Abstract
Context:
Periodontitis is an inflammatory disease of tooth-supporting tissue leading to bone destruction and tooth loss. Periodontitis affects almost 50% of adults greater than 30 years of age.
Objective:
This study evaluated the association between biomarkers linked to bone formation and resorption with the occurrence and progression of periodontal disease in older men (≥65 y).
Design:
The Osteoporotic Fractures in Men (MrOS) study is a prospective, observational study among men 65 years of age and older.
Setting:
This ancillary study, Oral and Skeletal Bone Loss in Older Men, was conducted at two of the six MrOS study sites (Birmingham, AL and Portland, OR).
Patients:
Patients underwent medical and dental evaluation. Diagnoses of periodontitis were based on clinical attachment loss, pocket depth, calculus, plaque, and bleeding on a random half-mouth. Bone metabolism biomarkers included serum levels of calcium, phosphate (Pi), alkaline phosphatase, albumin, carboxy-terminal collagen crosslinks (CTX), N-terminal propeptides of type I procollagen, isoform 5b of tartrate-resistant acid phosphatase, and urine alpha- carboxy-terminal collagen crosslinks (alpha-CTX) and beta-CTX and serum levels of calciotropic hormones vitamin D (25(OH)D) and PTH.
Main Outcome Measures:
The aim of this study is to correlate bone metabolism biomarkers with prevalence and progression of periodontal disease in older men.
Results:
Patients with more severe periodontitis had significantly higher levels of PTH (P trend = .0004), whereas 25(OH)D was lower (P trend = .001). In a subset of men reevaluated at a second dental visit, improvement of periodontitis was associated with lower alpha-CTX, beta-CTX, and CTX levels at baseline after adjusting for age, site, and body mass index.
Conclusion:
This study suggests that a distinct set of biomarkers of bone metabolism are associated with more severe periodontal disease (PTH, 25(OH)D) and periodontal progression (alpha-CTX, beta-CTX, and CTX) over time.
Periodontitis is an inflammatory disease of tooth-supporting tissue that affects approximately 30% of the adult population in the United States (1). More recently (2012), this estimate was increased to greater than 47% among adults age 30 years or older, with those over 65 years old accounting for most cases (2). Periodontitis is characterized by progressive loss of bone and periodontal attachment that ultimately leads to tooth loss if left untreated (3).
Historically, periodontal disease had been considered primarily an infectious disease caused by bacteria in the dental plaque. More recently, the individual host response has been shown to be crucial in the development and progression of the disease and thus focus has been placed on identifying the determinants of the local host response to bacteria and bacterial products such as lipopolysaccarides (4, 5). In fact, various chronic inflammatory disease states such as diabetes and obesity may modulate the local host response to periodontal pathogens and their lipopolysaccarides, resulting in specific forms and patterns of periodontal disease (6–8). Periodontal disease is not a static process, but rather includes periods of stability, remission, or progression (9) and a number of patient-related factors may influence these periods (10–16).
Given that one of the clinical features of periodontal disease is tissue destruction including bone that ultimately results in tooth loss, interest has been focused on whether bone metabolism and specifically bone remodeling are altered during development and progression of periodontal disease. One recent small population study in elderly Japanese people revealed an association between markers of bone metabolism and the occurrence of periodontal disease (17). It showed that the percentage of sites with ≥6-mm clinical attachment loss, a measure for the severity of periodontal disease, had a significant negative association with serum pro-osteoblastic bone remodeling marker osteocalcin that remained after adjusting for demographic variables. This was corroborated in other small population studies that showed that urinary and serum biomarkers could be used to help evaluate various aspects of periodontal disease (18, 19). Although a previous analysis of the Osteoporotic Fractures in Men Study (MrOS) did not find any significant correlation between periodontal disease and skeletal bone mineral density in older men (20), a recent study from the Buffalo OsteoPerio Study concluded that postmenopausal women with severe periodontitis or osteoporosis may exhibit a faster rate of oral bone loss over time (20, 21).
These data suggest that an individual's overall bone metabolism might be associated with periodontal disease and be reflective of its progression (17). We hypothesize that individuals who have increased bone turnover are more likely to present with periodontitis. Therefore, the aim of this study is to correlate bone metabolism biomarkers with prevalence and progression of periodontal disease in older men.
Materials and Methods
Participants
The MrOS study is a prospective, observational study to determine risk factors for osteoporosis among men 65 years and older. MrOS recruited 5994 men (March 2000 to April 2002 [Baseline]) at six clinical sites (Birmingham, AL; Minneapolis, MN; Palo Alto, CA; Pittsburgh, PA; Portland, OR; and San Diego, CA). Participants without bilateral hip replacements who were able to walk without assistance were included and informed consent was obtained (22, 23).
Between September 2002 and May 2003, participants at Portland and Birmingham partook in an ancillary study, Oral and Skeletal Bone Loss in Older Men, to evaluate dental and periodontal health. One thousand three hundred forty-seven participated in the first dental visit (Dental Visit 1), of whom 1107 had nonmissing values for calcium, 25(OH)D, PTH, Pi, alkaline phosphatase (ALP), and albumin, and were included in the cross-sectional analysis (Figure 1).
Figure 1.
Study flow diagram of individuals included in the MrOS study.
Of the 1107 patients analyzed above, 829 returned between March 2005 and May 2006 for a second dental visit (Dental Visit 2) and had complete periodontal measures at both dental visits. They were included in the longitudinal analysis. Furthermore, 141 participants, a subset of the 829, had nonmissing values for serum and urinary CTX, tartrate-resistant acid phosphatase (TRACP)5b, propeptides of type I procollagen (P1NP) and were included in a separate longitudinal analysis.
This study was approved by the Institutional Review Boards at Oregon Health & Science University and the University of Alabama.
Clinical assessment
Participants underwent clinical evaluation and provided information regarding medical and dental history (Supplemental Data).
Six calibrated examiners recorded missing teeth and implants for the full mouth (excluding third molars) and performed periodontal examination on a randomized half-mouth based on subject identification number using a UNC15 probe. clinical attachment loss, distance from the cemento-enamel-junction to the base of the pocket and pocket probing depth (PD), and distance from the free gingival margin to the pocket base were measured at six sites per tooth (disto-buccal, direct buccal, mesio-buccal, disto-lingual, direct lingual, and mesio-lingual). For plaque, calculus, and gingival bleeding, the worst score per tooth was recorded using the Silness and Loe Plaque and Gingival Index (24). Calculus was classified as none, supra-gingival, subgingival.
