Whole-grain and fiber intakes and periodontitis risk in men

¹ From the Population Health Research Institute and the Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Canada (ATM); the Department of Community Dentistry, Khon Kaen University, Khon Kaen, Thailand (WP); the Department of Nutrition, Harvard School of Public Health, Boston, MA (MF); and the University of Puerto Rico Research Center, San Juan, Puerto Rico (KJJ)

² Supported by National Institutes of Health research grants CA55075, HL35464, and DE12102.

³ Address correspondence to AT Merchant, Clinical Epidemiology and Biostatistics, Population Health Research Institute, McMaster University, 237 Barton Street East, Hamilton, ON L8L 2X2, Canada. E-mail: anwar.merchant{at}post.harvard.edu.

ABSTRACT  
Background: Diabetes and hyperglycemia increase periodontitis risk, severity, and extent. Increased whole-grain and fiber intakes are associated with improved insulin sensitivity and may therefore affect periodontitis risk.

Objective: The objective was to examine the associations between whole-grain and fiber intakes and periodontitis risk.

Design: We prospectively followed 34 160 male US health professionals aged 40–75 y at the outset. We updated medical and lifestyle information biennially with questionnaires and diet every 4 y by using a validated food-frequency questionnaire. We excluded men reporting periodontitis, myocardial infarction, stroke, and hypercholesterolemia before 1986 and those with incomplete dietary data. All diabetics were excluded. Periodontitis was determined by a report of professionally diagnosed disease and validated by a diagnosis of periodontitis by a periodontist from a blinded review of radiographs.

Results: Men in the highest quintile of whole-grain intake were 23% less likely to get periodontitis than were those in the lowest quintile (multivariate RR: 0.77; 95% CI: 0.66, 0.89; P for trend < 0.001) after adjustment for age, smoking, body mass index, alcohol intake, physical activity, and total energy intake. Periodontitis was not associated with refined-grain intake (multivariate RR comparing extreme quintiles of intake: 1.04; 95% CI: 0.89, 1.23; P for trend = 0.37). Cereal fiber was inversely related to periodontitis risk (multivariate RR comparing extreme quintiles of intake: 0.85; 95% CI: 0.73, 0.99; P for trend = 0.03), but the association was not significant after adjustment for whole-grain intake.

Conclusion: Increasing whole grain in the diet without increasing total energy intake may reduce periodontitis risk.

Key Words: Whole grain • nutrition • prospective study • periodontitis

INTRODUCTION  
Higher whole-grain (1) and fiber (2) intakes are associated with greater insulin sensitivity, lower BMI, and lower risk of diabetes and metabolic syndrome (3). Lower serum glucose concentrations are maintained over longer periods with whole-grain intake than with refined-grain intake (4). Hyperglycemia is associated with greater inflammatory cytokine production (5) and infection risk (6).

The risk, severity, and extent of peridontitis have been attributed to the hyperglycemic state (7). Improved glycemic control in diabetes reduces periodontitis risk (8). Because greater whole-grain and fiber intakes have been associated with improved insulin sensitivity (9) and consequently with improved glycemic control, we prospectively examined the associations of dietary whole-grain and fiber intakes with periodontitis risk in a large cohort of men.

SUBJECTS AND METHODS  
Study population
The Health Professionals Follow-up Study (HPFS) began in 1986 when 51 529 male US health professionals aged 40–75 y agreed to participate (10). The questionnaires were mailed to them at baseline and then biennially to ascertain lifestyle and medical conditions, and validated food-frequency questionnaires (FFQs) were sent every 4 y to evaluate diet (11).

At baseline we excluded men who reported periodontitis (8 845), myocardial infarction (1 598), stroke (251), diabetes (959), or hypercholesterolemia (4 227); those who left 70 items blank on the 131-item FFQ or reported caloric intake outside the range of 800 to 4200 kcal/d, because that was deemed to be implausible (1 030); and those who died before the start of follow-up (459), which left 34 160 men in the analysis. Deaths were reported by family members or the Postal Service or ascertained through state registries or the National Death Index. Ascertainment of death was > 98% complete (12).

