Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Feb 1.
Published in final edited form as: Nutr Metab Cardiovasc Dis. 2023 Oct 14;34(2):445–454. doi: 10.1016/j.numecd.2023.10.013

Time of eating and mortality in U.S. adults with heart failure: Analyses of the National Health and Nutrition Examination Survey 2003–2018

Hayley E Billingsley a,b, Marie-Pierre St-Onge c, Windy W Alonso d, Danielle L Kirkman a, Youngdeok Kim a,*, Salvatore Carbone a,b,**
PMCID: PMC10966516  NIHMSID: NIHMS1955290  PMID: 38155047

Abstract

Background and aims:

Promising associations have been demonstrated between delayed last eating occasion and cardiorespiratory fitness in adults with heart failure (HF), however, it is unknown if time of eating is associated with clinical endpoints such as mortality. This study aimed to examine associations between time of eating variables and all-cause and cardiovascular mortality in the National Health and Nutrition Examination Survey (NHANES).

Methods and results:

Participants self-disclosed HF diagnosis. Two dietary recalls were obtained and categorical variables were created based on mean time of first eating occasion (8:31 AM), last eating occasion (7:33 PM) and eating window (11.02 h). Mortality was obtained through linkage to the National Death Index. Covariate-adjusted Cox proportional hazard regression models were created examining the association between time of eating and mortality.

Participants (n = 991) were 68 (95 % CI 67–69) years of age, 52.6 (95 % CI 49.0–56.3)% men and had a body mass index of 32.5 (95 % CI 31.8–33.2) kg/m2 with follow up time of 68.9 (95 % CI 64.8–72.9) person-months. When models were adjusted for time of eating variables and all other covariates, extending the eating window beyond 11.02 h was associated with decreased risk of cardiovascular (HR 0.36 [95 % CI 0.16–0.81]), but not all-cause mortality. Time of first and last eating occasions were not associated with mortality.

Conclusions:

In adults with HF, an extended eating window is associated with reduced risk for cardiovascular mortality. Randomized controlled trials should examine if extending the eating window can improve prognostic indicators such as cardiorespiratory fitness in this population.

Keywords: NHANES, Heart failure, Time of eating, Eating window, Fasting, Cardiovascular mortality

1. Introduction

Prevalence of heart failure (HF) is rising in the US, with a projected 8 million Americans afflicted by 2030 [1]. Mortality rates for HF remain as high as 30 % in the year following hospitalization [2] and deaths where HF was an underlying cause increased over 50 % between 2009 and 2019 [1]. Significant advances have been made in the last decade in guideline directed medical therapy for HF, yet persistently high mortality and hospitalization rates [3] suggest a greater need for therapeutic solutions.

Dietary therapies offer a potentially powerful tool to improve outcomes in patients with HF [4,5], but dietary guidelines for HF are lacking [6-8]. Dietary sodium has remained the persistent focus in the dietary treatment of HF despite a lack of supportive evidence suggesting improved clinical outcomes with sodium restriction [9-12]. Therefore, many questions remain regarding the optimization of dietary intake for adults with HF, including the impact of time of eating. In healthy individuals or those with obesity alone, time of eating interventions such as time restricted eating (TRE) has recently been proposed as a therapeutic option not only for weight loss [13-25] but also for favorable effects on blood pressure [14,16,26,27], glycemia [17,24], and lipidemia [16,24].

TRE describes dietary interventions in which the individual shortens the eating window, defined as the period of time over which energy intake occurs, to between 6 and 10 h per day [28]. Larger randomized controlled trials have not been able to replicate the success of smaller TRE studies on weight loss and improvements in associated cardiometabolic risk factors [29-31]. Recent studies have noted reductions in physical activity [16,29,30,32] and a larger than expected proportion of lean mass loss [26,29,30,33] from these interventions. While these findings are concerning in healthy individuals, they warrant considerable caution in patients with HF for whom lean mass [34], of which skeletal muscle mass is a major component [35], and physical activity are major determinants of cardiorespiratory fitness, an independent prognostic indicator [36].

In adults with HF with preserved ejection fraction (HFpEF) and obesity, we have shown that peak oxygen consumption (VO2peak) was significantly higher for those who had their last eating occasion, i.e. last meal or snack, after the mean or median timepoint compared to those who had their last eating occasion prior to the mean or median. In addition, an association was observed between greater VO2peak and advancing time of last eating occasion [37,38]. Despite these promising associations between delayed last eating occasion and VO2peak, the relation between mortality and time of eating has never been examined in patients with HF. The aim of this study was to examine the association between time of last eating occasion, first eating occasion, and eating window with all-cause and cardiovascular mortality in participants with HF in the National Health and Nutrition Examination Survey (NHANES).

2. Methods

2.1. Study population

The NHANES is a complex, multistage, nationally-representative health survey which enrolls non-institutionalized civilian participants in the United States. Ethical approval for each two-year cycle of NHANES is obtained through the ethics review board of the National Center for Health Statistics and written informed consent is obtained for all participants. All data are publicly accessible (https://www.cdc.gov/nchs/nhanes/index.htm). For the primary mortality analyses, 8 cycles between the years 2003–2018 were included.

This analysis included participants aged ≥20 years, adults by NHANES standard analysis practice, with self-reported HF, mortality data and two dietary recalls. Participants self-identified as having HF by answering yes to the question “Has a doctor or health professional ever told you that you had congestive HF?”. Initially, a total of 1552 participants with HF were identified. Participants who had both dietary recalls coded as unreliable (n = 376), who did not report any eating occasions ≥50 kcals (n = 0) or reported extremely low or high total energy intake (<600 or >4400 kcals for women and <650 or >5700 kcals for men, [39]) (n = 18) were excluded. Participants who did not report at least one medication commonly used in the treatment of HF were excluded (n = 139); a detailed description of these medications is available in the supplemental materials. Participants who reported currently receiving dialysis (n = 27) or those who reported being currently pregnant and/or nursing were also excluded (n = 1) due to the accompanying marked physiological and dietary adjustments (Fig. S1).

