The real diagnostic value of serum iron in iron deficiency anemia among patients with inflammatory bowel disease: A retrospective study

Abstract

Objective

The diagnostic utility of serum iron in identifying iron deficiency anemia among patients with inflammatory bowel disease remains underreported. This study aimed to evaluate the true diagnostic value of serum iron.

Methods

A total of 160 patients with iron deficiency anemia from a grade-A general hospital in Sichuan Province between January 2022 and January 2025 were enrolled. Eighty patients with inflammatory bowel disease and iron deficiency anemia were included in the test group, while 80 patients with iron deficiency anemia without inflammation were included in the control group. Iron metabolism parameters, including serum iron and serum ferritin, complete blood count indices, and biochemical markers, were collected and compared. Differences in diagnosing iron deficiency anemia using serum iron and serum ferritin among patients with inflammatory bowel disease were compared using the chi-square test. Spearman correlation analysis was performed among serum iron, serum ferritin, and other indicators. Meanwhile, potential factors affecting serum iron levels were investigated using linear regression, and the significance of serum iron in diagnosing inflammatory bowel disease with iron deficiency was evaluated using the receiver operating characteristic curve.

Results

Among patients with inflammatory bowel disease, ulcerative colitis and Crohn’s disease showed no significant differences in serum iron and serum ferritin levels. The real clinical results showed that when using 1–6.6 times the lower limit of normal of serum ferritin as a threshold to diagnose patients with inflammatory bowel disease and iron deficiency anemia, a statistically significant difference was observed compared with the serum iron lower limit of 1 time (p < 0.01). The positivity rate of iron deficiency anemia was 88.75% (71/80) for serum iron and 61.25% (49/80) for 3.3-time serum ferritin. Correlation analysis revealed that serum iron was positively correlated with serum ferritin (r = 0.263, p = 0.018) and negatively correlated with most inflammatory parameters (p < 0.05). Multiple linear regression analysis showed that when serum iron was used as the dependent variable, the optimal regression model with the independent variables serum ferritin, hemoglobin, platelet count, and monocyte-to-lymphocyte ratio had regression coefficients (β) of 0.246, 0.231, −0.405, and −0.187, respectively, all of which were significantly different (p < 0.05). Moreover, according to the European Crohn’s and Colitis Organization diagnostic criteria for iron deficiency, patients with inflammatory bowel disease were divided into an iron deficiency group involving 41 patients and a non-iron deficiency group involving 39 patients. By comparing the parameters of the two groups and constructing the receiver operating characteristic curve, serum iron alone yielded the highest Youden index and the largest area under the curve for diagnosing iron deficiency (area under the curve = 0.764, p < 0.001). However, when the false-positive rate was controlled at 20%, the sensitivity of serum iron for diagnosing iron deficiency was 62.00%.

Conclusions

In patients with inflammatory bowel disease, serum iron is inversely associated with inflammation and appears to be more sensitive than serum ferritin in detecting iron deficiency anemia. These findings suggest that combined measurement of serum iron and serum ferritin may enhance the diagnosis of iron deficiency anemia in clinical practice.

Keywords

Serum iron ferritin iron metabolism inflammatory bowel disease iron deficiency anemia

Introduction

Iron deficiency anemia (IDA) is a common systemic complication of inflammatory bowel disease (IBD) and is the most common type of anemia in patients with IBD, particularly in inactive or quiescent phases of the disease. IDA significantly impairs patients’ quality of life¹ and is associated with increased healthcare costs.² Prior to the onset of anemia, when iron stores are moderately depleted, the recycled iron from daily red blood cell turnover is sufficient to support erythropoiesis and hemoglobin (Hb) synthesis—a condition referred to as iron deficiency (ID).³ The European Crohn’s and Colitis Organization (ECCO) guidelines⁴ emphasize the importance of early diagnosis of ID. The prevalence of ID among patients with IBD ranges from 36% to 90%⁵ and is characterized by inflammation-induced myelosuppression and restricted iron availability within enterocytes, hepatocytes, and macrophages. Symptoms associated with ID—both with and without anemia—include fatigue, reduced exercise tolerance, pica, restless legs syndrome, cognitive decline, and depression.¹ Severe ID may also impair ferritin synthesis in nonerythroid cells, leading to cellular dysfunction and extra-hematopoietic manifestations such as epithelial changes in the nails, tongue, and esophagus; deficits in cognitive and muscular performance; and compromised adaptive immune responses.³ Crohn’s disease (CD) is more prevalent than ulcerative colitis (UC) in many populations.⁶