Periodontal disease definitions
Periodontitis case definitions were based on the Biofilm-Gingival Interface (BGI) Level classification system by Offenbacher et al (25). Five levels of disease were defined based upon shallow (≤3 mm) or deep (≥4 mm) PDs in combination with bleeding on probing (BOP) scores (<10%, 10–50%, or ≥50%). PDs greater than 3 mm are classified as periodontally at risk because they could potentially provide the anaerobic conditions that are necessary for periodontal pathogens to thrive, therefore exposing them to greater risk for periodontal tissue break down (26). BOP can be used as a predictive measure that could foretell periodontal tissue breakdown (27). Two shallow PD groups were defined: BGI-healthy (BGI-H) was defined as all PDs ≤3 mm and <10% BOP; BGI-gingivitis (BGI-G) was defined as individuals with all PDs ≤3 mm and at least 10% BOP. Three BGI deep lesion (DL) groups were created: mild periodontal disease or BGI-deep lesion/low bleeding (BGI-DL/LB) was defined as one or more sites with PD ≥4 mm and <10% BOP; moderate periodontal disease or BGI-DL/moderate bleeding (BGI-DL/MB) was defined as one or more sites with PD ≥4 mm and 10–50% BOP; and severe periodontal disease or BGI-DL/severe bleeding (BGI-DL/SB) included subjects with one or more sites with PD ≥4 mm and ≥50% BOP. This case definition was most appropriate for a geriatric population because it identifies only cases with substantial extent and severity, and offers a common clinical feature of active periodontal disease to diagnoses. However, given the small number of subjects who were categorized as BGI-H or BGI-D, we combined the two groups. Furthermore, we included an edentulous subject category in the cross-sectional analysis (Table 1).
Table 1.
Characteristics of the MrOS Cohort by Prevalence of Periodontal Disease Groups
| Characteristics | Periodontal Disease Categories |
P Trend | |||||
|---|---|---|---|---|---|---|---|
| Overall | BGI-H BGI-G | BGI-DL/LB | BGI-DL/MB | BGI-DL/SB | Edentuloush | ||
| N | 1107 | 152 | 428 | 203 | 203 | 121 | |
| Age, y, Mean ± sd | 72.6 ± 5.5 | 71.8 ± 5.4 | 72.1 ± 5.2 | 73.1 ± 5.9 | 73.7 ± 5.4 | 73.0 ± 6.1 | .002 |
| BMI, kg/m,b mean ± sd | 27.3 ± 3.7 | 27.1 ± 3.9 | 27.2 ± 3.7 | 27.1 ± 3.2 | 27.2 ± 3.8 | 28.1 ± 4.0 | .09 |
| Weight, kg, mean ± sd | 89.5 ± 13.8 | 88.8 ± 13.7 | 89.4 ± 13.3 | 88.0 ± 11.9 | 89.2 ± 14.2 | 94.1 ± 16.8 | .03 |
| White, N (%) | 993 (89.7) | 139 (91.5) | 404 (94.4) | 175 (86.2) | 173 (85.2) | 102 (84.3) | .0006 |
| Smoking status, N (%) | |||||||
| Current | 41 (3.7) | 3 (2.0) | 12 (2.8) | 8 (3.9) | 7 (3.5) | 11 (9.1) | .008 |
| ≥20 packs/y, N (%) | 388 (35.1) | 40 (26.3) | 136 (31.8) | 60 (29.6) | 79 (38.9) | 73 (60.3) | <.0001 |
| Alcohol consumption/week, N (%) | |||||||
| None | 455 (41.1) | 76 (50.0) | 172 (40.2) | 67 (33.0) | 63 (31.0) | 77 (63.6) | .92 |
| Education, N (%) | <.0001 | ||||||
| College or more | 387 (35.0) | 58 (38.2) | 168 (39.3) | 83 (40.9) | 65 (32.0) | 13 (10.7) | |
| Physical activity, N (%) | |||||||
| PASE score, median (IQR) | 143 (106 192) | 150 (104, 194) | 145 (109, 190) | 147 (98, 198) | 139 (105, 180) | 147 (100, 192) | .69 |
| Walk activities ≥3 times/wk | 785 (70.9) | 107 (70.4) | 308 (72.0) | 147 (72.4) | 143 (70.4) | 80 (66.1) | .50 |
| Light sport ≥3 times/wk | 293 (26.5) | 33 (21.7) | 129 (30.1) | 52 (25.6) | 57 (28.1) | 22 (18.2) | .61 |
| Moderate sport ≥3 times/wk | 83 (7.5) | 8 (5.3) | 28 (6.5) | 27 (13.3) | 17 (8.4) | 3 (2.5) | .76 |
| Strenuous sport ≥3 times/wk | 195 (17.6) | 31 (20.4) | 84 (19.6) | 45 (22.2) | 31 (15.3) | 4 (3.3) | .0009 |
| Muscle exercise ≥3 times/wk | 344 (31.1) | 56 (36.8) | 137 (32.0) | 76 (37.4) | 51 (25.1) | 24 (19.8) | .002 |
| Comorbid disease, N (%) | |||||||
| CVD | 171 (15.5) | 22 (14.5) | 69 (16.1) | 35 (17.2) | 24 (11.8) | 21 (17.4) | .95 |
| Arthritis or gout | 561 (50.7) | 76 (50.0) | 216 (50.5) | 95 (46.8) | 102 (50.3) | 72 (59.5) | .31 |
| Osteoarthritis | 40 (3.6) | 6 (4.0) | 20 (4.7) | 4 (2.0) | 9 (4.4) | 1 (0.8) | .21 |
| Diabetes | 123 (11.1) | 15 (9.9) | 41 (9.6) | 23 (11.3) | 24 (11.8) | 20 (16.5) | .08 |