2.2. Time of eating occasions

The primary outcome variables related to the timing of eating occasions were extracted using information from the dietary assessment. NHANES participants are eligible to undergo two 24-h dietary recall interviews using the United States Department of Agriculture Automated Multiple Pass Method. During each recall, participants are asked to report the time of intake for each food and beverage consumed. A previously established cutoff of ≥50 kcals was used for identifying eating occasions [40]. Both dietary recalls were used to calculate mean dietary variables. Based on reported times of intake, weighted mean values were established for time of last eating occasion (7:33 PM, 95 % CI 7:27–7:39) and first eating occasion (8:31 AM, 95 % CI 8:23–8:39). Mean eating window was calculated as the difference in hours between first and last eating occasions (11.02 h, 95 % CI 10.87–11.18). Mean total daily energy intake and number of eating occasions were also identified.

2.3. Mortality

The National Center for Health Statistics has released linked mortality files through the National Death Index, a centralized database of all deaths in the United States, for NHANES participants from date of participation through December 31, 2019. For all included NHANES participants, mortality status and follow up time from interview were recorded. Categories are available for the leading cause of death based on recoding of the International Statistical Classification of Diseases, Injuries and Causes of Death (ICD-10) guidelines and mortality coded as “diseases of the heart” (ICD-10 codes I00-I78) and cerebrovascular mortality (ICD-10 codes 160–169) were considered cardiovascular mortality.

2.4. Study covariates

Analyses were progressively adjusted for relevant covariates identified in the literature [41-43] in 5 successive models. Model 1 included only the selected time of eating variable(s). Model 2 added age, race, sex, annual household income (<$25,000, $25,00–75,000 and >$75,000) and marital status (married or partnered/not married or partnered), education (less than high school, high school, some college or more). Model 3 further adjusted for body mass index (BMI) (kg/m2), current smoking status (yes/no) and reported medical history of cancer (yes/no), hypertension (yes/no), dyslipidemia (yes/no), myocardial infarction (yes/no), kidney disease (yes/no), and diabetes (yes/no). Model 4 added perceived difficulty walking score, time with HF (years) and sleep (hours) on weekdays or workdays to the previous variables. Time with HF was calculated as the difference between age at interview and reported age at diagnosis. Perceived difficulty walking score was calculated as the mean of the answer to 3 questions regarding mobility limitations collected during the interview [41]. From the 2005–2006 cycle onward, participants reported hours of sleep during the interview with the question “How much sleep do you get on weekdays or workdays?” therefore it should be noted that models 4 and 5 omit 2-year cycle 2003–2004. Finally, model 5 added dietary factors such daily energy intake and number of eating occasions per day.

2.5. Statistical analysis

The NHANES uses a complex, multistage, probability sampling design. All analyses utilized appropriately calculated sample weights, masked variance units and primary sampling units. Continuous variables are presented as mean with 95 % confidence intervals (CI). Categorical variables are specified as weighted or unweighted number and percentage (%).

Cox proportional hazard regression analysis were used to estimate the hazard ratios (HR) and 95 % CI for all-cause and cardiovascular mortality. Time of last eating occasion, first eating occasion and eating window were coded as either above or below mean values and served as independent categorical variables in the models. For Cox proportional hazard regression, individual models were first constructed in which only one time of eating variable was present per model. Then, combined models were constructed which included all three variables: time of first eating occasion, time of last eating occasion and length of eating window. Sensitivity analyses were performed on the combined model. The first sensitivity analysis included only participants prescribed 2 or more guideline directed medical therapies for HF [8,44-46], a detailed description of these medications is available in the supplemental materials. A second sensitivity analysis included an additional covariate in model 5 - dietary recall day of the week (coded as both dietary recalls taking place on weekdays or one or both dietary recalls on weekend days). This sensitivity analysis was performed as weekday vs. weekend dietary intake has been found to differ significantly [47].

Proportional hazard assumptions were checked for all variables on Cox proportional hazard regression analyses. The proportional hazard assumption was tested by comparing an individual model of each variable to an alternative model that included a time-interaction for the variable. If the proportional hazard assumption was not met, a time x variable interaction term was added to the model as appropriate. Collinearity was also checked for on Cox proportional hazard regression analyses by reviewing variance inflation factors for all variables as well as Pearson correlation coefficients for continuous variables (Table S1). The percentage of missing datapoints was assessed for all variables used in Cox proportional hazard analyses and was found to be less than 10 %. Despite this, the sample size did decline from model 1 (n = 991) to model 5 (n = 606) of Cox proportional hazard analyses due to listwise deletion for missing variables. For this reason, a third sensitivity analyses constrained to only participants with all datapoints available was performed on the combined model. All analyses were performed using SPSS 28.0 Complex Samples Package with a two-tailed P-value <0.05 considered significant.

3. Results

3.1. Baseline characteristics

Descriptive characteristics of the study population estimated from the 991 participants with HF who met inclusion criteria from the NHANES 2003–2018 are presented in Table 1.

Table 1.

Descriptive Characteristics (unweighted n = 991).

