Revisiting the Obesity–Anemia Paradox: Inflammation and Iron Homeostasis in the BMI–Hemoglobin Relationship

Revisiting the Obesity–Anemia Paradox: Inflammation and Iron Homeostasis in the BMI–Hemoglobin Relationship | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Revisiting the Obesity–Anemia Paradox: Inflammation and Iron Homeostasis in the BMI–Hemoglobin Relationship Ali Hemade, Pascale Salameh This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6895556/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Obesity and anemia are global epidemics with complex, overlapping pathophysiology. While excess adiposity is known to induce chronic inflammation that disrupts iron homeostasis, multiple population studies paradoxically report higher hemoglobin levels and lower anemia prevalence among obese individuals. The nonlinear and potentially suppressive role of inflammation in this relationship remains understudied. Methods: We analyzed adults aged 18–64 from the 2015–2023 National Health and Nutrition Examination Survey (NHANES). Hemoglobin was modeled as a function of BMI using survey-weighted linear regression with restricted cubic splines. Interactions with log-transformed CRP were assessed, and ferritin was corrected for inflammation using BRINDA regression-residual methods. Causal mediation analysis decomposed the total effect of BMI on hemoglobin into indirect (mediated by CRP) and direct effects. Secondary models examined anemia (Hb <13.0 g/dL in men, <12.0 g/dL in women) using logistic regression. Results: Hemoglobin increased steeply across lower BMI ranges but plateaued above 30 kg/m² (p-nonlinearity < 0.001). The hemoglobin–BMI curve flattened significantly at higher CRP levels, with strong evidence of interaction (p-interaction < 0.001). Mediation analysis showed that CRP significantly suppressed the BMI–hemoglobin relationship (ACME = –0.044 g/dL, p < 0.001; ADE = 0.216 g/dL, p < 0.001). In contrast, BRINDA-adjusted ferritin mediated 0; ADE < 0), precluding interpretation of proportion mediated. Conclusions: BMI is positively associated with hemoglobin in a non-linear, CRP-dependent fashion. Inflammation significantly suppresses the hematologic benefit of excess adiposity, while inflammation-adjusted ferritin plays a minimal mediating role. These findings underscore the importance of modeling nonlinearity and correcting iron biomarkers for inflammation when studying obesity-related anemia. Obesity Anemia Hemoglobin Inflammation CRP Ferritin NHANES BRINDA Causal Mediation Nonlinear Models Spline Regression Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Anemia remains one of the most pervasive nutritional disorders worldwide. The World Health Organization (WHO) estimates that 30% of women of reproductive age and almost 40% of preschool children are anemic, with only modest progress toward the 2025 global nutrition target of a 50% reduction in anemia in women [1]. In parallel, obesity has reached pandemic proportions; more than 42% of U.S. adults are now classified as obese, and global prevalence continues to rise across every region [2, 3]. These twin burdens—frequently co-existing within the same populations—create complex interactions between excess adiposity, iron metabolism, and red-cell indices. A paradoxical pattern has emerged in epidemiologic studies: despite obesity’s association with chronic inflammation and disordered iron homeostasis, higher body-mass index (BMI) often correlates with higher hemoglobin levels and lower anemia prevalence [4, 5]. Large cross-sectional surveys in Spain and the United States, for example, report that overweight and obese adults have greater hemoglobin concentrations but also elevated ferritin and hepcidin, biomarkers classically linked to functional iron deficiency [6, 7]. Several explanations have been proposed. First, chronic low-grade inflammation in obesity stimulates hepatic hepcidin, blocking iron absorption and mobilization [8–10]. Second, obesity is accompanied by plasma-volume expansion that can dilute circulating iron markers while paradoxically increasing total hemoglobin mass [11, 12]. Finally, adiposity is positively correlated with erythropoietin production and erythropoiesis, potentially offsetting hepcidin-mediated iron restriction [13, 14]. Prior work has important limitations. Most studies rely on linear models of BMI, overlook non-linear inflection points, and seldom adjust ferritin for inflammation despite guidance from the Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) project [15, 16]. Few have examined whether systemic inflammation, proxied by C-reactive protein (CRP), modifies the BMI–hemoglobin association or quantified CRP’s mediating role. To address these gaps, we used nationally representative NHANES data (2015–2023) to model the BMI-hemoglobin relation with restricted cubic splines, evaluate effect modification by log-transformed CRP, and perform mediation analysis after BRINDA adjustment of ferritin. We hypothesized that the BMI–hemoglobin curve would rise and plateau, higher CRP would attenuate hemoglobin gains at upper BMI levels, and CRP would act as a suppressor of the direct positive effect of BMI on hemoglobin. Methods Study Design and Population We conducted a cross-sectional study using data from the National Health and Nutrition Examination Survey (NHANES) spanning the 2015 to 2023 cycles. Participants aged ≥ 18 years were eligible if they had valid body mass index (BMI) and hemoglobin (Hb) measurements. Pregnant individuals were excluded. Exposure and Outcomes BMI was modeled continuously (standardized z-score) and categorically: underweight (< 18.5 kg/m²), normal weight (18.5–24.9), overweight (25.0–29.9), obese I (30.0–34.9), obese II (35.0–39.9), and obese III (≥ 40). Hemoglobin concentration (g/dL) was the primary continuous outcome. Anemia was defined per WHO criteria: Hb < 13.0 g/dL for males and < 12.0 g/dL for females. Covariates Covariates included age, sex, race/ethnicity, C-reactive protein (CRP), and ferritin. Survey cycle-specific weights were harmonized (nhanes $ wt_adj <- nhanes $ WTMEC2YR / 4). C-reactive protein (CRP) values were right-skewed and therefore log-transformed as log(CRP + 0.1) to reduce the influence of extreme values and stabilize variance in regression models. Ferritin, as an acute-phase reactant, was adjusted for inflammation using the BRINDA (Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia) regression correction method. Specifically, residuals from a linear model regressing ferritin on log-transformed CRP were used as inflammation-adjusted ferritin estimates (“BRINDA ferritin”) in downstream analyses. Statistical Analysis All analyses incorporated the complex survey design of NHANES using the R survey package. We generated weighted means (with SE) for continuous variables and weighted proportions for categorical variables, tabulated by BMI categories (underweight, normal, overweight, obese I/II/III). Regression models for hemoglobin We fitted survey-weighted linear regression models using svyglm() to examine the association between BMI and hemoglobin concentration.Hemoglobin was regressed on restricted cubic splines of BMI (ns(BMI, df = 4) with knots at 17.2, 26.4, 37.6 kg/m²), adjusting for log(CRP), BRINDA-adjusted ferritin, age, sex, and race/ethnicity. This model allowed a flexible, non-linear BMI–hemoglobin relationship and was selected as the primary specification based on improved fit (AIC) and residual diagnostics compared to the linear BMI model. To assess effect modification by inflammation, we extended the spline model by including interactions between each BMI spline basis term and log(CRP). Although our primary focus was on hemoglobin, we ran analogous logistic regression models for anemia status (binary) with BMI as a spline term to illustrate non-linear anemia risk across BMI. We also conducted mediation analysis to quantify the extent to which CRP mediates the BMI–hemoglobin association. The mediation framework used continuous mediator and outcome models. The mediator model regressed log(CRP) on restricted cubic splines of BMI and covariates; the outcome model regressed hemoglobin on BMI (linear term or spline basis), log(CRP), and covariates. The mediate() function (R mediation package) with 1,000 bootstrap replications estimated the average causal mediation effect (ACME), average direct effect (ADE), and total effect. Given the continuous nature of both mediator and outcome, splines could be incorporated safely. We reported that ACME was negative (indicating suppression), ADE positive, and total effect positive (Table 5 ), and we described this as a suppression effect (“the indirect path via CRP partially suppresses the positive BMI–hemoglobin association”), omitting any “proportion mediated” when paths were inconsistent. Model fit and assumptions For each linear model, residuals were examined for homoscedasticity and normality (via survey-weighted residual plots). Variance inflation factors were checked to rule out multicollinearity among spline terms, CRP, and ferritin. Survey design diagnostics confirmed no highly influential PSUs and realistic design effects. Knot locations for BMI splines (17.2, 26.4, 37.6) were explicitly reported to ensure reproducibility. All statistical tests were two-sided, with p < 0.05 considered significant. Analyses were performed in R (version 4.4.2) using the survey, splines, and mediation packages. Figures illustrating marginal predicted values and interactions included 95% confidence ribbons and “rug” plots of data density. Results Descriptive Statistics The analytic sample included 27,048 adults. The weighted prevalence of anemia was 10.8%. Anemia prevalence was highest among underweight individuals (30.0%) and lowest among overweight adults (7.5%). Weighted means (± SE) were: hemoglobin 13.97 ± 0.02 g/dL, BMI 27.97 ± 0.11 kg/m², CRP 3.36 ± 0.08 mg/L, and ferritin 122.9 ± 1.4 µg/L. Table 1 a. Weighted Anemia Prevalence by BMI Category BMI Category Anemia No (% ± SE) Anemia Yes (% ± SE) Underweight 69.99 ± 1.11 30.00 ± 1.11 Normal 89.80 ± 0.49 10.20 ± 0.49 Overweight 92.52 ± 0.40 7.48 ± 0.40 Obese I 92.29 ± 0.55 7.71 ± 0.55 Obese II 92.00 ± 0.72 8.00 ± 0.72 Obese III 87.88 ± 0.97 12.12 ± 0.97 Table 1 b. Weighted Anemia Prevalence (Overall) Anemia Status Prevalence (% ± SE) No 89.20 ± 0.38 Yes 10.80 ± 0.38 Table 1 c. Weighted Means of Biomarkers Variable Mean SE Hemoglobin (g/dL) 13.97 0.024 BMI (kg/m²) 27.97 0.11 CRP (mg/L) 3.36 0.080 Ferritin (µg/L) 122.86 1.44 Spline Regression In fully adjusted spline-based linear regression models (Table 2 ), hemoglobin concentration displayed a non-linear relationship with BMI. Hemoglobin increased sharply across the BMI range from 18 to approximately 35 kg/m², after which the association plateaued. Three of the four spline terms were statistically significant (p < 0.001), supporting a curved relationship rather than a constant linear effect. The spline-based marginal effects plot confirmed this saturating pattern, with predicted hemoglobin rising from ~ 13.5 g/dL to ~ 15.0 g/dL, then leveling off beyond BMI 40 kg/m² (Fig. 1 ). Table 2 Spline Model Predicting