling J, et al. Evidence summary: the relationship between oral health and dementia. Br Dent J. 2018;223(11):846–853. [DOI] [PubMed] [Google Scholar]
- 3.Teixeira FB, Saito MT, Matheus FC, et al. Periodontitis and Alzheimer’s Disease: A Possible Comorbidity between Oral Chronic Inflammatory Condition and Neuroinflammation. Frontiers in Aging Neuroscience. 2017;9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Watts A, Gatz M, Crimmins EM. Inflammation as a potential mediator for the association between periodontal disease and Alzheimer’s disease. Neuropsychiatric Disease and Treatment. 2008:865. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lee YL, Hu HY, Huang LY, Chou P, Chu D. Periodontal Disease Associated with Higher Risk of Dementia: Population-Based Cohort Study in Taiwan. J Am Geriatr Soc. 2017;65(9):1975–1980. [DOI] [PubMed] [Google Scholar]
- 6.Holmer J, Eriksdotter M, Schultzberg M, Pussinen PJ, Buhlin K. Association between periodontitis and risk of Alzheimer’s disease, mild cognitive impairment and subjective cognitive decline: A case-control study. J Clin Periodontol. 2018;45(11):1287–1298. [DOI] [PubMed] [Google Scholar]
- 7.Kaye EK, Valencia A, Baba N, Spiro A 3rd, Dietrich T, Garcia RI. Tooth loss and periodontal disease predict poor cognitive function in older men. J Am Geriatr Soc. 2010;58(4):713–718. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Shin HS, Shin MS, Ahn YB, Choi BY, Nam JH, Kim HD. Periodontitis Is Associated with Cognitive Impairment in Elderly Koreans: Results from the Yangpyeong Cohort Study. J Am Geriatr Soc. 2016;64(1):162–167. [DOI] [PubMed] [Google Scholar]
- 9.Sung CE, Huang RY, Cheng WC, Kao TW, Chen WL. Association between periodontitis and cognitive impairment: Analysis of national health and nutrition examination survey (NHANES) III. J Clin Periodontol. 2019;46(8):790–798. [DOI] [PubMed] [Google Scholar]
- 10.Naorungroj S, Schoenbach VJ, Beck J, et al. Cross-sectional associations of oral health measures with cognitive function in late middle-aged adults: a community-based study. J Am Dent Assoc. 2013;144(12):1362–1371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Demmer RT, Norby FL, Lakshminarayan K, et al. Periodontal Disease and Incident Dementia: The Atherosclerosis Risk in Communities Study (ARIC). Neurology. 2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Kamer AR, Pirraglia E, Tsui W, et al. Periodontal disease associates with higher brain amyloid load in normal elderly. Neurobiol Aging. 2015;36(2):627–633. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Dintica CS, Rizzuto D, Marseglia A, et al. Tooth loss is associated with accelerated cognitive decline and volumetric brain differences: a population-based study. Neurobiol Aging. 2018;67:23–30. [DOI] [PubMed] [Google Scholar]
- 14.Taguchi A, Miki M, Muto A, et al. Association between oral health and the risk of lacunar infarction in Japanese adults. Gerontology. 2013;59(6):499–506. [DOI] [PubMed] [Google Scholar]
- 15.Norby FL, Chen LY, Soliman EZ, Gottesman RF, Mosley TH, Alonso A. Association of left ventricular hypertrophy with cognitive decline and dementia risk over 20 years: The Atherosclerosis Risk In Communities–Neurocognitive Study (ARIC-NCS). American Heart Journal. 2018;204:58–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Atherosclerosis Risk in Communities Study Description. University of North Carolina at Chapel Hill. https://sites.cscc.unc.edu/aric/description. Published 2020. Accessed 02/25, 2020.
- 17.Cohort Characteristics-Table 1. Characteristics of ARIC Cohort at Baseline by Sex or Race. University of North Carolina at Chapel Hill. https://sites.cscc.unc.edu/aric/system/files/CohortCharacteristics.pdf. Published 2020. Accessed 02/25, 2020. [Google Scholar]
- 18.Knopman DS, Gottesman RF, Sharrett AR, et al. Mild cognitive impairment and dementia prevalence: The Atherosclerosis Risk in Communities Neurocognitive Study. Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring. 2016;2(1):1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Beck JD, Elter JR, Heiss G, Couper D, Mauriello SM, Offenbacher S. Relationship of Periodontal Disease to Carotid Artery Intima-Media Wall Thickness. Arteriosclerosis, Thrombosis, and Vascular Biology. 2001;21(11):1816–1822. [DOI] [PubMed] [Google Scholar]
- 20.Sen S, Giamberardino LD, Moss K, et al. Periodontal Disease, Regular Dental Care Use, and Incident Ischemic Stroke. Stroke. 2018;49(2):355–362. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Beck JD, Moss KL, Morelli T, Offenbacher S. Periodontal profile class is associated with prevalent diabetes, coronary heart disease, stroke, and systemic markers of C-reactive protein and interleukin-6. J Periodontol. 2018;89(2):157–165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Löe H, Silness J. Periodontal Disease in Pregnancy I. Prevalence and Severity. Acta Odontologica Scandinavica. 1963;21(6):533–551. [DOI] [PubMed] [Google Scholar]