Variable Weighted Mean
or % (Unweighted
Count)		 95 % CI
Age (years) 68 67–69
Sex
Male 52.6 (556) 49.0–56.3
Female 47.4 (435) 43.7–51.0
Race/Ethnicity
Non-Hispanic Black 14.6 (241) 12.0–17.6
Non-Hispanic White 74.6 (573) 70.7–78.1
Other 10.8 (177) 8.4–13.8
Education
Less than High School 27.2 (348) 22.9–31.9
High School 30.7 (267) 26.8–34.9
Some College or More 42.1 (376) 37.8–46.6
Household Income
<$25,000 41.2 (452) 36.3–46.3
$25,000–75,000 47.5 (379) 41.8–53.3
>$75,000 11.3 (75) 8.5–14.8
Marital/Partnership Status
Not Married or Partnered 46.1 (485) 41.3–50.8
Married or Partnered 53.9 (506) 49.2–58.7
Smoking Status
Not Current Smoker 81.4 (826) 77.0–85.1
Current Smoker 18.6 (165) 14.9–23.0
BMI (kg/m2) 32.5 31.8–33.2
Sleep (Hours) a 7.2 7.1–7.3
Perceived Difficulty Walking Score 1.2 1.1–1.2
Self-Reported Medical Hx
Cancer Hx
No 76.4 (772) 72.6–79.9
Yes 23.6 (214) 20.1–27.4
Chronic Kidney Disease Hx
No 86.6 (829) 84.1–88.8
Yes 13.4 (162) 11.2–15.9
Diabetes Hx
No 56.5 (526) 52.1–60.8
Yes 43.5 (431) 39.2–47.9
Myocardial Infarction Hx
No 54.5 (531) 49.5–59.4
Yes 45.5 (454) 40.6–50.5
Dyslipidemia
No 31.6 (327) 27.4–36.2
Yes 68.4 (624) 63.8–72.6
Hypertension
No 17.3 (166) 14.1–21.1
Yes 82.7 (823) 78.9–85.9
Guideline Directed Medical Therapy
Angiotensin-converting enzyme inhibitor 44.1 (441) 39.4–48.8
Angiotensin receptor blockers 22.8 (222) 19.3–26.8
Angiotensin receptor-neprilysin inhibitor 0.3 (5) 0.1–0.8
Beta Blocker 70.9 (693) 67.0–74.5
Aldosterone antagonist 9.6 (94) 7.3–12.6
Hydralazine 5.8 (56) 3.7–8.9
Sodium-glucose cotransporter 2 inhibitor 0.9 (4) 0.2–3.2
Isosorbide dinitrate 2.9 (26) 1.7–5.1
Loop diuretics 51.7 (504) 47.0–56.2
Thiazide 16.4 (136) 13.2–20.3
Digoxin 13.3 (125) 10.3–17.0
Moderate-Vigorous Physical Activity (minutes/week) b 370 273–468
Time with Heart Failure (years) 9.5 8.5–10.5
Time of First Eating Occasion (Hour:Minute) (AM) 8:31 8:23–8:39
Identified as Breakfast (Day 1) 74.3 (695) 70.2–78.0
Identified as Breakfast (Day 2) 78.3 (741) 74.5–81.6
Time of Last Eating Occasion (Hour:Minute) (PM) 7:33 7:27–7:39
Identified as Dinner (Day 1) 22.1 (210) 18.1–26.7
Identified as Dinner (Day 2) 26.2 (270) 22.2–30.7
Identified as Supper (Day 1) 21.9 (211) 18.1–26.3
Identified as Supper (Day 2) 20.9 (178) 17.2–25.1
Identified as Snack (Day 1) 46.9 (443) 41.9–51.9
Identified as Snack (Day 2) 42.6 (414) 37.8–47.5
Eating Window (Hours) 11.02 10.87–11.18
Total Kcals per Day 1831 1747–1915
Number of Eating Occasions per Day 4.1 4.0–4.2
Participants with both dietary recalls occurring on weekdays 48.7 (413) 42.7–54.8
Open in a new tab

Abbreviations: BMI, body mass index; Hx, history; Kcals, kilocalories.

a

Questions about sleep were not added until 2005–2006 therefore participants in cycle 2003–2004 are excluded.

b

The Global Physical Activity Questionnaire was administered beginning in 2007–2008 therefore participants in cycles 2003–2004 and 2005–2006 are excluded.

Of the 991 participants included, the mean age was 68 (95 % CI: 67–69) years, participants were 52.6 (95 % CI 49.0–56.3) % male, 74.6 (95 % CI 70.7–78.1) % white and mean BMI was 32.5 (95 % CI 31.8–33.2) kg/m2. Participants had been diagnosed with HF for a mean of 9.5 (95 % CI 8.5–10.5) years. Most participants reported comorbid hypertension (82.7 [95 % CI 78.9–85.9] %) and dyslipidemia (68.4 [95 % CI 63.8–72.6] %) and 45.5 (95 % CI 40.6–50.5) % reported a history of myocardial infarction. Participants were most commonly prescribed beta blockers (70.9 [95 % CI 67.0–74.5] %), loop diuretics (51.7 [95 % CI 47.0–56.2]%) and angiotensin-converting enzyme inhibitors (44.1 [95 % CI 39.4–48.8]%).

Participants reported their first eating occasion occurring at a mean of 8:31 (8:23–8:39) AM and their last eating occasion occurring at 7:33 (95 % CI 7:27–7:39) PM. Reported eating window spanned a mean of 11.02 (95 % CI 10.87–11.18) hours and participants had 4 eating occasions per day. Mean daily energy intake was 1831 (95 % CI 1747–1915) kcals/day.

3.2. Effects of time of eating on all-cause and cardiovascular mortality

Mean participant follow-up time was 68.9 (95 % CI 64.8–72.9) months over which an unweighted 52 % (n = 512) of participants died. Of these deaths, an unweighted count of 227 were classified as occurring from cardiovascular causes.

The results of the Cox proportional hazard regression analysis examining the individual effects of time of first eating occasion, time of last eating occasion and eating window on all-cause mortality are presented in Fig. 1 and Table S2. In models 1–3 and after full adjustment in model 5, an extended eating window ≥11.02 h was significantly associated with a decreased risk of all-cause mortality compared to a shorter eating window <11.02 h (Model 5 HR 0.62 [95 % CI 0.40–0.96]) (Fig. 1). Eating window was also associated with decreased cardiovascular mortality after full adjustment in model 5 (HR 0.40 [95 % CI 0.21–0.80]) (Fig. 2) (Table S2).

Figure 1. Individual Model Associations of Time of First Eating Occasion, Last Eating Occasion and Eating Window with All-Cause Mortality.