| Fracture history | 643 (58.1) | 80 (52.6) | 251 (58.6) | 112 (55.2) | 128 (63.1) | 72 (59.5) | .13 |
| Metabolic syndrome, N (%) | |||||||
| Obesity | 222 (20.1) | 27 (17.8) | 85 (19.9) | 33 (16.3) | 43 (21.2) | 34 (28.1) | .08 |
| Insulin resistance | 504 (45.5) | 69 (45.4) | 186 (43.5) | 86 (42.4) | 98 (48.3) | 65 (53.7) | .13 |
| Hypertension | 461 (41.6) | 60 (39.5) | 183 (42.8) | 76 (37.4) | 82 (40.4) | 60 (49.6) | .36 |
| Annual T injections, N (%) | 10 (0.9) | 1 (0.7) | 5 (1.2) | 1 (0.5) | 1 (0.5) | 2 (1.7) | .83 |
| Medication, N (%) | |||||||
| Benzodiazepine | 48 (4.3) | 8 (5.3) | 18 (4.2) | 6 (3.0) | 8 (3.9) | 8 (6.6) | .93 |
| Calcium channel blocker | 163 (14.7) | 19 (12.5) | 65 (15.2) | 26 (12.8) | 34 (16.8) | 19 (15.7) | .40 |
| HMG CoA reductase inhibitor (statin) | 275 (24.8) | 41 (27.0) | 116 (27.1) | 49 (24.1) | 41 (20.2) | 28 (23.1) | .11 |
| NSAID | 171 (15.5) | 32 (21.1) | 77 (18.0) | 25 (12.3) | 22 (10.8) | 15 (12.4) | .002 |
| Dental history | |||||||
| Annual dental visit, N (%) | 817 (73.8) | 120 (79.0) | 366 (85.5) | 165 (81.3) | 157 (77.3) | 9 (7.4) | <.0001 |
| Personal hygiene, N (%) | |||||||
| Daily brushing | 965 (87.2) | 147 (96.7) | 417 (97.4) | 199 (98.0) | 193 (95.1) | 9 (7.4) | .48 |
| Daily flossing | 302 (27.3) | 45 (29.6) | 147 (34.4) | 71 (35.0) | 38 (18.7) | 1 (0.8) | .03 |
| Permanent tooth pulled due to gum disease, N (%)a | 266 (24.0) | 30 (19.7) | 89 (20.8) | 53 (26.1) | 46 (22.7) | 48 (39.7) | <.0001 |
| No. pulled teeth, median (IQR) | 3 (1, 8) | 3.5 (2, 20) | 2.5 (1, 4) | 3 (1, 5) | 3 (1, 6) | 28 (12, 28) | <.0001 |
| Self-reported Gingivitis, N (%)b | 162 (14.6) | 19 (12.5) | 71 (16.6) | 25 (12.3) | 29 (14.3) | 18 (14.9) | .81 |
| Self-report Periodontitis, N (%)c | 193 (17.4) | 23 (15.1) | 87 (20.3) | 25 (12.3) | 34 (16.8) | 24 (19.8) | .67 |
| Ever referred to a periodontist, N (%)d | 188 (17.0) | 24 (15.8) | 91 (21.3) | 29 (14.3) | 34 (16.8) | 10 (8.3) | .06 |
| Surgery due to gum disease, N (%)e | 138 (12.5) | 18 (11.8) | 66 (15.4) | 22 (10.8) | 24 (11.8) | 8 (6.6) | .11 |
| Deep cleaning/root planning for treatment of gum disease, N (%)f | 169 (15.3) | 22 (14.5) | 83 (19.4) | 29 (14.3) | 29 (14.3) | 6 (5.0) | .02 |
| Missing any teeth, N (%) | 957 (86.5) | 131 (86.2) | 346 (80.8) | 172 (84.7) | 187 (92.1) | 121 (100.0) | <.0001 |
| No. missing teeth + implants, both side, median (IQR) | 4 (1, 11) | 5 (2, 14) | 2 (1,6) | 3 (1, 7) | 5 (2, 11) | 28 (28, 28) | <.0001 |
| Plaque index, Mean ± sd | 1.1 ± 0.5 | 0.9 ± 0.5 | 0.8 ± 0.4 | 1.2 ± 0.4 | 1.5 ± 0.4 | 1.5 ± 0.9 | <.0001 |
| Biomarkers | |||||||
| Serum albumin g/dL, mean ± sd | 4.27 ± 0.24 | 4.32 ± 0.24 | 4.28 ± 0.22 | 4.24 ± 0.25 | 4.23 ± 0.25 | 4.30 ± 0.26 | .02 |
| Serum calcium, mg/dL, mean ± sd | 9.3 ± 0.4 | 9.4 ± 0.4 | 9.3 ± 0.4 | 9.2 ± 0.4 | 9.3 ± 0.4 | 9.5 ± 0.4 | .14 |
| 25(OH)D, ng/mL, mean ± sd | 23.0 ± 8.0 | 23.5 ± 7.4 | 23.7 ± 7.4 | 22.8 ± 7.7 | 22.4 ± 8.9 | 21.2 ± 8.6 | .001 |
| Total intact PTH, pg/mL, mean ± sd | 31.2 ± 16.3 | 28.2 ± 14.4 | 30.6 ± 17.0 | 31.7 ± 14.3 | 32.5 ± 16.6 | 34.4 ± 17.9 | .0004 |
| Pi, mg/dL, mean ± sd | 3.2 ± 0.4 | 3.2 ± 0.4 | 3.2 ± 0.5 | 3.1 ± 0.4 | 3.1 ± 0.5 | 3.2 ± 0.4 | .24 |
| ALP, U/L, mean ± sd | 75.5 ± 22.6 | 74.1 ± 20.9 | 74.8 ± 21.1 | 73.1 ± 24.2 | 76.3 ± 21.1 | 82.0 ± 28.2 | .02 |
| Urine α CTX, μg/L, median (IQR)g | 4.2 (2.4–6.9) | 4.5 (2.1–6.8) | 4.5 (2.1–6.7) | 3.4 (2.1–5.8) | 4.4 (2.8–7.9) | 4.6 (2.9–9.1) | .21 |
| Urine β CTX, g/mL, median (IQR)g | 12.7 (7.6–19.7) | 11.7 (6.5–19.7) | 13.4 (7.6–19.6) | 10.1 (5.9–17.7) | 12.4 (8.1–22.6) | 13.1 (7.9–18.8) | .39 |
| Serum CTX, ng/mL, median (IQR)g | 0.4 (0.3–0.5) | 0.4 (0.3–0.5) | 0.4 (0.3–0.5) | 0.4 (0.3–0.5) | 0.4 (0.2–0.5) | 0.4 (0.2–0.5) | .91 |
| Serum P1NP (ng/mL), median (IQR)g | 33.2 (26.6–44.0) | 35.2 (28.6–39.9) | 32.7 (26.5–44.5) | 33.6 (23.0–37.6) | 34.8 (28.4–44.1) | 31.5 (25.2–47.2) | .65 |
| Serum TRACPb, U/L, median (IQR)g | 3.2 ± 1.0 | 3.5 ± 1.0 | 3.4 ± 0.9 | 3.0 ± 0.9 | 2.7 ± 0.7 | 3.8 ± 1.2 | .15 |
One hundred twenty-six men with missing value of extraction of permanent tooth due to gum disease.
