The publicly available data were collected from the 1999–2002 cycle of the National Health and Nutrition Examination Survey (NHANES), which is a program of studies designed to assess the health and nutritional status of adults and children in the United States. Each year, the survey assesses a sample of approximately 5,000 individuals from counties across the country, with 15 counties visited annually. NHANES is a significant program under the auspices of the National Center for Health Statistics (NCHS), which is a division of the Centers for Disease Control and Prevention (CDC), and is responsible for generating critical health and vital statistics for the country (20). Participants who had complete information on demographics, medical history (congestive heart failure, coronary heart disease, stroke, cancer, hypertension, diabetes and hospitalization), specific laboratory measurements (NT-proBNP, hemoglobin, serum iron, ferritin, transferrin saturation, creatinine, alanine aminotransferase, aspartate aminotransferase, C-reactive protein, segmented neutrophils percent and glycohemoglobin), dietary iron and total folate and corresponding sample weights were included. Of the 6,660 participants, 182 had HF. The survey protocol was approved by the NCHC (National Center for Health Statistics) Ethics Review Board and all participants provided informed written consent (21). All methods were carried out in accordance with Declaration of Helsinki.

The glomerular filtration rate was estimated by the CKD-EPI equation (22). In this study, the third quartile of the percentage of neutrophils in the general population was 64.6%. Based on this value, both the general population and HF patients were divided into high and low neutrophil percentage groups. In the present study, the median daily intake of iron and total folate in the general population were 13.42 mg and 345 mcg, respectively. Based on these values, the population was divided into high and low intake groups. Anemia was diagnosed in males with hemoglobin levels below 13 g/dl and females with hemoglobin levels below 12 g/dl. Participants with a ratio of family income to poverty less than 1.3 were classified as low-income group, while those with a ratio greater than or equal to 1.3 were classified as middle-to-high income group. According to the Physical Activity Questionnaire which was based on the Global Physical Activity Questionnaire and included questions related to daily activities, leisure time activities, and sedentary activities, the classification of cardiac function was divided into two groups: NYHA I/II and NYHA III/IV.

The in-person interview for participants was conducted in a private room in the NHANES mobile examination center in the United States to collect dietary data during the 24-h period prior to the interview. Some of the NHANES 1999–2000 dietary interview data were collected by a telephone dietary interview. In 2001, dietary intake data were collected using the NHANES computer-assisted dietary interview system, which was also used to collect dietary data for the 1999–2000 collection period. In 2002, data were collected using USDA's dietary data collection instrument, the Automated Multiple Pass Method. The sample weights Weighted Dietary Recall 4-Year (WTDR4YR) which were constructed by taking the Mobile Examination Center sample weights (WTMEC2YR) and further adjusting for the additional non-response and the differential allocation by day of the week for the dietary intake data collection were applied to produce national, representative estimates.

The National Center for Health Statistics has linked data collected from several NCHS population surveys with death certificate records from the National Death Index. The public-use LMF provide mortality follow-up data from the date of survey participation through December 31, 2019 (23).

The statistical analysis was performed using the R programming language (version 4.1.3). Continuous variables were presented as Median (IQR). Categorical variables were presented as n (unweighted) (%). To compare the baseline characteristics, the Wilcoxon rank-sum test for complex survey samples and the chi-squared test with Rao & Scott's second-order correction were used. Propensity score matching (PSM) was utilized to eliminate bias and control for demographics (age, sex, race, income and BMI) using the “MatchIt” package. The weighted Cox proportional hazards models and weighted logistic regression models were performed using the “survey” package. Variables with P ≤ 0.1 in univariable models were included in the multivariable models. Significance was defined as two-sided probability values less than 0.05.

During the period of 1999–2002, the NHANES database documented data from 21,004 individuals. However, only 6,663 individuals had complete baseline information, and an additional 3 individuals did not have survival data. Among the remaining 6,660 individuals, 182 self-reported that they had congestive heart failure (Figure 1).

For both HF patients and non-HF individuals, there were no significant differences in their iron indices (serum ferritin (99 vs. 85 ng/ml, P = 0.2), serum iron (14 vs. 15 μmol/L, P = 0.088) and TSAT (21% vs. 24%, P = 0.12) (Table 1). Besides, after adjusting for age, gender, income, and BMI using propensity score matching, there was no significant difference in the incidence of ID (as defined by all the aforementioned definitions) between individuals with HF and those without (Supplementary Table S1).

Among the cohort of 6,660 individuals, 373 decease cases were recorded after a 5-year follow-up, followed by 899 mortalities at the 10-year mark, and 1,416 mortalities at the 15-year follow-up period.

Within the group of 182 patients diagnosed with HF, 46 fatalities were recorded during the 5-year follow-up, followed by 94 mortalities at the 10-year follow-up, and 126 mortalities at the 15-year follow-up.

In the general population, the median daily intake of iron is 13.42 mg, while the median daily intake of total folate is 345 mcg. Based on the median intake, the entire population was categorized into high and low intake groups.