Currently, there is no universally accepted diagnostic criterion for ID in patients with IBD. Bone marrow biopsy remains the gold standard for confirming ID; however, due to its invasiveness, it is not routinely performed in clinical practice.³ Conventional hematologic parameters such as Hb, mean corpuscular volume (MCV), and mean corpuscular hemoglobin concentration (MCHC) are limited, and novel markers of iron metabolism—such as hepcidin, transferrin saturation (TSAT), and soluble transferrin receptor (sTfR)—have been proposed to assess body iron status. Hepcidin expression is regulated by multiple factors: it increases during inflammation and iron overload and decreases under conditions of hypoxia, enhanced erythropoiesis, and ID.⁷ Given the chronic inflammatory state in IBD, interpretation of hepcidin levels must account for confounding effects of inflammation. Due to its fluctuating nature, whether hepcidin reliably reflects true ID in inflamed individuals requires validation through prospective trials.⁸ TSAT is calculated as the ratio of serum iron (SI) to total iron-binding capacity. SI represents circulating iron bound to transferrin and is influenced by circadian rhythm.⁹ TSAT has been suggested as an adjunctive tool for diagnosing ID when serum ferritin (SF) levels fall between 100 and 300 µg/L.¹⁰ However, SI is a key variable contributing to TSAT variability, and its levels can fluctuate markedly based on dietary iron intake and clinical status.¹¹ Moreover, TSAT is sensitive to changes in both SI and transferrin levels. Transferrin may remain above 20% despite low transferrin levels,¹² potentially masking underlying ID. Transferrin decreases during inflammation and chronic disease, and the TSAT measurement process is technically complex. sTfR serves as a biomarker of erythropoietic activity; however, different assays for sTfR lack standardization, and reference thresholds vary across platforms, limiting its widespread clinical utility. Other emerging tests—such as reticulocyte hemoglobin content (CHr) and low hemoglobin density (LHD)¹³—have shown promise but are not yet widely adopted in routine clinical settings. Although these markers are useful for evaluating body iron status, they lack prospective validation, are not widely used in clinical practice, involve complex detection methods, and may increase costs for patients.

The 2015 ECCO guidelines⁴ recommend defining ID in patients with IBD as SF < 30 µg/L in the absence of inflammation or SF < 100 µg/L in the presence of inflammation. In contrast, the American Gastroenterological Association suggests using an SF cutoff of 45 µg/L for diagnosing ID.¹⁴ However, during inflammation, SF—an acute-phase reactant—may be elevated independently of iron stores, thereby failing to accurately reflect true iron status. Previous studies in other patient populations have indicated that SI is subject to diurnal variation and influenced by dietary intake, which may limit its diagnostic reliability.^15–17 In real clinical practice, patients with IBD who have reduced Hb often have normal or increased SF levels; however, their fasting morning SI levels remain persistently low, supporting the diagnosis of IDA. This observation highlights the potential supplementary role of SI in diagnosis. SI refers to iron bound to transferrin in circulation. In patients with IBD, iron is lost from the intestine due to factors such as inflammation and chronic bleeding. Hepcidin affects SI by promoting the degradation of iron transporters in intestinal absorptive cells and iron-recycling macrophages.¹⁸ Furthermore, there are limited studies proposing SI as an indicator for evaluating IDA in patients with IBD. Therefore, this study aims to evaluate and compare the diagnostic value of SI and SF in detecting IDA in patients with IBD.

Patients and methods

Patients

This study initially screened 300 patients with IBD from the Affiliated Hospital of North Sichuan Medical College between January 2022 and January 2025. A total of 82 patients with IBD with a clear diagnosis of IDA and complete iron metabolism indicators, including SI and SF, were selected. Among them, nine patients had previously received iron supplementation therapy, but none had received iron supplementation for more than 6 months prior to follow-up observation, while the remaining patients had not received iron supplementation previously. Two patients had missing data and were therefore excluded from the analysis. To compare differences in diagnosing patients with IBD with IDA using different multiples of the lower limit of normal (LLN) of SF versus the LLN of SI, with a binary outcome indicator, a preliminary analysis showed that 90.63% of patients had decreased SI levels and 68.75% had SF levels below three-times the LLN (which is close to the international guideline⁴ of SF < 100 µg/L). The sample size calculation formula was n = (Z_α/2 +Z_β)² × [p₁(1-p₁)+p₂(1-p₂)]/(p₁p₂)². With a β error of 0.1 and a two-sided α error of 0.05, the sample size was 71. Considering a 10% loss to follow-up rate, a total of 80 patients with IBD (39 patients with UC and 41 with CD) with IDA were included. Additionally, to explore the impact of inflammatory status on iron metabolism indicators in patients with IBD, 80 patients with IDA without acute or chronic active inflammatory diseases were included as a control group for comparison. Inclusion criteria were as follows: (a) confirmed diagnosis of IBD according to established international criteria (e.g. ECCO guidelines); (b) diagnosis of IDA defined as follows: anemia based on the 2024 World Health Organization criteria¹⁶ (Hb < 120 g/L in nonpregnant women and Hb < 130 g/L in men aged ≥15 years), concurrent evidence of ID per the 2015 ECCO guidelines,⁴ and documented improvement in hemoglobin following iron supplementation. Exclusion criteria included history of oral or intravenous iron therapy, antianemia medications, vitamin C supplementation, or blood transfusion within the past 3 months; prior major gastrointestinal surgery; active acute or chronic infections; suspected or confirmed malignancy; significant organ dysfunction (cardiac, hepatic, or renal); coagulopathy; presence of mixed or other types of anemia; pregnancy or lactation; psychiatric disorders or impaired consciousness; and incomplete medical records.

