Snacking may improve physical function among older Americans

Abstract

Background/Objectives

Snacking was reported to provide nutritional benefits among older adults, but the association between such dietary behavior and health outcomes has not been clearly established. The purpose of this study is to examine the associations between snacking and gait speed, a performance-based measure of physical function.

Design

Cross-sectional population-based survey.

Setting

The 1999–2002 National Health and Nutrition Examination Survey (NHANES).

Participants

A nationally representative sample of Americans aged 60 and older n = 2,333).

Measurements

Participants were classified by snacking frequency (0, 1, 2, 3, ≥4 snacks/d) and by the contribution of snacking to their daily energy intake (0 to <10%, 10% to <20%, 20% to <30%, ≥30%). Physical function was assessed by measurement of gait speed over 20 feet.

Results

After adjusting for age, gender, education, race/ethnicity, smoking status, and marital status, older adults who snacked four times or more daily had a faster gait speed (P = 0.033) than non-snackers. Snacking that contributed 20% to <30% (P = 0.017) of energy was associated with a faster gait speed than snacking that contributed 0 to <10% of energy. Similar associations were observed after further adjustment for potential confounders.

Conclusion

Both snacking frequency and percentage of energy from snacking are positively associated with gait speed among older adults. The benefits of snacking on older adults' physical function may warrant their inclusion in this population's diet.

Introduction

The number and the proportion of older Americans are increasing rapidly over the past few decades due to the longer lifespan and aging baby boomers (1). Poor physical function associated with aging is a primary concern for the older adult population, limiting their daily activities, predisposing them to falls, and increasing their mortality and morbidity risk (2., 3., 4.). There are many methods for measuring physical function. Of these, gait speed is commonly considered convenient, adequate and reliable; in fact, it has been recommended as a “geriatric vital sign” to assess the overall health status among older adults (4, 5). Gait speed is highly correlated with multiple features of lower body physical function such as balance, strength, and coordination (2, 6). Since gait speed tends to slow in early disabling processes, it can be monitored during interventions aimed at preventing or delaying disability, mortality, hospitalization, and other adverse events (3, 6). To achieve longer, high-quality, productive, and independent lives, there is an increasing public health interest to study and promote healthy dietary behaviors to improve physical function among the older adult population (7., 8., 9., 10., 11., 12.).

Older adults may have difficulty maintaining health-promoting diets because of multiple factors including gradual loss of appetite, taste, smell and vision, and various other physiological changes in dental, digestive, and other systems (13., 14., 15.). In addition, they often undergo psychosocial changes such as eating alone, depression, and loneliness (16, 17). All of these factors affect dietary intake and may result in nutrient deficiencies among older adults. Snacking, as a common dietary behavior, has been suggested as a method of providing shortfall nutrients that cannot be met with the consumption of only 3 meals a day (13., 14., 15.). Studies have found that snacking contributes a nutrient-dense diet, particularly among older adults, and can improve overall diet quality (13-15, 18). Snacking has also been suggested to be a target behavior to improve nutritional well-being among older adults (19). Since nutritional well-being has been associated with physical function, it is reasonable to hypothesize that snacking positively affects physical function among older adults (7). However, such association has not been thoroughly investigated. This study aims to examine the associations between snack patterns, as measured by frequency of daily snack occasions and percentage of total daily energy intake contributed from snacking, and gait speed, a performance-based measure of physical function, using a national sample of independent, community-living older adults.

Methods

Study Design and Participants

Data used in this study are from the U.S. Federal Government’s National Health and Nutrition Examination Survey (NHANES) for 1999-2002. Data from these earlier rounds were used because gait speed data were not collected in subsequent surveys. NHANES is a survey of the U.S. civilian, non-institutionalized population based on a complex, multistage sampling design. This survey provides information regarding participants’ dietary, demographic, socioeconomic, and health-related characteristics. Informed consent was obtained from all participants, and the CDC Institutional Review Board approved the protocol (20).