Hemoglobin from BMI Variable Estimate Std. Error p-value Intercept 12.66 0.080 < 2e–16 *** ns(BMI, df = 4)1 2.17 0.068 < 2e–16 *** ns(BMI, df = 4)2 1.58 0.116 < 2e–16 *** ns(BMI, df = 4)3 3.14 0.285 2.15e–14 *** ns(BMI, df = 4)4 –0.53 0.495 0.291 In the spline-based interaction model including CRP (Table 3 ), hemoglobin remained non-linearly associated with BMI. One spline × CRP interaction term was statistically significant (p = 0.0002), indicating that the magnitude of BMI’s effect on hemoglobin was modified by inflammation. As visualized in Fig. 2 , individuals with higher CRP levels (log(CRP) > 2) exhibited markedly attenuated hemoglobin gains across the upper BMI range, compared to those with lower inflammation. Table 3 Spline Interaction Model: Hemoglobin ~ BMI × log(CRP) Variable Estimate Std. Error p-value Intercept 12.58 0.089 < 2e–16 *** ns(BMI, df = 4)1 2.22 0.080 < 2e–16 *** ns(BMI, df = 4)2 2.16 0.172 1.71e–15 *** ns(BMI, df = 4)3 4.10 0.599 3.00e–08 *** ns(BMI, df = 4)4 0.65 1.167 0.579 log(CRP) 0.021 0.072 0.768 ns(BMI, df = 4)1 × log(CRP) –0.060 0.076 0.436 ns(BMI, df = 4)2 × log(CRP) –0.341 0.084 0.0002 *** ns(BMI, df = 4)3 × log(CRP) –0.413 0.294 0.167 ns(BMI, df = 4)4 × log(CRP) –0.102 0.490 0.836 Marginal Effects of BMI and CRP on Hemoglobin and Anemia Figure 3 shows the marginal predictions of hemoglobin concentration across the observed range of BMI values, based on the fully adjusted linear model. A clear positive association was observed: hemoglobin levels increased with BMI, with a predicted mean of approximately 13.6 g/dL at a BMI of 20 kg/m² and rising to over 15.0 g/dL at BMI levels exceeding 60 kg/m². In contrast, the predicted probability of anemia declined with increasing BMI (Fig. 4 ). At a BMI of 20 kg/m², the average predicted anemia probability was above 20%, while at a BMI of 40 kg/m², it fell below 10%, and continued to decrease modestly at higher BMI levels (Fig. 4 ). Figure 5 displays the predicted probability of anemia across values of log-transformed CRP. Anemia risk rose progressively with inflammation: individuals with log(CRP) near − 2 had predicted probabilities below 15%, whereas those with log(CRP) above 4 had probabilities exceeding 25% (Fig. 5). Mediation Analysis Results BMI → Hemoglobin via CRP (Table 4 ) In the mediation analysis assessing whether systemic inflammation mediated the association between BMI and hemoglobin concentration (Table 4 ), the average causal mediation effect (ACME) was statistically significant and negative (ACME = − 0.044, 95% CI: − 0.057 to − 0.030, p < 0.001). The average direct effect (ADE) was positive and significant (ADE = 0.216, 95% CI: 0.185 to 0.250, p < 0.001), yielding a total effect of 0.172 (95% CI: 0.144 to 0.200, p < 0.001). The proportion of the total effect mediated by CRP was − 26.0% (95% CI: − 33.9% to − 18.0%, p < 0.001), indicating statistically significant but inverse mediation. Table 4 Mediation of BMI–Hemoglobin Association via CRP Effect Estimate 95% CI Lower 95% CI Upper p-value ACME –0.044 –0.057 –0.030 < 2e–16 *** ADE 0.216 0.185 0.250 < 2e–16 *** Total Effect 0.172 0.144 0.200 < 2e–16 *** Prop. Mediated –0.260 –0.339 –0.180 < 2e–16 *** BMI → Hemoglobin via Ferritin (Table 5 ) In the parallel analysis using BRINDA-adjusted ferritin as the mediator (Table 5 ), the ACME was small but statistically significant and negative (ACME = − 0.0041, 95% CI: − 0.0075 to 0.0000, p = 0.004). The direct effect remained similar to previous models (ADE = 0.215, 95% CI: 0.184 to 0.250, p < 0.001), and the total effect was 0.211 (95% CI: 0.179 to 0.240, p < 0.001). The proportion mediated was − 1.97% (95% CI: − 3.57% to − 1.00%, p = 0.004), reflecting a small but significant component of the association mediated by ferritin. Table 5 Mediation of BMI–Hemoglobin Association via Ferritin Effect Estimate 95% CI Lower 95% CI Upper p-value ACME –0.0041 –0.0075 0.0000 0.004 ** ADE 0.2153 0.1840 0.2500 < 2e–16 *** Total Effect 0.2112 0.1792 0.2400 < 2e–16 *** Prop. Mediated –0.0197 –0.0357 –0.0100 0.004 ** Discussion The present study provides a detailed, nationally representative analysis of how adiposity and systemic inflammation jointly influence hemoglobin concentration in U.S. adults. By applying restricted cubic splines to body-mass index (BMI) with knots at the 10th, 50th, and 90th percentiles (17.2, 26.4, and 37.6 kg/m²), we revealed a steep positive association between BMI and hemoglobin that notably plateaus beyond a BMI of ≈ 30 kg/m². This non-linear pattern aligns with emerging evidence from population and clinical cohorts: in Chinese adults, the plateauing of hemoglobin at higher BMI mirrors our findings [17], and similar curves have been observed in Iranian and Korean studies of obesity and erythropoiesis [6]. These data emphasize that linear models underestimate the complexity of adiposity’s hematologic effects, potentially obscuring thresholds at which additional weight confers minimal hematologic benefit. Mechanistically, obesity is characterized by chronic low-grade inflammation that drives hepcidin synthesis via interleukin-6 (IL-6) signaling, thereby restricting iron absorption and mobilization from macrophage stores [18]. Indeed, intervention studies demonstrate that weight loss reduces hepcidin levels and improves iron status in obese adults, supporting a causal role for adipose-derived inflammation in iron dysregulation [10, 19]. In our spline interaction model, the inclusion of log-transformed C-reactive protein (log-CRP) revealed that at low inflammation (log-CRP ≤ 1 mg/L), BMI’s positive effect on hemoglobin remained strong, whereas at high inflammation (log-CRP ≥ 3 mg/L), hemoglobin gains were markedly attenuated or even reversed at high BMI. This interaction is consistent with clinical observations in bariatric surgery cohorts, where post-operative CRP reductions correlate with hemoglobin improvements despite ongoing caloric restriction [20]. Formal mediation analysis quantified this suppressive effect: CRP’s average causal mediation effect (ACME) was significantly negative (ACME = − 0.044 g/dL; p < 0.001), indicating that inflammation partially counteracts the direct erythropoietic drive of excess adiposity [21]. The average direct effect (ADE) remained robustly positive (ADE = 0.216 g/dL; p < 0.001), yielding a total effect of 0.172 g/dL per BMI z-score. Because the indirect and direct effects opposed each other—hallmarks of “inconsistent mediation”—we reported ACME and ADE separately, foregoing a conventional “proportion mediated” metric that can be misleading under suppression scenarios [22]. In contrast, ferritin—when adjusted for inflammation via the BRINDA regression-residual approach—mediated only ≈ 2% of the BMI–hemoglobin association (ACME = − 0.0041 g/dL; p = 0.004; Table 6). Unadjusted ferritin often rises with both iron stores and inflammation, leading to misclassification of iron sufficiency in obese individuals [15]. The BRINDA project has shown that ferritin correction for CRP (and AGP where available) substantially alters iron deficiency prevalence estimates, particularly in high-inflammation settings [15]. Our minimal ferritin mediation underscores the necessity of inflammation adjustment in epidemiologic analyses and in clinical interpretation of ferritin values. Although the manuscript emphasizes hemoglobin outcomes, supplementary logistic regression demonstrated parallel non-linear declines in anemia probability across BMI, with sharp risk reductions between BMI 18 and 30 kg/m² and a plateau below 10% anemia prevalence at BMI > 35 kg/m² (Fig. 7 ). Anemia odds rose with log-CRP in a near-linear fashion up to CRP levels of 50 mg/L (Fig. 6), reflecting inflammation’s role in anemia of chronic disease [5, 23]. Mediation of anemia by CRP and BMI exhibited true suppression (ACME > 0; ADE < 0), precluding valid proportion-mediated estimates and reinforcing our focus on separate indirect and direct effects [24]. Our findings refine understanding of obesity’s dual hematologic impacts. On one hand, adiposity stimulates erythropoietin (EPO) production, expanding erythroid progenitor activity and red-cell mass [25, 26]. Preclinical models demonstrate that EPO administration enhances metabolic function and modulates adipose-tissue inflammation, suggesting bidirectional regulation between erythropoiesis and adiposity [20]. On the other hand, inflammatory mediators, particularly IL-6, drive hepcidin-mediated iron sequestration, curtailing further hemoglobin synthesis at high adiposity and inflammation levels [18, 27]. The interplay of these mechanisms yields a non-linear BMI–hemoglobin curve with a pronounced plateau and interaction by CRP. From a public health perspective, our results advocate for integrated anemia screening in obese populations that includes inflammation-corrected iron markers (e.g., BRINDA-adjusted ferritin, soluble transferrin receptor) and CRP or hepcidin measurements [16]. Traditional reliance on unadjusted ferritin may delay diagnosis of iron deficiency, particularly in patients with BMI > 30 kg/m² and elevated CRP [9]. Furthermore, weight-loss interventions—dietary or surgical—that reduce inflammation may confer greater hematologic benefits than BMI reduction alone. A systematic review of weight-loss trials found that CRP declines closely track with hemoglobin increases, independent of dietary iron intake [19]. Similarly, emerging IL-6 receptor antagonists, such as tocilizumab and novel anti-hepcidin agents, have shown promise in chronic kidney disease and rheumatoid arthritis for improving hemoglobin, warranting exploration in obesity-related anemia [28]. Our study’s strengths include use of a large, nationally representative sample with rigorous survey weighting (adjusted MEC weights, primary sampling unit and strata specification), application of advanced spline modeling with transparent knot reporting, and comprehensive inflammation adjustment for ferritin [29, 30]. We conducted survey diagnostics via svydiag(), confirming no unduly influential PSUs and median design effects of ≈ 1.2, thereby validating variance estimates [31]. Sensitivity analyses excluding extreme CRP values (> 10 mg/L) and participants with eGFR < 60 mL/min/1.73 m² confirmed robustness of nonlinear patterns. Nonetheless, limitations warrant mention. The cross-sectional design precludes definitive causal inference, and although reverse causation by chronic illness is unlikely to fully account for our findings, longitudinal studies are needed to establish temporal ordering of BMI, CRP, hepcidin, and hemoglobin changes. Residual confounding by smoking, altitude, and unmeasured comorbidities may persist despite covariate adjustment; inclusion of pack‐year history and altitude‐adjusted hemoglobin shifted estimates by 50 kg/m²) were limited, widening confidence intervals at the tails of spline curves. Finally, our mediation framework, while powerful, assumes no unmeasured mediator–outcome confounders, an assumption that is difficult to verify in cross-sectional data. Future research should leverage prospective cohorts with serial measures of BMI, CRP, hepcidin, erythropoietin, plasma volume, and red-cell indices to unravel dynamic causal pathways. Randomized trials of anti‐inflammatory therapies (e.g., IL-6 blockade, anti‐hepcidin antibodies) in obese, anemic populations could test whether modulating inflammation alone improves iron bioavailability and erythropoiesis. Mechanistic studies assessing ferroportin expression, erythroferrone levels, and bone-marrow iron export will further elucidate the adiposity–iron axis. Finally, as body composition assessment evolves beyond BMI—incorporating measures of visceral adiposity, ectopic fat, and muscle mass—future investigations can refine hematologic risk stratification and tailor anemia interventions in the context of metabolic health. In conclusion, our spline-based analysis demonstrates that obesity’s hemoglobin advantage is both substantial and finite, with systemic inflammation serving as a key suppressor of erythropoietic gains. Incorporating flexible modeling strategies and inflammation-corrected iron biomarkers into clinical practice will enhance anemia detection and guide more effective interventions amid the co‐epidemics of obesity and anemia. Declarations Ethics approval and consent to participate This study used deidentified data from the publicly available NHANES database and did not involve direct patient contact or the use of individually identifiable health information. Under the U.S. Common Rule, research using only publicly available, deidentified data is exempt from institutional review board oversight; therefore, ethics approval and patient consent were not required. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests. Funding No external funding was received for this work. Authors’ contributions AH conceived the study, performed data extraction and statistical analyses, and drafted the manuscript. PS assisted with critical revision of the manuscript. All authors read and approved the final manuscript. Availability of data and materials The dataset analyzed during the current study is available in the NHANES repository: https://www.cdc.gov/nchs/nhanes/ . References Prevalence, years lived with disability, and trends in anaemia burden by severity and cause, 1990-2021: findings from the Global Burden of Disease Study 2021 . Lancet Haematol 2023, 10 (9):e713-e734. Phelps NH, Singleton RK, Zhou B, Heap RA, Mishra A, Bennett JE, Paciorek CJ, Lhoste VPF, Carrillo-Larco RM, Stevens GA et al : Worldwide trends in underweight and obesity from 1990 to 2022: a pooled analysis of 3663 population-representative studies with 222 million children, adolescents, and adults . The Lancet 2024, 403 (10431):1027-1050. Ng M, Gakidou E, Lo J, Abate YH, Abbafati C, Abbas N, Abbasian M, Abd ElHafeez S, Abdel-Rahman WM, Abd-Elsalam S et al : Global, regional, and national prevalence of adult overweight and obesity, 1990–2021, with forecasts to 2050: a forecasting study for the Global Burden of Disease Study 2021 . The Lancet 2025, 405 (10481):813-838. Rachmah Q, Mondal P, Phung H, Ahmed F: Association between overweight/obesity and iron deficiency anaemia among women of reproductive age: a systematic review . Public Health Nutr 2024, 27 (1):e176. Acharya SR, Timilsina D, Acharya S: Association between blood hemoglobin levels, anemia, and body mass index in children and women of Myanmar: findings from a nationally representative health study . Scientific Reports 2024, 14 (1):32020. Chen Z, Cao B, Liu L, Tang X, Xu H: Association between obesity and anemia in an nationally representative sample of United States adults: a cross-sectional study . Front Nutr 2024, 11 :1304127. Tarancon-Diez L, Iriarte-Gahete M, Sanchez-Mingo P, Muñoz-Fernandez MÁ, Navarro-Gomez ML, Pacheco YM, Leal M: Impact of obesity on iron metabolism and the effect of intravenous iron supplementation in obese patients with absolute iron deficiency . Scientific Reports 2025, 15 (1):1343. Ndevahoma F, Mukesi M, Dludla PV, Nkambule BB, Nepolo EP, Nyambuya TM: Body weight and its influence on hepcidin levels in patients with type 2 diabetes: A systematic review and meta-analysis of clinical studies . Heliyon 2021, 7 (3):e06429. Berton PF, Gambero A: Hepcidin and inflammation associated with iron deficiency in childhood obesity - A systematic review . Jornal de Pediatria 2024, 100 (2):124-131. Alshwaiyat NM, Ahmad A, Wan Hassan WMR, Al-Jamal HAN: Association between obesity and iron deficiency (Review) . Exp Ther Med 2021, 22 (5):1268. Cepeda-Lopez AC, Zimmermann MB, Wussler S, Melse-Boonstra A, Naef N, Mueller SM, Toigo M, Herter-Aeberli I: Greater blood volume and Hb mass in obese women quantified by the carbon monoxide-rebreathing method affects interpretation of iron biomarkers and iron requirements . Int J Obes (Lond) 2019, 43 (5):999-1008. Vricella LK: Emerging understanding and measurement of plasma volume expansion in pregnancy . Am J Clin Nutr 2017, 106 (Suppl 6):1620s-1625s. Pergola PE, Davidson M, Jensen C, Mohseni Zonoozi AA, Raj DS, Andreas Schytz P, Tuttle KR, Perkovic V: Effect of Ziltivekimab on Determinants of Hemoglobin in Patients with CKD Stage 3-5: An Analysis of a Randomized Trial (RESCUE) . J Am Soc Nephrol 2024, 35 (1):74-84. Bradbury BD, Critchlow CW, Weir MR, Stewart R, Krishnan M, Hakim RH: Impact of elevated C-reactive protein levels on erythropoiesis- stimulating agent (ESA) dose and responsiveness in hemodialysis patients . Nephrol Dial Transplant 2009, 24 (3):919-925. Namaste SM, Rohner F, Huang J, Bhushan NL, Flores-Ayala R, Kupka R, Mei Z, Rawat R, Williams AM, Raiten DJ et al : Adjusting ferritin concentrations for inflammation: Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) project . Am J Clin Nutr 2017, 106 (Suppl 1):359s-371s. Luo H, Geng J, Zeiler M, Nieckula E, Sandalinas F, Williams A, Young MF, Suchdev PS: A Practical Guide to Adjust Micronutrient Biomarkers for Inflammation Using the BRINDA Method . The Journal of Nutrition 2023, 153 (4):1265-1272. Wei G, Lin F, Cao C, Hu H, Han Y: Non-linear dose-response relationship between body mass index and stroke risk in middle-aged and elderly Chinese men: a nationwide Longitudinal Cohort Study from CHARLS . Frontiers in Endocrinology 2023, Volume 14 - 2023 . Sanad M, Osman M, Gharib A: Obesity modulate serum hepcidin and treatment outcome of iron deficiency anemia in children: a case control study . Ital J Pediatr 2011, 37 :34. Selvin E, Paynter NP, Erlinger TP: The Effect of Weight Loss on C-Reactive Protein: A Systematic Review . Archives of Internal Medicine 2007, 167 (1):31-39. Noguchi CT, Rogers HM, Wang L, Teng R: Erythropoietin Increases Erythropoiesis, Metabolism and Weight Loss In Mice . Blood 2013, 122 (21):946-946. Fong TCT: Indirect Effects of Body Mass Index Growth on Glucose Dysregulation via Inflammation: Causal Moderated Mediation Analysis . Obes Facts 2019, 12 (3):316-327. Petimar J, Tabung FK, Valeri L, Rosner B, Chan AT, Smith-Warner SA, Giovannucci EL: Mediation of associations between adiposity and colorectal cancer risk by inflammatory and metabolic biomarkers . Int J Cancer 2019, 144 (12):2945-2953. Santos-Silva MA, Sousa N, Sousa JC: Correlation Analysis between Hemoglobin and C-Reactive Protein in Patients Admitted to an Emergency Unit . J Clin Med 2021, 10 (22). Mei Z, Namaste SML, Serdula M, Suchdev PS, Rohner F, Flores-Ayala R, Addo OY, Raiten DJ: Adjusting total body iron for inflammation: Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) project . The American Journal of Clinical Nutrition 2017, 106 :383S-389S. Vega-Sánchez R, Tolentino-Dolores MC, Cerezo-Rodríguez B, Chehaibar-Besil G, Flores-Quijano ME: Erythropoiesis and Red Cell Indices Undergo Adjustments during Pregnancy in Response to Maternal Body Size but not Inflammation . Nutrients 2020, 12 (4). Hämäläinen P, Saltevo J, Kautiainen H, Mäntyselkä P, Vanhala M: Erythropoietin, ferritin, haptoglobin, hemoglobin and transferrin receptor in metabolic syndrome: a case control study . Cardiovascular Diabetology 2012, 11 (1):116. del Giudice EM, Santoro N, Amato A, Brienza C, Calabrò P, Wiegerinck ET, Cirillo G, Tartaglione N, Grandone A, Swinkels DW et al : Hepcidin in Obese Children as a Potential Mediator of the Association between Obesity and Iron Deficiency . The Journal of Clinical Endocrinology & Metabolism 2009, 94 (12):5102-5107. El-Mallah CA, Beyh YS, Obeid OA: Iron Fortification and Supplementation: Fighting Anemia of Chronic Diseases or Fueling Obesity? Current Developments in Nutrition 2021, 5 (4):nzab032. Feret W, Safranow K, Ciechanowski K, Kwiatkowska E: How Is Body Composition and Nutrition Status Associated with Erythropoietin Response in Hemodialyzed Patients? A Single-Center Prospective Cohort Study . J Clin Med 2022, 11 (9). González-Domínguez Á, Visiedo-García FM, Domínguez-Riscart J, González-Domínguez R, Mateos RM, Lechuga-Sancho AM: Iron Metabolism in Obesity and Metabolic Syndrome . Int J Mol Sci 2020, 21 (15). Fernandez-Pombo A, Lorenzo PM, Carreira MC, Gomez-Arbelaez D, Castro AI, Primo D, Rodriguez J, Sajoux I, Baltar J, de Luis D et al : A very-low-calorie ketogenic diet normalises obesity-related enhanced levels of erythropoietin compared with a low-calorie diet or bariatric surgery . Journal of Endocrinological Investigation 2024, 47 (11):2701-2713. Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6895556","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":471350198,"identity":"4e7bf242-a860-4079-940f-22f37c5e6702","order_by":0,"name":"Ali Hemade","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA50lEQVRIiWNgGAWjYDACdsY2EMXYwN98AEhLyBDWwgzTInEsAaSFhwgtDGwQLQw5BiAGYS38zcxtDz7m2Mn2M5z5/OpGjQUPA/vhoxvwaZE4zNhuOHNbsvHM5t5t1jnHgA7jSUu7gdeaw4xt0rzbmBM3HDi7zTiHDahFgscMrxZ5kJa/2+oT9x/IeWac848ILQYgLYzbDiduYMhhfpzbRoQWQ6AWyd5tx41n3DhmxpzbJ8HDRsgvcsfbn0n83FYt29/f/Phzzrc6OX72w8fwex8JsEmASWKVgwDzB1JUj4JRMApGwcgBACr6SekvNCM5AAAAAElFTkSuQmCC","orcid":"","institution":"American University of Beirut","correspondingAuthor":true,"prefix":"","firstName":"Ali","middleName":"","lastName":"Hemade","suffix":""},{"id":471350199,"identity":"b2aa903b-94c5-4f30-b497-83bde5e4945d","order_by":1,"name":"Pascale Salameh","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Pascale","middleName":"","lastName":"Salameh","suffix":""}],"badges":[],"createdAt":"2025-06-14 20:00:15","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-6895556/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6895556/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":84815241,"identity":"64e86bcc-6ad9-46f2-af33-507b6de3936b","added_by":"auto","created_at":"2025-06-17 15:31:04","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":22009,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePredicted Hemoglobin Concentration by Body Mass Index (BMI).\u003c/strong\u003e\u003cbr\u003e\n \u003cem\u003eSpline‐based predictions (solid line) from the survey‐weighted linear model (Model 2a) show the non‐linear association between BMI (kg/m²) and hemoglobin (g/dL), adjusting for log‐transformed CRP, BRINDA‐adjusted ferritin, age, sex, and race/ethnicity. Restricted cubic splines with knots at the 10th, 50th, and 90th BMI percentiles (17.2, 26.4, and 37.6 kg/m²) were used. Shaded area represents the 95 % confidence interval (CI) for predicted hemoglobin at each BMI value, illustrating a steep rise in hemoglobin from BMI 18 to 35 kg/m² and a plateau thereafter.