Responses to the questionnaires constituted written informed consent. The protocol was approved by the Institutional Review Board at the Harvard University School of Public Health.

Case ascertainment
Periodontitis was assessed every 2 y from 1986 to 1998 by asking the question, "Have you had professionally diagnosed periodontal disease with bone loss?" We previously validated the question in this cohort against periodontitis diagnosed by a periodontist from a blinded review of radiographs in a sample of dentists (positive predictive value: 76%; negative predictive value: 74%; 13) and other health professionals (positive predictive value: 83%; negative predictive value: 69%; 14).

Diet and exposure information
We first assessed diet with the use of a validated semiquantitative FFQ in 1986 and updated it every 4 y as described elsewhere (11). Whole-grain foods were classified by using the procedure described in other studies (15, 16). Whole-grain foods included brown rice, dark breads, whole-grain breakfast cereal, cooked cereal, popcorn, wheat germ, bran, and other grains. Breakfast cereals were considered whole grain if they had 25% whole-grain content by weight. Refined grains included white bread, white rice, English muffins, pancakes, waffles, cakes, sweet rolls, refined-grain breakfast cereals, muffins, biscuits, and pizza. Dietary fiber intake was calculated by multiplying the number of times a food portion was eaten by its average fiber content. The fiber content of foods was determined from data provided by the US Department of Agriculture (17) or the manufacturer. We adjusted fiber intake for total energy by using regression analysis (18). The energy-adjusted intake of fiber is interpreted as the composition of these substances in the diet independent of the amount of food eaten.

The FFQ was validated by comparing it to diet records maintained for 2-wk periods at 4 different times in the year by a subsample of men in this cohort. The adjusted correlation between intakes assessed by FFQ compared with the average intake ascertained from diet records was good: 0.86 for breakfast cereal, 0.68 for dietary fiber, 0.71 for white bread, and 0.77 for dark breads (11, 19).

The enrollment and follow-up questionnaires included information on age, smoking, cardiovascular disease, diabetes mellitus, hypercholesterolemia, supplement use, weight, and physical activity. Metabolic equivalents (METs), a measure of energy expenditure defined for each activity as multiples of time spent sitting quietly in a chair, were used to assess physical activity. Body mass index (BMI; in kg/m²) was ascertained at baseline and updated every 2 y.

Statistical analysis
We considered the average intakes of whole grain, refined grain, and fiber from all dietary assessments before the date of first report of periodontitis. For example, disease incidence from 1986 to 1988 was related to intakes measured in 1986, and disease incidence from 1990 to 1992 was related to the mean intakes from 1986 to 1990 and so on. With the use of this approach, we were able to assess long-term diet. Because periodontitis develops slowly (mean alveolar bone loss: 0.1 mm/y; 20), long-term rather than immediate intakes were more likely to be associated with disease. We stopped updating diet during follow-up if a person developed heart disease, stroke, diabetes, or hypercholesterolemia or underwent coronary artery bypass surgery, because the diet is likely to change as a result of the condition or procedure. If dietary data were missing for one questionnaire, we used the value derived from the immediately preceding questionnaire. We evaluated whole grains both as continuous variables and in quintiles.

We calculated person-time for each participant from the date of return of the questionnaire in 1986 to the date of report of periodontitis, death, or January 31, 2000, whichever came first. We excluded men who developed diabetes during follow-up. We used the Cox proportional hazard model to estimate RRs. In the multivariate Cox model, we controlled for age (in mo), smoking (never smokers, past smokers, or current smokers who smoked 1–14, 15–24, or 25 cigarettes/d), BMI (<21, 21–22.9, 23–24.9, 25–29.9, or 30), alcohol use (never, 0.1–4.9, 5.0–14.9, 15.0–29.9, or 30 g/d), physical activity (quintiles of METs), and total energy (kcal, continuous). To evaluate potential confounding by other factors, we further adjusted the analyses for intakes of total fiber, cereal fiber, fruit fiber, vegetable fiber, total fruit, total vegetables, nuts and multivitamins, number of teeth, and profession, and we also restricted the analyses to men who had 25 teeth. To further control for confounding, we calculated a propensity score by computing the probability that a person was in the highest category of whole-grain intake as compared with the lowest by entering all the measured confounders as predictors and whole-grain intake as the outcome (21). We then related whole-grain intake and the propensity score to periodontitis risk (21).