Methods

Data collection

Clinical data collected included medical history, age, sex, disease activity (Truelove–Witts score for patients with UC, Crohn’s Disease Activity Index for patients with CD), SI (our hospital reference range: males: 10.6–36.7 µmol/L; females: 7.8–32.2 µmol/L), SF (our hospital reference range: males: 30–400 µg/L; females: 13–150 µg/L), erythropoietin (EPO), vitamin B12, folic acid, Hb, hematocrit (HCT), MCV, MCHC, white blood cell count (WBC), platelet count (PLT), albumin (ALB), globulin (GLB), high-sensitivity C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), absolute neutrophil count (ANC), and absolute lymphocyte count (ALC).

Platelet-to-lymphocyte ratio (PLR =PLT/ALC), monocyte-to-lymphocyte ratio (MLR = AMC/ALC), and neutrophil-to-lymphocyte ratio (NLR = ANC/ALC) were calculated.

Laboratory examination methods

All participants fasted for at least 8 h prior to blood sampling. Morning fasting venous blood (3 mL) was collected in anticoagulant tubes, thoroughly mixed, and centrifuged. Hematologic parameters were analyzed using an automated hematology analyzer, while iron metabolism and biochemical markers were assessed using an automated biochemical analyzer. Equipment used included the Hitachi LABOSPECT 008α fully automated biochemical analyzer and BC-3000 hematology analyzer (Shenzhen Mindray Biomedical Electronics Co., Ltd.).

Statistical analysis

This study adopted a retrospective design, with the sample drawn from patients at a single center, which may introduce selection bias. Furthermore, the study included only patients with a definitive diagnosis of IDA, and the data relied on medical records, which may introduce information bias. We attempted to minimize these biases by excluding incomplete data; however, their impact could not be completely eliminated. Data were analyzed using SPSS 25.0. Normally distributed continuous variables were expressed as mean ±standard deviation (SD) and compared using Student’s t-test. Non-normally distributed variables were presented as M (P₂₅_, P₇₅) and analyzed using the Mann–Whitney U test. Categorical variables were reported as frequencies (%) and compared using the chi-square test. Spearman correlation analysis was performed to assess associations among variables, and a correlation heatmap was generated. Multiple linear regression models were constructed with SI and SF as dependent variables and statistically significant indicators as independent variables to identify potential predictors. Receiver operating characteristic (ROC) curve analysis was conducted to calculate the area under the curve (AUC), sensitivity, specificity, positive and negative predictive values, Youden index, and optimal cutoff values for markers in diagnosing ID in patients with IBD. A two-sided p-value <0.05 was considered statistically significant. The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.¹⁹

Results

Characteristics and laboratory findings of patients with IBD and non-IBD

Compared with patients with IDA without IBD, those with IBD had a younger median age and mainly had moderate anemia. Patients with IBD had higher SI and SF values, with the median SF value exceeding 30 µg/L; however, SI values in both groups were below the LLN. Student’s t-test was applied for normally distributed variables (Hb, HCT, PLT, WBC, MCV, MCHC, ALB, and GLB), while the Mann–Whitney U test was used for non-normally distributed variables (SI, SF, and EPO). The chi-square test was used for age and degree of anemia. Except for sex and vitamin B12, statistically significant differences were observed between the two groups (p < 0.05) (Table 1).

Table 1.

Characteristics and laboratory findings of IBD patients and non-IBD patients.

