Abstract
Poor dental status negatively relates to dietary intakes. However, this issue has not been researched among the most vulnerable groups in society. We aimed to investigate, in a national low-income sample, the association between dental status and fruit and vegetable consumption. We analyzed data on adults aged 50 years and older from the Low Income Diet and Nutrition Survey, a representative sample of deprived UK households. Considerable numbers reported difficulty eating specific foods, with significantly worse experience among edentate than dentate people. The mean daily fruits and vegetables consumption was low (256.5 g for dentate, 207.1 g for edentate). After adjustment for socio-demographic and behavioral variables, edentate individuals consumed 50.7 g (27.0, 74.3) fewer fruits/vegetables per day than the dentate. Over and above the effects of material deprivation on nutrient intake, edentulism negatively relates to eating fruits and vegetables and achieving a healthy diet in materially deprived older adults.
Introduction
The consumption of an adequate, well-balanced diet is crucial to the prevention of chronic diseases (World Health Organization, 2003). Many people fail to meet healthy eating recommendations, particularly in relation to the consumption of fruits, vegetables, and dietary fiber (Henderson et al., 2002, 2003). In the UK, fruit and vegetable consumption is used as an important nutritional indicator, with the recommended 5 portions a day being a public health target. There are stark inequalities in diet and nutritional intake between different socio-economic groups (Irala-Estevez et al., 2000; Dowler, 2001; Dowler et al., 2007). Food choice is influenced by an array of political, environmental, social, and individual factors (Dowler et al., 2007). In addition, dental status also affects the intake of nutritious foods, particularly in older people (Smith and Sheiham, 1979; Locker, 1992). However, very few studies have explored this association in national population samples (Nowjack-Raymer and Sheiham, 2003, 2007; Walls and Steele, 2004).
Both the UK National Diet and Nutrition Survey (NDNS) for people over 65 yrs and the US National Health and Nutrition Examination Survey (NHANES) showed that poor dental status adversely related to dietary intakes, with the edentulous being at a nutritional disadvantage compared with dentate individuals, even after adjustment for demographic factors (Sheiham et al., 2001; Nowjack-Raymer and Sheiham, 2003, 2007). However, no national studies have investigated the association between dental status and nutritional intakes among the most vulnerable sections of society, such as adults living in deprived households. The aim of this study was to investigate, in a nationally representative sample of low-income adults, the association between dental status and fruit and vegetable intakes. In particular, this study assessed whether being edentate negatively relates to the ability of low-income adults aged 50 yrs and over to chew specific foods, and assessed their consumption of fruits and vegetables.
Materials & Methods
This study analyzed data collected for the Low Income Diet and Nutrition Survey (LIDNS), which was designed to provide a comprehensive picture of food consumption and nutritional status of a nationally representative sample of respondents living in low-income households in the UK (Nelson et al., 2007). The underlying purpose was to provide evidence for developing food policy to reduce health inequalities. ‘Low-income’ was used in a broader sense to include wider aspects of material deprivation, rather than only low income per se. This wider approach to material deprivation was considered both in relation to sampling, by focusing on neighborhoods ranked among the most materially deprived on 2001 census data, and also during the screening process for determining eligibility for inclusion in the survey, through questions on benefits, car ownership, household composition, and employment status. The LIDNS target population was the 15% most deprived households in the UK.
In the study, we used a multi-stage clustered probability sample to select a nationally representative sample of low-income materially deprived people, with the different stages of cluster selection being wards (electoral areas), addresses, households, and respondents. In the first stage, wards were stratified by country, by region, and then by population density. Wards were selected with probability proportional to number of addresses multiplied by the deprivation index of the ward. This resulted in an unequal probability sample, with over-sampling for deprived wards as well as for Scotland, Wales, and Northern Ireland. At the second stage, addresses in each of 528 selected wards were selected from the small-users Postcode Address File. At the third stage, one household at each address was randomly selected. To determine whether a household was eligible for inclusion, a 10-item screening questionnaire was administered. Finally, at each selected household, up to two respondents were recruited. All household members were eligible for selection, apart from children younger than 2 yrs and pregnant women. In households with more than two eligible members, two respondents were randomly selected. In case the household consisted of both adults and children, one adult and one child were invited to participate in the study. Complete details of the LIDNS methodology are published elsewhere (Bates et al., 2007).
Data were collected through interviews and a nurse visit. The extensive face-to-face computer-assisted personal interviews (CAPI) provided information on household, socio-economic status, environment, circumstances, and attitudes that could affect dietary intake, and health characteristics. In addition, each respondent was asked to provide repeat 24-hour dietary recalls on 4 random days (including at least 1 weekend day) over a 10-day period. Information was also collected through self-completion questionnaires. Height and weight measurements were taken. Successful completion of the interview stage was followed by a visit from a nurse who collected information on medication and nutritional supplements and took objective health measurements and blood samples. The variables used in the current analysis refer to the interview stage of the study.