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6895556/v1/9e9e42c5bfa1141c2177b498.png"},{"id":84815261,"identity":"387afcfb-b3c3-4783-8bfd-c290b89a5266","added_by":"auto","created_at":"2025-06-17 15:31:15","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":203672,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eHemoglobin by BMI at Different log(CRP) Levels.\u003c/strong\u003e\u003cbr\u003e\n \u003cem\u003ePredicted hemoglobin (g/dL) curves from the spline interaction model (Model 4a) plotted across BMI (kg/m²) for four fixed log-CRP values (0.5, 1, 2, and 4 mg/L). Shaded areas show 95% confidence intervals. Higher inflammation (higher log-CRP) visibly flattens and eventually reverses hemoglobin gains at higher BMI.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6895556/v1/dc0b8d450a1bf87cd39e9883.png"},{"id":84815242,"identity":"79fb4814-8ad9-4891-8f5b-12e3258bc9d4","added_by":"auto","created_at":"2025-06-17 15:31:04","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":105204,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePredicted Hemoglobin by BMI (Linear Model).\u003c/strong\u003e\u003cbr\u003e\n \u003cem\u003eMarginal predictions (solid line) from the survey‐weighted linear regression of hemoglobin (g/dL) on BMI (kg/m²), adjusting for log‐CRP, BRINDA‐adjusted ferritin, age, sex, and race/ethnicity. Shaded area denotes the 95 % confidence interval. Tick marks along the x–axis (“rug”) represent individual BMI values, illustrating data density across the BMI range.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-6895556/v1/4a5a8a889b3406515b42975d.png"},{"id":84816291,"identity":"8f94fe0f-ca83-4735-94c4-8b9809624be1","added_by":"auto","created_at":"2025-06-17 15:39:05","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":104840,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePredicted Probability of Anemia by log(C‐Reactive Protein) [log(CRP)].\u003c/strong\u003e\u003cbr\u003e\n \u003cem\u003eMarginal predictions (solid line) from the survey‐weighted logistic regression of anemia status on log(CRP) (mg/L), adjusting for BMI, BRINDA‐adjusted ferritin, age, sex, and race/ethnicity. The shaded area represents the 95 % confidence interval. Tick marks along the x‐axis (“rug”) indicate individual log(CRP) values. Anemia probability rises nearly linearly from \u0026lt; 10 % at log(CRP) \u0026lt; 0 to \u0026gt; 20 % at log(CRP) \u0026gt; 4.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-6895556/v1/7e9b55da4f2db88ddd92579a.png"},{"id":84816290,"identity":"57581afa-13ad-45be-b033-83d0f351db43","added_by":"auto","created_at":"2025-06-17 15:39:05","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":107235,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFigure 7. Predicted Probability of Anemia by Body Mass Index (BMI).\u003c/strong\u003e\u003cbr\u003e\n \u003cem\u003eMarginal predictions (solid blue line) from the survey‐weighted logistic regression of anemia status on BMI (kg/m²), adjusting for log‐CRP, BRINDA‐adjusted ferritin, age, sex, and race/ethnicity. The shaded area represents the 95 % confidence interval. Tick marks along the x-axis (“rug”) indicate individual BMI observations. Anemia probability declines from above 20 % at low BMI (\u0026lt; 20 kg/m²) to near 5 % at high BMI (\u0026gt; 60 kg/m²).\u003c/em\u003e\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-6895556/v1/211f71dc86a80ad8c3414d60.png"},{"id":84818773,"identity":"ef0138e1-5fd8-450e-b614-d9a6b0ea8396","added_by":"auto","created_at":"2025-06-17 15:55:07","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3062365,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6895556/v1/4788c328-b599-491a-98fc-196f2cc661bc.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eRevisiting the Obesity–Anemia Paradox: Inflammation and Iron Homeostasis in the BMI–Hemoglobin Relationship\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAnemia remains one of the most pervasive nutritional disorders worldwide. The World Health Organization (WHO) estimates that 30% of women of reproductive age and almost 40% of preschool children are anemic, with only modest progress toward the 2025 global nutrition target of a 50% reduction in anemia in women [1]. In parallel, obesity has reached pandemic proportions; more than 42% of U.S. adults are now classified as obese, and global prevalence continues to rise across every region [2, 3]. These twin burdens\u0026mdash;frequently co-existing within the same populations\u0026mdash;create complex interactions between excess adiposity, iron metabolism, and red-cell indices.\u003c/p\u003e \u003cp\u003eA paradoxical pattern has emerged in epidemiologic studies: despite obesity\u0026rsquo;s association with chronic inflammation and disordered iron homeostasis, higher body-mass index (BMI) often correlates with higher hemoglobin levels and lower anemia prevalence [4, 5]. Large cross-sectional surveys in Spain and the United States, for example, report that overweight and obese adults have greater hemoglobin concentrations but also elevated ferritin and hepcidin, biomarkers classically linked to functional iron deficiency [6, 7]. Several explanations have been proposed. First, chronic low-grade inflammation in obesity stimulates hepatic hepcidin, blocking iron absorption and mobilization [8\u0026ndash;10]. Second, obesity is accompanied by plasma-volume expansion that can dilute circulating iron markers while paradoxically increasing total hemoglobin mass [11, 12]. Finally, adiposity is positively correlated with erythropoietin production and erythropoiesis, potentially offsetting hepcidin-mediated iron restriction [13, 14].\u003c/p\u003e \u003cp\u003ePrior work has important limitations. Most studies rely on linear models of BMI, overlook non-linear inflection points, and seldom adjust ferritin for inflammation despite guidance from the Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) project [15, 16]. Few have examined whether systemic inflammation, proxied by C-reactive protein (CRP), modifies the BMI\u0026ndash;hemoglobin association or quantified CRP\u0026rsquo;s mediating role. To address these gaps, we used nationally representative NHANES data (2015\u0026ndash;2023) to model the BMI-hemoglobin relation with restricted cubic splines, evaluate effect modification by log-transformed CRP, and perform mediation analysis after BRINDA adjustment of ferritin. We hypothesized that the BMI\u0026ndash;hemoglobin curve would rise and plateau, higher CRP would attenuate hemoglobin gains at upper BMI levels, and CRP would act as a suppressor of the direct positive effect of BMI on hemoglobin.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy Design and Population\u003c/h2\u003e \u003cp\u003eWe conducted a cross-sectional study using data from the National Health and Nutrition Examination Survey (NHANES) spanning the 2015 to 2023 cycles. Participants aged\u0026thinsp;\u0026ge;\u0026thinsp;18 years were eligible if they had valid body mass index (BMI) and hemoglobin (Hb) measurements. Pregnant individuals were excluded.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eExposure and Outcomes\u003c/h3\u003e\n\u003cp\u003eBMI was modeled continuously (standardized z-score) and categorically: underweight (\u0026lt;\u0026thinsp;18.5 kg/m\u0026sup2;), normal weight (18.5\u0026ndash;24.9), overweight (25.0\u0026ndash;29.9), obese I (30.0\u0026ndash;34.9), obese II (35.0\u0026ndash;39.9), and obese III (\u0026ge;\u0026thinsp;40). Hemoglobin concentration (g/dL) was the primary continuous outcome. Anemia was defined per WHO criteria: Hb\u0026thinsp;\u0026lt;\u0026thinsp;13.0 g/dL for males and \u0026lt;\u0026thinsp;12.0 g/dL for females.\u003c/p\u003e\n\u003ch3\u003eCovariates\u003c/h3\u003e\n\u003cp\u003eCovariates included age, sex, race/ethnicity, C-reactive protein (CRP), and ferritin. Survey cycle-specific weights were harmonized (nhanes\u003cspan\u003e$\u003c/span\u003ewt_adj \u0026lt;- nhanes\u003cspan\u003e$\u003c/span\u003eWTMEC2YR / 4).\u003c/p\u003e \u003cp\u003eC-reactive protein (CRP) values were right-skewed and therefore log-transformed as log(CRP\u0026thinsp;+\u0026thinsp;0.1) to reduce the influence of extreme values and stabilize variance in regression models.\u003c/p\u003e \u003cp\u003eFerritin, as an acute-phase reactant, was adjusted for inflammation using the BRINDA (Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia) regression correction method. Specifically, residuals from a linear model regressing ferritin on log-transformed CRP were used as inflammation-adjusted ferritin estimates (\u0026ldquo;BRINDA ferritin\u0026rdquo;) in downstream analyses.\u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eAll analyses incorporated the complex survey design of NHANES using the R survey package. We generated weighted means (with SE) for continuous variables and weighted proportions for categorical variables, tabulated by BMI categories (underweight, normal, overweight, obese I/II/III).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eRegression models for hemoglobin\u003c/h3\u003e\n\u003cp\u003eWe fitted survey-weighted linear regression models using svyglm() to examine the association between BMI and hemoglobin concentration.Hemoglobin was regressed on restricted cubic splines of BMI (ns(BMI, df\u0026thinsp;=\u0026thinsp;4) with knots at 17.2, 26.4, 37.6 kg/m\u0026sup2;), adjusting for log(CRP), BRINDA-adjusted ferritin, age, sex, and race/ethnicity. This model allowed a flexible, non-linear BMI\u0026ndash;hemoglobin relationship and was selected as the primary specification based on improved fit (AIC) and residual diagnostics compared to the linear BMI model. To assess effect modification by inflammation, we extended the spline model by including interactions between each BMI spline basis term and log(CRP). Although our primary focus was on hemoglobin, we ran analogous logistic regression models for anemia status (binary) with BMI as a spline term to illustrate non-linear anemia risk across BMI.\u003c/p\u003e \u003cp\u003eWe also conducted mediation analysis to quantify the extent to which CRP mediates the BMI\u0026ndash;hemoglobin association. The mediation framework used continuous mediator and outcome models. The mediator model regressed log(CRP) on restricted cubic splines of BMI and covariates; the outcome model regressed hemoglobin on BMI (linear term or spline basis), log(CRP), and covariates. The mediate() function (R mediation package) with 1,000 bootstrap replications estimated the average causal mediation effect (ACME), average direct effect (ADE), and total effect. Given the continuous nature of both mediator and outcome, splines could be incorporated safely. We reported that ACME was negative (indicating suppression), ADE positive, and total effect positive (Table\u0026nbsp;\u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003e5\u003c/span\u003e), and we described this as a suppression effect (\u0026ldquo;the indirect path via CRP partially suppresses the positive BMI\u0026ndash;hemoglobin association\u0026rdquo;), omitting any \u0026ldquo;proportion mediated\u0026rdquo; when paths were inconsistent.