Nutritional status of the older adult is associated with dentition status

Independent variables and covariates 

In our study, we describe dental status as the number of posterior occlusal pairs of teeth (molars and premolars, maximum=eight pairs), number of posterior teeth present, and total number of teeth. Except for full dentures, natural and prosthetic teeth were counted as present. Oral health examination consisted of visual and tactile oral and dental examination performed by a licensed dentist and administered in the MEC. In addition to the comprehensive dental examination, a prosthesis assessment and denture questionnaire was administered. Details of the protocol can be found elsewhere ((14)).

Total number and pairs of posterior teeth, as continuous and discreet variables, were examined in association with nutritional status indicators. Those participants who wore full dentures were treated as having no teeth when the variable was examined as continuous. This assumption was made because the chewing ability of these individuals is known to be similar to that of individuals with very few teeth. This would tend to err on the conservative side. To examine the data in more detail, discreet variables were created for total teeth and for posterior teeth. The discreet variable for total teeth included assigning the data of people into three categories; full dentures, fewer than 18 teeth, and 18 or more teeth. The discreet variable for pairs of posterior teeth included four categories: people with zero teeth, one to four pairs of posterior teeth, five or more pairs of posterior teeth, and full denture wearers.

Outcome variables 

The nutritional status outcomes used in our study were dietary, biochemical, and anthropometric measures. Dietary outcomes included absolute nutrient intake of vitamins C, E, A, folate, and carotenes, the HEI score, and its individual components.

Dietary intake data were collected via a 24-hour recall and analyzed for nutrient content ((13)). The dietary intake data were recoded to provide an HEI score, a measure of overall quality of an individual's diet. The HEI is composed of 10 components, each representing different aspects of a healthful diet as defined by the Dietary Guidelines for Americans, 1995 ((15)) and the United States Department of Agriculture's Food Guide Pyramid ((16)). Each HEI component is scored according to the consumed number of servings of each of the five food groups (grain, fruit, vegetable, meat, and dairy), the percentage of calories from fat and saturated fat, the consumed amount of cholesterol and sodium, and a measure of dietary variety. The total score has a maximum of 100 and is the sum of these 10 dietary components, weighted equally [12], [17]. It is important to note that a high score is given for an assigned minimum intake of the five food groups and the dietary variety component. However, a high score is given for low intakes of the fat, cholesterol, and sodium components. The HEI provides a picture of foods people are eating, the amount of variety in the diet, and compliance with specific dietary guidelines recommendations. The HEI scores were then classified as a good diet (score greater than 80), a needs improvement diet (score between 51 and 80), and a poor diet (score less than 51).

A comprehensive biochemical and nutrient profile was measured on serum, blood, and urine ((13)). In this study, serum vitamins C, E, folate, and beta carotene were selected as nutritional status indicators. The laboratory methods for measuring these nutrients are described elsewhere ((13)). Finally, the BMI, calculated as measured weight in kilograms divided by the height in meters squared, was used as the anthropometric outcome measure.

Subjects 

Of all subjects ages 50 years and over examined in the survey (n=8,654), 5,958 had data available for the current analysis. Excluded were 1,719 subjects who did not have a dental examination; 1 subject with no dietary, biochemical, or BMI data; 250 subjects who fit in the ″other” racial/ethnic category; 7 who did not have data on self-reported health condition; and 719 who did not have information on their poverty status. Of the 5,958 participants whose data were used in the analysis, 20 did not have body weight and height measures, 222 did not have dietary intake data, and 476 participants had missing serum data.

Statistical analysis 

The analyses were conducted using the SUDAAN software package (Release 5.50, 1991, Research Triangle Institute, Research Triangle Park, NC) to adjust for complex study design. We used descriptive statistics to obtain weighted mean and frequency distributions for sociodemographic and health status variables and t tests and χ2 to examine differences in selected characteristics among study participants and those of people excluded from the study for missing data. Linear regression analysis modeling was performed to examine associations between the dental variables used as discreet and continuous and the nutritional status indicator outcomes, adjusting for confounding variables. Least square adjusted means of HEI score, nutrients, and BMI were obtained within categories of dental status, and tests for trend were performed by fitting the variables in their continuous form in the regression models.