This study included 3,061 participants aged 60 years and older with reliable dietary data. Participants (n = 575) with missing variables for age, gender, education, race/ethnicity, smoking status, Body Mass Index (BMI), physical activity, co-morbidities, marital status, and dietary data were excluded. After excluding incomplete data for gait speed (n = 153), the final analytic sample size was 2,333.

Dietary Assessment

Dietary data were collected via a 24-hour recall, which included a multiple pass method. The multiple pass method has been validated with the doubly-labeled water technique (21). Detailed descriptions of the dietary interview methods are provided in the NHANES Dietary Interviews Procedure Manual, which includes pictures of the computer-assisted dietary interview system screens, measurement guides, and charts used to collect dietary information (20).

Detailed information about the identity and amount of each food and beverage item consumed was collected using probes. Participants were asked to report the time each food and beverage item was eaten and to classify the type of eating occasion using a pre-defined list of categories. This study classified snacks as those foods and beverages that had been reported as snack/beverage and extended consumption in 1999-2002 NHANES. All other eating occasions were considered as meals. For those participants who were interviewed in Spanish, those foods reported being consumed as a merienda, entre comida/bebida/tentempis, botana, or bocadillo were all considered snacks. All other eating occasions reported in Spanish were considered meals.

Foods and beverages consumed within 15 minutes of each other were considered one snack or meal occasion and subsequently summed for each participant to quantify snacking occasions over 24 hours (13). For this analysis, snacking frequency was categorized into five groups: 0, 1, 2, 3, and ≥4 times per day. In a subsequent analysis, participants were classified by the contribution of snacking to their total daily intake of energy, which reflects the energy density and portion sizes of foods and beverage consumed as snacks (22). Four groups were created where snacking accounted for 0 to <10%, 10% to <20%, 20% to <30%, and ≥30% of total daily energy intake. Participants were classified as either consuming < 3 or ≥3 meals per day.

Gait Speed

The timed 20-foot gait test was performed at the participant’s usual pace. Participants were allowed to use a walker or cane if needed, but individuals were excluded from the testing if they needed the assistance of another individual to walk. Gait speed (m/s) was calculated as the walking distance (20 feet = 6.15 m) divided by the time required to complete the task (in seconds).

Statistical Analysis

All descriptive and inferential statistics were estimated using STATA (version 10, 2007, Stata Corp, College Station, TX) to account for sample design and sampling weights. Associations across categories of snacking frequencies and percentage of energy intake from snacking were tested by using chi-square tests for categorical variables and linear regression analysis for continuous variables.

Multiple linear regression models were used to estimate the associations between snacking and gait speed. Potential confounders included age, gender, education, race/ethnicity, smoking status, marital status, energy from meals, consumption of ≥3 meals per day, presence of co-morbidities, BMI, physical activity, availability of emotional support, availability of financial support and cognitive function. Education was defined as less than a high school degree, a high school degree or equivalent, some college or an associate’s degree, or college graduate or more. Three race/ethnicity categories were examined: non-Hispanic white, non-Hispanic black, and Hispanics. Participants were classified as current, former, or never smokers based on self-reported information. Participants were asked if they had ever smoked at least 100 cigarettes, a pipe at least 20 times, or a cigar at least 20 times. Those who answered no to all these questions were classified as never smokers. If participants answered yes to any of the previous questions, they were designated former smokers or current smokers if they reported presently smoking. Physical activity levels were based on self-reported physical activities done during the participants’ leisure time over the past 30 days and classified as vigorous, moderate, or sedentary activity level. Presence of co-morbidities was based on participants reporting a doctor’s diagnosis of any of the following chronic diseases: diabetes, hypertension, congestive heart failure, heart disease, arthritis, angina, heart attack, stroke, emphysema, chronic bronchitis, and cancer. Married included those who reported married or living with their partner. Single included those who were widowed, separated, or never married. BMI was calculated by dividing a participant’s weight in kilograms by the square of his/her height in meters. BMI was also analyzed as a categorical variable (<18.5; 18.5 to <25.0; 25.0 to <30.0; and ≥30.0 kg/m²) according to NIH guidelines (23). Availability of emotional support was assessed by the presence of someone to talk over problems or to help make a difficult decision. Availability of financial support was assessed by the presence of someone to help pay bills, housing costs, hospital visits, or provide food or clothes. Cognitive function was assessed using the Digit Symbol Substitution Test (20). Age, energy contribution from meals and cognitive function were considered as continuous variables.

