Poor Appetite and Eating Difficulties Can Predict the Long-Term Risk of Falling: A Longitudinal Study in Middle-Aged and Older Adults

Abstract

Objectives:

This study investigates whether self-reported nutritional status affects falling among middle-aged and older adults.

Method:

We used 8-year follow-up data from the Taiwan Longitudinal Study on Aging. At baseline, respondents’ appetite, changes in amount of food intake, and eating difficulties were assessed in a questionnaire-based survey in addition to anthropometric measurements (body mass index, mid-arm circumferences, and involuntary body weight loss). Their associations with falls in the follow-up were examined using multivariable log-binomial regression.

Results:

The study included 2,519 respondents aged 50 years and older. Poor appetite (prevalence ratio [PR] = 1.25, 95% confidence interval [CI] = [1.07, 1.46]) and eating difficulties (PR = 1.16, 95% CI = [1.02, 1.32]) significantly predicted falling 8 years later with adjustments for sociodemographics, health behaviors, comorbidities, and anthropometric measures by taking into account probabilities of follow-up.

Conclusion:

Poor appetite and eating difficulties can predict falling in the long-term independent of anthropometric measurements among middle-aged and older adults.

Keywords

appetite food intake eating difficulties falls risk factor longitudinal methods nutrition

Introduction

Accidental falls among older adults are the leading cause of fatal and nonfatal injuries, emergency department visits, and hospital admissions in Australia, Canada, and the United Kingdom (Do et al., 2015; Hendrie et al., 2004; Scuffham et al., 2003). Among people aged 65 years and older in communities, 20% to 35% experience falls worldwide, and the incidence increases with aging populations (Peel, 2011; World Health Organization [WHO], 2008). In total, about 27 per 1,000 individuals’ disability-adjusted life years were lost due to hip fractures caused by falls based on data from six cohorts of participants aged 50 years or older from Europe and the United States (Papadimitriou et al., 2017).

The physical causes of falls are multiple, including deterioration in walking, balance, gait (Jia et al., 2019), chronic conditions, frailty (Cheng & Chang, 2017), and sarcopenia (Zhang et al., 2019). Malnutrition results in sarcopenia, frailty, and reduction in muscle strength, leading to difficulties in mobility and balance. Malnutrition’s role in accidental falls may be more fundamental than the decline in physical capacity and functions in older adults.

Recent studies have shown the impact of malnutrition on accidental falls among older adults, although the results among different studies are inconsistent. Low body mass index (BMI), unintentional weight loss, and reduced nutritional intake in older adults were found correlated with an increased risk of falls in residential long-term care (Neyens et al., 2013) and home care (Meijers et al., 2012). Malnutrition, defined by the composite measures of malnutrition screening tools such as Mini Nutritional Assessments (MNA), Malnutrition Screening Tool (MST), and Subjective Global Assessment (SGA), was associated with falls (Tsai & Lai, 2014; Vivanti et al., 2009). Improved nutritional status, measured by an MNA score increase, was associated with physical performance improvement during post-fall recovery among community-dwelling older adults (Conzade et al., 2019). However, no association was found between malnutrition and falls in a study of community-dwelling older adults due to low prevalence of malnutrition risk (Isenring et al., 2013).

In malnutrition screening, anthropometrics and self-reported indicators are used to evaluate nutritional status for older adults living in community and long-term care settings. Anthropometric indicators are more objective, measuring BMI, mid-arm circumference (MAC), or body weight, which are complex in measurement and calculation, and require evaluators with specialized training (Isenring et al., 2012). Self-reported indicators—subjectively assessed by patients or interviewees—include poor appetite or food intake and mastication impairment and are easier to obtain and timesaving, making them valuable markers in early detection of nutritional deficiency before the emergence of unfavorable sequelae.

Appetite, defined as the desire and motivation to acquire food, is associated with sight, smell, and taste and comprises sensations including hunger, satiation, and satiety, and can result in food intake (Mattes et al., 2005). The aging mechanisms responsible for appetite decline, and food intake are complex; well-known risk factors include functional impairments, chronic diseases, medication use, socioeconomic inequality (Landi et al., 2016), and oral health-related infirmities (Hung et al., 2019). Anorexia in aging, along with reduced energy intake that cannot meet nutrient needs, can result in malnutrition and weight loss (Giezenaar et al., 2016). In addition, anorexia is a component that causes aging-related sarcopenia that mediates poor intake and falls (Morley, 2017). Therefore, detecting self-reported poor intake before weight loss and nutritional deficiencies will allow intervention at an early stage, preventing health declines.

In previous research, the prediction of falls and nutrition status were based on malnutrition screening tools such as MNA (Conzade et al., 2019; Tsai & Lai, 2014), MNA-Short Form (Isenring et al., 2013; Tsai & Lai, 2014), SGA (Vivanti et al., 2009), or other malnutrition criteria (Meijers et al., 2012; Neyens et al., 2013), combining both anthropometric and self-report measurements; furthermore, most had a cross-sectional design (Isenring et al., 2013; Neyens et al., 2013; Vivanti et al., 2009). However, limited quantitative cohort studies exist focusing on self-reported nutritional status changes, such as appetite impairment and food intake decline, and their association with falls. Only a qualitative report found that recurrent fallers showed poor appetite (Mahler & Sarvimäki, 2012). This study aimed to investigate associations between self-reported nutritional measurements and falls in long-term follow-up in middle-aged and older adults using a representative cohort of the Taiwanese population. The hypothesis is that self-reported indicators could predict long-term falls independent of anthropometric measures. We focused on self-assessed appetite impairment, food intake reduction, and eating difficulty as self-reported nutritional indicators and examined their associations with falls, adjusting for anthropometric measures.