The potentially confounding variables, age, sex, race/ethnicity, poverty income ratio (PIR), education, smoking status, and self-reported health condition, were controlled for in the regression models. Vitamin/mineral use was also included as a potential confounder in the association between dental health and serum nutrient levels. Total calorie intake was entered in the model when the outcome variable was nutrient intake or BMI. Race/ethnicity was self-reported, and NHANES III was designed to provide reliable estimates for three race/ethnic groups: non-Hispanic whites, non-Hispanic blacks, and Mexican Americans. All ″other” race/ethnicities were excluded from this study. Attained levels of education were grouped into three categories: less than high school education, high school education, and some college/graduate school. Smoking was defined as current, former, and never. PIR was also included in the model to control for socioeconomic status. The PIR was computed by NHANES III as a ratio of two components. The numerator was the midpoint of the observed family income category in the Family Questionnaire, and the denominator was the poverty threshold, the age of the family reference person, and the calendar year in which the family was interviewed ((13)). Poverty threshold values are based on calendar years and adjusted for changes caused by inflation between calendar years. Vitamin/mineral use in the past month was recorded as a yes/no answer. The survey participants were asked if their health, in general, was excellent, very good, good, fair, or poor. This variable was treated as discreet, with good health encompassing the first three responses, and poor encompassing the last two responses.

The healthy eating index 

Table 2 shows the adjusted least square means of the HEI score and its individual components by the categories for posterior pairs of teeth.Wearing dentures or having fewer than five pairs of posterior teeth was significantly and inversely associated with the HEI score even after adjusting for the confounders of age, gender, ethnicity, smoking, education, self-reported health condition, and calorie intake. These results seem to be primarily attributable to significantly lower fruit intake by participants with fewer posterior pairs of teeth and to a lesser extent to the lower score for the cholesterol and variety components. These results were consistently observed when the other categorical and continuous dental variables were used, in turn, in the regression model. In addition, the HEI score for the components vegetable and variety were significantly lower among participants with no posterior pairs of teeth as compared with the referent group (five or more pairs). Although not statistically significant, there seems to be a trend toward lower scores for fat, saturated fat, and cholesterol among participants with fewer pairs of posterior teeth. This lower score implies a higher intake of fat, saturated fat, and cholesterol.

Table 2. Associations between number of posterior teeth categories and the scores for the HEI and componentsa and P for trend for the model (N=5,736),b,c NHANES III, 1988-1994

	HEI	Posterior pairs of teeth	N	LSM	SE	P valued	P for trende
	Total score
		0	352	64.3	1.4	.015	.048
		Full denture	1,345	66.5	0.4	.004
		1 to 4	1,084	66.5	0.7	.030
		5 to 8f	2,955	68.2	0.3
	Grain
		0	352	6.4	0.2	.460	.544
		Full denture	1,345	6.6	0.1	.610
		1 to 4	1,084	6.6	0.1	.373
		5 to 8	2,955	6.5	0.1
	Fruit
		0	352	4.4	0.3	.018	.005
		Full denture	1,345	4.8	0.1	.018
		1 to 4	1,084	4.3	0.2	.000
		5 to 8	2,955	5.2	0.1
	Vegetable
		0	352	5.9	0.3	.012	.137
		Full denture	1,345	6.6	0.1	.265
		1 to 4	1,084	6.6	0.2	.331
		5 to 8	2,955	6.7	0.1
	Dairy
		0	352	6.4	0.2	.547	.631
		Full denture	1,345	6.2	0.1	.547
		1 to 4	1,084	6.2	0.1	.905
		5 to 8	2,955	6.2	0.1
	Meat
		0	352	6.9	0.2	.645	.058
		Full denture	1,345	6.7	0.1	.824
		1 to 4	1,084	7.0	0.1	.070
		5 to 8	2,955	6.8	0.1
	Saturated fatty acid
		0	352	6.4	0.4	.360	.106
		Full denture	1,345	6.5	0.1	.220
		1 to 4	1,084	6.5	0.2	.154
		5 to 8	2,955	6.7	0.1
	Fat
		0	352	6.3	0.4	.416	.546
		Full denture	1,345	6.4	0.2	.175
		1 to 4	1,084	6.7	0.2	.836
		5 to 8	2,955	6.7	0.1
	Cholesterol
		0	352	7.8	0.2	.198	.027
		Full denture	1,345	8.2	0.1	.855
		1 to 4	1,084	7.8	0.2	.045
		5 to 8	2,955	8.2	0.1
	Sodium
		0	352	6.7	0.2	.214	.365
		Full denture	1,345	6.7	0.1	.025
		1 to 4	1,084	6.8	0.2	.138
		5 to 8	2,955	7.0	0.1
	Variety
		0	352	7.1	0.3	.003	.194
		Full denture	1,345	7.8	0.1	.054
		1 to 4	1,084	8.0	0.1	.404
		5 to 8	2,955	8.1	0.1