Eighty-eight men with missing value of self-reported gingivitis.
Ninety-four men with missing value of self-reported periodontitis.
Twenty-one men with missing value of self-reported referral to periodontist.
Twenty-six men with missing value of self-reported surgery experience due to gum disease.
Eighty-five men with missing value of self-reported experience of deep cleaning or root planning for treatment of gum disease.
Nine hundred fifteen men with missing value of urine α CTX, 917 men with missing value of urine β CTX, 915 men with missing value of serum CTX, 915 men with missing value of serum P1NP; 917 men with missing TRACP.
Edentulous: Periodontal assessment was completed to a random selected half-mouth. Measurement was not available if the subjects had no teeth, implants only, or canines only on that half-mouth;
Progression/improvement of periodontal disease
Patients were defined as having their periodontal disease status progress or improve if they changed any number of Offenbacher's diagnostic categories between Dental Visits 1 and 2 to either a more severe or less severe status. Patients were defined as having their periodontal disease status remain stable if they did not change their diagnostic periodontal disease category.
Serum/urine specimens
Samples were collected after an overnight fast. 25(OH)D and PTH were taken at Dental Visit 1, whereas ALP, albumin, calcium, Pi, urine α-CTX, urine β-CTX, serum-CTX, P1NP, and TRACPb were collected at MrOS Baseline Visit. Detailed methods for serum analysis have been reported (28–30) and can be found with information regarding urine assays in Supplemental Data.
Statistical analyses
Baseline characteristics were classified as categorical or continuous variables according to their nature and distribution. Descriptive statistics (mean and SD; median and interquartile range) were obtained on all continuous variables. Number and percentages of subjects belonging to each group of categorical variables were also computed. Cochran-Armitage trend tests for categorical variables and logistic regression models for continuous variables were performed to test whether these characteristics differed by baseline periodontitis status.
Poisson regression models with a robust variance estimation were used to assess the relative risk of periodontitis progression according to levels of the baseline biomarkers.
Multivariable models were manually constructed to assess and control for confounding factors. During descriptive analyses, we determined which potential confounding factors varied according to both periodontitis progression and levels of biomarkers. These identified confounders were added into the model individually and compared with the crude model that contained age, study site, and biomarkers level. Confounders were only included in the multivariate model if the associations were statistically significant (P < .05) or there was a ≥10% change in the measure of association. All adjusted models included variables known to potentially influence periodontal status and bone metabolism markers.
Analyses were performed at Oregon Health & Science University using SAS v.9.3 (SAS Institute, Cary, NC).
Results
Participant characteristics
Demographics of the cross-sectional dental study population (n = 1107) by severity of periodontal disease are presented in Table 1. At Dental Visit 1, 13.7% of the study population was categorized as healthy/having gingivitis, 38.7% as having mild periodontitis, 18.3% as moderate periodontitis, 18.3% as severe periodontitis, and 10.9% were edentulous.
Men with worse periodontitis were, on average, older, weighed more, and were not Caucasian (P trend = .002, .03, and .0006, respectively). The mean age was 71.8 ± 5.4 years in participants with healthy/gingivitis compared with 73.7 ± 5.4 years in those with severe periodontitis. The percentage of minorities ranged from 5.6% in the mild periodontitis group to 14.8% in the severe periodontitis group. Smokers were more likely to have more severe periodontitis or were often edentulous; the percentage of subjects with 20 or more pack-years of smoking was highest in the severe periodontitis and edentulous group (P = .008, < .0001). Subjects who did not attend college and those who do not exercise were also more likely to have periodontitis (P trend <.0001, .009, and .002, respectively).
Between periodontitis categories, individuals who visited the dentist with the least frequency had more severe disease (77.3% with annual dental visits) (P trend <.001). This was also observed in relation to individuals' oral hygiene habits at home (ie, flossing). For instance, 18.7% of individuals with severe periodontitis reported daily flossing compared with 29.6% of individuals with healthy/gingivitis (P trend = .03). Overall plaque accumulation was highest in patients with severe periodontal disease (P trend <.0001). In addition, individuals who previously had teeth extracted for periodontal reasons had a higher prevalence of more severe disease (P trend <.0001).
Levels of 25(OH)D were slightly lower (P trend .001), and levels of PTH were slightly higher (P trend .0004), in those with more severe periodontitis (Table 1). There were no differences in serum levels of bone resorption markers CTX and TRACPb, urinary CTX levels, nor bone formation marker P1NP by periodontal severity (Table 1).
Relationship between transitions of periodontal disease and markers of mineral metabolism
Approximately 10.3% (n = 85) who attended both dental visits had improvement in their periodontal disease, whereas 42.6% (n = 353) progressed and 47.2% (n = 391) remained in the same. In those who progressed, 256 worsened by one severity level, 84 worsened by two severity levels, and 13 worsened by three or more levels. At baseline, there were no significant differences in calcium, 25(OH)D, PTH, Pi, ALP, nor albumin levels between categories of periodontal transition groups (Table 2). Furthermore, there were no significant differences in risk ratios for disease transition in calcium, 25(OH)D, PTH, Pi, ALP, or albumin levels between men who improved vs stayed the same or progressed vs stayed the same (Table 3).
Table 2.