Method

Data

The data were extracted from two of seven survey rounds from the Taiwan Longitudinal Study on Aging (TLSA) in 1999 and 2007. TLSA comprises a nationally representative sample of the population conducted by the Health Promotion Administration of Taiwan. The sampling design was a three-stage equal probability sample. In the first stage, 56 of 331 primary sampling units (townships) were randomly selected across 27 strata of Taiwanese administrative areas. In addition, blocks within the selected townships, serving as clusters, were sampled for probabilities proportional to the townships’ sizes; finally, two respondents aged 60 years or older were systematically sampled from each selected block. The surveys investigated socioeconomic and health statuses of older adults through face-to-face interviews, and seven follow-up interviews were conducted between 1989 and 2011. In Round 1 (1989), 4,049 respondents completed the interview survey. During Round 3 (1996), a supplementary sample of 2,462 new participants was recruited to represent the 50- to 66-year-old population. In the Round 4 (1999), a range of questions were introduced to query participant eating behaviors including appetite and food intake. Fall experiences, as the primary outcome, were collected from Round 6 (2007). Other participants’ characteristics at baseline for controlling for potential confounding were obtained from Round 4. Additional details concerning the TLSA data are described in previous studies (Hsu & Jones, 2012; Leng & Wang, 2013).

Measurement

Self-reported nutritional indicators

There are three self-reported nutrition-related questions in the Round 4 questionnaire: subjective appetite was assessed by asking, “In general, how is your appetite?” (normal/sometimes poor/poor); changed food intake was assessed by asking, “Has there been an obvious increase or decrease in your food intake in recent days?” (no change/obvious increase/obvious decrease); and a question about self-assessed eating ability, following the question on respondents’ use of dentures, asked “How is your ability to eat?” (very good/good/fair/not good/very bad) regardless of whether the respondent was dentulous, edentulous, or wearing dentures. Considering dental status affect chewing and swallowing functions (Naka et al., 2014; Okamoto et al., 2012), respondents were prone to assess their ability to eat regarding chewing and swallowing in this context. For the appetite question, responses with “sometimes poor” were combined with “poor” because a change in appetite might represent a change in physiological or psychological status that we were interested in. The responses with an obvious increase in food intake were combined with those without any change, making them a counterpart to the group with reduced amount of food intake. We defined having eating difficulties as responses of “not good” or “very bad” to the eating ability question.

Anthropometric measurements

Three anthropometric measures for malnutrition were (a) MAC, (b) BMI, and (c) involuntary body weight loss. We employed cutoff points of MAC (<22.5 cm for men and <21 cm for women) and BMI (<18.5, 18.5–23.9, 24–26.9, ≥27 kg/m² as suggested by the Taiwan Health Promotion Administration, 2020), validated by a previous study (Tsai et al., 2007). Meanwhile, involuntary body weight loss was measured by asking, “Did you lose body weight more than 3 kilograms during the last year?” (yes/no).

Falls

In Round 6, falls were self-reported in response to the following question: “Have you experienced any falls in the past year? (e.g., falls or slips when walking, sitting, standing, lying down, or collapses due to dizziness, regardless of an injury)” (yes/no). Falls in Round 6 was the current study’s primary outcome.

Covariates

Other covariates included age (50–64, 65–74, 75+ years), sex (female/male), education years (>6/≤6), and marital status (i.e., married or cohabitation). Health-related behavior was measured using the following variables: current smoking status (yes/no) and alcohol consumption (yes/no). Chronic health conditions including hypertension, diabetes, heart disease, stroke, cancer, lung disease, arthritis, and hepatobiliary disease were recorded via self-report, and the number of these conditions were summed up to reflect comorbidities categorized at three levels: 0, 1–2, and ≥3. These potential confounders correlate to both nutritional status and falls in older adults (Deandrea et al., 2010; Faulkner et al., 2009; Van Bokhorst-De Van Der Schueren et al., 2013). As the decline or impairment in physical function may mediate the pathway between nutritional factors and falls, in the primary analysis—to avoid over-adjustment—functional measurements, such as activities of daily life (ADLs) or instrumental activities of daily life (IADLs), were not used as covariates to control for confounding (Schisterman et al., 2009). Physical activity—a descending proxy of physical function—and depression—a potential consequence of malnutrition—were not included in the main analysis for the same reason.

Statistical Analysis

In descriptive statistics, we used frequencies and percentages for categorical variables including all explanatory variables and falls. In the analytic statistics, log-binomial regressions were used to examine the associations between explanatory variables at baseline and falls 8 years later for two reasons. First, falls are common outcomes in the study cohort (22%). To evaluate the relative risk of common outcomes between exposure and non-exposure groups, the log-binomial model is preferred over other models for binary outcomes such as the logistic model (Skove et al., 1998). Second, the results of the log-binomial model are easier to interpret in policymaking and communications (Reichenheim & Coutinho, 2010; Thompson et al., 1998).

The analytic strategy comprised three main models. First, unadjusted models, estimated using log-binomial regressions, examined associations between each explanatory variable and falls without any adjustment. Next, multivariable log-binomial regressions were applied to assess the effects of three self-reported nutritional indicators in Round 4, including poor appetite, reduced food intake, and eating difficulty, on falls in Round 6 while adjusting for potential confounders such as sociodemographic variables, smoking, alcohol consumption, and comorbidities; we also employed parallel models to assess the anthropometric measurements’ effects, instead of self-reported nutritional indicators, on falls. Through a comparison of separate models with self-reported and anthropometric predictors, we could assess their significance in predicting long-term falls. Finally, we used a full model including self-rated nutritional indicators, anthropometric measurements, and potential confounders to understand whether self-rated nutritional indicators can predict falls independent of anthropometric measurement. Our approach included all three self-reported indicators in the adjusted models, which explored the independent effect of every single indicator on falls with adjustments for the other two indicators and potential confounders. Prevalence ratios (PRs) with 95% confidence intervals (CIs) and p values for chi-squared tests represented the effects of predictors on falls through log-binomial regressions.

