Numerous studies have found that fewer remaining teeth,^6–10 edentulism,^9–14 poorer masticatory function^6,15 and other oral problems^7,8 are associated with decreased nutrient intake. Hildebrandt and colleagues¹⁶ grouped 602 elderly subjects according to the number of opposing tooth pairs, and assessed reported chewing difficulties and food avoidance. Subjects with fewer opposing tooth pairs were more likely to avoid stringy foods such as meat, crunchy vegetables such as carrots and celery, and breads. The authors also reported that removable dentures did not correct these problems. Using a large survey of male health professionals, Joshipura and colleagues⁹ found that a significantly smaller proportion of subjects who were edentulous or who had few natural teeth consumed apples, pears or carrots than did subjects who had intact natural dentitions. The study also found that intakes of vegetables and dietary fiber were significantly lower among edentulous people than among those who had intact natural dentitions.

Limited food choices among the edentulous or those with few teeth appear to affect nutritional status in these people. Several studies have shown that intakes and serum levels of folate and vitamins A and C are significantly lower among those who were edentulous or had fewer teeth.^6,10,11,13

Few studies have assessed relationships among food choices, nutrient intake and oral health comprehensively. We hypothesized that oral health problems compromise food selection and nutrient intake in the very old. The Diet Variety Score tool provides a convenient quantitative summary of food types consumed.¹⁷ We recently demonstrated a strong relationship between limited diet variety and inadequate nutrient intake in a cohort of community-dwelling elderly people¹⁸; however, that analysis did not assess relationships among dietary variety, nutrient intake and oral health. This article describes a study in which we evaluated the relationships among dietary variety, nutrient intake and oral health measures in a cohort of community-dwelling elderly people aged 79 years and older.

Limited food choices among the edentulous or those with few teeth appear to affect nutritional status in these people.

	SUBJECTS, METHODS AND MATERIALS

TOP ABSTRACT SUBJECTS, METHODS AND MATERIALS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Study population and sample. The study, which was part of a larger cross-sectional study of oral health in elderly people, specifically focused on coronal and root caries, prevalence of soft-tissue lesions and periodontal variables.¹⁹ All members of the study sample were members of the Iowa 65+ Rural Health Study cohort, a census of all people aged 65 years and older residing in two rural Iowa counties, originally recruited in 1981 using 1980 U.S. census data. From the original census, 3,673 (84 percent) community-dwelling senior citizens participated in the longitudinal health interview study. Researchers followed the original study cohort until 1994, at which time 1,781 surviving members of the cohort were living in the study area. Another round of assessments that included oral health and dietary assessments began in 1996 and focused on the 745 surviving members. Of these subjects, 449 agreed to have oral examinations and participate in standardized interviews conducted in their homes by study team members. In general, those who declined participation cited poor health as a reason for nonparticipation. This study was approved by the institutional review board at the University of Iowa, Iowa City.

Oral examinations. Following University of Iowa protocols for studies involving human subjects, we obtained written informed consent from each subject (or from his or her guardian). Each clinical dental examination (which did not include the taking of radiographs) was conducted by one of four trained and calibrated examiners (who included J.J.W. and J.S.H.).¹⁹ Examinations were conducted using a dental mouth mirror, a color-coded periodontal probe and a no. 23 explorer. Halogen headlights provided illumination. The examinations included counting remaining teeth, coronal carious lesions and root carious lesions, as well as evaluating the presence of oral soft-tissue lesions. Denture retention and stability were evaluated based on published criteria.²⁰

We adapted examination criteria from those used by the National Institute of Dental and Craniofacial Research (at the time, the National Institute of Dental Research) ²¹ for coronal and root caries. One of the examiners (J.S.H.) trained the others to use these criteria, and calibration sessions were held before the study and approximately halfway through the study period, with eight to 10 elderly subjects at each session. Percentage agreement among the examiners was 94.0 percent overall for coronal decayed or filled surfaces and 94.8 percent for root decayed or filled surfaces. We computed statistics for each possible pair of examiners. values ranged from 0.84 to 0.91 for coronal caries. For root caries, five of the six pairwise values were between 0.42 and 0.66, with the remaining pair having a value of 0.21. These data have been published previously.¹⁹