Several model specifications were used to explore the associations between snacking and gait speed. Age and gender were included as confounders in model 1; model 1 confounders plus education, race/ethnicity, smoking status, and marital status were included in model 2; model 2 confounders plus energy from meals, consumption of ≥3 meals per day, presence of co-morbidities, BMI, physical activity, availability of emotional support, availability of financial support and cognitive function were included in model 3. Statistical significance was assigned at P < 0.05 (two-sided).

Results

Descriptive characteristics of the analytic sample by categories of snacking frequency are presented in Table 1. Participants who snacked more frequently tended to be younger and non-Hispanic white. In addition, they were less likely to consume ≥3 meals per day. They were more likely to have co-morbidities, a faster gait speed, and less energy from meals.

Table 1.

Descriptive characteristics of participants aged ≥60 y by categories of snacking frequency: NHANES 1999-2002 (n = 2,333)

				Snacking frequency			P value
	0 n = 441	1 (n = 680)	2 (n = 622)	3 (n = 342)	≥4 (n = 248)	Total (n = 2,333)
Male (%)	42.7	37.4	38.9	32.2	44.8	38.7	0.178
Education (%) < High school education	30.9	37.1	26.8	20.2	24.4	28.7	0.022
High school	27.9	26.3	34.0	29.9	24.6	29.2
Some college	19.0	20.7	21.6	26.2	24.5	22.1
≥college	22.1	15.9	17.6	23.7	26.5	20.0
Race/ethnicity (%)* Non-Hispanic white	81.6	82.2	87.5	86.2	89.7	85.3	0.008
Non-Hispanic black	8.9	7.8	7.8	5.4	5.8	7.3
Hispanic	9.5	10.0	4.7	8.3	4.5	7.4
Smoking status (%) Nonsmoker	48.7	51.0	49.7	42.7	44.8	48.1	0.440
Former smoker	39.6	34.4	36.2	38.4	39.5	37.1
Current smoker	11.7	14.7	14.1	18.8	15.7	14.8
Married (%)	59.6	62.9	65.4	62.3	73.3	64.3	0.140
Consumption of ≥ 3 meals/d (%)	82.4	81.2	70.7	75.9	69.2	76.0	0.002
Presence of co-morbidities (%)	77.7	88.1	82.4	80.9	86.3	83.4	0.006
BMI (%) <18.5 kg/m2	2.5	1.4	1.2	2.3	2.1	1.8	0.422
18.5 to <25 kg/m2	31.1	28.4	26.3	28.6	27.2	28.1
25 to < 30 kg/m2	39.2	36.1	44.6	32.6	38.5	38.7
≥30 kg/m2	27.2	34.2	27.9	36.5	32.2	31.4
Physical activity (%) Sedentary	52.1	55.2	47	43.1	39.6	48.3	0.024
Moderate	25.0	29.8	33.8	35.5	40.0	32.4
Vigorous	22.9	15.1	19.2	21.4	20.5	19.2
Availability of emotional support (%) 94.5	93.4	94.6	91.0	99.5	94.3	0.004
Availability of financial support (%) 79.5		80.6	83.8	81.2	81.6	81.6	0.661
Age (y) †	70.9 (0.5)	71.2 (0.5)	70.2 (0.5)	68.6 (0.5)	69.0 (0.6)	70.1 (0.3)	0.025
Cognitive function†	44.61(1.05)	44.30(1.15)	48.15(1.25)	52.47(1.48)	52.89(1.05)	47.91(0.77)	<0.001
Energy from meals (kcal) †	1448.94	1382.96	1360.55	1301.54	1290.67	1400.89
	(32.23)	(51.04)	(38.98)	(43.85)	(39.59)	(18.85)	0.082
Gait speed (m/s) †	0.95 (0.02)	0.93 (0.01)	0.99 (0.02)	0.99(0.03)	1.05 (0.02)	0.98 (0.01)	<0.001

Notes: For continuous variables, the P value is for the test of difference in means of the variables of interest. For categorical variables, the P value is for the chi-square test of association.

