Introduction

As critical micronutrients, vitamin A (retinoids) and vitamin E (tocopherols) demonstrate pleiotropic biological functions, particularly in immunomodulation, skeletal system homeostasis maintenance, and regulation of cellular proliferation and differentiation processes1,2,3,4. Their deficiencies are associated with a range of health issues, particularly in children, where growth and development are most critical5. Despite their importance, the relationship between these two vitamins and the influence of demographic, environmental, and physiological factors on their levels remain complex and not fully understood.

Vitamin E, a potent antioxidant, is crucial for protecting cell membranes, supporting immune function, and maintaining neurological health. Its importance is further underscored by its etymological origin as “tocopherol,” derived from the Greek words “toc” (related to child) and “phero” (to bring), highlighting its essential role in normal childhood development. It is primarily obtained through diet, with vegetable oils, nuts, and green leafy vegetables being rich sources6. Vitamin E deficiency can lead to neurological disorders, anemia, and increased susceptibility to infections7. Vitamin A, on the other hand, is vital for vision, immune function, and reproductive health. It is obtained through dietary sources such as liver, fish oils, and dairy products, or synthesized from provitamin A carotenoids found in fruits and vegetables8,9. Deficiency in Vitamin A is a leading cause of preventable blindness in children and can also result in increased susceptibility to infections and other health complications10.

The metabolism and function of Vitamin E and Vitamin A are closely interrelated. Both vitamins are fat-soluble and share common transport mechanisms in the body. Recent studies have suggested that there may be synergistic effects between these two vitamins, particularly in terms of their antioxidant functions and their roles in immune modulation11,12. However, the exact nature of their interaction and the potential threshold effects in children have not been adequately explored. Understanding this relationship is crucial for developing effective public health interventions aimed at addressing micronutrient deficiencies in pediatric populations.

Globally, micronutrient deficiencies are a significant public health concern, particularly in low- and middle-income countries where access to diverse diets and supplements may be limited13. In children, these deficiencies can have long-term consequences on growth, development, and overall health outcomes. Previous research has examined the individual roles of Vitamin E and Vitamin A, but there is a paucity of studies investigating their combined effects and potential threshold interactions in children. Identifying such thresholds could have substantial implications for optimizing the status of these vitamins through targeted interventions.

This cross-sectional investigation employs population-stratified sampling to elucidate the biphasic dose-response correlation between Vitamin E and Vitamin A concentrations in pediatric cohorts. We postulate a critical serum threshold beyond which retinol homeostasis demonstrates stabilization, as quantified through segmented regression modeling. The dual focus on deficiency epidemiology and micronutrient interaction dynamics aims to advance nutritional biomarker interpretation frameworks, with potential applications in:1 refining dietary reference intake algorithms2, optimizing combined supplementation protocols, and3 developing age-specific deficiency risk stratification matrices.

Methods

Study design

This cross-sectional study set out to explore the relationship between Vitamin E and Vitamin A levels in children aged 0–10.8 years (Clinical trial number: not applicable). The sample size was calculated based on the expected effect size, statistical power, and significance level. We aimed to detect a moderate effect size (Cohen’s d = 0.5) with a power of 80% and a significance level of 0.05. Based on these parameters, a minimum sample size of 344 participants was required to ensure sufficient statistical power for identifying significant associations, thereby underpinning the reliability of the results. Participant recruitment took place from January 1, 2018, to December 31, 2021. This timeframe allowed for a comprehensive capture of the participants’ health status at a specific moment, facilitating an accurate analysis of Vitamin E prevalence and its relationship with Vitamin A levels. The study obtained ethical approval from the Ethics Committee of the Affiliated Women and Children’s Hospital of Ningbo University (No:20131220). All participants and their parents or legal guardians provided signed, informed consent. Children found to have risk levels of Vitamin E or Vitamin A were referred for further medical evaluation and appropriate interventions. The raw dataset generated and analysed during the current study is provided as Supporting Information (S1 raw data).