Figure 1

Open in a new tab

After full adjustment for all covariates in model 5, an extended eating window ≥11.02 h was significantly associated with decreased risk of all-cause mortality vs. a shorter eating window. Time of first and last eating occasions were not significantly associated with all-cause mortality. Individual model indicates only one time of eating variable (first eating occasion, last eating occasion or eating window) was utilized per model.

Figure 2. Individual Model Associations of Time of First Eating Occasion, Last Eating Occasion and Eating Window with Cardiovascular Mortality.

Figure 2

Open in a new tab

When analyses were repeated for cardiovascular mortality, only an extended eating window ≥11.02 h was significantly associated with a decreased risk of mortality after full adjustment in model 5. Neither time of last eating occasion nor first eating occasion demonstrated a significant association with cardiovascular mortality. Individual model indicates only one time of eating variable (first eating occasion, last eating occasion or eating window) was utilized per model.

In model 1, time of last eating occasion ≥7:33 PM was associated with a decreased risk of all-cause mortality in comparison to time of last eating occasion <7:33 PM (HR 0.82 [95 % CI 0.69–0.98]) (Fig. 1). After adjusting for covariates in models 2–5, time of last eating occasion was no longer a predictor of all-cause mortality. When analyses were repeated for cardiovascular mortality, only unadjusted model 1 demonstrated a trend between time of last eating occasion ≥7:33 PM and decreased mortality risk (HR 0.74 [95 % CI 0.53–1.03], P = 0.072) (Fig. 2).

In models 1–3, an earlier time of first eating occasion <8:31 AM demonstrated a trend with reduced risk of all-cause mortality in comparison to time of first eating occasion ≥8:31 AM (Model 3 HR 0.76 [95 % CI 0.56–1.04] P = 0.082), however this association disappeared with further adjustment in models 4 and 5 (Fig. 1). Time of first eating occasion was not associated with cardiovascular mortality (Fig. 2).

A combined analysis was then performed, fitting models 1–5 with time of first eating occasion, last eating occasion and eating window. In the combined analyses, an extended eating window ≥11.02 h was associated with a decreased risk of all-cause mortality vs. an eating window <11.02 h in models 1–3 (model 3 HR 0.24 [95 % CI 0.08–0.72]) but significance was lost with further adjustment in models 4 and 5 (Fig. 3) (Table S3). Neither time of first eating occasion nor last eating occasion displayed associations with all-cause mortality in models 1–5.

Figure 3. Combined Model Associations of Time of First Eating Occasion, Last Eating Occasion and Eating Window with All-Cause Mortality.

Figure 3

Open in a new tab

After adding all time of eating variables to each model of all-cause mortality, an extended eating window ≥11.02 h was associated with an decreased risk of all-cause mortality in models 1–3 but significance was lost with further adjustment in models 4 and 5. Neither time of first eating occasion nor last eating occasion displayed significant associations with all-cause mortality. Combined model indicates that all time of eating variables (first eating occasion, last eating occasion or eating window) were used in each model.

In the combined analyses examining the effects of time of eating on cardiovascular mortality, an extended eating window ≥11.02 h did not display a favorable association with mortality in models 1–3 (Fig. 4) (Table S3). However, with further adjustment, an extended eating window ≥11.02 h demonstrated a trend towards decreased risk of cardiovascular mortality in model 4 (HR 0.50 [95 % CI 0.25–1.01], P = 0.052). With full adjustment in model 5, an extended eating window ≥11.02 h was associated with a decreased risk of cardiovascular mortality vs. an eating window <11.02 h (HR 0.36 [95 % CI 0.16–0.81]). Neither time of first nor last eating occasion displayed any association with cardiovascular mortality in the combined model.

Figure 4. Combined Model Associations of Time of First Eating Occasion, Last Eating Occasion and Eating Window with Cardiovascular Mortality.

Figure 4

Open in a new tab

After full adjustment for all covariates including all time of eating variables, extending the eating window ≥11.02 h was associated with decreased risk of cardiovascular mortality. Neither time of first eating occasion nor last eating occasion displayed significant associations with cardiovascular mortality. Combined model indicates that all time of eating variables (first eating occasion, last eating occasion or eating window) were used in each model.

3.3. Sensitivity analyses

Approximately 76 % of the included participants, n = 749, were prescribed 2 or more HF guideline directed medical therapies, a detailed list of these medications is given in the supplemental materials. Interestingly, in models 1–4, a delayed time of last eating occasion ≥7:33 PM was associated with a reduced risk of all-cause mortality (model 4 HR 0.20 [95 % CI 0.05–0.86]) (Table S4). With further adjustment for total energy intake and frequency of eating occasions in model 5, time of last eating occasion was no longer associated with all-cause mortality. Neither time of first eating occasion nor eating window were associated with all-cause mortality.

As in the original combined analysis, an extended eating window ≥11.02 h was associated with reduced cardiovascular mortality compared to a eating window <11:02 h (model 5 HR 0.46 [95 % CI 0.24–0.91]) (Table S4). Neither first nor last eating occasion displayed an association with cardiovascular mortality.

In the second sensitivity analyses which added dietary recall day of the week (weekday/weekend day) as a covariate to the combined, fully adjusted model, none of the time of eating variables was significantly associated with all-cause or cardiovascular mortality. Eating window, however, did maintain a trend with cardiovascular mortality (model 5 HR 0.51 [95 % CI 0.22–1.22], P = 0.130) (Table S5).

In the final sensitivity analysis in which analyses of the combined models were constrained to only participants with all datapoints available (n = 606), results were largely unchanged (Table S6). In models 1–3 of all-cause mortality, however, eating window ≥11.02 h was no longer associated with reduced mortality compared to an eating window <11.02 h (model 3 HR 0.75 [95 % CI 0.454–1.241] P = 0.259).