Relationship between Biomarkers and Periodontal Disease Progression
| Biomarkers Available in Participants Who Attended Both Dental Visit 1 and 2; n = 829 | Periodontal Progression |
P Value |
|||
|---|---|---|---|---|---|
| Improved (n = 85) | Same (n = 391) | Progression (n = 353) | Improved VS Same | Progression VS Same | |
| Calcium, mg/dL, mean ± SD | 9.31 ± 0.37 | 9.31 ± 0.41 | 9.29 ± 0.38 | .99 | .66 |
| 25(OH)D, ng/mL, mean ± SD | 24.00 ± 8.03 | 23.08 ± 7.81 | 23.68 ± 7.64 | .34 | .28 |
| Total intact PTH, pg/mL, mean ± SD | 30.25 ± 16.26 | 30.39 ± 17.22 | 29.33 ± 12.88 | .94 | .34 |
| Whole PTH, pg/mL, mean ± SD | 19.50 ± 9.94 | 20.10 ± 12.02 | 19.58 ± 8.32 | .59 | .47 |
| Pi, mg/dL, mean ± SD | 3.21 ± 0.41 | 3.14 ± 0.45 | 3.15 ± 0.41 | .14 | .76 |
| ALP, IU/L, mean ± SD | 74.31 ± 21.83 | 73.46 ± 19.71 | 74.41 ± 22.75 | .74 | .53 |
| Albumin, g/dL, mean ± SD | 4.24 ± 0.23 | 4.28 ± 0.24 | 4.27 ± 0.22 | .17 | .86 |
| Biomarkers Available in a Subset of Participants Who Attended Both Dental Visit 1 and 2 n = 141 | Improved (n = 11) | Same (n = 69) | Progression (n = 61) | Improved VS Same | Progression VS Same |
|---|---|---|---|---|---|
| α-CTX, ug/L, mean ± sd | 2.9 ± 1.2 | 5.7 ± 4.8 | 5.9 ± 5.5 | .005 | .85 |
| β-CTX, ug/L, mean ± sd | 8.9 ± 3.6 | 17.4 ± 14.4 | 17.4 ± 14.9 | .003 | .98 |
| CTX, ng/mL, mean ± sd | 0.27 ± 0.11 | 0.38 ± 0.17 | 0.41 ± 0.14 | .02 | .35 |
| P1NP, ng/mL, mean ± SD | 27.8 ± 13.2 | 37.4 ± 15.7 | 40.2 ± 30.9 | .04 | .28 |
| TRACP5b, U/L, mean ± SD | 3.0 ± 0.9 | 3.1 ± 1.0 | 3.3 ± 0.9 | .71 | .16 |
If Bonferroni corrections were performed for the number of biomarkers evaluated, P < .0028 in the whole study cohort (n = 829), and P < .005 in the subset cohort (n = 141), were considered to be significant, to achieve an overall Type 1 error rate of 0.05.
Because the evaluated biomarkers are highly correlated and the tests shown are not independent, statistical tests were performed separately and are presented as unadjusted P values.
Table 3.
Risk Ratio and 95% Confidence Intervals for Periodontal Transition in Dentate Men per SD Difference in Metabolism Variables
| Risk Ratio in Participants Who Attended Both Dental Visit 1 and 2 (N total = 829) | Progression Versus Same Risk of Progression (n = 353/391) | Improved Versus Same Risk of improvement (n = 85/391) |
|---|---|---|
| Calcium, mg/dL, per sd = 0.4 increase | 1.02 (0.94–1.10) | 0.92 (0.77–1.11) |
| 25(OH)D, ng/mL, per sd = 7.8 increase | 1.03 (0.96–1.11) | 1.11 (0.91–1.33) |
| Total Intact PTH, pg/mL, per sd = 15.4 increase | 0.99 (0.91–1.08) | 1.04 (0.89–1.21) |
| Whole PTH, pg/mL, per sd = 10.4 increase | 0.99 (0.91–1.07) | 1.01 (0.86–1.19) |
| Pi, mg/dL, per sd = 0.4 increase | 1.04 (0.96–1.12) | 1.14 (0.96–1.35) |
| ALP, IU/L, per sd = 21.3 increase | 1.04 (0.97–1.11) | 1.01 (0.82–1.26) |
| Albumin, g/dL, per sd = 0.2 increase | 1.01 (0.93–1.09) | 0.87 (0.74–1.02) |
| Risk Ratio in a Subset of Participants Who Attended Both Dental Visit 1 and 2 (N total = 141) | Progression vs Same Risk of Progression (n = 61/69) | Improved vs Same Risk of Improvement (n = 11/69) |
|---|---|---|
| α-CTX, ug/L, per sd = 5.0 increase | 1.01 (0.87–1.18) | 0.29 (0.14–0.62) |
| β-CTX, ug/L, per sd = 14.2 increase | 1.00 (0.84–1.20) | 0.29 (0.14–0.62) |
| CTX, ng/mL, per sd = 0.2 increase | 1.06 (0.89–1.27) | 0.48 (0.27–0.86) |
| P1NP, per sd = 23.5 increase | 1.04 (0.96–1.14) | 0.35 (0.11–1.18) |
| TRACPb, per sd = 0.9 increase | 1.17 (0.96–1.43) | 0.78 (0.44–1.40) |
Fully adjusted log-binomial regression models: 1. Calcium was adjusted for age, site, arthritis or gout, fracture history, obesity, and plaque index; 2. 25(OH)D was adjusted for age, site, most weighed, arthritis or gout, fracture history, obesity, annual dental visit, daily flossing, and plaque index; 3. Total intact PTH, whole PTH were adjusted for age, site, most weighed, smoked ≥20 packs/y, CVD, arthritis or gout, fracture history, obesity, annual dental visit, and plaque index; 4. Pi was adjusted for age, site, most weighed, arthritis or gout, fracture history, and obesity; 5. ALP was adjusted for age, site, and obesity; 6. Albumin was adjusted for age, site, smoked ≥20 packs/y, walk activity ≥3 times/wk, HMG coenzyme A reductase inhibitor (stain), and plaque index; 7. α-CTX, β-CTX, and P1NP were adjusted for age and site; 8. CTX and TRACPb were adjusted for age, site, and BMI.
Relationship between transitions of periodontal disease and markers of bone metabolism
A subset of subjects who attended both dental visits (n = 141) had serum and urinary α-CTX and β-CTX available for analyses. The nonisomerized (alpha) forms are indicative of new bone resorption and the β-isomerized (beta) forms are standing for old bone degradation (31). Further, serum P1NP, TRACP5b levels were available for analyses. In this subset, 7.8% (n = 11) had improvement in their periodontal disease, whereas 43.3% (n = 61) progressed and 49.0% (n = 69) remained in the same category of periodontal disease at both Dental Visits 1 and 2.
Participants who had improved periodontitis had lower α-CTX (ug/L), β-CTX (ug/L), and total CTX (ng/mL) levels at baseline compared with those who remained the same (Table 2). For example, α-CTX baseline levels, on average, were 2.9 ± 1.2 ug/L in the improved group compared with 5.7 ± 4.8 ug/L in participants who stayed in the same category of periodontal disease (P < .05). Similar associations were observed for β-CTX (8.9 ± 3.6 ug/L vs 17.4 ± 14.4 ug/L; P < .05) and total CTX (0.27 ± 0.11 vs 0.38 ± 0.17 ng/mL; P < .05). Furthermore, bone formation marker P1NP was slightly lower at baseline in subjects who improved in their periodontal status compared with subjects that remained stable. There were no statistically significant differences in CTX, P1NP, or TRACP5b baseline levels between men who progressed or stayed in the same periodontal group (Table 2).