HEI=Healthy Eating Index. LSM=Least square mean. SE=Standard error. aThe maximum HEI score is 100 and is the sum of 10 dietary components, weighted equally. Each HEI component has a score of 0 to 10, and its score is assigned proportionately to the needs of the individual for that component ((12)). bAdjusted LSM, SE, and P values. cAdjusted for age, sex, ethnicity, education, smoking, poverty income ratio, self-reported health status, and calorie intake. dComparison to referent category. eTrend for continuous variable. fReferent category (presence of 5 to 8 pairs of posterior teeth).

When the HEI score was examined by its three criteria, good, needs improvement, and poor, by pairs of posterior teeth, the percentage of participants consuming diets in the poor category was 26.3% (zero posterior pairs of teeth) vs 8.8% (five to eight pairs of teeth) and the percentage of participants consuming diets in the good category was 9.7% vs 24.2%. The percent of participants with dentures or one to four pairs of teeth and who had HEI scores in the good and poor category fit in between (13.3% and 15.4% good and 12.9% and 14.3% poor, respectively).

Nutrient intakes 

Mean dietary intakes of carotenes, folacin, and ascorbic acid of participants with one to four pairs of posterior teeth were significantly lower than the intakes of those with five to eight pairs of teeth or of those with dentures (Table 3).Edentulous participants had lower intakes of these dietary nutrients, although the results did not reach statistical significance. Also, vitamin A intake of edentulous participants and of denture wearers was significantly lower than among the referent population. No significant differences were seen in vitamin E intakes among the groups.

Table 3. Associations between number of posterior teeth categories and dietary nutrient intakes and P for trend for the model (N=5,736),a,b NHANES III, 1988-1994

	Dietary nutrients	Posterior pairs of teeth	N	LSM	SE	P valuec	P for trendd
	Vitamin A (RE)
		0	352	1,068	83	.046	.466
		Full denture	1,345	1,052	49	.005
		1 to 4	1,084	1,083	70	.052
		5 to 8e	2,955	1,259	47
	Vitamin B12 (mg)
		0	352	4.7	0.4	.467	.896
		Full denture	1,345	4.5	0.3	.173
		1 to 4	1,084	4.8	0.3	.424
		5 to 8	2,955	5.1	0.2
	Vitamin E (mg)
		0	352	9.1	1.1	.892	.918
		Full denture	1,345	9.0	0.5	.712
		1 to 4	1,084	9.0	0.4	.734
		5 to 8	2,955	9.2	0.2
	Ascorbic acid (mg)
		0	352	102.8	6.0	.242	.143
		Full denture	1,345	104.1	4.2	.158
		1 to 4	1,084	98.9	3.6	.006
		5 to 8	2,955	110.3	2.7
	Folacin (mg)
		0	352	286.0	22.8	.413	.416
		Full denture	1,345	286.5	8.5	.086
		1 to 4	1,084	276.9	9.3	.006
		5 to 8	2,955	305.0	6.2
	Carotenes (RE)
		0	352	504.8	55.0	.057	.000
		Full denture	1,345	562.7	35.2	.168
		1 to 4	1,084	531.1	34.0	.037
		5 to 8	2,955	616.2	22.4

NHANES III=Third National Health and Nutrition Survey. LSM=Least square mean. SE=Standard error. aAdjusted LSM, SE, and P values. bAdjusted for age, sex, ethnicity, education, smoking, poverty income ratio, self-reported health status, and calorie intake. cComparison to referent category. dTrend for continuous variable. eReferent category (presence of 5 to 8 pairs of posterior teeth).