We stratified the analysis by age (65 y compared with >65 y), smoking (current compared with previous or never), BMI (<25 compared with 25), and physical activity (below median compared with above median). We computed multiplicative terms between the stratifying variables and whole-grain intake as a continuous variable and used the Wald test to assess interaction. We used the Mantel extension test to calculate tests for trend for categorical variables by using their respective median values and SAS software (version 9.1; SAS Inc, Cary, NC) for all data analyses.

RESULTS  
Men who consumed more whole grains were older, more physically active, and less likely to smoke and had lower BMI than did those who consumed less whole grains. By contrast, men who consumed more refined grains were younger than those who consumed less refined grains (Table 1). Men in the highest category of whole-grain intake ate more fruit, vegetables, and nuts than did those in the lowest category of whole-grain intake. An opposite trend for vegetable fiber and fruit fiber was observed with respect to refined-grain intake (Table 1). In 14 y of follow-up (220 491 person-years), there were 1897 new cases of periodontitis. Men who consumed more (median intake: 3.4 servings/d) whole grains were 23% less likely to get periodontitis than were those who consumed less (median intake: 0.3 servings/d) whole grains (multivariate RR: 0.77; 95% CI: 0.66, 0.89; P for trend < 0.001; Table 2). In contrast, periodontitis was not associated with refined-grain intake (multivariate RR comparing extreme quintiles of intake: 1.04; 95% CI: 0.89, 1.23; P for trend = 0.37; Table 2).

View this table:
TABLE 1. Characteristics of men by intakes of whole grain, refined grain, and cereal fiber¹

 

View this table:
TABLE 2. Intakes of whole grain and refined grain and periodontitis risk in men¹

 
There was a significant inverse linear trend of whole-grain intake and periodontitis risk. We did not observe any such relation with refined-grain intake. Periodontitis risk declined by 6% for every 1.0 g/d increase in whole-grain intake (multivariate RR: 0.94; 95% CI: 0.90, 0.97) but not with similar increasees in refined-grain intake (multivariate RR: 1.01; 95% CI: 0.98, 1.04).

Cereal fiber was inversely related to periodontitis risk (multivariate RR comparing extreme quintiles: 0.85; 95% CI: 0.73, 0.99; P for trend = 0.03; Table 3).However, when whole grain and cereal fiber were entered in the same multivariate model, whole grain remained inversely associated with periodontitis risk (multivariate RR: 0.73; 95% CI: 0.60, 0.89; P for trend = 0.001), but cereal fiber did not (multivariate RR: 1.04; 95% CI: 0.85, 1.26; P for trend = 0.79). Total fiber, fruit fiber, and vegetable fiber were not related to periodontitis risk (Table 3).

View this table:
TABLE 3. Intakes of fiber and periodontitis risk in men¹

 
The association between whole-grain intake and periodontitis risk was materially unchanged after additional adjustment for intakes of fruit, vegetables, nuts, and multivitamins; profession; and the number of teeth present (data not shown). The inverse relation between whole-grain intake and periodontitis risk remained significant when we restricted the analysis to men who had 25 teeth (multivariate RR: 0.74; 95% CI: 0.63, 0.87; P for trend < 0.001) and in the analysis comparing extreme categories after adjustment for propensity score (multivariate RR: 0.72; 95% CI: 0.59, 0.88). The relation was similar within strata of age, smoking, BMI, and physical activity.

DISCUSSION  
This is the first study to evaluate whole-grain intake in relation to periodontitis risk. Periodontitis risk decreased with higher whole-grain intake after adjustment for age, smoking, BMI, physical activity, and alcohol and energy intakes. The association was specific to whole-grain intake. There was no relation between refined-grain intake and periodontitis risk.