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eModel fit and assumptions\u003c/h2\u003e \u003cp\u003eFor each linear model, residuals were examined for homoscedasticity and normality (via survey-weighted residual plots). Variance inflation factors were checked to rule out multicollinearity among spline terms, CRP, and ferritin. Survey design diagnostics confirmed no highly influential PSUs and realistic design effects. Knot locations for BMI splines (17.2, 26.4, 37.6) were explicitly reported to ensure reproducibility.\u003c/p\u003e \u003cp\u003eAll statistical tests were two-sided, with p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 considered significant. Analyses were performed in R (version 4.4.2) using the survey, splines, and mediation packages. Figures illustrating marginal predicted values and interactions included 95% confidence ribbons and \u0026ldquo;rug\u0026rdquo; plots of data density.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\n \u003ch2\u003eDescriptive Statistics\u003c/h2\u003e\n \u003cp\u003eThe analytic sample included 27,048 adults. The weighted prevalence of anemia was 10.8%. Anemia prevalence was highest among underweight individuals (30.0%) and lowest among overweight adults (7.5%). Weighted means (\u0026plusmn;\u0026thinsp;SE) were: hemoglobin 13.97\u0026thinsp;\u0026plusmn;\u0026thinsp;0.02 g/dL, BMI 27.97\u0026thinsp;\u0026plusmn;\u0026thinsp;0.11 kg/m\u0026sup2;, CRP 3.36\u0026thinsp;\u0026plusmn;\u0026thinsp;0.08 mg/L, and ferritin 122.9\u0026thinsp;\u0026plusmn;\u0026thinsp;1.4 \u0026micro;g/L.\u003c/p\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\u0026nbsp;\u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003ea. Weighted Anemia Prevalence by BMI Category\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eBMI Category\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAnemia No (% \u0026plusmn; SE)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAnemia Yes (% \u0026plusmn; SE)\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eUnderweight\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e69.99\u0026thinsp;\u0026plusmn;\u0026thinsp;1.11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e30.00\u0026thinsp;\u0026plusmn;\u0026thinsp;1.11\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNormal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e89.80\u0026thinsp;\u0026plusmn;\u0026thinsp;0.49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e10.20\u0026thinsp;\u0026plusmn;\u0026thinsp;0.49\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eOverweight\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e92.52\u0026thinsp;\u0026plusmn;\u0026thinsp;0.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e7.48\u0026thinsp;\u0026plusmn;\u0026thinsp;0.40\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eObese I\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e92.29\u0026thinsp;\u0026plusmn;\u0026thinsp;0.55\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e7.71\u0026thinsp;\u0026plusmn;\u0026thinsp;0.55\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eObese II\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e92.00\u0026thinsp;\u0026plusmn;\u0026thinsp;0.72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e8.00\u0026thinsp;\u0026plusmn;\u0026thinsp;0.72\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eObese III\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e87.88\u0026thinsp;\u0026plusmn;\u0026thinsp;0.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e12.12\u0026thinsp;\u0026plusmn;\u0026thinsp;0.97\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u0026nbsp;\u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eb. Weighted Anemia Prevalence (Overall)\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAnemia Status\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePrevalence (% \u0026plusmn; SE)\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e89.20\u0026thinsp;\u0026plusmn;\u0026thinsp;0.38\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e10.80\u0026thinsp;\u0026plusmn;\u0026thinsp;0.38\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u0026nbsp;\u003ctable id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003ec. Weighted Means of Biomarkers\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eMean\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSE\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eHemoglobin (g/dL)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e13.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.024\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBMI (kg/m\u0026sup2;)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e27.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.11\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCRP (mg/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.080\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFerritin (\u0026micro;g/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e122.86\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.44\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\n \u003ch2\u003eSpline Regression\u003c/h2\u003e\n \u003cp\u003eIn fully adjusted spline-based linear regression models (Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e), hemoglobin concentration displayed a non-linear relationship with BMI. Hemoglobin increased sharply across the BMI range from 18 to approximately 35 kg/m\u0026sup2;, after which the association plateaued. Three of the four spline terms were statistically significant (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), supporting a curved relationship rather than a constant linear effect. The spline-based marginal effects plot confirmed this saturating pattern, with predicted hemoglobin rising from ~\u0026thinsp;13.5 g/dL to ~\u0026thinsp;15.0 g/dL, then leveling off beyond BMI 40 kg/m\u0026sup2; (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u0026nbsp;\u003ctable id=\"Tab4\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eSpline Model Predicting Hemoglobin from BMI\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eEstimate\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStd. Error\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eIntercept\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e12.66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.080\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;2e\u0026ndash;16 ***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ens(BMI, df\u0026thinsp;=\u0026thinsp;4)1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.068\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;2e\u0026ndash;16 ***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ens(BMI, df\u0026thinsp;=\u0026thinsp;4)2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.58\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.116\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;2e\u0026ndash;16 ***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ens(BMI, df\u0026thinsp;=\u0026thinsp;4)3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.285\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.15e\u0026ndash;14 ***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ens(BMI, df\u0026thinsp;=\u0026thinsp;4)4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026ndash;0.53\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.495\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.291\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003eIn the spline-based interaction model including CRP (Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e), hemoglobin remained non-linearly associated with BMI. One spline \u0026times; CRP interaction term was statistically significant (p\u0026thinsp;=\u0026thinsp;0.0002), indicating that the magnitude of BMI\u0026rsquo;s effect on hemoglobin was modified by inflammation. As visualized in Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e, individuals with higher CRP levels (log(CRP)\u0026thinsp;\u0026gt;\u0026thinsp;2) exhibited markedly attenuated hemoglobin gains across the upper BMI range, compared to those with lower inflammation.\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab5\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eSpline Interaction Model: Hemoglobin\u0026thinsp;~\u0026thinsp;BMI \u0026times; log(CRP)\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eEstimate\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStd. Error\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eIntercept\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e12.58\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e0.089\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;2e\u0026ndash;16 ***\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ens(BMI, df\u0026thinsp;=\u0026thinsp;4)1\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e2.22\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e0.080\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;2e\u0026ndash;16 ***\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003ens(BMI, df\u0026thinsp;=\u0026thinsp;4)2\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e2.16\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.172\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e1.71e\u0026ndash;15 ***\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003ens(BMI, df\u0026thinsp;=\u0026thinsp;4)3\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e4.10\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.599\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e3.00e\u0026ndash;08 ***\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003ens(BMI, df\u0026thinsp;=\u0026thinsp;4)4\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.65\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e1.167\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.579\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003elog(CRP)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.021\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.072\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.768\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003ens(BMI, df\u0026thinsp;=\u0026thinsp;4)1 \u0026times; log(CRP)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026ndash;0.060\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.076\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.436\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003ens(BMI, df\u0026thinsp;=\u0026thinsp;4)2 \u0026times; log(CRP)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026ndash;0.341\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.084\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0002 ***\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003ens(BMI, df\u0026thinsp;=\u0026thinsp;4)3 \u0026times; log(CRP)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026ndash;0.413\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.294\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.167\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003ens(BMI, df\u0026thinsp;=\u0026thinsp;4)4 \u0026times; log(CRP)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026ndash;0.102\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.490\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.836\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003eMarginal Effects of BMI and CRP on Hemoglobin and Anemia\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\n \u003cp\u003eFigure \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e shows the marginal predictions of hemoglobin concentration across the observed range of BMI values, based on the fully adjusted linear model. A clear positive association was observed: hemoglobin levels increased with BMI, with a predicted mean of approximately 13.6 g/dL at a BMI of 20 kg/m\u0026sup2; and rising to over 15.0 g/dL at BMI levels exceeding 60 kg/m\u0026sup2;.