Fiber intake was highest among participants with five or more pairs of posterior teeth (17.4 g) and lowest among those with no pairs of teeth (12.7 g). The average intake of fiber among participants with one to four pairs of posterior teeth or those wearing dentures was in between (15.4 g and 15.0 g, respectively). However, when the regression analysis was adjusted for potential confounders, only the group with no pairs of posterior teeth had a significantly lower intake of fiber as compared with participants with five or more pairs of posterior teeth.

Serum levels 

Serum vitamin C and beta carotene levels were significantly and positively associated with number of pairs of posterior teeth (Table 4).These associations were consistently found for beta carotene when the other dental variables were used in the model. Serum vitamin C level was also significantly lower among participants with fewer than 18 teeth compared with those with more than 18 teeth, but not when the dental variables were examined as continuous. No associations were found with vitamin E, but serum folate level was lower among denture wearers than among the referent group.

Table 4. Associations between number of posterior teeth categories and serum nutrients and P for trend for the model (N=5,482),a,b NHANES III, 1988-1994

	Serum nutrients	Posterior pairs of teeth	N	LSM	SE	P valuec	P for trendd
	Vitamin C (mmol/L)e
		0	337	42.82	2.81	.008	.336
		Full denture	1,271	45.13	1.55	.000
		1 to 4	1,019	46.98	1.45	.030
		5 to 8f	2,855	50.11	1.07
	Vitamin E (μmol/L)g
		0	337	30.84	1.49	.178	.38
		Full denture	1,271	31.73	0.54	.071
		1 to 4	1,019	32.56	0.54	.476
		5 to 8	2,855	33.03	0.47
	Beta carotene (μmol/L)h
		0	337	0.38	0.02	.000	.22
		Full denture	1,271	0.42	0.02	.000
		1 to 4	1,019	0.42	0.01	.000
		5 to 8	2,855	0.51	0.01
	Folate (nmol/L)i
		0	337	19.18	1.67	.397	.128
		Full denture	1,271	19.08	0.58	.030
		1 to 4	1,019	19.40	0.56	.083
		5 to 8	2,855	20.67	0.50

NHANES III=Third National Health and Nutrition Survey. LSM=Least square mean. SE=Standard error. aAdjusted LSM, SE, and P values. bAdjusted for age, sex, ethnicity, education, smoking, poverty income ratio, self-reported health status, and vitamin/mineral supplement use. cComparison to referent category. dTrend for continuous variable. eTo convert mmol/L vitamin C to mg/dL, multiply mmol/L by 0.0176. To convert mg/dL vitamin C to mmol/L, multiply mg/dL by 56.78. fReferent category (presence of five to eight pairs of posterior teeth). gTo convert μmol/L vitamin E to μg/dL, multiply μmol/L by 43. To convert μg/dL vitamin E to μmol/L, multiply μg/dL by 0.02322. hTo convert μmol/L beta carotene to μg/dL, multiply μmol/L by 54. To convert μg/dL beta carotene to μmol/L, multiply μg/dL by 0.01863. iTo convert nmol/L folate to ng/mL, multiply nmol/L by 0.441. To convert ng/mL folate to nmol/L, multiply ng/mL by 2.266.

Anthropometric measures 

Table 5 shows that there was a significantly higher BMI among individuals with fewer number of posterior teeth.This was also seen among participants with one to four pairs of posterior teeth as compared with the referent group (five or more pairs of teeth), although not among edentulous individuals or denture wearers. Edentulous individuals had significantly lower calorie intakes as compared with the individuals in the other dental categories (data not shown).

Table 5. Associations between number of posterior teeth categories and body mass index and P for trend for the model (N=5,938),a,b NHANES III, 1988-1994

	Posterior pairs of teeth	N	LSM	SE	P valuec	P for trend
	0	378	27.8	0.5	.222	.032
	Full denture	1,395	27.5	0.3	.193
	1 to 4	1,113	28.0	0.3	.009
	5 to 8d	3,052	27.2	0.1

NHANES III=Third National Health and Nutrition Survey. LSM=Least square mean. SE=Standard error. aAdjusted LSM, SE, and P values of body mass index. bAdjusted for age, sex, ethnicity, education, smoking, poverty income ratio, self-reported health status, and energy intake. cComparison to referent category. dReferent category (presence of 5 to 8 pairs of posterior teeth).