We used data in relation to the respondent’s reported difficulty in chewing in general, as well as in eating certain foods. The latter question, previously used in the UK National Diet and Nutrition Survey (NDNS) of individuals aged 65 yrs and over (Steele et al., 1998), was administered to those aged 45 yrs and over and referred to foods of various textures, such as raw carrots, apples, and nuts, but also tomatoes and lettuce. In addition, based on the dietary recalls, information was gathered on the actual daily consumption, in grams (g), of fruits and vegetables. In terms of dental status, respondents were not clinically examined, but were asked whether they had any natural teeth (dentate) or not (edentate).
To account for seasonal variation in food consumption and eating patterns, fieldwork was spread equally over a 12-month period. The methodology of the LIDNS and the validity of the repeat 24-hour recall had been tested in a prior feasibility study (Bates et al., 2004). The study was ethically approved by the London Multi-Centre Research Ethics Committee (MREC), while research governance was obtained for all participating hospital laboratories.
Data Analysis
Analysis was conducted with Stata software, using the survey command. To assess the influence of dental status on eating ability and actual consumption of nutritious food such as fruits and vegetables, we used appropriate regression analyses—linear when the outcome was continuous (fruit and vegetable consumption) and logistic when it was categorical (chewing ability and difficulty eating foods). The initial unadjusted models for the association of the outcome with dental status (dentate vs. edentate) were further adjusted for the effects of age group, sex, and years of education. Linear regression models for fruit and vegetable consumption were also adjusted for smoking (never, past, and current smokers). Since there were only very small numbers of edentate people aged fewer than 50 yrs, ‘data analysis’ refers to participants aged 50 yrs and over. In addition, analyses are also presented for age strata (50-64 yrs; 65 yrs and over) because an independent effect of age was a common feature of most of the regression models, and the distribution of dental status varied considerably by age. We used the appropriate weight throughout the analyses to account for the previously mentioned over-sampling as well as to reduce non-response bias.
Results
Overall, 3728 individuals from 2477 households were included in the final LIDNS data set. This report refers to 1398 participants aged 50 yrs and over: 795 (57.3%) dentate and 603 (42.7%) edentate. Edentate participants were older, less-educated, and had proportionately more females than the dentate (Table 1). The sample had low education levels; only 17.6% of dentate and 10.2% of edentate had attended full-time education for more than 10 yrs. High proportions of the sample were overweight (37.3% of dentate and 40.5% of edentate) and obese (33.8% of dentate and 32.9% of edentate).
Characteristics of the LIDNS Sample, Aged 50 Years and Older, by Dental Status: Numbers (unweighted) and Proportions (weighted)
Due to missing information, the bases for education are n = 597 for dentate and n = 517 for edentate, and for BMI n = 691 for dentate and n = 490 for edentate.
BMI categorizations in the LIDNS based on previous National Diet and Nutrition Surveys (Herrick and Holmes, 2007).
Difficulty in chewing foods was reported by 21.9% of dentate and 36.3% of edentate (Table 1). After adjustment for age group, sex, and education, edentate participants were 1.76 (1.21, 2.56) times more likely to report difficulty in chewing than the dentate. The difference in difficulty in chewing between dentate and edentate was much higher in the younger age group (50-64 yrs), with the dentate participants having a lower, and the edentate a higher prevalence of difficulty in chewing in comparison with the respective groups among participants aged 65 yrs and over (Table 2).
Difficulty in Chewing According to Dental Status in the Whole LIDNS Sample Aged 50 Years and Older, Stratified by Age Group
Adjusted for age group, sex, years of education (n = 1114).
Adjusted for sex and years of education (n = 417 for those aged 50-64 yrs and n = 697 for those aged 65 yrs or over).
Considerable proportions of the sample had difficulty in eating specific foods (Table 3). A much higher proportion of edentate than dentate people experienced difficulty in eating all foods, regardless of texture, though greater differences were observed for hard foods, such as raw carrots, apples, and nuts. The differences between dental status groups were significant after adjustment for age, sex, and education. The odds ratios ranged from 1.77 (1.23, 2.53) to 2.28 (1.57, 3.30). While proportions of difficulty in eating specific foods were not excessively different across the age strata for edentate, there was a distinct pattern among dentate participants, with much higher levels of difficulty in eating specific foods among the dentate participants aged 65 yrs and older compared with dentate persons aged 50 to 64 yrs old (Table 3).
Difficulty in Eating Specific Foods, by Dental Status, in the Whole LIDNS Sample Aged 50 Years and Older, Stratified by Age Group: Proportions Affected, Adjusted Odds Ratios (95% Confidence Intervals)
Adjusted for age group, sex, years of education (bases for analyses vary between 1107 and 1114 for different foods).
Adjusted for sex and years of education (bases for analyses vary between 413 and 417 for different foods for those aged 50-64 yrs and between 692 and 697 for those aged 65 yrs or over).
The overall consumption of fruits and vegetables was low in the sample and showed consistent and significant variation between dentate and edentate respondents. The mean daily consumption was 256.5 g for dentate and 207.1 g for edentate subjects (Table 4). After adjustment for age, sex, education, and smoking status, dentate participants consumed 50.7 (27.0, 74.3) more grams of fruits and vegetables per day than the edentate (p < 0.001): 28.3 (10.6, 46.0) grams of fruit and 22.3 (10.1, 34.6) grams of vegetables. The differences, by dental status, were also evident among age strata (Table 4).