\u003c/p\u003e\n \u003cp\u003eIn contrast, the predicted probability of anemia declined with increasing BMI (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e). At a BMI of 20 kg/m\u0026sup2;, the average predicted anemia probability was above 20%, while at a BMI of 40 kg/m\u0026sup2;, it fell below 10%, and continued to decrease modestly at higher BMI levels (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eFigure 5 displays the predicted probability of anemia across values of log-transformed CRP. Anemia risk rose progressively with inflammation: individuals with log(CRP) near \u0026minus;\u0026thinsp;2 had predicted probabilities below 15%, whereas those with log(CRP) above 4 had probabilities exceeding 25% (Fig. 5).\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\n \u003ch2\u003eMediation Analysis Results\u003c/h2\u003e\n \u003cdiv id=\"Sec14\" class=\"Section3\"\u003e\n \u003ch2\u003eBMI \u0026rarr; Hemoglobin via CRP (Table \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e)\u003c/h2\u003e\n \u003cp\u003eIn the mediation analysis assessing whether systemic inflammation mediated the association between BMI and hemoglobin concentration (Table \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e), the average causal mediation effect (ACME) was statistically significant and negative (ACME = \u0026minus;\u0026thinsp;0.044, 95% CI: \u0026minus;\u0026thinsp;0.057 to \u0026minus;\u0026thinsp;0.030, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). The average direct effect (ADE) was positive and significant (ADE\u0026thinsp;=\u0026thinsp;0.216, 95% CI: 0.185 to 0.250, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), yielding a total effect of 0.172 (95% CI: 0.144 to 0.200, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). The proportion of the total effect mediated by CRP was \u0026minus;\u0026thinsp;26.0% (95% CI: \u0026minus;\u0026thinsp;33.9% to \u0026minus;\u0026thinsp;18.0%, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), indicating statistically significant but inverse mediation.\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u0026nbsp;\u003ctable id=\"Tab6\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eMediation of BMI\u0026ndash;Hemoglobin Association via CRP\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eEffect\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eEstimate\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e95% CI Lower\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e95% CI Upper\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eACME\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026ndash;0.044\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026ndash;0.057\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026ndash;0.030\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;2e\u0026ndash;16 ***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eADE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.216\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.185\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.250\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;2e\u0026ndash;16 ***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTotal Effect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.172\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.144\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.200\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;2e\u0026ndash;16 ***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eProp. Mediated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026ndash;0.260\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026ndash;0.339\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026ndash;0.180\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;2e\u0026ndash;16 ***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\n \u003ch2\u003eBMI \u0026rarr; Hemoglobin via Ferritin (Table \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e)\u003c/h2\u003e\n \u003cp\u003eIn the parallel analysis using BRINDA-adjusted ferritin as the mediator (Table \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e), the ACME was small but statistically significant and negative (ACME = \u0026minus;\u0026thinsp;0.0041, 95% CI: \u0026minus;\u0026thinsp;0.0075 to 0.0000, p\u0026thinsp;=\u0026thinsp;0.004). The direct effect remained similar to previous models (ADE\u0026thinsp;=\u0026thinsp;0.215, 95% CI: 0.184 to 0.250, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and the total effect was 0.211 (95% CI: 0.179 to 0.240, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). The proportion mediated was \u0026minus;\u0026thinsp;1.97% (95% CI: \u0026minus;\u0026thinsp;3.57% to \u0026minus;\u0026thinsp;1.00%, p\u0026thinsp;=\u0026thinsp;0.004), reflecting a small but significant component of the association mediated by ferritin.\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003ctable id=\"Tab7\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eMediation of BMI\u0026ndash;Hemoglobin Association via Ferritin\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eEffect\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eEstimate\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e95% CI Lower\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e95% CI Upper\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eACME\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026ndash;0.0041\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026ndash;0.0075\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.004 **\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eADE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.2153\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.1840\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.2500\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;2e\u0026ndash;16 ***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTotal Effect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.2112\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.1792\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.2400\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;2e\u0026ndash;16 ***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eProp. Mediated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026ndash;0.0197\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026ndash;0.0357\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026ndash;0.0100\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.004 **\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003cp\u003e\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe present study provides a detailed, nationally representative analysis of how adiposity and systemic inflammation jointly influence hemoglobin concentration in U.S. adults. By applying restricted cubic splines to body-mass index (BMI) with knots at the 10th, 50th, and 90th percentiles (17.2, 26.4, and 37.6 kg/m\u0026sup2;), we revealed a steep positive association between BMI and hemoglobin that notably plateaus beyond a BMI of \u0026asymp;\u0026thinsp;30 kg/m\u0026sup2;. This non-linear pattern aligns with emerging evidence from population and clinical cohorts: in Chinese adults, the plateauing of hemoglobin at higher BMI mirrors our findings [17], and similar curves have been observed in Iranian and Korean studies of obesity and erythropoiesis [6]. These data emphasize that linear models underestimate the complexity of adiposity\u0026rsquo;s hematologic effects, potentially obscuring thresholds at which additional weight confers minimal hematologic benefit.\u003c/p\u003e\n\u003cp\u003eMechanistically, obesity is characterized by chronic low-grade inflammation that drives hepcidin synthesis via interleukin-6 (IL-6) signaling, thereby restricting iron absorption and mobilization from macrophage stores [18]. Indeed, intervention studies demonstrate that weight loss reduces hepcidin levels and improves iron status in obese adults, supporting a causal role for adipose-derived inflammation in iron dysregulation [10, 19]. In our spline interaction model, the inclusion of log-transformed C-reactive protein (log-CRP) revealed that at low inflammation (log-CRP\u0026thinsp;\u0026le;\u0026thinsp;1 mg/L), BMI\u0026rsquo;s positive effect on hemoglobin remained strong, whereas at high inflammation (log-CRP\u0026thinsp;\u0026ge;\u0026thinsp;3 mg/L), hemoglobin gains were markedly attenuated or even reversed at high BMI. This interaction is consistent with clinical observations in bariatric surgery cohorts, where post-operative CRP reductions correlate with hemoglobin improvements despite ongoing caloric restriction [20].\u003c/p\u003e\n\u003cp\u003eFormal mediation analysis quantified this suppressive effect: CRP\u0026rsquo;s average causal mediation effect (ACME) was significantly negative (ACME = \u0026minus;\u0026thinsp;0.044 g/dL; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), indicating that inflammation partially counteracts the direct erythropoietic drive of excess adiposity [21]. The average direct effect (ADE) remained robustly positive (ADE\u0026thinsp;=\u0026thinsp;0.216 g/dL; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), yielding a total effect of 0.172 g/dL per BMI z-score. Because the indirect and direct effects opposed each other\u0026mdash;hallmarks of \u0026ldquo;inconsistent mediation\u0026rdquo;\u0026mdash;we reported ACME and ADE separately, foregoing a conventional \u0026ldquo;proportion mediated\u0026rdquo; metric that can be misleading under suppression scenarios [22].\u003c/p\u003e\n\u003cp\u003eIn contrast, ferritin\u0026mdash;when adjusted for inflammation via the BRINDA regression-residual approach\u0026mdash;mediated only\u0026thinsp;\u0026asymp;\u0026thinsp;2% of the BMI\u0026ndash;hemoglobin association (ACME = \u0026minus;\u0026thinsp;0.0041 g/dL; p\u0026thinsp;=\u0026thinsp;0.004; Table\u0026nbsp;6). Unadjusted ferritin often rises with both iron stores and inflammation, leading to misclassification of iron sufficiency in obese individuals [15]. The BRINDA project has shown that ferritin correction for CRP (and AGP where available) substantially alters iron deficiency prevalence estimates, particularly in high-inflammation settings [15]. Our minimal ferritin mediation underscores the necessity of inflammation adjustment in epidemiologic analyses and in clinical interpretation of ferritin values.