At follow-up after 8 years, malnutrition and other factors determining falls may have a common effect, that is, death, the main cause of loss to follow-up. The observed association between malnutrition and falls was conditioned on survival, which caused selection bias. The inverse probability weighting (IPW) method—a two-step process—can be used to control for selection bias due to loss to follow-up (Howe et al., 2016). First, we calculated the probability that each participant was selected (successfully followed up) in the study. Second, weights, the inverse of the probability of being selected, were calculated and implemented in the full model.

The statistical analysis was performed using SAS, version 9.4 (SAS Institute, Inc. Cary, NC, USA). The p value < .05 was considered significant. The need to obtain informed consent was waived by the institutional review board due to the use of an anonymous and de-identified database.

Results

The descriptive statistics of study variables are summarized in Table 1. In Round 4, 4,440 respondents aged 50 years or older completed interviews. After 8 years, 1,526 participants were lost to follow-up due to death (n = 1,475) or other reasons (n = 51) and 395 provided incomplete responses, leaving a total of 2,519 respondents (1,283 men and 1,236 women) for analysis. Of 2,519 participants, 11.0% reported poor appetite, 6.0% perceived reductions in food intake, and 18.7% had eating difficulties. Half the sample had a normal BMI at baseline, followed by overweight (29.3%); only 4.2% were underweight. Very few (2.4%) had MACs smaller than the cutoff point (22.5 cm for men and 21 cm for women). One of 10 lost body weight more than 3 kg during the past year. The prevalence of any falls in the past year was 22.3% in Round 6.

Table 1.

Characteristics of Participants of the Taiwan Longitudinal Study on Aging at Round 4.

Variable	N	%
Sex
Male	1,283	50.9
Female	1,236	49.1
Age (years)
<65	1,148	45.6
65–74	952	37.8
≥75	419	16.6
Smoking	678	26.9
Alcohol drinking	165	6.6
Married	1,866	74.1
Education >6 years	623	24.7
Comorbidity number
0	1,081	42.9
1–2	1,179	46.8
≥3	259	10.3
Self-reported nutritional factors
Poor appetite	277	11.0
Reduced food intake	151	6.0
Eating difficulty	471	18.7
Anthropometric measures
Body mass index (BMI)
BMI < 18.5	105	4.2
18.5 ≤ BMI < 24	1,283	50.9
24 ≤ BMI < 27	739	29.3
27 ≤ BMI < 45	392	15.6
Mid-arm circumference <22.5 cm (male) or <21 cm (female)	60	2.4
Body weight loss >3 kg/year	267	10.6

Table 2 presents the results of log-binomial analysis. Without adjustments, an increased risk of any falls in the past year in Round 6 was significantly associated with poor appetite (PR = 1.11, 95% CI = [1.06, 1.17]) and eating difficulties (PR = 1.09, 95% CI = [1.04, 1.13]) at baseline, but not with reduced food intake (PR = 1.05, 95% CI = [0.98, 1.13]). The associations were not significant between falls and anthropometrics such as BMI, low MAC, and body weight loss.

Table 2.

Results of Log-Binomial Regression Models of the Relationship Between Nutritional Factors at Round 4 and Falls at Round 6 Among Participants of the Taiwan Longitudinal Study on Aging.

N = 2,519	Unadjusted model		Adjusted model 1		Adjusted model 2		Adjusted model 3
N = 2,519	PR	95% CI	PR	95% CI	PR	95% CI	PR	95% CI
Female	1.08	[1.04, 1.11]***	1.22	[1.03, 1.45]*	1.25	[1.06, 1.49]**	1.22	[1.03, 1.45]*
Age (years)
<65	Ref.		Ref.		Ref.		Ref.
65–74	1.07	[1.03, 1.11]***	1.27	[1.07, 1.50]**	1.30	[1.10, 1.54]**	1.27	[1.07, 1.50]**
≥75	1.10	[1.05, 1.15]***	1.31	[1.06, 1.61]*	1.35	[1.10, 1.67]**	1.31	[1.06, 1.62]*
Alcohol drinking	0.98	[0.92, 1.05]	0.93	[0.68, 1.26]	0.94	[0.69, 1.28]	0.94	[0.69, 1.28]
Smoking	0.96	[0.93, 1.00]	1.06	[0.87, 1.28]	1.06	[0.87, 1.28]	1.10	[0.93, 1.30]
Married	0.93	[0.90, 0.97]	0.91	[0.78, 1.08]	0.90	[0.76, 1.06]	0.91	[0.77, 1.07]
Education >6 years	0.93	[0.90, 0.97]***	0.81	[0.66, 0.99]*	0.80	[0.65, 0.97]*	0.82	[0.67, 1.00]
Comorbidity number
0	Ref.		Ref.		Ref.		Ref.
1–2	1.07	[1.03, 1.11]***	1.25	[1.06, 1.48]**	1.27	[1.07, 1.49]**	1.24	[1.05, 1.47]*
≥3	1.13	[1.07, 1.19]***	1.41	[1.12, 1.78]**	1.48	[1.18, 1.87]***	1.40	[1.11, 1.78]**
Self-reported nutritional factors
Poor appetite	1.11	[1.06, 1.17]***	1.24	[1.01, 1.52]*			1.25	[1.01, 1.54]*
Reduced food intake	1.05	[0.98, 1.13]	0.93	[0.70, 1.24]			0.92	[0.69, 1.24]
Eating difficulty	1.09	[1.04, 1.13]***	1.19	[1.00, 1.41]*			1.19	[1.01, 1.41]*
Anthropometric measures
Body mass index (BMI)
BMI < 18.5	1.02	[0.94, 1.10]			1.07	[0.74, 1.54]	1.00	[0.69, 1.45]
18.5 ≤ BMI < 24	Ref.				Ref.		Ref.
24 ≤ BMI < 27	0.99	[0.95, 1.03]			0.93	[0.79, 1.10]	0.94	[0.79, 1.11]
27 ≤ BMI < 45	1.02	[0.97, 1.07]			0.99	[0.81, 1.21]	1.02	[0.83, 1.25]
Mid-arm circumference <22.5 cm (male) or <21 cm (female)	0.93	[0.83, 1.03]			0.60	[0.32, 1.13]	0.60	[0.32, 1.12]
Body weight loss >3 kg/year	1.05	[0.99, 1.10]			1.12	[0.90, 1.39]	1.09	[0.87, 1.36]

Note. PR = prevalence ratio; CI = confidence interval.

p < .05. **p < .01. ***p < .001.