Characteristics	IBD (n = 80)	Non-IBD (n = 80)	Statistical significance	p-value
Age (M(P₂₅, P₇₅), years)	36.00 (21.25–54.75)	55.00 (42.00–70.00)	−5.111^a	<0.001
Sex (n, %)
Male	43 (26.88%)	31 (19.38%)	3.620^c	0.057
Female	37 (23.13%)	49 (3.06%)	3.620^c	0.057
Degree of anemia (n, %)
Mild anemia	17 (10.63%)	3 (1.88%)	70.809^c	<0.001
Moderate anemia	46 (28.75%)	9 (5.63%)
Severe anemia	17 (10.63%)	68 (42.50%)
Hb (mean ± SD, g/L)	94.08 ± 20.43	66.32 ± 20.48	−8.557^b	<0.001
HCT (mean ± SD)	0.305 ± 0.064	0.233 ± 0.051	−7.848^b	<0.001
SI (M(P₂₅, P₇₅), µmol/L)	4.50 (2.65, 7.08)	3.20 (1.90, 4.90)	−3.220^a	0.001
SF (M(P₂₅, P₇₅), µg/L)	49.15 (13.20, 134.0)	5.00 (2.90, 13.90)	−8.011^a	<0.001
PLT counts (mean ± SD, ×10¹¹/L)	345.89 ± 163.79	252.37 ± 100.99	−4.340^b	<0.001
WBC counts (M(P₂₅, P₇₅), ×10⁹/L)	6.89 (5.31, 8.81)	5.07 (4.09, 6.63)	−4.198^a	<0.001
MCV (mean ± SD, fL)	81.68 ± 9.99	75.76 ± 11.11	−3.530^b	0.001
MCHC (mean ± SD, g/L)	306.85 ± 17.76	284.56 ± 20.45	−7.335^b	<0.001
ALB (M(P₂₅, P₇₅), g/L)	38.35 (33.50, 42.55)	41.40 (37.10, 43.50)	−3.694^a	<0.001
GLB (M(P₂₅, P₇₅), g/L)	32.70 (30.23, 35.20)	28.00 (25.70, 30.30)	5.733^a	<0.001
EPO (M(P₂₅, P₇₅), mIU/L)	26.62 (14.44, 54.53)	334.60 (103.53, 746.00)	−8.254^a	<0.001
Vitamin B12 (M(P₂₅, P₇₅), pg/mL)	322.50 (216.50, 439.25)	283.00 (230.00, 395.00)	−0.687^a	0.492
Folic acid (M(P₂₅, P₇₅), ng/mL)	8.37 (4.97, 14.60)	11.31 (7.80, 16.96)	−2.585^a	0.010

Data are presented as median (first quartile, third quartile) for non-normally distributed continuous variables, mean ± SD for normally distributed variables, and frequency (n, %) for categorical variables. Comparisons were performed using the Mann–Whitney U test.

ALB: albumin; GLB: globulin; Hb: hemoglobin; HCT: hematocrit; MCHC: mean corpuscular hemoglobin concentration; MCV: mean corpuscular volume; PLT: platelet; SF: serum ferritin; SI: serum iron; WBC: white blood cell.

Independent-samples t-test.

Chi-square test.

As appropriate.

Characteristics and laboratory findings of patients with UC and CD

Among the 80 patients with IBD, 39 had UC and 41 had CD. With regard to anemia severity, 17 patients had mild anemia (9 UC, 8 CD), 46 had moderate anemia (21 UC, 25 CD), and 17 had severe anemia (9 UC, 8 CD). All patients with IBD were in the moderate-to-severe active phase of disease. Student’s t-test was applied for normally distributed variables (Hb, HCT, PLT, WBC, MCV, and GLB). The chi-square test was used for age, degree of anemia, and disease activity. The Mann–Whitney U test was used for non-normally distributed variables (SI, SF, MCHC, and EPO). Significant differences were observed between groups in age, disease activity, PLT, MCV, MCHC, vitamin B12, folic acid, PLR, and MLR (p < 0.05), whereas no significant differences were found for other variables (Table 2).

Table 2.

Characteristics and laboratory findings of patients with UC and patients with CD.