Study population

The study enlisted children aged 0–10.8 years who routinely visited community health service centers for health check-ups in Ningbo, Zhejiang, China. Participants were selected from those actively participating in the community health service centers’ regular health check-up programs. This cohort was particularly relevant for assessing the prevalence and impact of Vitamin A and E statuses in children. Eligibility criteria included children present at the health service centers on the examination day, with no known chronic conditions that might affect Vitamin A or Vitamin E status. To ensure the study’s validity, certain exclusions applied. Children exposed to recognized environmental contaminants were excluded to focus on the effects of Vitamins A and E. Specifically, 3 children were excluded because their blood lead (Pb) or copper (Cu) concentrations exceeded the age-specific reference limits (lead ≥ 5 µg/dL; copper ≥ 140 µg/dL). Also excluded were children on medications affecting bone metabolism and those with conditions impacting Vitamin E absorption or metabolism. 11 children were excluded due to the use of medications such as long-term corticosteroids and antiepileptic drugs, and 5 children were excluded due to conditions such as abetalipoproteinemia and chronic liver diseases. This recruitment strategy aimed to reflect the general pediatric population’s Vitamin A and E statuses accurately. It enabled a practical examination of the relationship between these micronutrients and health outcomes, offering insights beneficial for public health and clinical practice.

Measurement of Vitamin E and Vitamin A

During the physical examination, blood samples were obtained from participants, immediately stored at -20 °C, and subsequently transported to a central laboratory for detailed analysis. The serum concentrations of Vitamin E and Vitamin A were measured using Inductively Coupled Plasma Mass Spectrometry (ICP-MS), specifically with the AB SCIEX Triple Quad 4500MD Analyzer (Sciex, part of Danaher Corporation, located in Framingham, Massachusetts, USA). Serum (100 µL) was mixed with ethanol containing deuterated internal standards, extracted twice with hexane, dried under N₂ and re-dissolved in methanol/isopropanol 80:20 for injection.To ensure the accuracy and reliability of the measurements, stringent daily quality control protocols were implemented, utilizing control materials supplied by Hehe (Hefei Hehe Medical Technology Co., Ltd., Anhui, China). Quality-control pools (low, medium, high) were analysed in every batch; only runs within ± 10% of target were accepted. Across 42 batches the inter-assay CV was 4.8% for vitamin A and 5.2% for vitamin E, and NIST-spike recovery ranged 97–103%, confirming consistent measurement throughout the 4-year study.

The results for each vitamin were evaluated using age-specific reference intervals. For Vitamin A, levels below the lower limit of the reference interval were classified as “Deficiency,” those within the interval were considered “Normal,” and levels above the upper limit were labeled as “Excess.” The specific reference intervals for Vitamin A are as follows:

  • 0–6 years: 113.00-647.00 ng/mL.

  • 7–12 years: 128.00-812.00 ng/mL.

  • 13–17 years: 144.00-977.00 ng/mL.

  • ≥ 18 years: 325.00-780.00 ng/mL.

Similarly, Vitamin E levels were categorized based on the same criteria of “Deficiency,” “Normal,” and “Excess” relative to the reference intervals:

  • 0–1 years: 1.00–5.00 µg/mL.

  • 2–12 years: 3.00–9.00 µg/mL.

  • 13–19 years: 6.00–10.00 µg/mL.

  • >19 years: 5.00–18.00 µg/mL.

These reference intervals were determined through extensive evaluation of serum vitamin levels in a diverse and representative sample of healthy individuals, ensuring their applicability to the general population. The laboratory’s quality control measures, which include the use of control materials and regular calibration, play a crucial role in maintaining the accuracy and reliability of these reference intervals.