In the multivariate analysis, participants with poor appetite and eating difficulties were more likely to fall in Round 6, when adjusting for sociodemographics, health-related behaviors, and comorbidities (Model 1: PR = 1.24, 95% CI = [1.01, 1.52] in poor appetite; PR = 1.19, 95% CI = [1.00, 1.41] in eating difficulties, respectively); in contrast, reduced food intake had no association with falls 8 years later (PR = 0.93, 95% CI = [0.70, 1.24]). In Model 2, three anthropometric measurements were used to replace self-reported nutritional factors. The results revealed that BMI (reference: 18.5 ≤ BMI < 24; BMI < 18.5: PR = 1.07, 95% CI = [0.74, 1.54]; 24 ≤ BMI < 27: PR = 0.93, 95% CI = [0.79, 1.10]; BMI > 27: PR = 0.99, 95% CI = [0.81, 1.21]), low MAC (PR = 0.60, 95% CI = [0.32, 1.13]), and body weight loss (PR = 1.12, 95% CI = [0.90, 1.39]) were not associated with falls in Round 6 with adjustments for covariates. In Model 3, after adjusting for anthropometric measurements and other covariates, a 25% increase in any falls in Round 6 was associated with poor appetite versus normal appetite (PR = 1.25, 95% CI = [1.01, 1.54]) and a 19% increase in that with eating difficulties versus no eating difficulties (PR = 1.19, 95% CI = [1.01, 1.41]). Again, reduced food intake had no significant effect on falls in Round 6 (PR = 0.92, 95% CI = [0.69, 1.24]). The participants’ characteristics in Rounds 4 and 6 were different due to loss to follow-up (Supplementary Table S3). In the IPW analysis, after adjusting for participants’ probabilities of being observed in Round 6, the effect of poor appetite on falls remained (PR = 1.25, 95% CI = [1.07, 1.46]), and that of eating difficulties was slightly attenuated (PR = 1.16, 95% CI = [1.02, 1.32]).

We used a sensitivity analysis to explore the role of depression between self-reported nutritional factors and falls, including depressive symptoms in the models (Supplementary Table S1). The results were consistent with our primary analysis of the impact of poor appetite (full model without IPW: PR = 1.24, 95% CI = [0.99, 1.55], p = .066; full model with IPW: PR = 1.23, 95% CI = [1.04, 1.44], p = .013) and eating difficulty (full model without IPW: PR = 1.19, 95% CI = [1.00, 1.41], p = .051; full model with IPW: PR = 1.16, 95% CI = [1.01, 1.31], p = .029) on falls.

However, the self-report indicators may be inherently related, raising the issue of multicollinearity. Results from separate models with single self-reported indicators, after adjusting for anthropometric measurements and potential confounders, revealed similar coefficient estimates as those in Model 3 (Supplementary Table S4).

Discussion

This is the first study to quantify the relationship between self-reported nutrition status and long-term falling events in older adults, and it reveals that people with appetite loss and eating difficulties—rather than decreasing food intake—were prone to falling in the long-term follow-up even after adjusting for sociodemographic factors, health behaviors, chronic diseases, and anthropometric measures. This finding supports that self-reported nutritional indicators are determinants of falls in the long-term independent of anthropometric measures.

The associations between appetite impairment and falling are multifactorial and include physical and psychological factors (Mahler & Sarvimäki, 2012). Unlike normal appetites, decreased appetites were more prevalent in older people, females, those less educated, single people, and those with higher comorbidities in this study, which is consistent with previous findings (Landi et al., 2016; Pilgrim et al., 2015). Previous research had mentioned anorexia in aging as the cause of malnutrition and sarcopenia—composed of poor physical function, loss of muscle strength and muscle mass—and the principal determinant of falling (Morley, 2017; Zhang et al., 2019). Hip-fracture patients were more likely to suffer from undernutrition before a falling accident (Lumbers et al., 2001). Furthermore, pro-inflammatory cytokines secretions in acute and chronic inflammation lead to appetite changes (Langhans, 2007), and such inflammation-related diseases, including circulatory diseases, lung diseases, and arthritis, have been linked with increased risks of falling (Lawlor et al., 2003).

There were considerable differences between appetite and food intake. While food intake correlated with consumption volume, appetite reflects more on general eating patterns, including food variety and daily mealtimes (Wilson et al., 2005). For example, senile appetite decline was characterized by the shortened diet variety (reduced consumption of “meats, fishes, and eggs” and “fruit and vegetables,” while milk and cereals were unaffected) in the Italian population (Donini et al., 2013). Compared with reduced food intake amount, poor appetite may be a more comprehensive and better indicator of nutrition status. Based on further analysis of when appetite and food intake were all presented in models predicting falls 8 years later (Table 2, Adjusted Models 1 and 3), our study revealed that appetite impairment was an independent and predominant predictor of falls in the long-term even when including food intake status (PR = 1.25, 95% CI = [1.01, 1.54]).