Characteristics	UC (n = 39)	CD (n = 41)	Statistical significance	p-value
Age (M(P₂₅, P₇₅), years)	55.00 (32.00, 65.00)	27.00 (19.00, 36.50)	−5.046^a	<0.001
Sex (n, %)
Male	18 (22.50%)	25 (31.25%)	1.776^c	0.184
Female	21 (26.25%)	16 (20.00%)	1.776^c	0.184
Degree of anemia (n, %)
Mild anemia	9 (11.25%)	8 (10.00%)	0.416^c	0.812
Moderate anemia	21 (26.25%)	25 (31.25%)
Severe anemia	9 (11.25%)	8 (10.00%)
Active degree (n, %)
Moderate active	10 (12.50%)	35 (43.75%)	28.971^c	<0.001
Severe active	29 (36.25%)	6 (7.50%)	28.971^c	<0.001
Hb (mean ± SD, g/L)	93.26 ± 20.67	94.85 ± 20.42	−0.348^b	0.729
HCT (mean ± SD)	0.295 ± 0.049	0.315 ± 0.076	−1.394^b	0.168
SI (M(P₂₅, P₇₅), µmol/L)	5.10 (3.00, 8.00)	4.30 (2.50, 5.85)	−1.550^a	0.121
SF (M(P₂₅, P₇₅), µg/L)	49.90 (13.00, 190.80)	46.70 (17.90, 124.00)	−0.135^a	0.893
PLT counts (mean ± SD, ×10¹¹/L)	300.46 ± 171.322	389.10 ± 145.51	−2.498^b	0.015
WBC counts (mean ± SD, ×10⁹/L)	7.53 ± 2.86	6.80 ± 2.35	1.248^b	0.216
MCV (mean ± SD, fL)	84.82 ± 10.63	78.69 ± 8.41	2.869^b	0.005
MCHC (M(P₂₅, P₇₅), g/L)	310.00 (301.00, 321.00)	303.00 (295.00, 312.50)	−2.494^a	0.013
ALB (M(P₂₅, P₇₅), g/L)	40.60 (32.20–43.20)	37.10 (33.30–42.05)	−0.953^a	0.341
GLB (mean ± SD, g/L)	32.22 ± 5.62	33.49 ± 5.13	−1.057^b	0.294
EPO (M(P₂₅, P₇₅), mIU/L)	27.60 (13.42, 54.65)	26.50 (14.60, 51.83)	−0.265^a	0.791
Vitamin B12 (M(P₂₅, P₇₅), pg/mL)	398.00 (286.00, 591.00)	249.00 (194.0, 355.50)	−2.965^a	0.003
Folic acid (M(P₂₅, P₇₅), ng/mL)	9.70 (6.85, 15.18)	6.98 (4.33, 12.35)	−2.445^a	0.014
CRP (M(P₂₅, P₇₅), mg/L)	18.41 (1.85, 36.22)	21.24 (12.13,57.62)	−1.574^a	0.116
ESR (M(P₂₅, P₇₅), mm/h)	29.00 (17.00, 69.00)	49.00 (24.00, 75.00)	−1.497^a	0.134
PLR (M(P₂₅, P₇₅))	184.87 (101.41, 273.85)	258.46 (199.64, 432.04)	−3.431^a	0.001
MLR (M(P₂₅, P₇₅))	0.27 (0.16, 0.44)	0.35 (0.25, 0.45)	−1.988^a	0.047
NLR (M(P₂₅, P₇₅))	2.84 (1.81, 4.20)	3.31 (2.62, 4.60)	−1.401^a	0.161

ALB: albumin; CRP: C-reactive protein; ESR: erythrocyte sedimentation rate; GLB: globulin; Hb: hemoglobin; HCT: hematocrit; MCHC: mean corpuscular hemoglobin concentration; MCV: mean corpuscular volume; MLR: monocyte-to-lymphocyte ratio; NLR: neutrophil-to-lymphocyte ratio; PLR: platelet-to-lymphocyte ratio; PLT: platelet; SF: serum ferritin; SI: serum iron; WBC: white blood cell.

Independent-samples t-test.

Chi-square test.

As appropriate.

SI and SF at various multiples of LLN to identify IDA in IBD

Comparison of the diagnostic performance of SF with 1–6.6 times the LLN and SI with 1 time the LLN showed significant differences (p < 0.01). Among these, the positivity rate of SI for diagnosing IDA was 88.75% (71/80), whereas the positivity rate of SF at 3.3 times for diagnosing IDA was 61.25% (49/80). However, when SF thresholds reached eight times the LLN, no significant difference was observed. The true diagnostic value of SI in detecting IDA in patients with IBD appears to be more sensitive than that of SF. Using ECCO criteria⁴ based on SF levels to identify IDA in patients with IBD may be insufficient (Table 3).

Table 3.

Comparison of the diagnostic performance of SI and SF at various multiples of their LLN.