Statistical analysis

For the statistical analysis, continuous variables were represented as mean ± standard deviation (SD), while categorical variables were presented as frequencies or percentages. Intergroup differences were assessed using the Student t-test or one-way ANOVA for continuous variables, and the Chi-square (χ²) test or Fisher’s exact test for categorical variables, as appropriate. To investigate the relationship between Vitamin E and Vitamin A levels, smooth curve fitting was employed, followed by multivariate linear regression analyses. These analyses included both unadjusted and multivariate-adjusted models to evaluate the stability of the relationship. In Model 2, adjustments were made for age and gender, while Model 3 further accounted for seasonal variations. The findings are reported as beta coefficients (β) with corresponding 95% confidence intervals (CIs). To evaluate the threshold effect, two-piecewise regression models were utilized to examine the dose-response relationship between Vitamin E and Vitamin A levels, with the likelihood ratio test validating the threshold effect’s significance.

All statistical analyses were executed using the R statistical software (http://www.R-project.org, The R Foundation) and Free Statistics software (version 2.1.1, Beijing Free Clinical Medical Technology Co., Ltd.). A p-value of less than 0.05 (two-tailed) was considered statistically significant.

Results

Baseline characteristics

The baseline characteristics of the study participants stratified by Vitamin E status categories are presented in Table 1. Among the total of 4,752 participants, 88 were categorized as having Vitamin E deficiency, 3,950 had normal levels, and 714 were classified as having excess Vitamin E. Significant variations were observed in the median age across the Vitamin E status categories, with the deficiency group showing a higher median age (1.5 years) compared to the normal and excess groups (0.7 and 0.6 years respectively; p < 0.001). Seasonal distribution also exhibited significant differences, with the highest proportion of participants in the summer season (35.2% in the deficiency group) and the lowest in the winter season (13.6% in the deficiency group; p < 0.001).

Mean Vitamin E levels were significantly different across the categories, with the lowest mean level recorded in the deficiency group (2.4 ± 0.5 µg/mL) and the highest in the excess group (7.0 ± 1.6 µg/mL; p < 0.001). Similarly, mean Vitamin A levels showed significant variation, with the deficiency group having the lowest mean Vitamin A level (236.9 ± 60.4 ng/mL) and the excess group the highest (292.4 ± 72.3 ng/mL; p < 0.001). Gender distribution revealed a significant difference across the Vitamin E status categories (p < 0.001), with a slightly higher proportion of males in the deficiency and normal groups (53.4% and 53.5% respectively) compared to the excess group (45.9%). Conversely, females constituted a higher proportion in the excess group (54.1%) than in the deficiency and normal groups (46.6% and 46.5% respectively).

Table 1 Descriptive characteristics of participants by Vitamin E status categories.
Full size table

The descriptive characteristics of participants by Vitamin A status categories are delineated in Table 2. Among the total of 4,752 participants, 40 were identified as having Vitamin A deficiency, 4,710 had normal levels, and 2 were classified as excess. Owing to the very small number of participants classified as vitamin A excess (n = 2, 0.04%), this stratum was merged with the normal group to preserve statistical stability and clinical interpretability. A statistically significant variation in gender distribution was observed across the Vitamin A status categories, with a markedly higher proportion of females in the deficiency group relative to males (70.0% females versus 30.0% males in the deficiency group, p = 0.004). Notably, there was a statistically significant difference in the median age of participants across the Vitamin A status categories (p < 0.001). The median age was 0.5 years (IQR: 0.5–0.7) in the deficiency group and 0.7 years (IQR: 0.5–2.0) in the combined normal and excess group. In terms of seasonal distribution, a significant difference was evident, with the deficiency group exhibiting a greater proportion of participants during the summer and autumn seasons (p = 0.006). Furthermore, Vitamin E levels were found to be significantly lower in the deficiency group when compared to the normal and excess groups (p < 0.001). The mean Vitamin A level was considerably lower in the deficiency group at 98.4 ± 15.1 ng/mL, in contrast to the normal or excess group, where it was recorded at 261.1 ± 72.9 ng/mL, underscoring a statistically significant difference across the categories (p < 0.001).

