The regulatory effect of blood group on ferritin levels in aging: a retrospective study

Study design and participants

The biochemical examination data were collected from 3,843 individuals aged 40 and above who underwent blood type and ferritin testing at Beijing Zhongguancun Hospital between July 2019 and July 2024. A total of 883 middle-aged and elderly individuals with missing a large number of examination data were excluded from the study. Ultimately, 2,960 healthy participants were included in this retrospective study. The Ethics Committee of the Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences, approved the study protocol (ZS2024001). The study was conducted in accordance with the Declaration of Helsinki and its amendments. All participants provided written informed consent prior to enrollment. Clinical trial number: not applicable.

Measures

Age, sex, self-reported health status, and medical history were recorded. Blood samples were obtained under fasting conditions and analyzed within 2 h of centrifugation. The blood chemistry studied was assessed by standard automated techniques, including Architect C 8000 auto-analyzer and Axsyme Third-Generation Immunoassay System (Abbott, Abbott Park, IL, USA). This analysis included indicators of liver function, glycometabolism, lipid metabolism, myocardial function, thyroid function, and inflammatory factors (Table S1).

Statistical analysis

Descriptive statistics and Spearman correlation analysis were conducted using SPSS version 27.0 (SPSS Inc., Armonk, NY, USA). Continuous variables are presented as mean ± standard deviation, while categorical variables are presented as frequency (percentage). Then, assumption testing is employed to assess the normal distribution of continuous variables in the context of regression analysis. The normal distribution of continuous variables was assessed using the Kolmogorov–Smirnov and Shapiro–Wilk tests. The Mann–Whitney test was employed to compare the normal distribution between the two gender groups. The Kruskal-Wallis test was utilized to compare median differences among multiple blood type groups. P-values less than 0.05 in assumption testing were considered not consistent with the normal distribution. Spearman correlation analysis was performed to evaluate the relationship between biochemical indices and ferritin levels. Finally, age was treated as the independent variable, with gender and biochemical indicators associated with ferritin serving as control variables. Blood type was analyzed as a moderating variable to determine its effect on the relationship between age and ferritin levels. P-values less than 0.05 were considered statistically significant.

Descriptive statistical analysis of each biochemical index

We collected biochemical examination data from 3,843 individuals aged 40 and above who underwent health check-ups at Beijing Zhongguancun Hospital. A total of 2,960 healthy participants, with an average age of 53.96 ± 12.19 years, were included in this study, of which 900 (30.41%) were male. Among these participants, 859 (29.02%) had blood type O, 831 (28.07%) had type A, 978 (33.04%) had type B, and 292 (9.86%) had type AB (Table 1). There were individual differences in various biochemical indices, particularly in liver and kidney function, metabolic function, and myocardial damage. Notably, the mean ferritin level was 288.16, with a standard deviation of 99.89, indicating significant variability in ferritin levels among individuals. However, the median values of all indicators were close to the mean (Table 1), suggesting that the distribution of these indicators is relatively concentrated and suitable for further correlation and regression analyses. The flowchart of this study was shown in Fig. 2.

Correlation and univariate analysis of ferritin and other indicators

Since all variables did not conform to a normal distribution (p < 0.05) (Table S2). Spearman correlation analysis was employed for continuous variables, while a non-parametric test was utilized for categorical variables.

The results indicated a negative correlation between ferritin levels and age (ρ = −0.099, p < 0.001), suggesting that ferritin levels tend to decrease as age increases (Table 2). Additionally, the difference in ferritin levels between men and women was statistically significant (p = 0.005) (Table 3), highlighting a notable disparity in ferritin levels between the sexes. Furthermore, there were statistically significant differences in ferritin levels across various blood groups (p < 0.001) (Table 4).

Alanine aminotransferase (ALT) (ρ = 0.078, p < 0.001), aspartate aminotransferase (AST) (ρ = 0.039, p < 0.05), γ-glutamyl transpeptidase (γ-GT) (ρ = 0.074, p < 0.001), creatinine (Cr) (ρ = 0.105, p < 0.001), alkaline phosphatase (ALP) (ρ = 0.113, p < 0.001), and lipoprotein(a) (LP(a)) (ρ = 0.059, p < 0.01) demonstrated significant positive correlations with ferritin (Table 2). This suggests that abnormalities in liver, bile, and renal function are associated with increased ferritin levels.