We entered examination data directly into a laptop computer, using data entry software developed by personnel at the University of Iowa College of Dentistry specifically for this project. From the examination data, we selected several variables for the present analysis:

– number of natural teeth;

– number of functional teeth, defined as natural teeth with an opposing natural or artificial tooth surface;

– presence of a complete natural dentition, defined as 24 or more natural teeth without use of a partial denture;

– presence of an incomplete natural dentition, defined as less than 24 natural teeth without use of a partial or complete denture;

– presence of coronal caries, defined as the number of coronal surfaces with carious lesions;

– presence of root caries, defined as the number of root surfaces with carious lesions;

– use of partial dentures, defined as use of two partial dentures;

– use of mixed partial and complete dentures, defined as use of one partial denture, one complete denture or both;

– use of complete dentures, defined as use of two complete dentures.

Nutritional assessment. In addition to the clinical examination, we posed to the subjects several subjective questions adapted from the Oral Health Impact Profile^22,23 and the Nutrition Screening Initiative²⁴ or created for this study. The topics included self-perceived food avoidance owing to oral problems, difficulty in chewing, difficulty in eating solid foods and tooth sensitivity to hot or cold. We assessed intellectual function using the Short Portable Mental Status Questionnaire.²⁵ Near the conclusion of the assessments, the examiners distributed three-day food and beverage record forms for subjects to complete. We excluded subjects (n = 35) with extenuating circumstances (such as refusal to accept questionnaire, significant confusion, feeling of overwhelming interview burden), as well as people who resided in nursing homes (n = 30). The examiners provided oral instructions for completion of the records to the subjects, and written instructions were included as part of the records. The subjects were instructed to return the records to us, and each of them received a preaddressed, postage-paid envelope for this purpose. Of the 384 dietary records distributed, 261 (68 percent) were returned.

We determined dietary variety by counting the number of food types consumed within each food group.

A registered dietitian (either T.A.M. or P.J.S.) reviewed the dietary records for completeness. Subjects returning illegible, incomplete or blank records (n = 41) were excluded from diet analysis. Diet responders (n = 220) were defined as subjects returning complete, usable records. Registered dietitians using the Minnesota Nutrient Data System, or NDS, Version 2.92 (Nutrition Coordinating Center, Division of Epidemiology, University of Minnesota, Minneapolis) analyzed diets through the University of Iowa College of Medicine’s General Clinical Research Center. The NDS has data for more than 16,000 foods and provides values for 112 components (nutrients and nonnutrients).

We compared nutrient intakes with age- and sex-specific nutrient guidelines defined by the Food and Nutrition Board of the National Academy of Sciences in the recently released Dietary Reference Intakes, or DRIs.^26,27 DRIs are a set of nutrient-based reference values designed for use in planning and evaluating diets; they include estimated average requirements, or EARs; recommended dietary allowances, or RDAs; adequate intakes, or AIs; and tolerable upper intake levels. ^26,27 EARs are the nutrient intakes thought to define adequacy for 50 percent of given age- and sex-specific populations.^26,27 EARs are recommended for estimating the adequacy of nutrient intake for healthy people. Not all nutrients have EARs; if sufficient scientific evidence does not support establishment of an EAR, then an AI is defined.^26,27 The AIs are similar to RDAs in that both are designed to meet the needs of almost all healthy people. Because AIs and RDAs are thought to meet 98 percent or more of the population’s requirements, they overestimate nutrient requirements for most people. As previously described,¹⁸ we used the EAR or a cutoff value at 67 percent of the AI or RDA to define adequate nutrient intake. The Sum EAR represents the number of nutrients consumed at adequate nutrient intake levels and is a measure of overall dietary quality.

We determined dietary variety, another measure of dietary quality, by counting the number of food types consumed within each food group using a classification scheme developed by Drewnowski and colleagues¹⁷ based on a food frequency questionnaire developed by Block and colleagues²⁸; we had used this scheme in previous research.¹⁸ The scheme defines food types as closely related foods with similar nutrient compositions (for example, apples and applesauce, pancakes and waffles, cookies and cake), while it defines food groups as categories with similar dietary roles (such as fruits, breakfast foods and sweets). One of the authors (T.A.M.) individually reviewed each three-day dietary record and assigned foods to the appropriate food type and group. We defined each subject’s Diet Variety Score as the total number of food types consumed during the three-day period; we also determined the number of food types per major food group. Our research team¹⁸ previously reported a mean (± standard deviation) Diet Variety Score of 21.8 ± 4.3 (range of 10–32) for this cohort.