Table 2 Descriptive characteristics of participants by Vitamin A status categories.
Full size table

Association between Vitamin E and Vitamin A

Table 3 presents the regression outcomes assessing the relationship between serum vitamin E and vitamin A concentrations. Three hierarchical models were implemented with incremental covariate adjustment to evaluate this association.

Model 1 demonstrated a significant positive association between serum Vitamin E and Vitamin A concentrations, with each 1 µg/mL increment in Vitamin E corresponding to a β-coefficient of 14.47 (95% CI:13.47–15.47) for Vitamin A. This association persisted in Model 2 following covariate adjustment for sex and age (β = 14.4, 95% CI:13.35–15.46). Additional adjustment incorporating seasonal variation in Model 3 showed negligible impact on effect estimates (β = 14.5, 95% CI:13.45–15.56).

Stratified analysis by Vitamin E quintiles revealed dose-dependent increases in serum Vitamin A concentrations. Relative to the lowest Quintile (Q1: <3.15 µg/mL), covariate-adjusted models demonstrated progressive elevations: Q2 (3.15–4.18 µg/mL) β = 38.36 (95%CI:32.85–43.87); Q3 (4.18–5.82 µg/mL) β = 54.94 (49.43–60.44); Q4 (≥ 5.82 µg/mL) β = 77.94 (72.44–83.45). Significant linear trends persisted across all analytical models (p for trend < 0.001).

Gender-stratified analyses revealed statistically significant positive correlations in both sexes, demonstrating elevated β coefficients in fully adjusted models. Age-stratified evaluations maintained significance across all subgroups, with peak effect magnitudes observed in children aged 6–10.8 years. Significant seasonal variation was evident, with maximal β values recorded during summer and autumn meteorological periods.

Table 3 Association between Vitamin E and Vitamin A.
Full size table

Figure 1 and Table 4 showcase the findings from the analysis of the relationship between Vitamin E and Vitamin A levels, particularly emphasizing the threshold effect.

Fig. 1
figure 1

Relationship between Vitamin E and Vitamin A. Solid and dashed lines represent the predicted value and 95% confidence intervals. They were adjusted for gender, age and season.

Full size image
Table 4 Threshold effect analysis of the relationship of Vitamin E and Vitamin A. Adjusted for gender, age and season.
Full size table

In the restricted cubic spline model (Fig. 1), the correlation between Vitamin E levels and Vitamin A was linear (p < 0.001). The two-piecewise regression models identified a significant threshold effect in their relationship. As shown in Table 4, after adjusting for gender, age, and season, the threshold effect analysis indicates a critical threshold at a Vitamin E level of 3.579 µg/mL. Below this threshold, there is a robust correlation between Vitamin E and Vitamin A levels, with a beta coefficient (β) of 35.829 (95% CI: 30.217, 41.441). However, once Vitamin E levels reach or exceed this threshold, the beta coefficient drops substantially to 9.828 (95% CI: 8.250, 11.406), indicating a weaker association. The likelihood ratio test also confirms the significance of this threshold effect (p < 0.001), further supporting the two-phase linear relationship between Vitamin E and Vitamin A levels. Across the 1,000 bootstrap iterations, the median threshold remained 3.579 µg/mL (2.5th–97.5th percentile: 3.462–3.698 µg/mL), confirming the stability of the inflection point originally identified.

Discussion

Our study offers an extensive analysis of the relationship between Vitamin E and Vitamin A levels in children aged 0–10.8 years, revealing a two-phase linear pattern with a critical threshold in Vitamin E levels where the correlation with Vitamin A stabilizes. This threshold is highly relevant for public health strategies aiming to enhance the status of these vitamins in children. The baseline characteristics in Table 1 show significant variations in Vitamin E and Vitamin A levels across different deficiency categories. The differences in these vitamins’ mean levels by gender, age, and season highlight the impact of demographic and environmental factors on micronutrient status.