Parameter	Below the lower limit (n, %)	Above the lower limit (n, %)	χ²	p-value
SI	71 (88.75%)	9 (11.25%)	55.104	<0.001
1 time SF	25 (31.25%)	55 (68.75%)	55.104	<0.001
SI	71 (88.75%)	9 (11.25%)	29.789	<0.001
2-time SF	39 (48.75%)	41 (51.25%)	29.789	<0.001
SI	71 (88.75%)	9 (11.25%)	16.133	<0.001
3.3-time SF	49 (61.25%)	31 (38.75%)	16.133	<0.001
SI	71 (88.75%)	9 (11.25%)	12.692	0.001
4-time SF	52 (65.00%)	28 (35.00%)	12.692	0.001
SI	71 (88.75%)	9 (11.25%)	7.656	0.006
5-time SF	57 (71.25%)	23 (28.75%)	7.656	0.006
SI	71 (88.75%)	9 (11.25%)	6.762	0.009
6.6-time SF	58 (72.50%)	22 (27.50%)	6.762	0.009
SI	71 (88.75%)	9 (11.25%)	3.609	0.057
8-time SF	62 (77.50%)	18 (22.50%)	3.609	0.057

Comparisons were performed using the chi-square test (χ²). The normal reference ranges in our institution are as follows: SI: males, 10.6–36.7 µmol/L; females, 7.8–32.2 µmol/L; SF: males, 30–400 µg/L; females, 13–150 µg/L.

LLN: lower limit of normal; SF: serum ferritin; SI: serum iron.

Spearman correlation analysis and multiple linear regression analysis

Spearman correlation analysis revealed a statistically significant association between SI and SF. SI was negatively correlated with WBC, CRP, ESR, PLT, PLR, MLR, and NLR and positively correlated with SF, Hb, MCV, and MCHC. Additionally, SF was negatively correlated with ALB and positively correlated with MCHC (Figure 1). Using SI as the dependent variable, all significantly associated parameters (p < 0.05) were entered as independent variables into a multiple linear regression model. Through backward stepwise elimination, the final model included SF, PLT, Hb, and MLR as independent predictors. Their regression coefficients (β) were 0.246, 0.231, −0.405, and −0.187, respectively. The model demonstrated an R² of 0.433 and an adjusted R² of 0.402, with a Durbin–Watson (D-W) statistic of 1.803, indicating acceptable residual independence. All included variables satisfied the assumption of independence and independently influenced SI levels (p < 0.05). Considering the large numerical differences among independent variables, we created simple linear regression plots of each independent variable against the dependent variable SI (Figure 2). The variance inflation factor for each predictor was <5, confirming the absence of significant multicollinearity. Taking SF as the dependent variable and incorporating SI, ALB, and MCHC into a linear regression analysis, it was ultimately determined that only ALB, as an independent variable, had a significant impact on SF (p < 0.05).

Figure 1.

Spearman correlation analysis of serum iron, serum ferritin, and other indices. *Statistically significant (p < 0.05); ** statistically significant (p < 0.01). ALB: albumin; CRP: C-reactive protein; EPO: erythropoietin; ESR: erythrocyte sedimentation rate; GLB: globulin; Hb: hemoglobin; HCT: hematocrit; MCHC: mean corpuscular hemoglobin concentration; MCV: mean corpuscular volume; MLR: monocyte-to-lymphocyte ratio; NLR: neutrophil-to-lymphocyte ratio; PLR: platelet-to-lymphocyte ratio; PLT: platelet; SF: serum ferritin; SI: serum iron; WBC: white blood cell.

Figure 2.

Linear regression plots with SI as the dependent variable and SF, Hb, PLT, and MLR as independent variables. Hb: hemoglobin; MLR: monocyte-to-lymphocyte ratio; PLT: platelet; SF: serum ferritin; SI: serum iron.

Diagnostic efficacy analysis of different parameters for IBD with ID

Significant differences were observed in SI, MCV, and MCHC between the two groups (Table 4). Using ID status as the binary outcome variable (ID = 1, non-ID = 0), according to the ECCO guideline,⁴ ROC curve analysis was performed for the three significantly different parameters. Results indicated that among individual markers, SI yielded the largest AUC and Youden index for diagnosing ID in patients with IBD. The optimal cutoff value for SI was 5.30 µmol/L, which is below the institutional LLN. MLR showed the highest sensitivity, while MCHC demonstrated the highest specificity. However, the specificity of SI alone was relatively low. SI showed the highest positive predictive value (PPV), while MLR showed the lowest negative predictive value (NPV). As shown in Figure 3, when the false-positive rate was controlled at 20%, the sensitivity of SI remained the highest but was only 62% (Tables 4 and 5, Figure 3).

Table 4.

Characteristics and laboratory findings of patients with IBD with ID and without ID.