Amount of Daily Consumption of Fruits and Vegetables by Dental Status in the Whole LIDNS Sample Aged 50 Years and Older, Stratified by Age Group: Adjusted Regression Coefficients (95% Confidence Intervals)
Adjusted for age group, sex, years of education, and smoking status in the whole sample (n = 1114) and for sex, years of education, and smoking status in the age strata (n = 417 in those aged 50-64 yrs and n = 697 in those aged 65 yrs and over).
Discussion
This study demonstrated that being edentate contributes in a negative and significant way to the ability of low-income people aged 50 yrs and over in the UK to eat a variety of foods. Edentate people reported higher levels of chewing impairment and experienced greater difficulty in eating common foods than dentate people. The negative impact of dental status on nutritional intake is over and above that attributed to their material deprivation. After adjustment for socio-demographic and behavioral factors, edentate older adults consumed considerably and significantly lower quantities of fruits and vegetables than dentate. The differences between dentate and edentate were consistent throughout the analysis, for all outcome measures, and across age strata.
Previous studies on older adults have also shown that edentate participants had lower consumption of vegetables and some fruits (Halling et al., 1988; Moynihan et al., 1994; Sheiham et al., 1999). This first national study on a low-income population has not only confirmed these differences, but also shown that the differences are large among the more deprived groups in society.
An interesting finding is the age-related pattern of differences between dentate and edentate people, primarily for overall chewing ability, but, to a lesser extent, also for difficulty in eating certain foods. The differences were large in general, as well as for each age group, though they were considerably larger among the 50- to 64-year-olds and tended to decrease in those aged 65 yrs and over. This can be partly attributed to two factors. First, older dentate participants tended to report higher prevalence of difficulties (both chewing overall and eating certain foods) than the 50- to 64-year age group. Second, in contrast, older edentate participants reported lower prevalence of overall chewing difficulty than younger edentate participants, though the differences in the difficulty in eating certain foods were marginal.
The age-related pattern of differences for the dentate can be linked to the accumulation of oral diseases over time throughout the life course. As dentate people get older, their oral health gradually declines. Despite the overall improvement in oral health and edentulousness rates (Kelly et al., 2000; Steele et al., 2000), older dentate people have fewer natural teeth and the teeth are in a worse condition than in younger dentate adults (Centers for Disease Control and Prevention, 2003). Having fewer natural teeth has been linked to increased difficulty, particularly in the eating of hard foods, and to deficient nutrient intakes (Krall et al., 1998; Sheiham et al., 1999, 2001; Sahyoun et al., 2003; Nowjack-Raymer and Sheiham, 2007). The deterioration in chewing ability and consumption of nutritious foods as people age was not apparent in edentate adults in our sample. While there was considerable variation in oral health status and oral functioning, particularly in relation to number of teeth among dentate people, there was considerably less variation among edentate people who wear full dentures. Therefore, it is to be expected that chewing difficulty among the edentate would not be affected by age to the same extent as for the dentate. Better chewing ability among older edentate compared with the younger age group may partly be due to older people adjusting their eating patterns after wearing dentures.
The LIDNS older adults had average energy intakes below their estimated requirements (Holmes, 2007). However, the majority were overweight or obese. Considerable numbers of respondents reported difficulty eating a wide range of nutritious foods (e.g., apples and raw carrots). This difficulty was further reflected in their low overall fruit and vegetable consumption. Consequently, both dentate and edentate materially deprived adults were consuming far below the recommended 5 portions a day, equivalent to 400 grams. The findings provide clear evidence of the nutritional disadvantage of this low-income population (Nelson et al., 2007) and suggest targeting health promotion interventions to them. The co-existence of obesity and low fruit and vegetable intake in this population may be indicative of a diet rich in energy-dense, but not nutritious, foods. Public health action for this group should put emphasis on the consumption of nutritious foods that are easy to chew (e.g., baking apples). Furthermore, edentate low-income people are doubly disadvantaged, since they consume even lower quantities of fruits and vegetables. This has clear policy implications for the prevention of edentulousness, which is prevalent in low-income older people.
Being a nationally representative sample of the materially deprived population, the LIDNS sample has considerably worse socio-economic background characteristics than the general UK population, as well as higher prevalence of edentulousness (Tsakos et al., 2007). A methodological limitation refers to data collection through dietary recalls, which are prone to underreporting, particularly in overweight and obese individuals. Other limitations include the crude measurement of dental status. Data on number of teeth and clinical status for the dentate, years of edentulousness, and quality and use of dentures for the edentate, as well as self-rated oral health measures for both dentate and edentate, were not collected. The use of subjective chewing ability scales (Feine and Lund, 2006) would have provided more sensitive measurements. However, this is the first national study on a low-income UK population and demonstrated that, over and above the nutritional disadvantage associated with material deprivation, low-income edentate older adults consumed considerably fewer fruits and vegetables than their dentate counterparts.
Footnotes
Acknowledgements
The LIDNS was funded by the Food Standards Agency and conducted by the National Centre for Social Research, in collaboration with researchers from King’s College London and University College London.