The elevated mean age in the severe deficiency group and seasonal fluctuations suggest a complex interplay between Vitamin E synthesis, dietary intake, and other environmental factors that merits further exploration. Older children might have lower Vitamin E levels due to accelerated growth and increased demand for skeletal and muscular development, which could deplete body stores of Vitamin E14,15. The observed seasonal variations align with findings that UVB radiation and dietary patterns vary with seasons, influencing Vitamin E status16,17. Gender differences in Vitamin E levels could be influenced by variations in metabolic rates, body composition, and dietary habits. This balance is crucial as excessive intake of Vitamin E can have pro-oxidant effects, while insufficient intake leads to inadequate Vitamin E status, underscoring the need for public health strategies that optimize Vitamin E intake while avoiding excessive supplementation18,19.

Association analysis reveals a strong link between Vitamin E and Vitamin A levels, even after controlling for multiple covariates. The progressively increasing beta coefficients across the three models indicate that higher Vitamin E levels are positively correlated with higher Vitamin A levels. This relationship may be explained by several potential mechanisms. Both Vitamin E and Vitamin A are fat-soluble antioxidants that work synergistically to protect cell membranes and reduce oxidative damage20. Their absorption, transport, and storage rely on lipid metabolic pathways, which means their levels in the body are often influenced by similar factors21. A deficiency in either vitamin can impair antioxidant defenses, increasing the risk of cellular damage and disease21. The relatively high β-coefficient in the 6–10.8-year group stems mainly from the pubertal growth spurt, when expanding adipose depots and sex-hormone-driven lipolysis transiently elevate circulating retinol, amplifying vitamin E–dependent stabilization of RBP. The positive association observed in this study may reflect a shared mechanism where both vitamins contribute to critical physiological processes such as antioxidant function and cell membrane stability. Future research should further explore these interactions to better understand their implications for health and disease prevention.

Our study detected a threshold phenomenon, with a significant inflection point at 3.579 µg/mL of Vitamin E. This implies the Vitamin E-Vitamin A relationship varies with Vitamin E concentrations. When Vitamin E levels are below this point, the two vitamins’ association strengthens considerably. It hints that Vitamin E might be more crucial in regulating Vitamin A levels under low - concentration conditions.

Multiple biological mechanisms can clarify this. As a potent antioxidant22, Vitamin E protects cell membranes and averts Vitamin A oxidation. Vitamin E deficiency can escalate oxidative stress, indirectly impacting Vitamin A status by changing key enzymes in Vitamin A metabolism. Vitamin E’s impact on immunity is well - established23. It modulates cytokine expression, influencing Vitamin A metabolism and retinoid signaling. In infection or inflammation, the Vitamin E-Vitamin A interaction is more vital at low Vitamin E levels.

Genetic factors also matter. For instance, α - tocopherol transfer protein gene variants can alter Vitamin E metabolism24. Genetic elements affecting Vitamin E absorption and transport can influence its levels and subsequent Vitamin A interactions25. Vitamin E synthesis and distribution depend on dietary intake26. In regions with sparse Vitamin E - rich food access, Vitamin E deficiency is more common, intensifying the Vitamin E-Vitamin A interaction27. Environmental factors like cooking and food processing also affect Vitamin E and Vitamin A28. These factors may contribute to the threshold effect by influencing the balance between Vitamin E intake and metabolic demands.

Above the identified threshold of 3.579 µg/mL, the vitamin E–vitamin A association weakens, and two complementary mechanisms likely underlie this plateau. First, intestinal α-tocopherol absorption becomes saturable once luminal concentrations exceed the transport capacity of NPC1L1, SR-BI and CD36, causing fractional uptake to fall from ≈ 70% to < 35% and blunting further increases in circulating vitamin E29. Second, hepatic export mediated by α-tocopherol transfer protein (α-TTP) reaches an upper limit because excess vitamin E stabilises the protein against degradation without additional transcriptional up-regulation, thereby flattening the dose-response curve and attenuating downstream effects on retinoid-binding proteins30. Collectively, these processes explain why additional vitamin E beyond 3.579 µg/mL confers diminishing influence on vitamin A status.