Characteristics	ID (n = 49)	Non-ID (n = 31)	Statistical significance	p-value
Age (M(P₂₅, P₇₅), years)	35.00 (24.50, 52.00)	37.00 (20.00, 51.00)	−0.178^c	0.859
Sex (n, %)			0.166^c	0.684
Male	23 (28.75%)	16 (20.00%)
Female	26 (32.50%)	15 (18.75%)
Active degree (n, %)			0.068^c	0.795
Moderate active	27 (33.75%)	18 (22.50%)
Severe active	22 (27.50%)	13 (16.25%)
SI (M(P₂₅, P₇₅), µmol/L)	3.30 (2.50, 5.10)	7.20 (4.80, 9.80)	−4.332^a	<0.001
HCT (mean ± SD)	0.302 ± 0.05	0.311 ± 0.08	0.551^b	0.585
EPO (M(P_25, P₇₅), mIU/L)	30.47 (17.33, 65.00)	20.48 (10.41, 45.05)	−2.099^a	0.036
MCV (mean ± SD, fL)	79.53 ± 9.75	85.05 ± 9.54	2.483^a	0.015
MCHC (M(P₂₅, P₇₅), mIU/L)	302.00 (295.00, 311.50)	313.00 (305.00, 321.00)	−3.444^a	0.001
PLT (mean ± SD, ×10¹¹/L)	367.76 ± 153.39	311.32 ± 176.02	−1.514^b	0.134
WBC (mean ± SD, ×10⁹/L)	7.41 ± 2.68	6.80 ± 2.33	−2.085^b	0.040
CRP (M(P₂₅, P₇₅), mg/L)	21.24 (10.89, 50.14)	18.30 (2.25, 41.27)	−1.027^a	0.304
ESR (M(P₂₅, P₇₅), mm/h)	52.00 (23.50, 76.50)	27.00 (11.00, 61.00)	−2.455^a	0.014
ALB (M(P₂₅, P₇₅), g/L)	39.90 (35.00, 42.60)	37.10 (27.40, 42.40)	−1.521^a	0.128
GLB (mean ± SD, g/L)	33.70 ± 5.37	31.55 ± 5.21	−1.762^b	0.082
PLR (M(P₂₅, P₇₅))	231.53 (180.40, 307.11)	223.94 (96.92, 383.78)	−0.193^a	0.847
MLR (M(P₂₅, P₇₅))	0.34 (0.22, 0.46)	0.24 (0.15, 0.39)	−2.612^a	0.009
NLR (M(P₂₅, P₇₅))	3.31 (2.19, 4.53)	3.07 (1.72, 4.36)	−0.667^a	0.505

ALB: albumin; CRP: C-reactive protein; EPO: erythropoietin; ESR: erythrocyte sedimentation rate; GLB: globulin; HCT: hematocrit; MCHC: mean corpuscular hemoglobin concentration; MCV: mean corpuscular volume; MLR: monocyte-to-lymphocyte ratio; NLR: neutrophil-to-lymphocyte ratio; PLR: platelet-to-lymphocyte ratio; PLT: platelet count; SI: serum iron; WBC: white blood cell count.

Student’s t-test.

Chi-square test.

As appropriate.

Figure 3.

ROC curve analysis of single-parameter diagnosis of patients with IBD and ID. EPO: erythropoietin; ESR: erythrocyte sedimentation rate; MCHC: mean corpuscular hemoglobin concentration; MCV: mean corpuscular volume; MLR: monocyte-to-lymphocyte ratio; SI: serum iron; WBC: white blood cell count.

Table 5.

Diagnostic efficacy analysis of seven parameters for ID.

Variable	Cutoff	AUC ± SE	CI （95%）	p-value	Sensitivity	Specificity	PPV	NPV	Youden index
SI	5.30	0.789 ± 0.057	0.676–0.901	<0.001	0.816	0.742	0.976	0.795	0.558
MCHC	302.50	0.729 ± 0.057	0.617–0.841	<0.001	0.510	0.903	0.610	0.923	0.413
MCV	86.20	0.692 ± 0.063	0.568–0.816	0.004	0.837	0.581	0.780	0.692	0.418
MLR	0.19	0.672 ± 0.066	0.543–0.802	0.010	0.898	0.419	0.878	0.333	0.317
ESR	33.50	0.664 ± 0.062	0.543–0.785	0.141	0.674	0.613	0.732	0.718	0.287
EPO	24.22	0.640 ± 0.066	0.511–0.769	0.036	0.674	0.645	0.780	0.718	0.319
WBC	5.86	0.635 ± 0.064	0.509–0.760	0.043	0.776	0.484	0.634	0.282	0.260

AUC: area under the curve; EPO: erythropoietin; ESR: erythrocyte sedimentation rate; MCHC: mean corpuscular hemoglobin concentration; MCV: mean corpuscular volume; MLR: monocyte-to-lymphocyte ratio; NPV: negative predictive value; PPV: positive predictive value; SI: serum iron; WBC: white blood cell count.