Contrarily, older adults reporting reduced food intake showed no significant difference in falls during follow-up (PR = 0.92, 95% CI = [0.69, 1.24]). A possible reason for this result is that people often eat beyond their nutritional needs when food is widely available, leading to obesity and cardiometabolic risks (Pan et al., 2011). Hence, intention to reduce food intake, such as caloric restriction, has beneficial effects on preventing aging-associated chronic diseases (López-Lluch & Navas, 2016; Walford et al., 2002; Willcox et al., 2009). In general, decreased food intake may represent a protective factor for metabolic syndromes and cardiovascular diseases and thus, subsequently offset its harmful effect on falling (Laudisio et al., 2017; Liao et al., 2012).

Eating difficulties may be due to participants’ impaired chewing capacity. Chewing difficulty and tooth loss were determinants of lower food variety and insufficient food intake (Kimura et al., 2013), resulting in less muscle mass and compromised physical capacity (Yokoyama et al., 2016). Some studies show that chewing capacity is related to sarcopenia (Murakami et al., 2015), which is a strong risk factor of falling in older people (Landi et al., 2012; Visser & Schaap, 2011; Zhang et al., 2019). In our study, eating difficulties, generally self-assessed by a participant, were associated with long-term risks of falls. Our findings echoed the observed association between impaired masticatory or swallowing functions and malnutrition. In other words, oral function of occlusal force, tongue motor function, and tongue pressure, along with other bodily functions, deteriorate with aging, which might affect eating dependencies (Iyota et al., 2020).

Nutrition has a significant effect on the onset, severity, and duration of depression (Rao et al., 2008); the latter is one risk factor for falls among older adults (Briggs et al., 2018; Whooley, 1999). Compromised nutritional status, such as reduced appetite, skipping meals, or decreased chewing ability, may precede or co-exist with depression (Hwang et al., 2013; Rao et al., 2008). In a sensitivity analysis, as shown in Supplementary Table S1, the significance of their effects becomes marginal in the full model without IPW adjustments, whereas the effects remain significant in the full model with IPW adjustments. Because the effect of depression itself on falls was not significant, we speculate that depression has no intermediary role between self-reported nutritional factors and falls.

It is beneficial to identify malnutrition early in clinical practice based simply on self-reported poor intake. With early signs of malnutrition, compromised appetite, and eating difficulties, certain nutrition interventions such as food assistance programs, food consistency modification, and dietary supplements addressed in previous research can offer nutritional benefits and reduce the fall risks for older adults in the early stages of malnutrition (Esquivel, 2018). For example, early initiatives with vitamin D supplements along with exercise and physical therapy are recommended to prevent falls for community-dwelling older adults by the U.S. Preventive Services Task Force (USPSTF; Moyer & U.S. Preventive Services Task Force, 2012).

The strength of this study was its investigation into the relationship between self-assessed nutritional indicators on falling based on a large and nation-wide sample of the middle-aged and older adults of the Taiwanese population. Second, our study distinguishes self-reported nutritional factors from anthropometric measurements. Previous studies often adopted composite measurements to predict health outcomes. In addition, we employed a longitudinal analysis, which included data from a follow-up after 8 years; these data allowed us to identify the cause-effect relationships of interest.

This study had several limitations. First, appetite, food intake, and eating difficulties were measured only at baseline: Any declines or increases after the baseline may lead to or be shaped by physical changes throughout the follow-up period. Along these lines, the data did not fully capture the effects of the self-reported nutritional factors on falling. In addition, falling experiences were queried solely for a 1-year duration. Fall events during periods not surveyed may have been influenced by baseline nutritional status, yet these outcomes were not considered. To understand whether substantial physical function changes occurred between the baseline and the 8-year follow-up, we investigated whether participants had ADL and IADL declines and self-reported new hip fractures (a surrogate of severe falls) within the period (Supplementary Table S2). Compared with their counterparts, participants with poor appetite, reduced food intake, or eating difficulty had a higher proportion of IADL and ADL decline over this period, with eating difficulty related to subsequent hip fractures. By looking at changes in functional status and the relationship with nutritional status, we can recognize that the self-report nutritional factors at baseline were associated with changes in function over the 8-year period, which could be one of the potential explanations for falls. However, this functional change could also be a consequence of falls, and we must be cautious in interpreting the relationship of functional change between nutritional status and falls. Further research is warranted to clarify the relationship between them.

Second, 1,475 participants were lost to follow-up during the 8 years; a majority of them (96.7%) were due to death. The deceased who were lost to follow-up had poorer nutritional indicators, including self-reported and anthropometric measurements (Supplementary Table S3); thus, the estimates of the effects of poor nutritional indicators on falling—both objective and subjective—may be underestimated in the long-term analysis. Nevertheless, IPW models accounting for participants lost to follow-up revealed a robust analysis of results (Table 3). Finally, factors confounding the relationship of interest were not fully adjusted for. For example, environmental and neighborhood issues may play a significant role on this relationship (Chung et al., 2012; Nicklett et al., 2017). Given the lack of accurate residency (even at the township level), we cannot link TLSA participants to their neighborhood localities. More detailed data with zip codes may allow us to evaluate the influence of living environments on appetite and falling in participants.

Table 3.

Results of Log-Binomial Regression Models by Inverse Probability Weighting of the Relationship Between Nutritional Factors at Round 4 and Falls at Round 6 Among Participants of the Taiwan Longitudinal Study on Aging.

Variable	IPW model^a
Variable	PR	95% CI
Female	1.15	[1.00, 1.31]*
Age (years)
<65	Ref.
65–74	1.26	[1.09, 1.46]**
≥75	1.30	[1.11, 1.53]**
Alcohol drinking	1.06	[0.91, 1.23]
Smoking	0.81	[0.62, 1.05]
Married	1.12	[0.99, 1.27]
Education >6 years	0.82	[0.70, 0.96]*
Comorbidity number
0	Ref.
1–2	1.28	[1.11, 1.47]***
≥3	1.51	[1.26, 1.79]***
Self-reported nutritional factors
Poor appetite	1.25	[1.07, 1.46]**
Reduced food intake	0.94	[0.77, 1.16]
Eating difficulty	1.16	[1.02, 1.32]*
Anthropometric measures
Body mass index (BMI)
BMI <18.5	0.85	[0.66, 1.11]
18.5 ≤ BMI < 24	Ref.
24 ≤ BMI < 27	0.91	[0.80, 1.04]
27 ≤ BMI < 45	1.01	[0.85, 1.20]
Mid-arm circumference <22.5 cm (male) or <21 cm (female)	0.87	[0.58, 1.30]
Body weight loss >3 kg/year	1.15	[0.98, 1.35]

Note. IPW = inverse probability weighting; PR = prevalence ratio; CI = confidence interval.