Cross-regional and age-stratified evidence corroborate the threshold we observed. In a cohort of 1,823 healthy individuals aged 0–59 years from southern Sichuan, China, Li et al. reported a vitamin E–vitamin A correlation that was strongest in children ≤ 6 years and gradually attenuated with age; although no formal inflection point was calculated, the slope change occurred at a serum vitamin E concentration of ≈ 3.3 µg/mL, closely aligned with our 3.579 µg/mL threshold31. Similarly, analysis of the US NHANES 2005–2018 cycles (ages 8–19 y) showed that each 1 µg/mL increment in vitamin E corresponded to a 12 ng/mL rise in vitamin A among children, whereas the increment fell to 5 ng/mL among adolescents, again suggesting saturation beyond ≈ 3 µg/mL19. These data confirm consistency across ages and regions.

The identification of a nonlinear correlation and a threshold point in children’s Vitamin E levels in our study carries substantial clinical weight. This highlights the critical nature of preserving Vitamin E within an optimal bandwidth to bolster children’s health, particularly among high - risk cohorts, such as ≥ 6 years, rural residence, or < 1 serving week⁻¹ of vitamin E-rich foods—the groups with the highest deficiency rates. Our data supply a definitive threshold, aiding clinical observation and intervention. This enables healthcare providers to direct Vitamin E supplementation and dietary guidance, guaranteeing children’s Vitamin E levels are appropriately balanced. Public health initiatives ought to advocate for adequate Vitamin E consumption via diet while curbing the perils of over - supplementation. Personalized interventions should cater to children with Vitamin E deficiency, taking into account the interplay with Vitamin A. The threshold point represents a pivotal juncture where Vitamin E’s influence on health outcomes plateaus, presenting a focal point for clinical action. This understanding can underpin targeted efforts to optimize Vitamin E status, possibly enhancing pediatric health outcomes.

Our study provides novel insights into the dynamic between Vitamin E and Vitamin A levels, yet several constraints must be acknowledged. The cross - sectional nature of the study implies that data collection occurred at a solitary time point. This restricts our capacity to deduce causation or track the temporal evolution of these micronutrients. Moreover, our sample was restricted to children in Ningbo, Zhejiang, an area with unique climatic and dietary characteristics. This may circumscribe the applicability of our findings to regions with disparate climates, latitudes, or eating patterns. Lastly, despite excluding children with chronic conditions or specific medications, unmeasured confounders like genetic diversity, dietary intake, socioeconomic standing, or micronutrient supplementation habits might still skew the observed relationships. Future research should aim to overcome these limitations by employing longitudinal methodologies and factoring in a wider array of potential confounders.

Conclusions

Our research solidifies the understanding of the significant interplay between Vitamin E and Vitamin A in children, pinpointing a pivotal threshold at 3.579 µg/mL for Vitamin E, after which its linkage to Vitamin A levels stabilizes. This threshold accentuates the necessity of adequate Vitamin E levels for maintaining proper Vitamin A status, especially in vulnerable groups like older children or those with limited access to Vitamin E-rich foods. While our study lays a foundational understanding, there’s a clear need for future longitudinal research. Such studies could validate our findings and delve into whether changes in Vitamin E levels can forecast Vitamin A deficiencies or other health-related issues over time. This line of inquiry has the potential to clarify the cause-and-effect relationships involved and to guide the development of more precise and effective interventions for children at risk. Our work, therefore, not only contributes to the current knowledge base but also opens avenues for more detailed and mechanistic explorations into how these essential micronutrients interact and influence each other.