Urine α-CTX and β-CTX exhibited significant associations between decreased values at baseline and an increased likelihood of improving vs being stable when adjusted for age and site (Table 3). In addition, serum CTX demonstrated a significant correlation between decreased baseline values and an increased likelihood of improving vs being stable when adjusted for age, site, and body mass index (BMI).
Discussion
In this study we examined the association between measures of mineral metabolism and markers of bone metabolism and the prevalence and progression of periodontal disease in a cohort of older American men. At the initial dental visit, men with more severe periodontitis had lower levels of vitamin D and higher levels of PTH; but these measures were not associated with progression. In contrast, markers of bone metabolism were not related to the severity of periodontitis cross sectionally, but men who had improvement in periodontitis had lower levels of bone remodeling markers at baseline than those who remained stable or progressed.
The prevalence of periodontitis in this population was similar to that previously described in other large-scale epidemiologic studies. In general we detected that men with worse periodontitis were, on average, older, weighed more, were not Caucasian, and included a higher percentage of subjects with 20 or more pack-years of smoking. This corresponds to what has been documented in numerous studies (32).
Although tests for trend showed significance between periodontal disease categories, some of the measures analyzed as part of patient characteristics did not increase or decrease in relation to worsening of the periodontal condition linearly. For instance, history of smoking is slightly higher in patients with mild vs moderate periodontitis, nevertheless, it does not reach the level of the severe and edentulous patient group. This suggests that there may not be a direct relationship between each patient characteristic and disease severity, even if a specific characteristic showed a significant trend for association. Instead, it may be that other influences play a roll in these trends, as periodontitis is a complex, multifactorial disease. Still, in general, demographic information that is typically associated with negative health outcomes (ie, smoking status, low education, poor oral hygiene, etc.,) were highest in participants with severe periodontitis or edentulism.
After adjusting for potential confounders and covariates, decreased alpha and β-CTX and total CTX measurements demonstrated significant risk ratios for improvement of disease compared with men who remained stable. Serum CTX (C-telopeptide) is a measure of fragments derived from degradation of type I collagen and levels are indicative of osteoclastic activity. CTX can be found in nonisomerized (alpha) or β-isomerized (beta) forms depending on the age of bone. Typically, new bone is correlated to the alpha form, and old bone is related to the beta form. Despite our entire patient population being over the age of 65 years (average age, 72.6 ± 5.5 y), both markers were significantly different between periodontal transition groups. Nevertheless, further investigation must be performed to determine whether these values could be used in a younger population, and to what their role of all of the bone metabolism biomarkers might be in predicting progression of periodontitis in the general population. It may be that α-CTX is a better indicator in risk of progression in a younger population, whereas β-CTX is superior for an older population (31).
Our data from the cross-sectional analysis show that men with more severe periodontal disease have lower 25(OH)D and higher PTH serum levels. An association between periodontal disease occurrence and lower 25(OH)D levels has been reported previously in the literature (33–35). Men and women over the age of 50 years exhibited 0.39 and 0.26 mm more periodontal attachment loss, respectively, when serum 25(OH)D concentrations were in the lowest compared with the highest quintile of the population (34). Further, gingival inflammation as assessed by bleeding on probing was increased as well in the subjects in the lowest quintile (36). PTH (an indirect modulator of osteoclastic activity) typically varies inversely with 25(OH)D (37). Our data from Dental Visit 1 confirm this opposite relationship. Patients with more severe periodontal disease have lower levels of 25(OH)D and higher levels of PTH.
Bone formation marker P1NP gets secreted mainly from bone into circulation. However, because it reflects the amount of newly synthesized collagen, these propeptides can also come from skin, tendons, ligaments, cornea, blood vessels, fibrocartilage, and many other tissues (38). This might explain higher levels in participants who had progression in periodontitis. Given that P1NP levels were lowest in the improvement group, it also might suggest that disease improvement is associated with a decrease in bone resorption rather than an increase in bone formation. It has been shown in other studies that inflammatory disease progression results in an uncoupling between bone resorption and formation and that it favors excessive bone resorption (39).
There exists a lack of uniformity within the literature with respect to the criteria used to define a periodontitis case, and in turn, a number of definitions are available for use, each with different diagnostic criteria and thresholds. In this study, Offenbacher et al's Biofilm-Gingival Interface categories (25) were used to define severity of periodontal disease. With the age of the population, this seemed to provide the most accurate representation of active periodontal disease as it included a clinical indicator of inflammation. Although this is not the most commonly used definition of periodontitis, we believed that it provided a better model for our older population, who may have had incidental bone loss throughout life unrelated to periodontal disease, or may have had a previous case of periodontal disease that was treated and arrested. Further investigation should be performed to help validate our results using other case definitions, including the European Workshop in Periodontology, the Centers for Disease Control and Prevention, and American Academy of Periodontology, in a more diverse and larger patient population.
The study had several strengths and limitations. We were able to study a large number of community-dwelling men who were extensively characterized for overall health status and potential confounders, nevertheless, there might have been additional factors that were unidentified and not accounted for in our study cohort. Seventy five percent of the participants returned for longitudinal assessments of periodontal status although not all were available for follow-up exams because many were deceased or had missing data (20, 22, 23). Cross-sectional analysis revealed significantly lower 25(OH)D levels and higher levels of PTH in those with more severe periodontitis at the time of their first dental visit. However, 25(OH)D levels across all periodontitis groups are considered to be low. Studies that evaluated thresholds for serum 25(OH)D concentrations in relation to mineral density, lower-extremity function, dental health, fractures, and colorectal cancer describe that advantageous serum concentrations of 25(OH)D begin at 75 nmol/L (30 ng/mL) (40).
To examine periodontal progression, we analyzed calcium, vitamin D, PTH, and alkaline phosphatase separately from alpha, beta, and total CTX, which were only available in a smaller subset of participants. Moreover, CTX and some of the other biomarkers were collected at the baseline MrOS visit before the first dental examination. To what extent those levels were different during the period of the dental examinations is unknown. Ideally, additional blood/urine samples should have been taken at the dental follow-up visit, to help correlate mineral and bone metabolism biomarkers to progression of periodontal disease. Further, no information was available about whether subjects received periodontal treatment after Dental Visit 1. However, we assume that disease progression was not due to an intervention because most the subjects either remained within their disease category or progressed to a more severe one. All study subjects underwent periodontal evaluation performed by a total of six examiners that were calibrated to ensure evaluation consistency, however, there might have been still slight differences between the examiners. Despite all these limitations, significant differences were still detectable and future studies should determine further whether they are clinically significant as well.