4. Discussion

The data presented demonstrate that, in adults with HF, an extended eating window is associated with a reduced risk of cardiovascular mortality in comparison to a shorter eating window. This association is independent of socio-demographic factors, comorbidities, sleep duration, length of HF diagnosis, self-reported functional status, frequency of eating and energy intake. Moreover, the association between extended eating window and reduced risk of cardiovascular mortality persisted when time of first and last eating occasion were added to the model, suggesting that regardless of when the eating window begins or ends, spreading energy intake over a greater number of waking hours may be cardioprotective for adults with HF. When only participants that were prescribed ≥2 guideline directed medical therapies for HF were included in the models, the association between extended eating window and reduced cardiovascular mortality was also sustained. This not only suggests that the individuals included did have HF but also that the association between extended eating window and reduced cardiovascular mortality is not eliminated with guideline directed medical therapy.

Time of eating is a novel risk factor to consider in adults with HF. The healthy human heart demonstrates considerable flexibility in its use of metabolic substrates, primarily using fatty acids but can switch to glucose, amino acids, ketones and lactate when necessary. In HF, the ability of the myocardium and whole body to switch between metabolic substrates is substantially disrupted, referred to as a state of metabolic inflexibility [48-52]. These derangements have resulted in descriptions of the failing myocardium as being ‘starved’ for metabolic substrate. It is possible that rather than improving metabolic flexibility, extended fasting periods may put greater stress on an already dysfunctional system. Additionally, recent preclinical studies in mice at high risk for atherosclerosis [53] and receiving chemotherapy [54] have suggested that extended fasting strategies may activate adverse pathways, such as recruitment of leukocytes to large arteries and stimulation of nuclear translocation of transcription factor EB (TFEB) resulting in left ventricular atrophy in mice at high cardiac risk. Future mechanistic work should examine how extending the eating window impacts whole body and myocardial substrate utilization in individuals with HF. The association of eating window with all-cause and cardiovascular mortality should also be explored in other populations, particularly those diagnosed with or at high risk for other cardiovascular diseases [28].

Outside of metabolic considerations, some, but not all, TRE interventions have demonstrated reduced physical activity [16,29,32]. It is reasonable to consider whether compensatory reductions in physical activity may be a consequence of extended fasting periods [55], and if these compensatory reductions over time in individuals could impact the risk for cardiovascular mortality. Recently, our group has demonstrated that self-reported moderate-vigorous physical activity is significantly associated with a reduced risk of all-cause mortality in NHANES participants with HF [41], a finding also observed in the HF Adherence and Retention Trial (HART) [56].

Although self-reported moderate-vigorous physical activity was available from 2007 onward in this sample, physical activity was not included as a covariate in Cox proportional hazard as less than half of the included participants (n = 365) reported engaging in any moderate-vigorous physical activity. This is in accordance with our previous finding that NHANES participants with HF engage in very little moderate physical activity when objectively measured with accelerometry [57], however, it is notable the mean minutes per week of moderate vigorous physical activity reported in our current sample was 370 (95 % CI 273–468), despite including all individuals who reported no activity. This likely indicates significant over-reporting for some participants as objective physical activity measured in NHANES 2003–2006 for adults aged 60–69 was estimated from 11 to 74 min per week depending on the cutoff utilized [58]. Although objective physical activity does not carry the limitations of self-report such as over- and underreporting, accelerometry data was only collected in select participants in cycles 2003–2006 and 2011–2014, which would have significantly limited the available sample size if it was included as a covariate. Future work should examine the impact of altering the length of the eating window, independent of energy intake, on amount, type and time distribution of daily physical activity in adults with HF.

NHANES is a large, nationally representative cohort with extremely well characterized participants for whom long term mortality outcomes are available. The observations made in this manuscript, however, do have limitations. First, the presence of HF was self-reported. Although we limited inclusion to participants who were able to report age of HF diagnosis, we cannot be certain that participants had HF. It should be noted that all included participants were prescribed at least one medication typically used in HF, and over 75 % of included participants brought in 2 or more HF guideline directed medical therapy prescriptions for interviewers to record. It is also worth noting that mortality reached nearly 50 % over the follow-up period, which is in accordance with the high mortality rates observed in HF.

It is important to note that HF type was not available. Patients with HF with reduced ejection fraction (HFrEF) and HFpEF respond in markedly different ways to medical therapy [8,59]. Moreover, in our initial analyses of patients with HFpEF, the favorable associations observed were between cardiorespiratory fitness and time of last eating occasion, while eating window displayed a non-significant positive trend with cardiorespiratory fitness [38]. Between HFrEF and HFpEF, some differences in fuel utilization have also been noted, but recent work has suggested more similarities in metabolism that previously thought [60]. It is possible that associations observed herein between mortality and time of eating would be influenced by HF type.

Decreasing statistical power may have also attenuated or masked some associations. To better understand the associations between time of eating and mortality, multiple models were constructed which decreased sample size considerably. As sample size decreased in-part due to list-wise deletion for missing variables, we did perform sensitivity analyses constrained to only those participants with all variables available. The only notable difference observed between this sensitivity analysis and analysis of the overall cohort was the loss of significance for extended eating window in combined models 1–3 of all-cause mortality. It is important to note that Models 1–3 captured NHANES cycles 2003–2018 while models 4 and 5 reflected NHANES cycles 2005–2018, as sleep was not collected as a variable until cycle 2005–2006 resulting in a decrease in sample size (n = 125). This finding suggests that with a greater sample size, an association between extended eating window and decreased all-cause mortality may also have been observed in the fully adjusted models.

In these analyses, neither differences in overall dietary quality nor nutrients at specific eating occasions were explored nor included as covariates. It is possible that these variables impact the associations observed and future analyses should examine whether dietary quality modifies the association between time of eating and mortality in patients with HF. As weekend and weekday dietary quality tends to differ, we conducted a sensitivity analysis with the day of dietary recall (weekend/weekday) included as a covariate. While extended eating window was no longer significantly associated with lower cardiovascular mortality, a trend remained and it is important to note the inclusion of this covariate dropped the size of the final sample to n = 540, reducing power.