Adjusted for variables the same as Adjusted Model 3 in Table 2.

p < .05. **p < .01. ***p < .001.

Conclusion

Significant differences found in the long-term risk of falling by appetite impairment and eating difficulties may reflect a predominant effect independent of objective nutritional indices. These subjective assessments, which are quick, easy-to-use, and valuable, may help in the early identification of a high-risk group for falling in middle-aged and older people. However, further research will be needed to investigate the beneficial effects of interventions that improve appetite and eating capacity.

Supplemental Material

sj-pdf-1-jag-10.1177_0733464820976439 – Supplemental material for Poor Appetite and Eating Difficulties Can Predict the Long-Term Risk of Falling: A Longitudinal Study in Middle-Aged and Older Adults

Supplemental material, sj-pdf-1-jag-10.1177_0733464820976439 for Poor Appetite and Eating Difficulties Can Predict the Long-Term Risk of Falling: A Longitudinal Study in Middle-Aged and Older Adults by Yu-Chun Lin and Yu-Hung Chang in Journal of Applied Gerontology

Footnotes

Acknowledgements

This study used data from the Taiwan Longitudinal Study on Aging (TLSA) which was provided and managed by the Health and Welfare Data Science Center (HWDC), Ministry of Health and Welfare (), Taiwan. The conclusions expressed here do not represent those of the administration. The authors gratefully acknowledge Ms Ya-Chun Hung and Ms Yi-Fung Lin who provided early processing of data.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the China Medical University (grant number CMU102-N-07) and China Medical University Hospital (grant number DMR-109-006).

IRB Protocol/Human Subjects Approval Numbers

This study was approved by the China Medical University & Hospital Research Ethics Committee, Taichung, Taiwan (CMUH103-REC3-057). Data collection, data analysis, and manuscript writing comply with the current laws of the country in which they were performed.

ORCID iDs

Yu-Chun Lin

Yu-Hung Chang

Supplemental Material

Supplemental material for this article is available online.

References

Briggs

Kennelly

S. P.

Kenny

R. A.

(2018). Does baseline depression increase the risk of unexplained and accidental falls in a cohort of community-dwelling older people? Data from The Irish Longitudinal Study on Ageing (TILDA). International Journal of Geriatric Psychiatry, 33(2), e205–e211.

Cheng

M. H.

Chang

S. F.

(2017). Frailty as a risk factor for falls among community dwelling people: Evidence from a meta-analysis. Journal of Nursing Scholarship, 49(5), 529–536.

Chung

W. T.

Gallo

W. T.

Giunta

Canavan

M. E.

Parikh

N. S.

Fahs

M. C.

(2012). Linking neighborhood characteristics to food insecurity in older adults: The role of perceived safety, social cohesion, and walkability. Journal of Urban Health, 89(3), 407–418.

Conzade

Phu

Vogrin

Bani Hassan

Sepúlveda-Loyola

Thorand

Duque

(2019). Changes in nutritional status and musculoskeletal health in a geriatric post-fall care plan setting. Nutrients, 11(7), Article 1551.

Deandrea

Lucenteforte

Bravi

Foschi

La Vecchia

Negri

(2010). Risk factors for falls in community-dwelling older people. Epidemiology, 21(5), 658–668.

M. T.

Chang

V. C.

Kuran

Thompson

(2015). Fall-related injuries among Canadian seniors, 2005–2013: An analysis of the Canadian Community Health Survey. Health Promotion and Chronic Disease Prevention in Canada, 35(7), 99–108.

Donini

L. M.

Poggiogalle

Piredda

Pinto

Barbagallo

Cucinotta

Sergi

(2013). Anorexia and eating patterns in the elderly. PLOS ONE, 8(5), Article e63539.

Esquivel

M. K.

(2018). Nutritional assessment and intervention to prevent and treat malnutrition for fall risk reduction in elderly populations. American Journal of Lifestyle Medicine, 12(2), 107–112.

Faulkner

K. A.

Cauley

J. A.

Studenski

S. A.

Landsittel

D. P.

Cummings

S. R.

Ensrud

K. E.

Donaldson

M. G.

Nevitt

M. C.

(2009). Lifestyle predicts falls independent of physical risk factors. Osteoporosis International, 20(12), 2025–2034.

10.

Giezenaar

Chapman

Luscombe-Marsh

Feinle-Bisset

Horowitz

Soenen

(2016). Ageing is associated with decreases in appetite and energy intake—A meta-analysis in healthy adults. Nutrients, 8(1), Article 28.

11.

Hendrie

Hall

S. E.

Arena

Legge

(2004). Health system costs of falls of older adults in Western Australia. Australian Health Review, 28(3), 363–373.

12.

Howe

C. J.

Cole

S. R.

Lau

Napravnik

Eron

J. J.

(2016). Selection bias due to loss to follow up in cohort studies. Epidemiology, 27(1), 91–97.

13.

Hsu

H. C.

Jones

B. L.

(2012). Multiple trajectories of successful aging of older and younger cohorts. Gerontologist, 52(6), 843–856.

14.

Hung

Wijnhoven

H. A. H.

Visser

Verbeke

(2019). Appetite and protein intake strata of older adults in the European Union: Socio-demographic and health characteristics, diet-related and physical activity behaviours. Nutrients, 11(4), Article 777.