Another potential drawback was that random half-mouth evaluations were used for periodontal examinations. This study design typically underreports disease prevalence (41, 42). However, 41% of our participants had either moderate or severe forms of periodontitis using our disease definition, which is similar to 47% of subjects over 30 years old who had periodontitis in the most recent NHANES study that used a full-mouth periodontal examination. This supports that our choice of periodontal disease definition and examination method corresponded to an accurate representation of periodontal disease prevalence on a national scale (2). Due to sample size, we combined participants who had periodontal disease transition into either the improved or progressed group, regardless of how many periodontal disease categories were changed. This might have under- or overestimated associations given that groups included subjects that transitioned between one and three categories. Improvement in periodontitis associated with lower levels of bone remodeling markers at baseline could suggest a less pronounced inflammatory response. However, our study design only allows for identifying associations and we, therefore, can only speculate about underlying biological mechanism.
Additional studies should be performed to expand on our findings and to help correlate bone metabolism biomarkers to periodontal disease occurrence and determine when periodontal disease might progress further. This information could be used to establish better clinical guidelines for the ideal treatment(s) among patients diagnosed with various forms of periodontal disease, and to help determine the ideal time to provide this treatment.
Acknowledgments
This work was supported by Osteoporotic Fractures in Men Study (National Institutes of Health): U01 AR45580, U01 AR45614, U01 AR45632, U01 AR45657, U01 AR45654 U01 AR 45583, U01 AG18197, U01 AG027810; and UL1 TR000128, R01 DE14386, R01 AR52862; U.S.-S.: K08DE018968; T.-T.D.: KL2 TR000081, K23AG040168; P.C.S.: P30 HL101272, R01 HL114813, UL1 RR024156.
Disclosure Summary: The authors have nothing to disclose.
Footnotes
- ALP
- alkaline phosphatase
- BGI
- Biofilm-Gingival Interface
- BGI-H
- BGI-healthy
- BGI-G
- BGI-gingivitis
- BMI
- body mass index
- BOP
- bleeding on probing
- CTX
- carboxy-terminal collagen crosslinks
- DL
- deep lesion
- LB
- low bleeding
- MB
- moderate bleeding
- MrOS
- Osteoporotic Fractures in Men Study
- P1NP
- propeptides of type I procollagen
- PD
- probing depth
- Pi
- phosphate
- SB
- severe bleeding
- TRACP
- tartrate-resistant acid phosphatase.
References
- 1. Oliver RC, Brown LJ, Löe H. Periodontal diseases in the United States population. J Periodontol. 1998;69:269–278. [DOI] [PubMed] [Google Scholar]
- 2. Eke PI, Dye BA, Wei L, et al. Prevalence of periodontitis in adults in the United States: 2009 and 2010. J Dent Res. 2012;91:914–920. [DOI] [PubMed] [Google Scholar]
- 3. Papapanou PN, Wennström JL, Gröndahl K. A 10-year retrospective study of periodontal disease progression. J Clin Periodontol. 1989;16:403–411. [DOI] [PubMed] [Google Scholar]
- 4. Armitage GC. Diagnosis of periodontal diseases. J Periodontol. 2003;74:1237–1247. [DOI] [PubMed] [Google Scholar]
- 5. Schenkein HA. Host responses in maintaining periodontal health and determining periodontal disease. Periodontol 2000. 2006;40:77–93. [DOI] [PubMed] [Google Scholar]
- 6. Pischon N, Heng N, Bernimoulin JP, Kleber BM, Willich SN, Pischon T. Obesity, inflammation, and periodontal disease. J Dent Res. 2007;86:400–409. [DOI] [PubMed] [Google Scholar]
- 7. Tsai C, Hayes C, Taylor GW. Glycemic control of type 2 diabetes and severe periodontal disease in the US adult population. Community Dent Oral Epidemiol. 2002;30:182–192. [DOI] [PubMed] [Google Scholar]
- 8. Graves DT, Li J, Cochran DL. Inflammation and uncoupling as mechanisms of periodontal bone loss. J Dent Res. 2011;90:143–153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Faddy MJ, Cullinan MP, Palmer JE, Westerman B, Seymour GJ. Ante-dependence modeling in a longitudinal study of periodontal disease: The effect of age, gender, and smoking status. J Periodontol. 2000;71:454–459. [DOI] [PubMed] [Google Scholar]
- 10. Baelum V, Luan WM, Chen X, Fejerskov O. A 10-year study of the progression of destructive periodontal disease in adult and elderly Chinese. J Periodontol. 1997;68:1033–1042. [DOI] [PubMed] [Google Scholar]
- 11. Beck JD, Cusmano L, Green-Helms W, Koch GG, Offenbacher S. A 5-year study of attachment loss in community-dwelling older adults: Incidence density. J Periodontal Res. 1997;32:506–515. [DOI] [PubMed] [Google Scholar]
- 12. Ismail AI, Morrison EC, Burt BA, Caffesse RG, Kavanagh MT. Natural history of periodontal disease in adults: Findings from the Tecumseh Periodontal Disease Study, 1959–87. J Dent Res. 1990;69:430–435. [DOI] [PubMed] [Google Scholar]
- 13. Lindhe J, Okamoto H, Yoneyama T, Haffajee A, Socransky SS. Longitudinal changes in periodontal disease in untreated subjects. J Clin Periodontol. 1989;16:662–670. [DOI] [PubMed] [Google Scholar]
- 14. Machtei EE, Hausmann E, Dunford R, et al. Longitudinal study of predictive factors for periodontal disease and tooth loss. J Clin Periodontol. 1999;26:374–380. [DOI] [PubMed] [Google Scholar]
- 15. Schätzle M, Land NP, Anerud A, Boysen H, Bürgin W, Löe H. The influence of margins of restorations of the periodontal tissues over 26 years. J Clin Periodontol. 2001;28:57–64. [DOI] [PubMed] [Google Scholar]