We were not able to exclude those who participated in shift work, which substantially alters typical times of sleeping and eating, as this information was not collected in all cycles. Moreover, information on participants’ chronotype was not available, making it impossible to interpret effects of time of eating relative to the biological circadian clock. Reduced cardiovascular mortality was observed in this sample with an eating window greater than the mean of 11.02 h which is less than the mean eating window of the overall NHANES cohort ≥20 years of age from 2003 to 2018, which was 11.45 (95 % CI 11.40–11.50) h. Moreover, we acknowledge that selecting mean values as our dichotomous cutoff values for time of eating variables is somewhat arbitrary and further work is needed to establish clinically meaningful time of eating cutoffs for patients with heart failure and to examine whether an upper limit for benefit may exist for the eating window in this population. Finally, the causality of all included observations cannot be determined. Future randomized control trials should be conducted to examine the impact of time of eating alterations on clinically-important risk factors in patients with HF.

5. Conclusion

In adults with HF, an extended eating window is associated with reduced cardiovascular mortality in comparison with a shorter eating window. These novel findings warrant a randomized controlled trial examining whether extending the eating window in patients with HF can favorably impact important prognostic indicators such as cardiorespiratory fitness.

Supplementary Material

1

Acknowledgements

Statement of authors’ contributions to the manuscript.

HEB conceived and performed the statistical analyses. HEB, SC, YK, WA, DK and MPSO wrote the manuscript. HEB has primary responsibility for the final content. All authors have read and approved the final manuscript.

Support

MPSO is funded by the National Heart, Lung, and Blood Institute (NHLBI), grants R01HL142648 and R35HL155670. WA is funded by the NHLBI, grant R01HL163288.

Abbreviations:

BMI

body mass index

CI

confidence intervals

HF

heart failure

HFpEF

heart failure with preserved ejection fraction

HFrEF

heart failure with reduced ejection fraction

HR

hazard ratio

ICD-10

International Statistical Classification of Diseases

NHANES

National Health and Nutrition Examination Survey

TRE

Time Restricted Eating

VO2peak

peak oxygen consumption

Footnotes

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.numecd.2023.10.013.

Data sharing

All data utilized in this manuscript is publicly and freely available without restriction at https://www.cdc.gov/nchs/nhanes/index.htm.