There was a significant negative correlation between apolipoprotein A1 (APOA1) and ferritin (ρ = −0.086, p < 0.001) (Table 2), suggesting that an increase in ferritin levels is associated with a decrease in high-density lipoprotein (HDL) metabolism.

There were significant positive correlations between ferritin and creatine kinase (CK) (ρ = 0.156, p < 0.001), lactate dehydrogenase (LDH) (ρ = 0.224, p < 0.001), and myoglobin (Mb) (ρ = 0.207, p < 0.001) (Table 2). These findings suggested that elevated ferritin levels may be closely associated with myocardial injury.

Thyroid-stimulating hormone (TSH) (ρ = 0.126, p < 0.001) exhibited a positive correlation with ferritin, whereas free triiodothyronine (FT3) (ρ = −0.174, p < 0.001) and total triiodothyronine (TT3) (ρ = −0.168, p < 0.001) demonstrated negative correlations with ferritin (Table 2). These findings suggested a relationship between thyroid function and ferritin levels.

Interleukin-6 (IL-6) (ρ = 0.103, p < 0.001) exhibited a positive correlation with ferritin (Table 2), suggesting that ferritin levels are closely associated with the elevation of inflammatory factors.

Glucose (Glu) (ρ = 0.042, p < 0.05) and small and dense low-density lipoprotein cholesterol (sd LDL-C) (ρ = 0.038, p < 0.05) exhibited a weak positive correlation with ferritin (Table 2). This finding suggested that metabolic disorders may be associated with ferritin levels.

The regulatory effect of blood group on ferritin levels with aging

We included indicators related to ferritin and analyzed the regulatory effects with age as the independent variable, gender and other biochemical indicators as control variables, and blood type as the moderating variable.

In Model 1, which analyzed the direct effects of age and several control variables (including gender, ALT, AST, creatinine, glucose, etc.) on ferritin levels, age exhibited a significant negative effect on ferritin levels (t = −2.340, p = 0.019). Specifically, as age increased, ferritin levels decreased significantly (Table 5). These results indicate that age is an independent factor influencing ferritin levels.

Model 2 incorporates blood type as a regulatory variable based on Model 1 to investigate whether different blood types influence the relationship between age and ferritin levels. The results indicated that blood group B (p = 0.042) and blood group AB (p = 0.048) had significant effects on ferritin levels compared to blood group O, while blood group A did not show a significant difference (Table 5). This suggests that blood groups B and AB may modulate the effect of age on ferritin levels to some extent.

Model 3 further incorporated the interaction term of age and blood type to elucidate the specific regulatory effects of different blood types on the relationship between age and ferritin levels. The results indicated that various blood groups indeed play a significant regulatory role in this relationship. In blood group A (p = 0.003, β=−0.072) and blood group B (p < 0.001, β=−0.110), the negative impact of age on ferritin levels was more pronounced, demonstrating a stronger regulatory effect compared to other blood groups. In populations with blood group AB (p = 0.033, β =−0.044), the negative effect of age on ferritin was greater than that observed in blood group O, although the regulatory effect was relatively weak (Table 5).

Strength and limitation

The strength of this study lies in its analysis of the regulatory effect of blood groups on ferritin levels in middle-aged and elderly individuals as a cohesive group. This approach helps to mitigate the loss of continuity associated with ageing that can occur when dividing participants into distinct age groups. Consequently, both the independent and dependent variables in this study are treated as continuous variables.

Our study has several limitations. First, the cross-sectional analysis of the data does not permit a causal assessment of the relationship under investigation. Second, while we have accounted for some confounding factors, our study lacks data on diet and lifestyle, which may influence ferritin levels. Third, although we included and analyzed some relevant biochemical indicators as control variables, the study did not comprehensively cover all biochemical indicators, and the results may be affected by other unmeasured biochemical factors. Therefore, we plan to incorporate questionnaires and follow-up assessments, as well as include additional biochemical indicators, to enhance the comprehensiveness and rigor of the study in the future.