Statistical analysis. We converted oral health examination data to SPSS format and initially analyzed them using SPSS software (SPSS Base 7.5 for Windows, SPSS, Chicago). For the present analysis, we analyzed data using another software package (SAS, Version 8, SAS Institute, Cary, N.C.). We reported subjects’ characteristics and dental variables as means and standard deviations when continuous and as percentages when categorical. We evaluated differences between means using t test or analysis of variance with the Tukey test. Differences between categorical variables were evaluated using ² analysis. To evaluate associations among continuous variables, we used the Pearson product moment correlation. The number of teeth and number of functional teeth variables were not distributed normally; therefore, we analyzed them using nonparametric tests (Kruskal-Wallis test). We adjusted protein and carbohydrate intakes for energy intake using general linear models. Finally, we tried stepwise multiple regression models to predict dietary measures (that is, energy, protein, calcium, Diet Variety Score and Sum EAR) using oral health measures and demographic variables including age, sex and educational status.

Mean daily nutrient intake did not differ between subjects with complete and incomplete natural dentitions.

	RESULTS

TOP ABSTRACT SUBJECTS, METHODS AND MATERIALS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Responders vs. nonresponders. All subjects were white and aged 79 years or older. As reported previously,¹⁸ responders (subjects with complete, usable dietary records) were slightly younger and had higher scores of cognitive function than did nonresponders, but did not differ from nonresponders in terms of sex, living arrangements or marital status (Table 1). Responders retained significantly more teeth, including functional teeth, than did nonresponders and had different overall dentition statuses from them as well (Table 1). Neither numbers of crown, root and total caries nor denture retention or stability differed between responders and nonresponders.

Number of teeth and nutrition measures. Table 2 (page 1374) presents correlations among mean daily nutrient intake, dietary quality index scores, number of teeth and number of functional teeth. Mean daily intake of protein, vitamin C, calcium, phosphorus and zinc was associated with both the number of teeth and number of functional teeth. Adjustment of protein for energy intake did not affect these relationships. The Diet Variety Score, a measure of overall dietary quality, also was weakly associated with number of teeth and number of functional teeth. A final multivariate model for prediction of calcium intake (P = .005) included the number of teeth (P = .03) and the presence of coronal caries (P = .03). Multivariate analysis including number of teeth did not improve on bivariate analyses for other dietary variables.

Mean daily nutrient intake did not differ between subjects with complete and incomplete natural dentitions; however, fewer subjects with complete dentitions consumed adequate amounts of vitamin A (78 percent vs. 97 percent, P = .017).

Caries status and nutrition measures. The number of coronal carious lesions was positively associated with mean daily intakes of protein (r = 0.152; P = .024) and calcium (r = 0.165; P = .014), but not with adequacy of nutrient intake. The relationship between coronal caries and protein intake did not remain after we made adjustments for energy intake. The number of root carious lesions was not associated with mean daily nutrient intake, but was higher in subjects with inadequate zinc intakes (0.73 ± 2.3 vs. 0.23 ± 0.72 carious lesions; P = .020). Neither total carbohydrate nor total sucrose intake was associated with coronal or root caries. Total caries was positively associated with mean daily calcium intake (r = 0.138; P = .042). Although we included coronal caries (P = .03) with number of teeth (P = .03) in the final multivariate model predicting calcium intake, inclusion of coronal or root caries in multivariate analysis did not improve on bivariate analysis for other dietary variables.

Denture status and nutrition measures. Retention and stability of the mandibular denture also was associated with higher mean daily nutrient intake and higher scores on dietary quality indexes (such as Sum EAR and Diet Variety Score) (Table 3, page 1375). Subjects with either retention or stability problems of the mandibular denture had lower mean daily intakes of protein, thiamin, riboflavin, pantothenic acid, vitamin D, calcium, iron, magnesium and phosphorus, as well as lower dietary quality scores than did subjects without mandibular denture retention and stability problems. The relationship between retention and stability of the mandibular denture and protein intake did not remain after adjustment for energy intake.