References

  • [1].Tsao CW, Aday AW, Almarzooq ZI, Alonso A, Beaton AZ, Bittencourt MS, et al. Heart disease and stroke Statistics 2022 update: a report from the American heart association. Circulation 2022;145:e153–639. [DOI] [PubMed] [Google Scholar]
  • [2].Chang PP, Wruck LM, Shahar E, Rossi JS, Loehr LR, Russell SD, et al. Trends in hospitalizations and survival of acute decompensated heart failure in four US communities (2005–2014). Circulation 2018;138:12–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [3].Khan MS, Sreenivasan J, Lateef N, Abougergi MS, Greene SJ, Ahmad T, et al. Trends in 30- and 90-day readmission rates for heart failure. Circulation: Heart Fail 2021;14:e008335. [DOI] [PubMed] [Google Scholar]
  • [4].Bonilla-Palomas JL, Gámez-López AL, Castillo-Domínguez JC, Moreno-Conde M, López Ibáñez MC, Alhambra Expósito R, et al. Nutritional intervention in malnourished hospitalized patients with heart failure. Arch Med Res 2016;47:535–40. [DOI] [PubMed] [Google Scholar]
  • [5].Hummel SL, Karmally W, Gillespie BW, Helmke S, Teruya S, Wells J, et al. Home-delivered meals postdischarge from heart failure hospitalization. Circ Heart Fail 2018;11:e004886. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [6].Vest AR, Chan M, Deswal A, Givertz MM, Lekavich C, Lennie T, et al. Nutrition, obesity, and cachexia in patients with heart failure: a consensus statement from the heart failure society of America scientific statements committee. J Card Fail 2019;25: 380–400. [DOI] [PubMed] [Google Scholar]
  • [7].McDonagh TA, Metra M, Adamo M, Gardner RS, Baumbach A, Bohm M, et al. 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J 2021;42:3599–726. [DOI] [PubMed] [Google Scholar]
  • [8].Heidenreich Paul A, Bozkurt B, Aguilar D, Allen Larry A, ByunJoni J, Colvin Monica M, et al. 2022 AHA/ACC/HFSA guideline for the management of heart failure. J Am Coll Cardiol 2022;79:e263–42. [DOI] [PubMed] [Google Scholar]
  • [9].Billingsley HE, Hummel SL, Carbone S. The role of diet and nutrition in heart failure: a state-of-the-art narrative review. Prog Cardiovasc Dis 2020;63:538–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [10].Ezekowitz JA, Colin-Ramirez E, Ross H, Escobedo J, Macdonald P, Troughton R, et al. Reduction of dietary sodium to less than 100 mmol in heart failure (SODIUM-HF): an international, open-label, randomised, controlled trial. Lancet 2022;399:1391–400. [DOI] [PubMed] [Google Scholar]
  • [11].Allen Kathleen E, Billingsley Hayley E, Carbone S. Nutrition, heart failure, and quality of life. JACC Heart Fail 2020;8:765–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [12].Montgomery RA, Mauch J, Sankar P, Martyn T, Engelman T, Martens P, et al. Oral sodium to preserve renal efficiency in acute heart failure: a randomized, placebo-controlled, double-blinded study. J Card Fail 2023;29(7):986–96. [DOI] [PubMed] [Google Scholar]
  • [13].Gill S, Panda S. A smartphone app reveals erratic diurnal eating patterns in humans that can Be modulated for health benefits. Cell Metabol 2015;22:789–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [14].Prasad M, Fine K, Gee A, Nair N, Popp CJ, Cheng B, et al. A smartphone intervention to promote time restricted eating reduces body weight and blood pressure in adults with overweight and obesity: a pilot study. Nutrients 2021;13:2148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [15].Anton SD, Lee SA, Donahoo WT, McLaren C, Manini T, Leeuwenburgh C, et al. The effects of time restricted feeding on overweight, older adults: a pilot study. Nutrients 2019;11:1500. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [16].Wilkinson MJ, Manoogian ENC, Zadourian A, Lo H, Fakhouri S, Shoghi A, et al. Ten-hour time-restricted eating reduces weight, blood pressure, and atherogenic lipids in patients with metabolic syndrome. Cell Metabol 2020;31:92–104 e105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [17].Kesztyüs D, Cermak P, Gulich M, Kesztyüs T. Adherence to time-restricted feeding and impact on abdominal obesity in primary care patients: results of a pilot study in a pre-post design. Nutrients 2019;11:2854. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [18].Li C, Xing C, Zhang J, Zhao H, Shi W, He B. Eight-hour time-restricted feeding improves endocrine and metabolic profiles in women with anovulatory polycystic ovary syndrome. J Transl Med 2021;19:148. 148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [19].Park SJ, Yang JW, Song YJ. The effect of four weeks dietary intervention with 8-hour time-restricted eating on body composition and cardiometabolic risk factors in young adults. Nutrients 2021;13:2164. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [20].Xie Z, Sun Y, Ye Y, Hu D, Zhang H, He Z, et al. Randomized controlled trial for time-restricted eating in healthy volunteers without obesity. Nat Commun 2022;13:1003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [21].Brady AJ, Langton HM, Mulligan M, Egan B. Effects of 8 wk of 16:8 time-restricted eating in male middle- and long-distance runners. Med Sci Sports Exerc 2021;53:633–42. [DOI] [PubMed] [Google Scholar]
  • [22].Gabel K, Hoddy KK, Haggerty N, Song J, Kroeger CM, Trepanowski JF, et al. Effects of 8-hour time restricted feeding on body weight and metabolic disease risk factors in obese adults: a pilot study. Nutr Healthy Aging 2018;4:345–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [23].Chow LS, Manoogian ENC, Alvear A, Fleischer JG, Thor H, Dietsche K, et al. Time-restricted eating effects on body composition and metabolic measures in humans who are overweight: a feasibility study. Obesity 2020;28:860–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [24].Che T, Yan C, Tian D, Zhang X, Liu X, Wu Z. Time-restricted feeding improves blood glucose and insulin sensitivity in overweight patients with type 2 diabetes: a randomised controlled trial. Nutr Metab 2021;18:88. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].Kotarsky CA-OX, Johnson NR, Mahoney SJ, Mitchell SL, Schimek RL, Stastny SN, et al. Time-restricted eating and concurrent exercise training reduces fat mass and increases lean mass in overweight and obese adults. Phys Rep 2021;9:e14868. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [26].Lin YJ, Wang YT, Chan LC, Chu NF. Effect of time-restricted feeding on body composition and cardio-metabolic risk in middle-aged women in Taiwan. Nutrition 2022;93:30. [DOI] [PubMed] [Google Scholar]
  • [27].Sutton EF, Beyl R, Early KS, Cefalu WT, Ravussin E, Peterson CM. Early time-restricted feeding improves insulin sensitivity, blood pressure, and oxidative stress even without weight loss in men with prediabetes. Cell Metabol 2018;27:1212–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [28].Billingsley HE. The effect of time of eating on cardiometabolic risk in primary and secondary prevention of cardiovascular disease. Diabetes Metabol Res Rev 2023:e3633. [DOI] [PubMed] [Google Scholar]
  • [29].Lowe DA, Wu N, Rohdin-Bibby L, Moore AH, Kelly N, Liu YE, et al. Effects of time-restricted eating on weight loss and other metabolic parameters in women and men with overweight and obesity: the TREAT randomized clinical trial. JAMA Intern Med 2020;180:1491–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [30].Templeman I, Smith HA, Chowdhury E, Chen Y-C, Carroll H, Johnson-Bonson D, et al. A randomized controlled trial to isolate the effects of fasting and energy restriction on weight loss and metabolic health in lean adults. Sci Transl Med 2021;13:eabd8034. [DOI] [PubMed] [Google Scholar]