15.

Hwang

S. H.

Park

S. G.

Min

J. Y.

(2013). Relationship between chewing ability and depressive symptoms. Community Dental Health, 30(4), 254–256.

16.

Isenring

E. A.

Baker

Kerr

(2013). Malnutrition and falls risk in community-dwelling older adults. The Journal of Nutrition, Health and Aging, 17(3), 277–279.

17.

Isenring

E. A.

Banks

Ferguson

Bauer

J. D.

(2012). Beyond malnutrition screening: Appropriate methods to guide nutrition care for aged care residents. Journal of the Academy of Nutrition and Dietetics, 112(3), 376–381.

18.

Iyota

Mizutani

Oku

Asao

Futatsuki

Inoue

Imai

Kashiwazaki

(2020).A cross-sectional study of age-related changes in oral function in healthy Japanese individuals. International Journal of Environmental Research and Public Health, 17(4), Article 1376.

19.

Jia

Lubetkin

E. I.

DeMichele

Stark

D. S.

Zack

M. M.

Thompson

W. W.

(2019). Prevalence, risk factors, and burden of disease for falls and balance or walking problems among older adults in the U.S. Preventive Medicine, 126, Article 105737.

20.

Kimura

Ogawa

Yoshihara

Yamaga

Takiguchi

Wada

. . . Matsubayashi

(2013). Evaluation of chewing ability and its relationship with activities of daily living, depression, cognitive status and food intake in the community-dwelling elderly. Geriatrics & Gerontology International, 13(3), 718–725.

21.

Landi

Calvani

Tosato

Martone

A. M.

Ortolani

Savera

Sisto

Marzetti

(2016). Anorexia of aging: Risk factors, consequences, and potential treatments. Nutrients, 8(2), 69–69.

22.

Landi

Liperoti

Russo

Giovannini

Tosato

Capoluongo

Bernabei

Onder

(2012). Sarcopenia as a risk factor for falls in elderly individuals: Results from the ilSIRENTE study. Clinical Nutrition, 31(5), 652–658.

23.

Langhans

(2007). Signals generating anorexia during acute illness. Proceedings of the Nutrition Society, 66(3), 321–330.

24.

Laudisio

Lo Monaco

M. R.

Vetrano

D. L.

Pisciotta

M. S.

Bentivoglio

A. R.

Bernabei

Zuccalà

(2017). Association of metabolic syndrome with falls in patients with Parkinson’s disease. Clinical Nutrition, 36(2), 559–563.

25.

Lawlor

D. A.

Patel

Ebrahim

(2003). Association between falls in elderly women and chronic diseases and drug use: Cross sectional study. BMJ, 327(7417), 712–717.

26.

Leng

C. H.

Wang

J. D.

(2013). Long term determinants of functional decline of mobility: An 11-year follow-up of 5464 adults of late middle aged and elderly. Archives of Gerontology and Geriatrics, 57(2), 215–220.

27.

Liao

K. C.

S. J.

Lin

C. H.

Chang

H. J.

Chen

Y. J.

Liu

M. S.

(2012). Association between the metabolic syndrome and its components with falls in community-dwelling older adults. Metabolic Syndrome and Related Disorders, 10(6), 447–451.

28.

López-Lluch

Navas

(2016). Calorie restriction as an intervention in ageing. The Journal of Physiology, 594(8), 2043–2060.

29.

Lumbers

New

S. A.

Gibson

Murphy

M. C.

(2001). Nutritional status in elderly female hip fracture patients: Comparison with an age-matched home living group attending day centres. British Journal of Nutrition, 85(6), 733–740.

30.

Mahler

Sarvimäki

(2012). Appetite and falls: Old age and lived experiences. International Journal of Qualitative Studies on Health and Well-Being, 7(1), 11540.

31.

Mattes

R. D.

Hollis

Hayes

Stunkard

A. J.

(2005). Appetite: Measurement and manipulation misgivings. Journal of the American Dietetic Association, 105(5 Suppl 1), S87–S97.

32.

Meijers

J. M.

Halfens

R. J.

Neyens

J. C.

Luiking

Y. C.

Verlaan

Schols

J. M.

(2012). Predicting falls in elderly receiving home care: The role of malnutrition and impaired mobility. The Journal of Nutrition, Health and Aging, 16(7), 654–658.

33.

Morley

J. E.

(2017). Anorexia of ageing: A key component in the pathogenesis of both sarcopenia and cachexia. Journal of Cachexia, Sarcopenia and Muscle, 8(4), 523–526.

34.

Moyer

V. A.

, & U.S. Preventive Services Task Force. (2012). Prevention of falls in community-dwelling older adults: U.S. Preventive Services Task Force recommendation statement. Annals of Internal Medicine, 157(3), 197–204.

35.

Murakami

Hirano

Watanabe

Sakai

Kim

Katakura

(2015). Relationship between chewing ability and sarcopenia in Japanese community-dwelling older adults. Geriatrics & Gerontology International, 15(8), 1007–1012.

36.

Naka

Anastassiadou

Pissiotis

(2014). Association between functional tooth units and chewing ability in older adults: A systematic review. Gerodontology, 31(3), 166–177.

37.

Neyens

Halfens

Spreeuwenberg

Meijers

Luiking

Verlaan

Schols

(2013). Malnutrition is associated with an increased risk of falls and impaired activity in elderly patients in Dutch residential long-term care (LTC): A cross-sectional study. Archives of Gerontology and Geriatrics, 56(1), 265–269.

38.

Nicklett

Lohman

Smith

(2017). Neighborhood environment and falls among community-dwelling older adults. International Journal of Environmental Research and Public Health, 14(2), Article 175.

39.

Okamoto

Tomioka

Saeki

Iwamoto

Morikawa

Harano

Kurumatani

(2012). Relationship between swallowing problems and tooth loss in community-dwelling independent elderly adults: The Fujiwara-kyo study. Journal of the American Geriatrics Society, 60(5), 849–853.