Table 4 (page 1376) presents mean daily nutrient intake and dietary quality index scores according to dentition status; we excluded subjects with denture problems from these analyses. Neither intakes of individual nutrients nor scores of dietary quality differed among subjects who had complete dentures, mixed partial/complete dentures, partial dentures, complete natural dentitions or incomplete natural dentitions. However, the number of subjects achieving an adequate intake of vitamin A differed by dentition status: 67 percent of subjects who had full dentures, 85 percent of subjects who had full/ partial dentures, 95 percent of subjects who had partial dentures, 78 percent of subjects with full natural dentitions and 97 percent of subjects with incomplete natural dentitions (P = .038). Multivariate models to predict nutrient intake that included these categories of dentition status did not improve on bivariate analysis.

	DISCUSSION

TOP ABSTRACT SUBJECTS, METHODS AND MATERIALS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Our results support observations of other investigators^6–16 that loss of natural teeth or ill-fitting dentures reduce dietary quality and nutrient intake. Specifically, we found that both intake and adequacy of intake for multiple nutrients and scores on dietary quality indexes were lower in subjects with fewer natural teeth or poorly retained or unstable mandibular dentures. Subjects with well-fitting dentures had nutrient intake and scores on dietary quality indexes similar to those of subjects with natural teeth. These data suggest that tooth loss in itself may not pose a nutritional problem, but rather that impaired nutrient intakes arise when teeth are not replaced or when denture fit and stability are inadequate. Regular dental care to replace missing teeth and maintain denture function may be critical to the maintenance of dietary quality and adequate nutrient intake in the oldest old.

The finding that poor retention or stability of mandibular dentures was related to inadequate intakes of several nutrients and poor dietary quality has important implications for dentists. It stresses not only the need for adequate prostheses, but also for regular maintenance care for partially dentate and edentulous patients.

Mandibular dentures, particularly complete mandibular dentures, are difficult to fabricate to both the patient’s and dentist’s satisfaction under the best of circumstances. The very old often have minimal mandibular ridge support, making it even harder to fabricate retentive and stable prostheses. However, given the nutritional deficiencies found in our study, it clearly is important to provide elderly patients with adequate mandibular dentures not only to maintain function, but also to support general health. Thus, dentists should give serious consideration to use of implant-retained prostheses, overdentures and other, more retentive, mandibular prostheses for not only elderly patients, but also younger patients, to prevent nutritional risk later in life.

Dietary variety, a measure of dietary quality, was lower in subjects with fewer total teeth, fewer functional teeth or ill-fitting mandibular dentures. These subjects likely limited their food choices, thus reducing mean nutrient intake and overall dietary quality. In subjects with fewer teeth or ill-fitting dentures, recommendations to increase consumption of meat, legumes and dairy products would improve protein, iron, zinc, phosphorus and magnesium intakes; fruits and vegetables would improve vitamin C intakes; and a variety of foods would improve folate, thiamin, riboflavin and pantothenic acid intakes. Unlike other investigators,^6,11,13 we did not find differences in mean nutrient intake; adequacy of nutrient intake, with the exception of vitamin A; or dietary quality among subjects with different dentition statuses. Our results may differ from those of previous reports because we did not include subjects with poorly retained or unstable dentures in this comparison. Alternatively, it is possible that after we removed subjects with ill-fitting dentures from the analysis, our statistical power was too low to detect differences. Shinkai and colleagues²⁹ also reported no differences in dietary quality among measures of dentition status using the Healthy Eating Index, a measure of dietary quality that includes dietary risk factors for systemic disease.³⁰

Reduced intake of nutrients associated with fewer total natural teeth or ill-fitting dentures may increase morbidity and hasten death in the oldest old.