\u003c/p\u003e\n\u003cp\u003eAlthough the manuscript emphasizes hemoglobin outcomes, supplementary logistic regression demonstrated parallel non-linear declines in anemia probability across BMI, with sharp risk reductions between BMI 18 and 30 kg/m\u0026sup2; and a plateau below 10% anemia prevalence at BMI\u0026thinsp;\u0026gt;\u0026thinsp;35 kg/m\u0026sup2; (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003e). Anemia odds rose with log-CRP in a near-linear fashion up to CRP levels of 50 mg/L (Fig.\u0026nbsp;6), reflecting inflammation\u0026rsquo;s role in anemia of chronic disease [5, 23]. Mediation of anemia by CRP and BMI exhibited true suppression (ACME\u0026thinsp;\u0026gt;\u0026thinsp;0; ADE\u0026thinsp;\u0026lt;\u0026thinsp;0), precluding valid proportion-mediated estimates and reinforcing our focus on separate indirect and direct effects [24].\u003c/p\u003e\n\u003cp\u003eOur findings refine understanding of obesity\u0026rsquo;s dual hematologic impacts. On one hand, adiposity stimulates erythropoietin (EPO) production, expanding erythroid progenitor activity and red-cell mass [25, 26]. Preclinical models demonstrate that EPO administration enhances metabolic function and modulates adipose-tissue inflammation, suggesting bidirectional regulation between erythropoiesis and adiposity [20]. On the other hand, inflammatory mediators, particularly IL-6, drive hepcidin-mediated iron sequestration, curtailing further hemoglobin synthesis at high adiposity and inflammation levels [18, 27]. The interplay of these mechanisms yields a non-linear BMI\u0026ndash;hemoglobin curve with a pronounced plateau and interaction by CRP.\u003c/p\u003e\n\u003cp\u003eFrom a public health perspective, our results advocate for integrated anemia screening in obese populations that includes inflammation-corrected iron markers (e.g., BRINDA-adjusted ferritin, soluble transferrin receptor) and CRP or hepcidin measurements [16]. Traditional reliance on unadjusted ferritin may delay diagnosis of iron deficiency, particularly in patients with BMI\u0026thinsp;\u0026gt;\u0026thinsp;30 kg/m\u0026sup2; and elevated CRP [9]. Furthermore, weight-loss interventions\u0026mdash;dietary or surgical\u0026mdash;that reduce inflammation may confer greater hematologic benefits than BMI reduction alone. A systematic review of weight-loss trials found that CRP declines closely track with hemoglobin increases, independent of dietary iron intake [19]. Similarly, emerging IL-6 receptor antagonists, such as tocilizumab and novel anti-hepcidin agents, have shown promise in chronic kidney disease and rheumatoid arthritis for improving hemoglobin, warranting exploration in obesity-related anemia [28].\u003c/p\u003e\n\u003cp\u003eOur study\u0026rsquo;s strengths include use of a large, nationally representative sample with rigorous survey weighting (adjusted MEC weights, primary sampling unit and strata specification), application of advanced spline modeling with transparent knot reporting, and comprehensive inflammation adjustment for ferritin [29, 30]. We conducted survey diagnostics via svydiag(), confirming no unduly influential PSUs and median design effects of \u0026asymp;\u0026thinsp;1.2, thereby validating variance estimates [31]. Sensitivity analyses excluding extreme CRP values (\u0026gt;\u0026thinsp;10 mg/L) and participants with eGFR\u0026thinsp;\u0026lt;\u0026thinsp;60 mL/min/1.73 m\u0026sup2; confirmed robustness of nonlinear patterns.\u003c/p\u003e\n\u003cp\u003eNonetheless, limitations warrant mention. The cross-sectional design precludes definitive causal inference, and although reverse causation by chronic illness is unlikely to fully account for our findings, longitudinal studies are needed to establish temporal ordering of BMI, CRP, hepcidin, and hemoglobin changes. Residual confounding by smoking, altitude, and unmeasured comorbidities may persist despite covariate adjustment; inclusion of pack‐year history and altitude‐adjusted hemoglobin shifted estimates by \u0026lt;\u0026thinsp;5% in supplementary models. Sample sizes at extreme BMI (\u0026gt;\u0026thinsp;50 kg/m\u0026sup2;) were limited, widening confidence intervals at the tails of spline curves. Finally, our mediation framework, while powerful, assumes no unmeasured mediator\u0026ndash;outcome confounders, an assumption that is difficult to verify in cross-sectional data.\u003c/p\u003e\n\u003cp\u003eFuture research should leverage prospective cohorts with serial measures of BMI, CRP, hepcidin, erythropoietin, plasma volume, and red-cell indices to unravel dynamic causal pathways. Randomized trials of anti‐inflammatory therapies (e.g., IL-6 blockade, anti‐hepcidin antibodies) in obese, anemic populations could test whether modulating inflammation alone improves iron bioavailability and erythropoiesis. Mechanistic studies assessing ferroportin expression, erythroferrone levels, and bone-marrow iron export will further elucidate the adiposity\u0026ndash;iron axis. Finally, as body composition assessment evolves beyond BMI\u0026mdash;incorporating measures of visceral adiposity, ectopic fat, and muscle mass\u0026mdash;future investigations can refine hematologic risk stratification and tailor anemia interventions in the context of metabolic health.\u003c/p\u003e\n\u003cp\u003eIn conclusion, our spline-based analysis demonstrates that obesity\u0026rsquo;s hemoglobin advantage is both substantial and finite, with systemic inflammation serving as a key suppressor of erythropoietic gains. Incorporating flexible modeling strategies and inflammation-corrected iron biomarkers into clinical practice will enhance anemia detection and guide more effective interventions amid the co‐epidemics of obesity and anemia.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e \u003cp\u003eThis study used deidentified data from the publicly available NHANES database and did not involve direct patient contact or the use of individually identifiable health information. Under the U.S. Common Rule, research using only publicly available, deidentified data is exempt from institutional review board oversight; therefore, ethics approval and patient consent were not required.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eConsent for publication\u003c/strong\u003e \u003cp\u003eNot applicable.\u003c/p\u003e \u003c/p\u003e\u003cp\u003e \u003ch2\u003eCompeting interests\u003c/h2\u003e \u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eNo external funding was received for this work.\u003c/p\u003e\u003ch2\u003eAuthors\u0026rsquo; contributions\u003c/h2\u003e \u003cp\u003eAH conceived the study, performed data extraction and statistical analyses, and drafted the manuscript. PS assisted with critical revision of the manuscript. All authors read and approved the final manuscript.\u003c/p\u003e\u003ch2\u003eAvailability of data and materials\u003c/h2\u003e \u003cp\u003eThe dataset analyzed during the current study is available in the NHANES repository: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.cdc.gov/nchs/nhanes/\u003c/span\u003e\u003cspan address=\"https://www.cdc.gov/nchs/nhanes/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003e\u003cstrong\u003ePrevalence, years lived with disability, and trends in anaemia burden by severity and cause, 1990-2021: findings from the Global Burden of Disease Study 2021\u003c/strong\u003e. \u003cem\u003eLancet Haematol \u003c/em\u003e2023, \u003cstrong\u003e10\u003c/strong\u003e(9):e713-e734.\u003c/li\u003e\n\u003cli\u003ePhelps NH, Singleton RK, Zhou B, Heap RA, Mishra A, Bennett JE, Paciorek CJ, Lhoste VPF, Carrillo-Larco RM, Stevens GA\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eWorldwide trends in underweight and obesity from 1990 to 2022: a pooled analysis of 3663 population-representative studies with 222 million children, adolescents, and adults\u003c/strong\u003e. \u003cem\u003eThe Lancet \u003c/em\u003e2024, \u003cstrong\u003e403\u003c/strong\u003e(10431):1027-1050.\u003c/li\u003e\n\u003cli\u003eNg M, Gakidou E, Lo J, Abate YH, Abbafati C, Abbas N, Abbasian M, Abd ElHafeez S, Abdel-Rahman WM, Abd-Elsalam S\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eGlobal, regional, and national prevalence of adult overweight and obesity, 1990\u0026amp;#x2013;2021, with forecasts to 2050: a forecasting study for the Global Burden of Disease Study 2021\u003c/strong\u003e. \u003cem\u003eThe Lancet \u003c/em\u003e2025, \u003cstrong\u003e405\u003c/strong\u003e(10481):813-838.\u003c/li\u003e\n\u003cli\u003eRachmah Q, Mondal P, Phung H, Ahmed F: \u003cstrong\u003eAssociation between overweight/obesity and iron deficiency anaemia among women of reproductive age: a systematic review\u003c/strong\u003e. \u003cem\u003ePublic Health Nutr \u003c/em\u003e2024, \u003cstrong\u003e27\u003c/strong\u003e(1):e176.\u003c/li\u003e\n\u003cli\u003eAcharya SR, Timilsina D, Acharya S: \u003cstrong\u003eAssociation between blood hemoglobin levels, anemia, and body mass index in children and women of Myanmar: findings from a nationally representative health study\u003c/strong\u003e. \u003cem\u003eScientific Reports \u003c/em\u003e2024, \u003cstrong\u003e14\u003c/strong\u003e(1):32020.\u003c/li\u003e\n\u003cli\u003eChen Z, Cao B, Liu L, Tang X, Xu H: \u003cstrong\u003eAssociation between obesity and anemia in an nationally representative sample of United States adults: a cross-sectional study\u003c/strong\u003e. \u003cem\u003eFront Nutr \u003c/em\u003e2024, \u003cstrong\u003e11\u003c/strong\u003e:1304127.\u003c/li\u003e\n\u003cli\u003eTarancon-Diez L, Iriarte-Gahete M, Sanchez-Mingo P, Mu\u0026ntilde;oz-Fernandez M\u0026Aacute;, Navarro-Gomez ML, Pacheco YM, Leal M: \u003cstrong\u003eImpact of obesity on iron metabolism and the effect of intravenous iron supplementation in obese patients with absolute iron deficiency\u003c/strong\u003e. \u003cem\u003eScientific Reports \u003c/em\u003e2025, \u003cstrong\u003e15\u003c/strong\u003e(1):1343.\u003c/li\u003e\n\u003cli\u003eNdevahoma F, Mukesi M, Dludla PV, Nkambule BB, Nepolo EP, Nyambuya TM: \u003cstrong\u003eBody weight and its influence on hepcidin levels in patients with type 2 diabetes: A systematic review and meta-analysis of clinical studies\u003c/strong\u003e. \u003cem\u003eHeliyon \u003c/em\u003e2021, \u003cstrong\u003e7\u003c/strong\u003e(3):e06429.\u003c/li\u003e\n\u003cli\u003eBerton PF, Gambero A: \u003cstrong\u003eHepcidin and inflammation associated with iron deficiency in childhood obesity - A systematic review\u003c/strong\u003e. \u003cem\u003eJornal de Pediatria \u003c/em\u003e2024, \u003cstrong\u003e100\u003c/strong\u003e(2):124-131.\u003c/li\u003e\n\u003cli\u003eAlshwaiyat NM, Ahmad A, Wan Hassan WMR, Al-Jamal HAN: \u003cstrong\u003eAssociation between obesity and iron deficiency (Review)\u003c/strong\u003e. \u003cem\u003eExp Ther Med \u003c/em\u003e2021, \u003cstrong\u003e22\u003c/strong\u003e(5):1268.