40.

Pan

W. H.

H. J.

Yeh

C. J.

Chuang

S. Y.

Chang

H. Y.

Yeh

N. H.

Hsieh

Y.-T.

(2011). Diet and health trends in Taiwan: Comparison of two nutrition and health surveys from 1993–1996 and 2005–2008. Asia Pacific Journal of Clinical Nutrition, 20(2), 238–250.

41.

Papadimitriou

Tsilidis

K. K.

Orfanos

Benetou

Ntzani

E. E.

Soerjomataram

. . .Trichopoulou

(2017). Burden of hip fracture using disability-adjusted life-years: A pooled analysis of prospective cohorts in the CHANCES consortium. Lancet Public Health, 2(5), e239–e246.

42.

Peel

N. M.

(2011). Epidemiology of falls in older age. Canadian Journal on Aging, 30(1), 7–19.

43.

Pilgrim

A. L.

Robinson

S. M.

Sayer

A. A.

Roberts

H. C.

(2015). An overview of appetite decline in older people. Nursing Older People, 27(5), 29–35.

44.

Rao

T. S.

Asha

M. R.

Ramesh

B. N.

Rao

K. S.

(2008). Understanding nutrition, depression and mental illnesses. Indian Journal of Psychiatry, 50(2), 77–82.

45.

Reichenheim

M. E.

Coutinho

E. S.

(2010). Measures and models for causal inference in cross-sectional studies: Arguments for the appropriateness of the prevalence odds ratio and related logistic regression. BMC Medical Research Methodology, 10(1), Article 66.

46.

Schisterman

E. F.

Cole

S. R.

Platt

R. W.

(2009). Overadjustment bias and unnecessary adjustment in epidemiologic studies. Epidemiology, 20(4), 488–495.

47.

Scuffham

Chaplin

Legood

(2003). Incidence and costs of unintentional falls in older people in the United Kingdom. Journal of Epidemiology and Community Health, 57(9), 740–744.

48.

Skove

Deddens

Petersen

M. R.

Endahl

(1998). Prevalence proportion ratios: Estimation and hypothesis testing. International Journal of Epidemiology, 27(1), 91–95.

49.

Taiwan Health Promotion Administration. (2020, November). Online Quiz: Body Mass Index. https://health99.hpa.gov.tw/onlineQuiz/bmi

50.

Thompson

M. L.

Myers

J. E.

Kriebel

(1998). Prevalence odds ratio or prevalence ratio in the analysis of cross sectional data: What is to be done? Occupational and Environmental Medicine, 55(4), 272–277.

51.

Tsai

A. C.

C. S.

Chang

M. C.

(2007). Population-specific anthropometric cut-points improve the functionality of the Mini Nutritional Assessment (MNA) in elderly Taiwanese. Asia Pacific Journal of Clinical Nutrition, 16(4), 656–662.

52.

Tsai

A. C.

Lai

M. Y.

(2014). Mini Nutritional Assessment and short-form Mini Nutritional Assessment can predict the future risk of falling in older adults—Results of a national cohort study. Clinical Nutrition, 33(5), 844–849.

53.

Van Bokhorst-De Van Der Schueren

M. A. E.

Lonterman-Monasch

De Vries

O. J.

Danner

S. A.

Kramer

M. H. H.

Muller

. (2013). Prevalence and determinants for malnutrition in geriatric outpatients. Clinical Nutrition, 32(6), 1007–1011.

54.

Visser

Schaap

L. A.

(2011). Consequences of sarcopenia. Clinics in Geriatric Medicine, 27(3), 387–399.

55.

Vivanti

A. P.

McDonald

C. K.

Palmer

M. A.

Sinnott

(2009). Malnutrition associated with increased risk of frail mechanical falls among older people presenting to an emergency department. Emergency Medicine Australasia, 21(5), 386–394.

56.

Walford

R. L.

Mock

Verdery

MacCallum

(2002). Calorie restriction in biosphere 2: Alterations in physiologic, hematologic, hormonal, and biochemical parameters in humans restricted for a 2-year period. The Journals of Gerontology Series A, 57(6), B211–B224.

57.

Whooley

M. A.

(1999). Depression falls, and risk of fracture in older women. Study of Osteoporotic Fractures Research Group. Archives of Internal Medicine, 159(5), 484–490.

58.

Willcox

D. C.

Willcox

B. J.

Todoriki

Suzuki

(2009). The Okinawan diet: Health implications of a low-calorie, nutrient-dense, antioxidant-rich dietary pattern low in glycemic load. Journal of the American College of Nutrition, 28 Suppl, 500s–516s. https://doi.org/10.1080/07315724.2009.10718117

59.

Wilson

M. M.

Thomas

D. R.

Rubenstein

L. Z.

Chibnall

J. T.

Anderson

Baxi

Diebold

M. R.

Morley

J. E.

(2005).Appetite assessment: Simple appetite questionnaire predicts weight loss in community-dwelling adults and nursing home residents. The American Journal of Clinical Nutrition, 82(5), 1074–1081. https://doi.org/10.1093/ajcn/82.5.1074

60.

World Health Organization. (2008). WHO global report on falls prevention in older age. https://apps.who.int/iris/handle/10665/43811

61.

Yokoyama

Nishi

Murayama

Amano

Taniguchi

Nofuji

Narita

Matsuo

Seino

Kawano

Shinkai

(2016). Association of dietary variety with body composition and physical function in community-dwelling elderly Japanese. The Journal of Nutrition Health and Aging, 20(7), 691–696.

62.

Zhang

Huang

Dou

Wang

Zhang

Yang

Wang

Xie

Zhou

Zeng

Y. C.

(2019). Falls among older adults with sarcopenia dwelling in nursing home or community: A meta-analysis. Clinical Nutrition, 39(1), 33–39.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.17 MB