^*Because of the small sizes, Mexican Americans and other Hispanics were combined to form the Hispanics category and the other race was combined with non-Hispanic whites. f Values are means; SE in parentheses.

Descriptive characteristics of the analytic sample by categories of percentage of energy from snacking are available from the author. We also found that completion or non-completion of the gait speed test did not predict snacking frequency (P = 0.556) or percentage of energy from snacking (P = 0.608).

Adjusted mean values for gait speed relative to snacking frequency are presented in Table 2. After adjusting for age and gender, older adults who snacked four times or more daily had a faster gait speed (P = 0.006). After adjusting for age, gender, education, race/ethnicity, smoking status, and marital status, older adults who snacked four times or more daily had a faster gait speed (P = 0.033) than those who were non-snackers. The association between snacking frequency and gait speed remained after further adjustment for energy from meals, consumption of ≥3 meals per day, presence of co-morbidities, BMI, physical activity, availability of emotional support, availability of financial support and cognitive function. Participants who snacked four times or more daily had faster gait speeds (P = 0.049) than older adults who were non-snackers.

Table 2.

Predicted values (LS means (SE)) for gait speed among those aged ≥60 y by categories of daily snacking frequency

	n	Model 1	Model 2	Model 3
Daily snacking frequency 0*	441	0.95(0.01)	0.96(0.01)	0.97(0.01)
1	680	0.94(0.01)	0.95(0.01)	0.97(0.01)
2	622	0.98(0.01)	0.99(0.01)	0.99(0.01)
3	342	0.97(0.02)	0.97(0.02)	0.96(0.02)
≥4	248	1.02(0.02) f	1.01(0.02) f	1.01(0.01) f

Notes: Model 1 was adjusted for age and gender. Model 2 was adjusted for variables in Model 1 plus education, race/ethnicity, smoking status, and marital status. Model 3 was adjusted for variables in Model 1 and 2 plus energy from meals, consumption of ≥3 meals per day, presence of co-morbidities, BMI, physical activity, availability of emotional support, availability of financial support and cognitive function.

^*The reference group. fStatistical significance at P < 0.05.

Table 3 shows the results for the associations between percentage of energy from snacking and gait speed. After adjusting for age and gender, participants whose snacking contributed 20% to <30% (P = 0.007) of daily energy had a faster gait speed than individuals whose snacking contributed 0 to <10% of daily energy. After adjusting for age, gender, education, race/ethnicity, smoking status, and marital status, those participants whose snacking contributed ≥30% (P = 0.017) of their daily energy had a faster gait speed than those older adults whose snacking contributed to 0 to 10% of their daily energy. After further adjustment for energy from meals, consumption of ≥3 meals per day, presence of co-morbidities, BMI, physical activity, availability of emotional support, availability of financial support and cognitive function, participants whose snacking contributed 20% to <30% (P = 0.024) and ≥30% (P = 0.028) of their daily energy had a faster gait speed than those participants whose snacking contributed 0 to <10% of their daily energy intake.

Table 3.

Predicted values (LS means (SE)) for gait speed among those aged ≥60 y by categories of percentage of energy from snacking

	n	Model 1	Model 2	Model 3
Percentage of energy from snacking
0 to <10% *	953	0.96 (0.01)	0.96 (0.01)	0.97 (0.01)
10 to <20%	507	0.96 (0.02)	0.97 (0.02)	0.96 (0.02)
20 to <30%	392	1.01 (0.01) f	1.00 (0.01) f	1.00 (0.01) f
≥30%	481	0.96 (0.01)	0.97 (0.01)	0.99 (0.01) f

Notes: Model 1 was adjusted for age and gender. Model 2 was adjusted for variables in Model 1 plus education, race/ethnicity, smoking status, and marital status. Model 3 was adjusted for variables in Model 1 and 2 plus energy from meals, consumption of ≥3 meals per day, presence of co-morbidities, BMI, physical activity, availability of emotional support, availability of financial support and cognitive function.