- 16. Thomson WM, Slade GD, Beck JD, Elter JR, Spencer AJ, Chalmers JM. Incidence of periodontal attachment loss over 5 years among older South Australians. J Clin Periodontol. 2004;31:119–125. [DOI] [PubMed] [Google Scholar]
- 17. Yoshihara A, Deguchi T, Hanada N, Miyazaki H. Relation of bone turnover markers to periodontal disease and jaw bone morphology in elderly Japanese subjects. Oral Dis. 2009;15:176–181. [DOI] [PubMed] [Google Scholar]
- 18. Gibert P, Tramini P, Sieso V, Piva MT. Alkaline phosphatase isozyme activity in serum from patients with chronic periodontitis. J Periodontal Res. 2003;38:362–365. [DOI] [PubMed] [Google Scholar]
- 19. Takaishi Y, Okamoto Y, Ikeo T, et al. Correlations between periodontitis and loss of mandibular bone in relation to systemic bone changes in postmenopausal Japanese women. Osteoporos Int. 2005;16:1875–1882. [DOI] [PubMed] [Google Scholar]
- 20. Phipps KR, Chan BK, Madden TE, et al. Longitudinal study of bone density and periodontal disease in men. J Dent Res. 2007;86:1110–1114. [DOI] [PubMed] [Google Scholar]
- 21. LaMonte MJ, Hovey KM, Genco RJ, Millen AE, Trevisan M, Wactawski-Wende J. Five-year changes in periodontal disease measures among postmenopausal females: The Buffalo OsteoPerio study. J Periodontol. 2013;84:572–584. [DOI] [PubMed] [Google Scholar]
- 22. Blank JB, Cawthon PM, Carrion-Petersen ML, et al. Overview of recruitment for the osteoporotic fractures in men study (MrOS). Contemp Clin Trials. 2005;26:557–568. [DOI] [PubMed] [Google Scholar]
- 23. Orwoll E, Blank JB, Barrett-Connor E, et al. Design and baseline characteristics of the osteoporotic fractures in men (MrOS) study—A large observational study of the determinants of fracture in older men. Contemp Clin Trials. 2005;26:569–585. [DOI] [PubMed] [Google Scholar]
- 24. Loe H. The Gingival Index, the Plaque Index and the Retention Index Systems. J Periodontol. 1967;38:Suppl:610–616. [DOI] [PubMed] [Google Scholar]
- 25. Offenbacher S, Barros SP, Singer RE, Moss K, Williams RC, Beck JD. Periodontal disease at the biofilm-gingival interface. J Periodontol. 2007;78:1911–1925. [DOI] [PubMed] [Google Scholar]
- 26. Machtei EE, Dunford R, Hausmann E, et al. Longitudinal study of prognostic factors in established periodontitis patients. J Clin Periodontol. 1997;24:102–109. [DOI] [PubMed] [Google Scholar]
- 27. Lang NP, Joss A, Orsanic T, Gusberti FA, Siegrist BE. Bleeding on probing. A predictor for the progression of periodontal disease? J Clin Periodontol. 1986;13:590–596. [DOI] [PubMed] [Google Scholar]
- 28. Orwoll E, Nielson CM, Marshall LM, et al. Vitamin D deficiency in older men. J Clin Endocrinol Metab. 2009;94:1214–1222. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Bauer DC, Garnero P, Harrison SL, et al. Biochemical markers of bone turnover, hip bone loss, and fracture in older men: The MrOS study. J Bone Miner Res. 2009;24:2032–2038. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Cawthon PM, Parimi N, Barrett-Connor E, et al. Serum 25-hydroxyvitamin D, parathyroid hormone, and mortality in older men. J Clin Endocrinol Metab. 2010;95:4625–4634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. de la Piedra C, Calero JA, Traba ML, Asensio MD, Argente J, Munoz MT. Urinary alpha and beta C-telopeptides of collagen I: Clinical implications in bone remodeling in patients with anorexia nervosa. Osteoporos Int. 1999;10:480–486. [DOI] [PubMed] [Google Scholar]
- 32. Genco RJ, Borgnakke WS. Risk factors for periodontal disease. Periodontol 2000. 2013;62:59–94. [DOI] [PubMed] [Google Scholar]
- 33. Bischoff-Ferrari HA, Dietrich T, Orav EJ, Dawson-Hughes B. Positive association between 25-hydroxy vitamin D levels and bone mineral density: A population-based study of younger and older adults. Am J Med. 2004;116:634–639. [DOI] [PubMed] [Google Scholar]
- 34. Dietrich T, Joshipura KJ, Dawson-Hughes B, Bischoff-Ferrari HA. Association between serum concentrations of 25-hydroxyvitamin D3 and periodontal disease in the US population. Am J Clin Nutr. 2004;80:108–113. [DOI] [PubMed] [Google Scholar]
- 35. Dietrich T, Garcia RI. Associations between periodontal disease and systemic disease: Evaluating the strength of the evidence. J Periodontol. 2005;76:2175–2184. [DOI] [PubMed] [Google Scholar]
- 36. Dietrich T, Nunn M, Dawson-Hughes B, Bischoff-Ferrari HA. Association between serum concentrations of 25-hydroxyvitamin D and gingival inflammation. Am J Clin Nutr. 2005;82:575–580. [DOI] [PubMed] [Google Scholar]
- 37. Steingrimsdottir L, Gunnarsson O, Indridason OS, Franzson L, Sigurdsson G. Relationship between serum parathyroid hormone levels, vitamin D sufficiency, and calcium intake. JAMA. 2005;294:2336–2341. [DOI] [PubMed] [Google Scholar]
- 38. Singer FR, Eyre DR. Using biochemical markers of bone turnover in clinical practice. Cleve Clin J Med. 2008;75:739–750. [DOI] [PubMed] [Google Scholar]
- 39. Hardy R, Cooper MS. Bone loss in inflammatory disorders. J Endocrinol. 2009;201:309–320. [DOI] [PubMed] [Google Scholar]
- 40. Bischoff-Ferrari HA, Giovannucci E, Willett WC, Dietrich T, Dawson-Hughes B. Estimation of optimal serum concentrations of 25-hydroxyvitamin D for multiple health outcomes. Am J Clin Nutr. 2006;84:18–28. [DOI] [PubMed] [Google Scholar]
- 41. Hunt RJ, Fann SJ. Effect of examining half the teeth in a partial periodontal recording of older adults. J Dent Res. 1991;70:1380–1385. [DOI] [PubMed] [Google Scholar]
- 42. Kingman A, Morrison E, Löe H, Smith J. Systematic errors in estimating prevalence and severity of periodontal disease. J Periodontol. 1988;59:707–713. [DOI] [PubMed] [Google Scholar]