  • [31].Liu D, Huang Y, Huang C, Yang S, Wei X, Zhang P, et al. Calorie restriction with or without time-restricted eating in weight loss. N Engl J Med 2022;386:1495–504. [DOI] [PubMed] [Google Scholar]
  • [32].Tinsley GM, Moore ML, Graybeal AJ, Paoli A, Kim Y, Ju Gonzales, et al. Time-restricted feeding plus resistance training in active females: a randomized trial. Am J Clin Nutr 2019;110:628–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [33].Parr EB, Kouw IWK, Wheeler MJ, Radford BE, Hall RC, Senden JM, et al. Eight-hour time-restricted eating does not lower daily myofibrillar protein synthesis rates: a randomized control trial. Obesity 2022;22:23637. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [34].Billingsley H, Del Buono MG, Canada JM, Kim YA, Damonte JI, Trankle C, et al. Sarcopenic obesity is associated with reduced cardiorespiratory fitness compared with nonsarcopenic obesity in patients with heart failure with reduced ejection fraction. Circ heart fail; 2022. p. e009518. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [35].Earthman CP. Body composition tools for assessment of adult malnutrition at the bedside: a tutorial on Research considerations and clinical applications. JPEN - J Parenter Enter Nutr 2015;39:787–822. [DOI] [PubMed] [Google Scholar]
  • [36].Del Buono MG, Arena R, Borlaug BA, Carbone S, Canada JM, Kirkman DL, et al. Exercise intolerance in patients with heart failure: JACC state-of-the-art review. J Am Coll Cardiol 2019;73: 2209–25. [DOI] [PubMed] [Google Scholar]
  • [37].Billingsley HE, Dixon DL, Canada JM, Kirkman DL, Rotelli B, Kadariya D, et al. Time of eating and cardiorespiratory fitness in patients with heart failure with preserved ejection fraction and obesity. Nutr Metabol Cardiovasc Dis 2021;31:2471–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [38].Billingsley HE, Canada JM, Dixon DL, Kirkman DL, Bohmke N, Rotelli B, et al. Midpoint of energy intake, non-fasting time and cardiorespiratory fitness in heart failure with preserved ejection fraction and obesity. Int J Cardiol 2022;355:23–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [39].HHS. Reviewing and cleaning ASA24 data. https://epi.grants.cancer.gov/asa24/resources. [Google Scholar]
  • [40].Murakami K, Livingstone MBE. Associations between meal and snack frequency and diet quality in US adults: national health and nutrition examination survey 2003–2012. J Acad Nutr Diet 2016;116:1101–13. [DOI] [PubMed] [Google Scholar]
  • [41].Kim Y, Canada JM, Kenyon J, Billingsley H, Arena R, Lavie CJ, et al. Physical activity, sedentary behaviors and all-cause mortality in patients with heart failure: findings from the NHANES 2007–2014. PLoS One 2022;17:e0271238. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [42].DiBello JR, Miller R, Khandker R, Bourgeois N, Galwey N, Clark RV. Association between low muscle mass, functional limitations and hospitalisation in heart failure: NHANES 1999–2004. Age Ageing 2015;44:948–54. [DOI] [PubMed] [Google Scholar]
  • [43].Zhang H, Lin Z, Chen J, Gan D, Zhou H, Ma Z, et al. Association between dietary fiber intake and heart failure among adults: national health and nutrition examination survey 2009-2018. Front Cardiovasc Med 2022;9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [44].Yancy CW, Jessup M, Bozkurt B, Butler J, Casey DE, Colvin MM, et al. 2017 ACC/AHA/HFSA focused update of the 2013 ACCF/AHA guideline for the management of heart failure: a report of the American college of cardiology/American heart association task force on clinical practice guidelines and the heart failure society of America. Circulation 2017;136:e137–61. [DOI] [PubMed] [Google Scholar]
  • [45].Yancy CW, Jessup M, Bozkurt B, Butler J, Casey DE, Drazner MH, et al. 2013 ACCF/AHA guideline for the management of heart failure: executive summary. Circulation 2013;128:1810–52. [DOI] [PubMed] [Google Scholar]
  • [46].Heart Failure Society of A. Executive summary: HFSA 2010 comprehensive heart failure practice guideline. J Card Fail 2010; 16:475–539. [Google Scholar]
  • [47].An R. Weekend-weekday differences in diet among U.S. adults, 2003–2012. Ann Epidemiol 2016;26:57–65. 10.1016/j.annepidem.2015.10.010. [DOI] [PubMed] [Google Scholar]
  • [48].Selvaraj S, Kelly DP, Margulies KB. Implications of altered ketone metabolism and therapeutic ketosis in heart failure. Circulation 2020;141:1800–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [49].McCommis KS, Kovacs A, Weinheimer CJ, Shew TM, Koves TR, Ilkayeva OA-O, et al. Nutritional modulation of heart failure in mitochondrial pyruvate carrier-deficient mice. Nat Metab 2020;2:1232–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [50].Lopaschuk GD, Karwi QG, Tian R, Wende AR, Abel ED. Cardiac energy metabolism in heart failure. Circ Res 2021;128:1487–513. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [51].Wende AR, Brahma MK, McGinnis GR, Young ME. Metabolic origins of heart failure. JACC Basic Transl Sci 2017;2:297–310. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [52].Honka H, Solis-Herrera C, Triplitt C, Norton L, Butler J, DeFronzo Ralph A. Therapeutic manipulation of myocardial metabolism. J Am Coll Cardiol 2021;77:2022–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [53].Pan C, Herrero-Fernandez B, Borja Almarcha C, Gomez Bris R, Zorita V, Sáez A, et al. Time-restricted feeding enhances early atherosclerosis in hypercholesterolemic mice. Circulation 2023;147:774–7. [DOI] [PubMed] [Google Scholar]
  • [54].Ozcan M, Guo Z, Valenzuela Ripoll C, Diab A, Picataggi A, Rawnsley D, et al. Sustained alternate-day fasting potentiates doxorubicin cardiotoxicity. Cell Metabol 2023;35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [55].Betts JA, Chowdhury EA, Gonzalez JT, Richardson JD, Tsintzas K, Thompson D. Is breakfast the most important meal of the day? Proc Nutr Soc 2016;75:464–74. [DOI] [PubMed] [Google Scholar]
  • [56].Doukky R, Mangla A, Ibrahim Z, Poulin M-F, Avery E, Collado FM, et al. Impact of physical inactivity on mortality in patients with heart failure. Am J Cardiol 2016;117:1135–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [57].Kim Y, Canada JM, Kenyon J, Billingsley HE, Arena R, Lavie CJ, et al. Effects of replacing sedentary time with physical activity on mortality among patients with heart failure: national health and nutrition examination survey follow-up study. Mayo Clin Proc 2022;97:1897–903. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [58].Evenson KR, Buchner DM, Morland KB. Objective measurement of physical activity and sedentary behavior among US adults aged 60 years or older. Prev Chronic Dis 2012;9:E26. [PMC free article] [PubMed] [Google Scholar]
  • [59].Bohmke NJ, Billingsley HE, Kirkman DL, Carbone S. Non-pharmacological strategies in heart failure with preserved ejection fraction. Cardiol Clin 2022;40:491–506. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [60].Hahn VS, Petucci C, Kim M-S, Bedi KC, Wang H, Mishra S, et al. Myocardial metabolomics of human heart failure with preserved ejection fraction. Circulation 2023;147:1147–61. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1955290-supplement-1.pdf (695.1KB, pdf)

Data Availability Statement

All data utilized in this manuscript is publicly and freely available without restriction at https://www.cdc.gov/nchs/nhanes/index.htm.

RESOURCES