^*The reference group. fStatistical significance at P < 0.05.

Discussion

To our knowledge, this is the first study to report the association between snacking and gait speed among older Americans. Both higher snacking frequency and percentage of energy intake from snacking were found to positively associate with a faster gait speed and similar associations were observed after controlling for potential confounders.

Snacking was recently reported to moderately improve overall diet quality, assessed by Healthy Eating Index-2005 scores (18), which have been associated with physical function among older Americans (7). Previous studies also found snacking was associated with greater intakes of both macronutrients and micronutrients (13., 14., 15.). One study reported that snacking has been associated with greater intakes of protein and energy among older adults (13). Inadequate protein and energy intakes may contribute to functional decline through an increase in the loss of the muscle mass, a decrease in the energy reserves, decreased immune function, increased skin fragility and poor healing (24). In addition, beverages consumed during snacking occasions may prevent dehydration among older adults (14). Dehydration results in negative effects on physiological function including impaired cardiovascular, thermoregulatory, metabolic, and central nervous system function, and thus impaired physical function among older adults who are vulnerable to dehydration due to insufficient water intake and increased water excretion (11, 14).

Snacking frequency is also positively associated with daily intakes of vitamin A, C, and E, beta carotene, and magnesium, copper, and potassium among older Americans (15). one study reported that a low summary musculoskeletal nutrient intake including calcium, vitamin D, magnesium, and phosphorus intakes was associated with poorer performance in balance, gait speed, and repeated chair stands (10). Furthermore, vitamin C and E, beta carotene, and copper have antioxidant potential and low antioxidant intakes are associated with higher risk of cardiovascular disease and neurodegenerative disorders, which in turn lead to poor physical function (25, 26). However, antioxidants as supplements have been reported to have no protective effect in the prevention of myocardial infarction or stroke, and beta-carotene and vitamin A supplements may increase the risk of lung cancer and mortality (27, 28). Positive associations between snacking frequency and fruit, whole grains, and milk were reported in one recent study (18). Therefore, improvement of snacking behavior to increase intakes of fruits and vegetables containing large amounts of antioxidants should be encouraged to prevent age-related functional decline.

While many biological and environmental factors have been found to associate with poor physical function, our finding of an association between snacking and physical function is important from a public health perspective, because dietary behaviors are amenable to change. Snacking as a dietary intervention is under-studied among the older adult population (29). Studies are needed to understand more fully if snacking plays a role in physical function among older adults. Contrary to traditional opinions about the negative aspects of snacking behaviors, snacking can be associated with a more nutrient-dense diet with healthy food choices and prevent adverse health outcomes.

The primary limitation of this study is its cross-sectional design, which prevents establishing a causal relationship between snacking and gait speed. It is possible that poor gait speed induces a vicious cycle of reducing physical activity and deconditioning that will limit an individual’s access to snack food. In addition, the NHANES study does not include institutionalized individuals who probably have a higher prevalence of poor functional status than community-dwelling older adults. Furthermore, dietary patterns in this study were based on a single 24-hour dietary recall, which may be subject to under- or over-reporting (30). Under-reporting is a concern for foods and beverages consumed as snacks (31); however, the multiple-pass approach has been suggested as a way to improve under-reporting (32). Finally, a single 24-hour dietary recall may not capture participants’ usual dietary patterns.

This study fills an important gap in the literature by demonstrating that snacking is positively associated with gait speed, a performance-based measurement of physical function, using a nationally representative sample of older Americans. Recommendations to maintain physical function for older adults should include the provision of snack foods and beverages in their daily diet. Future work that evaluates longitudinally whether snacking patterns contribute to improved physical function is warranted. Other areas of research should address the influence that snacking has on other important health outcomes such as age-related diseases. Such investigations may provide greater evidence in support of nutritional interventions targeting older adults.

Conflict of interest

None of the authors have any conflicts of interest to declare.