The research utilized data from the National Health and Nutrition Examination Survey (NHANES) spanning the years 1999 to 2014, which is a comprehensive nationwide survey aimed at evaluating the health and nutritional status of both adults and children across the United States. Participants diagnosed with diabetes were identified based on self-reported medical history, usage of hypoglycemic medications, or specific biochemical criteria, including serum HbA1c levels exceeding 6.5% or fasting serum glucose levels surpassing 7.0 mmol/L, as outlined below. Following the exclusion of individuals below 18 years of age (n = 109) and those lacking records of iron intake (n = 26), a total of 5970 diabetic participants were included in the analysis. Approval for the study was obtained from the institutional review board of the National Center for Health Statistics (Protocol #98-12), and written informed consent was obtained from all participants prior to their involvement in the study.

Detailed dietary intake information, encompassing the types and quantities of foods and beverages consumed, was gathered through a 24-h dietary recall method. To ensure accuracy, a multi-pass approach was employed to meticulously document iron intake. This method involved multiple passes of questioning to clarify and validate the reported intake of iron-rich foods and beverages. (https://www.cdc.gov/nchs/nhanes/measuring_guides_dri/measuringguides.htm). Trained interviewers conducted the interviews prior to the physical examination, utilizing the Computer-Assisted Personal Interview system, within the participants’ homes. During this process, detailed dietary information was collected. Subsequently, trained nutritionists meticulously reviewed the gathered data, cross-referencing reported dietary supplement entries with known supplements listed in the in-house NCHS Product Label Database to ensure accuracy and completeness (https://wwwn.cdc.gov/Nchs/Nhanes/2013-2014/DS1TOT_H.htm). The total energy and iron intake were calculated according to the dietary intake.

The primary outcome of interest in the study was all-cause mortality, while secondary outcomes included mortality specifically attributed to cardiovascular disease and malignant neoplasms. Mortality status was determined through linkage to the National Death Index, with data collected up to December 31, 2015. Cardiovascular disease-related mortality was defined using International Classification of Diseases, Tenth Revision (ICD-10) codes I00-I09, I11, I13, or I20-I51, while malignant neoplasm-related mortality was defined by ICD-10 codes C00-C97.

Baseline characteristics of participants were obtained through a combination of questionnaires and physical examinations, encompassing various aspects such as demographics (including sex, age, race, educational level, family poverty income ratio), lifestyle factors (such as physical activity, alcohol consumption, smoking), medical history (including hypertension and cardiovascular diseases), and medication usage (such as antihypertensive drugs, hypoglycemic drugs, and lipid-lowering drugs). Biomedical measurements were conducted as follows: Plasma fasting glucose levels were measured using an enzymatic method, while HbA1c levels were determined through an HPLC method. Triglycerides and LDL cholesterol levels were enzymatically assessed using the Roche Modular P chemistry analyzer. Estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease-Epidemiology Collaboration (CKD-EPI) equation. Race was categorized as non-Hispanic white, non-Hispanic black, Mexican American, other Hispanic, or others. Educational attainment was grouped into less than high school, high school or equivalent, and college or above. Poverty income ratio (PIR) was calculated as the ratio of household income reported by the participant to the appropriate poverty threshold for household size, categorized as <1, 1–3, and >3. Smoking status was classified as current, past, or never. Physical activity status was categorized as vigorous, moderate, or inactive. Hypertension was defined as a history of hypertension, blood pressure readings ≥140/90 mm Hg, or the use of antihypertensive medications. Cardiovascular disease (CVD) was defined based on self-reported conditions, including congestive heart failure, coronary heart disease, angina pectoris, heart attack, and stroke. For variables with missing values in Cox proportional hazards regression models, imputation was performed using the predictive mean matching method, where the observation with the closest predicted transformed value served as the donor.

Continuous variables were expressed as mean ± standard deviation, while categorical variables were presented as counts and proportions (percentage). Iron intake was divided into five groups based on quintiles, and differences between these groups were assessed using one-way analysis of variance (ANOVA) for continuous variables or chi-square tests for categorical variables. Univariate survival analysis was conducted using Kaplan–Meier analysis and the Log-rank test. Multivariate Cox proportional hazards regression models were utilized to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for all-cause, cardiovascular, and cancer mortality. Three models were employed for adjustment: Model 1 included age and gender; Model 2 further adjusted for race, education level, poverty income ratio (PIR), body mass index (BMI), smoking status, alcohol consumption, and physical activity; Model 3 additionally adjusted for hypertension, cardiovascular disease (CVD), estimated glomerular filtration rate (eGFR), baseline iron levels, and serum iron levels. The nonlinear relationship between iron intake and all-cause mortality was described using restricted cubic splines. Subgroup analysis was conducted to investigate the effect of baseline serum iron levels on this relationship. All statistical analyses were performed using R version 3.6, with statistical significance set at P < 0.05.