- 23.Silness J, Löe H. Periodontal Disease in Pregnancy II. Correlation Between Oral Hygiene and Periodontal Condition. Acta Odontologica Scandinavica. 1964;22(1):121–135. [DOI] [PubMed] [Google Scholar]
- 24.Beck JD, Moss KL, Morelli T, Offenbacher S. In search of appropriate measures of periodontal status: The Periodontal Profile Phenotype (P(3)) system. J Periodontol. 2018;89(2):166–175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Morelli T, Moss KL, Beck J, et al. Derivation and Validation of the Periodontal and Tooth Profile Classification System for Patient Stratification. Journal of Periodontology. 2017;88(2):153–165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Gottesman RF, Schneider ALC, Zhou Y, et al. The ARIC-PET amyloid imaging study. Neurology. 2016;87(5):473–480. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Knopman DS, Griswold ME, Lirette ST, et al. Vascular Imaging Abnormalities and Cognition. Stroke. 2015;46(2):433–440. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Lutsey PL, Norby FL, Gottesman RF, et al. Sleep Apnea, Sleep Duration and Brain MRI Markers of Cerebral Vascular Disease and Alzheimer’s Disease: The Atherosclerosis Risk in Communities Study (ARIC). PLOS ONE. 2016;11(7):e0158758. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Wu A, Sharrett AR, Gottesman RF, et al. Association of Brain Magnetic Resonance Imaging Signs With Cognitive Outcomes in Persons With Nonimpaired Cognition and Mild Cognitive Impairment. JAMA Network Open. 2019;2(5):e193359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Schneider ALC, Selvin E, Sharrett AR, et al. Diabetes, Prediabetes, and Brain Volumes and Subclinical Cerebrovascular Disease on MRI: The Atherosclerosis Risk in Communities Neurocognitive Study (ARIC-NCS). Diabetes Care. 2017;40(11):1514–1521. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Palta P, Sharrett AR, Wei J, et al. Central Arterial Stiffness Is Associated With Structural Brain Damage and Poorer Cognitive Performance: The ARIC Study. Journal of the American Heart Association. 2019;8(2). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Gottesman RF, Schneider ALC, Zhou Y, et al. Association Between Midlife Vascular Risk Factors and Estimated Brain Amyloid Deposition. JAMA. 2017;317(14):1443. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Beck JD, Eke P, Heiss G, et al. Periodontal Disease and Coronary Heart Disease. Circulation. 2005;112(1):19–24. [DOI] [PubMed] [Google Scholar]
- 34.Rosamond WD, Folsom AR, Chambless LE, et al. Stroke Incidence and Survival Among Middle-Aged Adults : 9-Year Follow-Up of the Atherosclerosis Risk in Communities (ARIC) Cohort. 1999;30(4):736–743. [DOI] [PubMed] [Google Scholar]
- 35.Weuve J, Tchetgen Tchetgen EJ, Glymour MM, et al. Accounting for bias due to selective attrition: the example of smoking and cognitive decline. Epidemiology. 2012;23(1):119–128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Gottesman RF, Rawlings AM, Sharrett AR, et al. Impact of Differential Attrition on the Association of Education With Cognitive Change Over 20 Years of Follow-up: The ARIC Neurocognitive Study. American Journal of Epidemiology. 2014;179(8):956–966. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Kamer AR, Craig RG, Dasanayake AP, Brys M, Glodzik-Sobanska L, De Leon MJ. Inflammation and Alzheimer’s disease: Possible role of periodontal diseases. Alzheimer’s & Dementia. 2008;4(4):242–250. [DOI] [PubMed] [Google Scholar]
- 38.Joshipura KJ, Hung H-C, Rimm EB, Willett WC, Ascherio A. Periodontal Disease, Tooth Loss, and Incidence of Ischemic Stroke. Stroke. 2003;34(1):47–52. [DOI] [PubMed] [Google Scholar]
- 39.Wu T, Trevisan M, Genco RJ, Dorn JP, Falkner KL, Sempos CT. Periodontal Disease and Risk of Cerebrovascular Disease. Archives of Internal Medicine. 2000;160(18):2749. [DOI] [PubMed] [Google Scholar]
- 40.Grau AJ, Becher H, Ziegler CM, et al. Periodontal Disease as a Risk Factor for Ischemic Stroke. Stroke. 2004;35(2):496–501. [DOI] [PubMed] [Google Scholar]
- 41.Lee H-J, Garcia RI, Janket S-J, et al. The Association Between Cumulative Periodontal Disease and Stroke History in Older Adults. Journal of Periodontology. 2006;77(10):1744–1754. [DOI] [PubMed] [Google Scholar]
- 42.Kim H-D, Sim S-J, Moon J-Y, Hong Y-C, Han D-H. Association Between Periodontitis and Hemorrhagic Stroke Among Koreans: A Case-Control Study. Journal of Periodontology. 2010;81(5):658–665. [DOI] [PubMed] [Google Scholar]
- 43.Dominy SS, Lynch C, Ermini F, et al. Porphyromonas gingivalisin Alzheimer’s disease brains: Evidence for disease causation and treatment with small-molecule inhibitors. Science Advances. 2019;5(1):eaau3333. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Eke PI, Page RC, Wei L, Thornton-Evans G, Genco RJ. Update of the case definitions for population-based surveillance of periodontitis. J Periodontol. 2012;83(12):1449–1454. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
ARIC data are available, with appropriate permissions, via the NIH NHLBI-sponsored Biologic Specimen and Data Repository Information Coordinating Center (BioLINCC) https://biolincc.nhlbi.nih.gov/.