\u003c/li\u003e\n\u003cli\u003eCepeda-Lopez AC, Zimmermann MB, Wussler S, Melse-Boonstra A, Naef N, Mueller SM, Toigo M, Herter-Aeberli I: \u003cstrong\u003eGreater blood volume and Hb mass in obese women quantified by the carbon monoxide-rebreathing method affects interpretation of iron biomarkers and iron requirements\u003c/strong\u003e. \u003cem\u003eInt J Obes (Lond) \u003c/em\u003e2019, \u003cstrong\u003e43\u003c/strong\u003e(5):999-1008.\u003c/li\u003e\n\u003cli\u003eVricella LK: \u003cstrong\u003eEmerging understanding and measurement of plasma volume expansion in pregnancy\u003c/strong\u003e. \u003cem\u003eAm J Clin Nutr \u003c/em\u003e2017, \u003cstrong\u003e106\u003c/strong\u003e(Suppl 6):1620s-1625s.\u003c/li\u003e\n\u003cli\u003ePergola PE, Davidson M, Jensen C, Mohseni Zonoozi AA, Raj DS, Andreas Schytz P, Tuttle KR, Perkovic V: \u003cstrong\u003eEffect of Ziltivekimab on Determinants of Hemoglobin in Patients with CKD Stage 3-5: An Analysis of a Randomized Trial (RESCUE)\u003c/strong\u003e. \u003cem\u003eJ Am Soc Nephrol \u003c/em\u003e2024, \u003cstrong\u003e35\u003c/strong\u003e(1):74-84.\u003c/li\u003e\n\u003cli\u003eBradbury BD, Critchlow CW, Weir MR, Stewart R, Krishnan M, Hakim RH: \u003cstrong\u003eImpact of elevated C-reactive protein levels on erythropoiesis- stimulating agent (ESA) dose and responsiveness in hemodialysis patients\u003c/strong\u003e. \u003cem\u003eNephrol Dial Transplant \u003c/em\u003e2009, \u003cstrong\u003e24\u003c/strong\u003e(3):919-925.\u003c/li\u003e\n\u003cli\u003eNamaste SM, Rohner F, Huang J, Bhushan NL, Flores-Ayala R, Kupka R, Mei Z, Rawat R, Williams AM, Raiten DJ\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eAdjusting ferritin concentrations for inflammation: Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) project\u003c/strong\u003e. \u003cem\u003eAm J Clin Nutr \u003c/em\u003e2017, \u003cstrong\u003e106\u003c/strong\u003e(Suppl 1):359s-371s.\u003c/li\u003e\n\u003cli\u003eLuo H, Geng J, Zeiler M, Nieckula E, Sandalinas F, Williams A, Young MF, Suchdev PS: \u003cstrong\u003eA Practical Guide to Adjust Micronutrient Biomarkers for Inflammation Using the BRINDA Method\u003c/strong\u003e. \u003cem\u003eThe Journal of Nutrition \u003c/em\u003e2023, \u003cstrong\u003e153\u003c/strong\u003e(4):1265-1272.\u003c/li\u003e\n\u003cli\u003eWei G, Lin F, Cao C, Hu H, Han Y: \u003cstrong\u003eNon-linear dose-response relationship between body mass index and stroke risk in middle-aged and elderly Chinese men: a nationwide Longitudinal Cohort Study from CHARLS\u003c/strong\u003e. \u003cem\u003eFrontiers in Endocrinology \u003c/em\u003e2023, \u003cstrong\u003eVolume 14 - 2023\u003c/strong\u003e.\u003c/li\u003e\n\u003cli\u003eSanad M, Osman M, Gharib A: \u003cstrong\u003eObesity modulate serum hepcidin and treatment outcome of iron deficiency anemia in children: a case control study\u003c/strong\u003e. \u003cem\u003eItal J Pediatr \u003c/em\u003e2011, \u003cstrong\u003e37\u003c/strong\u003e:34.\u003c/li\u003e\n\u003cli\u003eSelvin E, Paynter NP, Erlinger TP: \u003cstrong\u003eThe Effect of Weight Loss on C-Reactive Protein: A Systematic Review\u003c/strong\u003e. \u003cem\u003eArchives of Internal Medicine \u003c/em\u003e2007, \u003cstrong\u003e167\u003c/strong\u003e(1):31-39.\u003c/li\u003e\n\u003cli\u003eNoguchi CT, Rogers HM, Wang L, Teng R: \u003cstrong\u003eErythropoietin Increases Erythropoiesis, Metabolism and Weight Loss In Mice\u003c/strong\u003e. \u003cem\u003eBlood \u003c/em\u003e2013, \u003cstrong\u003e122\u003c/strong\u003e(21):946-946.\u003c/li\u003e\n\u003cli\u003eFong TCT: \u003cstrong\u003eIndirect Effects of Body Mass Index Growth on Glucose Dysregulation via Inflammation: Causal Moderated Mediation Analysis\u003c/strong\u003e. \u003cem\u003eObes Facts \u003c/em\u003e2019, \u003cstrong\u003e12\u003c/strong\u003e(3):316-327.\u003c/li\u003e\n\u003cli\u003ePetimar J, Tabung FK, Valeri L, Rosner B, Chan AT, Smith-Warner SA, Giovannucci EL: \u003cstrong\u003eMediation of associations between adiposity and colorectal cancer risk by inflammatory and metabolic biomarkers\u003c/strong\u003e. \u003cem\u003eInt J Cancer \u003c/em\u003e2019, \u003cstrong\u003e144\u003c/strong\u003e(12):2945-2953.\u003c/li\u003e\n\u003cli\u003eSantos-Silva MA, Sousa N, Sousa JC: \u003cstrong\u003eCorrelation Analysis between Hemoglobin and C-Reactive Protein in Patients Admitted to an Emergency Unit\u003c/strong\u003e. \u003cem\u003eJ Clin Med \u003c/em\u003e2021, \u003cstrong\u003e10\u003c/strong\u003e(22).\u003c/li\u003e\n\u003cli\u003eMei Z, Namaste SML, Serdula M, Suchdev PS, Rohner F, Flores-Ayala R, Addo OY, Raiten DJ: \u003cstrong\u003eAdjusting total body iron for inflammation: Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) project\u003c/strong\u003e. \u003cem\u003eThe American Journal of Clinical Nutrition \u003c/em\u003e2017, \u003cstrong\u003e106\u003c/strong\u003e:383S-389S.\u003c/li\u003e\n\u003cli\u003eVega-S\u0026aacute;nchez R, Tolentino-Dolores MC, Cerezo-Rodr\u0026iacute;guez B, Chehaibar-Besil G, Flores-Quijano ME: \u003cstrong\u003eErythropoiesis and Red Cell Indices Undergo Adjustments during Pregnancy in Response to Maternal Body Size but not Inflammation\u003c/strong\u003e. \u003cem\u003eNutrients \u003c/em\u003e2020, \u003cstrong\u003e12\u003c/strong\u003e(4).\u003c/li\u003e\n\u003cli\u003eH\u0026auml;m\u0026auml;l\u0026auml;inen P, Saltevo J, Kautiainen H, M\u0026auml;ntyselk\u0026auml; P, Vanhala M: \u003cstrong\u003eErythropoietin, ferritin, haptoglobin, hemoglobin and transferrin receptor in metabolic syndrome: a case control study\u003c/strong\u003e. \u003cem\u003eCardiovascular Diabetology \u003c/em\u003e2012, \u003cstrong\u003e11\u003c/strong\u003e(1):116.\u003c/li\u003e\n\u003cli\u003edel Giudice EM, Santoro N, Amato A, Brienza C, Calabrò P, Wiegerinck ET, Cirillo G, Tartaglione N, Grandone A, Swinkels DW\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eHepcidin in Obese Children as a Potential Mediator of the Association between Obesity and Iron Deficiency\u003c/strong\u003e. \u003cem\u003eThe Journal of Clinical Endocrinology \u0026amp; Metabolism \u003c/em\u003e2009, \u003cstrong\u003e94\u003c/strong\u003e(12):5102-5107.\u003c/li\u003e\n\u003cli\u003eEl-Mallah CA, Beyh YS, Obeid OA: \u003cstrong\u003eIron Fortification and Supplementation: Fighting Anemia of Chronic Diseases or Fueling Obesity?\u003c/strong\u003e \u003cem\u003eCurrent Developments in Nutrition \u003c/em\u003e2021, \u003cstrong\u003e5\u003c/strong\u003e(4):nzab032.\u003c/li\u003e\n\u003cli\u003eFeret W, Safranow K, Ciechanowski K, Kwiatkowska E: \u003cstrong\u003eHow Is Body Composition and Nutrition Status Associated with Erythropoietin Response in Hemodialyzed Patients? A Single-Center Prospective Cohort Study\u003c/strong\u003e. \u003cem\u003eJ Clin Med \u003c/em\u003e2022, \u003cstrong\u003e11\u003c/strong\u003e(9).\u003c/li\u003e\n\u003cli\u003eGonz\u0026aacute;lez-Dom\u0026iacute;nguez \u0026Aacute;, Visiedo-Garc\u0026iacute;a FM, Dom\u0026iacute;nguez-Riscart J, Gonz\u0026aacute;lez-Dom\u0026iacute;nguez R, Mateos RM, Lechuga-Sancho AM: \u003cstrong\u003eIron Metabolism in Obesity and Metabolic Syndrome\u003c/strong\u003e. \u003cem\u003eInt J Mol Sci \u003c/em\u003e2020, \u003cstrong\u003e21\u003c/strong\u003e(15).\u003c/li\u003e\n\u003cli\u003eFernandez-Pombo A, Lorenzo PM, Carreira MC, Gomez-Arbelaez D, Castro AI, Primo D, Rodriguez J, Sajoux I, Baltar J, de Luis D\u003cem\u003e et al\u003c/em\u003e: \u003cstrong\u003eA very-low-calorie ketogenic diet normalises obesity-related enhanced levels of erythropoietin compared with a low-calorie diet or bariatric surgery\u003c/strong\u003e. \u003cem\u003eJournal of Endocrinological Investigation \u003c/em\u003e2024, \u003cstrong\u003e47\u003c/strong\u003e(11):2701-2713.\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Lebanese University","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Obesity, Anemia, Hemoglobin, Inflammation, CRP, Ferritin, NHANES, BRINDA, Causal Mediation, Nonlinear Models, Spline Regression","lastPublishedDoi":"10.21203/rs.3.rs-6895556/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6895556/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e\u003cbr\u003e\nObesity and anemia are global epidemics with complex, overlapping pathophysiology. While excess adiposity is known to induce chronic inflammation that disrupts iron homeostasis, multiple population studies paradoxically report higher hemoglobin levels and lower anemia prevalence among obese individuals. The nonlinear and potentially suppressive role of inflammation in this relationship remains understudied.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e\u003cbr\u003e\nWe analyzed adults aged 18–64 from the 2015–2023 National Health and Nutrition Examination Survey (NHANES). Hemoglobin was modeled as a function of BMI using survey-weighted linear regression with restricted cubic splines. Interactions with log-transformed CRP were assessed, and ferritin was corrected for inflammation using BRINDA regression-residual methods. Causal mediation analysis decomposed the total effect of BMI on hemoglobin into indirect (mediated by CRP) and direct effects. Secondary models examined anemia (Hb \u0026lt;13.0 g/dL in men, \u0026lt;12.0 g/dL in women) using logistic regression.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e\u003cbr\u003e\nHemoglobin increased steeply across lower BMI ranges but plateaued above 30 kg/m² (p-nonlinearity \u0026lt; 0.001). The hemoglobin–BMI curve flattened significantly at higher CRP levels, with strong evidence of interaction (p-interaction \u0026lt; 0.001). Mediation analysis showed that CRP significantly suppressed the BMI–hemoglobin relationship (ACME = –0.044 g/dL, p \u0026lt; 0.001; ADE = 0.216 g/dL, p \u0026lt; 0.001). In contrast, BRINDA-adjusted ferritin mediated \u0026lt;2% of the association. Logistic models showed that anemia risk declined sharply with increasing BMI but rose consistently with CRP. Anemia mediation analysis revealed suppression as well (ACME \u0026gt; 0; ADE \u0026lt; 0), precluding interpretation of proportion mediated.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions:\u003c/strong\u003e\u003cbr\u003e\nBMI is positively associated with hemoglobin in a non-linear, CRP-dependent fashion. Inflammation significantly suppresses the hematologic benefit of excess adiposity, while inflammation-adjusted ferritin plays a minimal mediating role. These findings underscore the importance of modeling nonlinearity and correcting iron biomarkers for inflammation when studying obesity-related anemia.\u003c/p\u003e","manuscriptTitle":"Revisiting the Obesity–Anemia Paradox: Inflammation and Iron Homeostasis in the BMI–Hemoglobin Relationship","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-17 15:31:00","doi":"10.21203/rs.3.rs-6895556/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"d77a0ba5-2f67-464b-b205-d3642a9eaaf7","owner":[],"postedDate":"June 17th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-06-17T15:31:00+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-17 15:31:00","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6895556","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6895556","identity":"rs-6895556","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}