Reduced intake of nutrients associated with fewer total natural teeth or ill-fitting dentures may increase morbidity and hasten death in the oldest old. Both protein and zinc deficiencies decrease immune function, thereby increasing elderly people’s susceptibility to infection.^31–33 Low serum folate concentrations are associated with decreased cognitive function^34,35 and are a risk factor for cardiovascular disease.^35–37 Marginal vitamin C intakes and status have been associated with decreased cognitive function, increased peripheral arterial disease and periodontal disease, particularly in smokers.^38–40

Both mean intake and adequacy of intake of calcium were positively associated with the presence of teeth and well-retained, stable mandibular dentures in our subjects, which is consistent with data reported by Sheiham and colleagues,¹⁰ who associated higher calcium intakes with the presence of natural teeth and increased numbers of natural teeth. The primary dietary source of calcium is milk.⁴¹ Obviously, one does not require the presence of teeth or functional dentures to drink milk. Rather, dietary calcium intake likely reflects a person’s lifelong pattern of milk consumption. Milk intake may be indicative of a healthy lifestyle and inversely associated with behaviors, including smoking and excessive alcohol consumption, that are associated with tooth loss. Furthermore, intakes of milk and sugared beverages are negatively associated^42,43; replacement of milk with sugared beverages may have increased the incidence of caries and related tooth loss in our subjects. The positive association between current coronal caries and calcium intake in our subjects is not easily explained. Although we hypothesized that current calcium intake reflects lifelong milk consumption, calcium intake also may be associated with a recent dietary change that increases caries risk. For instance, age-related changes in dietary patterns might include, in lifelong milk consumers, an increased frequency of cookie or dessert snacks accompanying their customary milk consumption.

Screening tools developed to identify older people at risk of experiencing poor nutrition typically include an oral health component, as the impact of oral health on dietary quality, nutrient intake and nutritional status is widely recognized. Oral health questions have successfully identified people whose nutritional status is at risk.^44–48 Elderly Bostonians responding positively to the item "I have tooth or mouth problems that make it hard for me to eat" in the Nutrition Screening Initiative Checklist had a significantly increased risk of mortality during the following eight years.⁴⁶ Positive responses to the item "I find it hard to bite or chew food ... " in the Risk Evaluation for Eating and Nutrition component of the Seniors in the Community study and a dietitian’s perception of nutritional risk were strongly correlated.⁴⁶ Jensen and colleagues⁴⁷ reported that perceived "eating problems" were a sensitive indicator of functional limitations and a predictor of hospitalization within one year of screening.⁴⁸ We did not find associations between subjects’ perceptions of food avoidance, chewing difficulties or temperature sensitivity and either nutrient intake or adequacy of nutrient intake.

However, these data should be interpreted cautiously, because we had insufficient statistical power to detect associations, given the small number of positive responses to these questions in our sample. The low number of positive responses to food avoidance, chewing difficulties and temperature sensitivity reported herein may reflect our subjects’ acceptance of such difficulties as normal or as an adaptation of dietary habits. Alternatively, perceptions of oral difficulties may be associated with risk of experiencing poor nutrition, but not with nutrient intake. In this case, a more objective nutrition-screening item that measures tooth or denture status may be necessary to identify people at risk of experiencing low nutrient intake.

Although responders were slightly younger and more cognitively intact than were nonresponders, other demographic variables did not differ between responders and nonresponders. In comparison with nonresponders, responders had more natural teeth and more functional teeth, were less likely to have complete dentures, and were less likely to avoid certain foods. We previously reported that nonresponders were at higher risk of experiencing poor nutrition than were responders.¹⁶ In addition, 40 percent of potential subjects in the two counties declined participation, frequently citing health reasons, so the sample cannot be considered representative of the oldest old in the two counties. Thus, responders were not representative of the sample from which they were drawn or of community-dwelling elderly people residing in other geographical areas.

Our study was limited by the cross-sectional nature of data collection and difficulties inherent in dietary data collection. Cross-sectional studies allow for associations to be identified but do not support causal associations. All dietary data were self-reported. This means that responders may have changed dietary habits in response to the recording process, or that actual nutrient intake may have differed from recorded intake.

	CONCLUSION

TOP ABSTRACT SUBJECTS, METHODS AND MATERIALS RESULTS DISCUSSION CONCLUSION REFERENCES

 
In summary, our results support the hypothesis that poor oral health is associated with limited dietary variety and lower nutrient intake. Aggressive preventive oral health care to maintain natural dentition throughout life and regular dental care to ensure adequate denture fit and function may decrease nutritional risk in elderly people.