Whole grains contain a number of substances known to affect health, including antioxidants, fiber, lignans, and phytoestrogens (22). Whole grains slow digestion and absorption of carbohydrates from the gut, which results in lower serum concentrations of glucose that are maintained over a long period (4). Higher fasting blood glucose is associated with greater inflammation (higher blood C-reactive protein concentration) in nondiabetic persons (23), and greater inflammation (higher C-reactive protein concentration) is related to a higher risk of diabetes and insulin resistance (24). A diet rich in whole grains lowers inflammatory biomarkers (25), and fiber intake is inversely related to C-reactive protein (26). Whole-grain intake is associated with greater insulin sensitivity (9) and less risk of diabetes (27). Improved insulin sensitivity in turn is associated with less production of advanced glycation end-products, and consequently less oxidative stress, cytokine production, and possibly less periodontitis (28). It is plausible, therefore, that high intakes of whole grains would reduce the risk of periodontitis by improving insulin sensitivity.

A limitation of this study is that we evaluated periodontitis on the basis of a question about a professional diagnosis of periodontitis. However, we validated this question against a diagnosis of periodontal bone loss made from radiographs by a periodontist who was blinded to the periodontal status of the participant, and we found that our measure had good sensitivity and specificity (13, 14). Misclassification of outcome or exposure is still possible; however, they would most probably attenuate any relation but not create a spurious association (29).

Another potential limitation is that of residual confounding. To minimize this possibility, we first adjusted for a number of established risk factors in the multivariate model. Second, we additionally adjusted for dietary factors and multivitamin use. Third, we adjusted for the number of teeth present in the mouth because periodontitis is a cause of tooth loss in adults (30). For the same reason, we also restricted the analysis to men with minimal tooth loss. Fourth, because 58% of the participants in this study were dentists, we adjusted for profession (dentist compared with other health professionals) to account for possible differences due to greater awareness of the disease. These additional adjustments did not change the association between whole-grain intake and periodontitist risk. Fifth, we measured a propensity score by putting all the measured potential confounders in a model predicting whole-grain intake. We then ran a multivariate model with whole-grain intake and the propensity score to predict periodontitis risk. Whole-grain intake was significantly associated with periodontitis risk in all these analyses.

High fiber intake is associated with improved insulin sensitivity (2). Because whole grain is a source of fiber in the diet, we further evaluated these associations. There was no association between intakes of total fiber, fruit fiber, or vegetable fiber and periodontitis risk. Cereal fiber intake was inversely associated with periodontitis risk after multivariate adjustment. However, when whole grain and cereal fiber were entered in the same model, whole grain but not cereal fiber (the main source of dietary fiber from whole grain) was associated with periodontitis risk. Moreover, refined grain was not related to periodontitis risk.

Tooth loss is associated with changes in diet, particularly a decrease in fruit and vegetable intakes (31, 32). Because periodontitis is associated with tooth loss (30), it is possible that persons with periodontitis and tooth loss changed their diets as a consequence of missing teeth. To assess whether this could have an effect on our findings, we conducted the following analyses. First, we further adjusted the multivariate association for tooth loss. Second, we adjusted the analyses for fruit and vegetable intakes (because they change with tooth loss). Third, we restricted our analyses to men with minimal tooth loss (those who had lost 25 teeth). Whole-grain intake remained significantly associated with reduced periodontitis risk in all these analyses. The prospective study design minimized the possibility of recall bias, and a follow-up rate of >90% reduced selection bias due to differential loss to follow-up. The specificity and robustness of our findings support the hypothesis that whole-grain intake reduces periodontitis risk.

The findings of this study are consistent with recommendations to improve general health (33). Eating 4 servings whole grain/d (3/4 cup whole-grain cereal or one slice of whole wheat bread constitutes 1 serving) without increasing total calories may reduce periodontitis risk.

ACKNOWLEDGMENTS  
We thank the participants of the Health Professionals Follow-up Study for their cooperation and participation.

ATM was responsible for data analysis; ATM, WP, MF, and KJJ were responsible for drafting the manuscript; and KJJ was responsible for procuring funding. None of the authors had any personal or financial conflict of interest.

REFERENCES