From the data of NHANES III, the prevalence of ID (defined by heart failure guideline criteria) in 574 participants (male: 50.2%) with self-reported HF was 61.3% (5). In a prospective observational study of 546 patients (male: 88%) with stable systolic chronic HF, the prevalence of ID (defined by heart failure guideline criteria) was 37% in the entire population of systolic chronic HF patients (4). In a survey of 832 hospitalized patients with decompensated chronic HF, the prevalence of ID (defined by heart failure guideline criteria) was found to be 68.6% in male and 75.3% in female (3). In three European population-based cohorts (male: 45.2%), 60% of individuals had ferritin levels below 100 ng/ml, 16.4% had ferritin levels below 30 ng/ml, and 64.3% met the heart failure guideline criteria for ID (6). In the present study, the prevalence of ID defined by guideline criteria in HF patients (male: 52%) was 66%. It was noteworthy that the prevalence of ID was comparable in the non-HF population (Table 1, Supplementary Table S1), which contradicted our previous assumptions. A large proportion of HF patients had impaired nutritional status (24), which may affect iron metabolism. Besides, the proportion of anemia in HF patients is relatively high. These factors make us unconsciously believe that the proportion of ID in HF patients is relatively high. Alternatively, it was also hypothesized that HF patients would demonstrate a lower incidence of ID due to the pro-inflammatory state associated with HF, which could lead to elevated levels of ferritin (25, 26). In terms of disease prevalence, the existing diagnostic criteria for ID are ineffective in distinguishing between individuals with HF and those without.

In line with the findings reported by Masini et al. (7), our study revealed a significant association between TSAT <20% and all-cause mortality over a five-year period in HF patients, whereas levels of ferritin did not demonstrate a significant association (Table 2). However, our findings demonstrated a significant association between low levels of ferritin (<30 ng/ml) and an elevated risk of long-term all-cause mortality in both HF patients and the general population. Hence, TSAT emerges as a predictor of short-term all-cause mortality in HF patients, whereas ferritin demonstrates greater suitability for predicting long-term all-cause mortality, possibly due to the inherent fluctuations of TSAT and ferritin levels in the human body. In patients with HF, 55.5% of patients with iron ≤13 μmol/L at baseline remained ID after 1 year. Applying TSAT <20% as the definition of ID, 58.2% of patients remained ID after 1 year. When applying the heart failure guideline criteria, this proportion rose to 79.4% (27). Of the 537 patients with baseline ID (TSAT <20%), 53% (284 individuals) still had ID after 6 months. Meanwhile, 20% (128 individuals) of the 451 patients without baseline ID developed ID by the 6-month follow-up (15). Another possible reason is that the definitions of ID mentioned above were overly inclusive, encompassing a substantial proportion of individuals who did not truly exhibit ID. Therefore, the observed changes in iron status may have been only normal fluctuations within the normal range. A randomized, placebo-controlled, double-blind trial (IronIC) found that heart transplant recipients with ID defined by heart failure guideline criteria had no improvement in peak oxygen consumption 6 months after intravenous iron treatment while the peak oxygen consumption was significantly improved in patients with ferritin <30 ng/ml at baseline (4.4 ml/kg/min, 95% confidence interval 0.01–8.72, P = 0.0495) (28). Hence, a diagnostic threshold of ferritin <30 ng/ml may be more appropriate, as it is associated with the prognosis of both HF and non-HF individuals. Nevertheless, additional research and validation are warranted to investigate the stability of iron status in individuals presenting ferritin levels below this threshold.

In the present study, it is recommended that when the ferritin level is <30 ng/ml or the TSAT level is <20%, treatment should be considered for the general population aged 45 years and above. For patients with HF, intravenous iron replacement therapy should be considered when the ferritin level is <30 ng/ml (15%) or the TSAT level is <20% (45%). In our study, a ferritin level <300 ng/ml is also associated with all-cause mortality in patients with HF over a period of five years, therefore, a ferritin level >300 ng/ml may be used as a treatment target, and revious research has demonstrated the feasibility of achieving a ferritin level >400 ng/ml through intravenous iron replacement therapy (8).

Iron from dietary sources exists in two primary forms: haem iron and non-haem iron. Haem iron is absorbed from the gastrointestinal tract with notably higher efficiency (29). Notably, the consumption of haem iron, rather than total iron intake, exhibits a significant correlation with ferritin levels (16, 17). Extensive research has demonstrated the enhanced effectiveness of a combination of folic acid and iron, as opposed to iron supplementation alone, in addressing ID (18, 19). In the present study, after adjusting for confounding factors such as gender, there was no significant relationship between dietary iron intake and ID. It is not surprising that there is no significant association between dietary iron intake and ID when oral iron therapy failed to achieve the desired effect (30). However, for HF patients, after adjusting for covariates such as anemia, gender and anemia-related treatment, a significant correlation was observed between higher total folate intake and an increased risk of ferritin <30 ng/ml. That might be explained by the tendency of HF patients with low levels of ferritin to receive recommendations or to spontaneously increase their consumption of iron and folate-rich foods. After excluding these severe ID patients (ferritin <30 ng/ml), it is promising that an increase in the consumption of folate-rich foods seems to contribute to a reduction in the incidence of ID (ferritin <100 ng/ml, OR: 0.80; 95% CI: 0.65–1.00; P = 0.047). Given the suboptimal outcomes observed in the treatment of ID in HF patients using oral iron supplementation, oral folic acid may represent a promising avenue for intervention (30). However, it is imperative to elucidate the underlying mechanism through which increased folate intake mitigates iron deficiency, without considering iron intake, warranting further exploration.