Discussion

IDA is the most prevalent form of anemia in patients with IBD. A multicenter Italian study reported that the prevalence of IDA among anemic patients with IBD reached 75.9%.²⁰ In IBD, chronic intestinal inflammation leads to mucosal blood loss, resulting in iron losses that exceed dietary absorption and thereby establishing a negative iron balance, which ultimately contributes to the development of IDA.²¹ In healthy individuals, total body iron stores range from 3 to 4 g, with daily iron losses of 1–2 mg that must be replenished through dietary intake, referred to as dietary iron absorption.²² Iron deficiency without anemia (IDWA) refers to a state in which iron stores are depleted despite normal hemoglobin levels.²³ Evidence suggests that intravenous iron supplementation in patients with IBD and IDWA not only restores iron parameters but also alleviates symptoms and improves quality of life.²⁴ Therefore, early identification of ID is crucial for patients with IBD.

According to ECCO criteria⁴ in patients with SF levels exceeding 100 µg/L, a multicenter study demonstrated that only 25% of patients with IBD were diagnosed with ID.²⁵ Among the remaining patients, one-third exhibited SF levels between 30 and 100 µg/L, and fewer than one-fifth had levels below 30 µg/L.²⁵ Notably, only 61.25% of 80 patients with IBD and IDA were correctly diagnosed based on ECCO guidelines⁴ in this study. However, SF, being an acute-phase reactant, is elevated during inflammation and thus may not accurately reflect true iron stores. Additionally, liver disease and metabolic syndrome can independently increase SF levels.⁷ Consequently, relying solely on SF for diagnosing IDA in patients with IBD is insufficient. This retrospective study showed that compared with patients with ordinary noninflammatory IDA, median SF levels (49.15 µg/L) in patients with IBD were significantly elevated and exceeded 30 µg/L. However, SI levels in both groups were below the lower limit of the institutional reference normal range. Additionally, most patients with IBD presented with moderate anemia, whereas patients with IDA without inflammation were primarily characterized by severe anemia. However, there was no significant difference between patients with UC and patients with CD in SI and SF. Spearman correlation analysis showed that SI was negatively correlated with inflammatory indicators, including CRP (r = −0.272), ESR (r = −0.292), PLR (r = −0.371), NLR (r = −0.260), and MLR (r = −0.347) (p < 0.05). However, SF was not correlated with these inflammatory indicators. In recent years, inflammatory indices such as MLR, PLR, and NLR have been validated as markers of systemic inflammation and immune response, with prognostic value in various conditions, including cardiovascular diseases, severe infections, malignancies, and autoimmune disorders.²⁶ A meta-analysis²⁷ suggested that both PLR and NLR may correlate with IBD disease activity. Specifically, PLR has emerged as a potential predictor of disease severity, while NLR may predict recurrence and endoscopic response. Julio et al.²⁸ found that MLR has high sensitivity, NPV, and odds ratio in the pre-endoscopic diagnosis of pediatric patients with UC. In this study, multiple linear regression showed that MLR influenced SI (β = −0.187, p = 0.048) and had diagnostic value for ID (AUC = 0.672, p = 0.01). Whether inflammatory indicators can predict IBD complicated by IDA requires further investigation. Both Spearman correlation and multiple linear regression analyses demonstrated a negative association between albumin and SF levels, suggesting that elevated SF may reflect disease severity rather than adequate iron stores.

In this study, we evaluated various multiples (1–8 times) of the LLN for SF versus one time the LLN for SI as hypothetical diagnostic thresholds for IDA in IBD. Significant differences were observed across 1–6.6 times the LLN for SF. In circulation, SI exists as iron bound to transferrin, forming a complex with two Fe³⁺ ions.²⁹ Since the 1940s, SI has been recognized to exhibit diurnal variation, peaking during the daytime and reaching a nadir at night.³⁰ It is primarily used to calculate TSAT and is influenced by dietary iron intake and supplementation.³¹ Despite fluctuations due to circadian rhythm and nutritional factors, SI decreases in response to both inflammation and ID. Low SI has been implicated in other chronic conditions, such as heart failure, where it aids in identifying ID,³² and in chronic kidney disease (stages 1–4), where it serves as an independent risk factor for anemia.¹¹ In patients with IBD, increased iron excretion due to epithelial shedding and active mucosal bleeding as well as acute and chronic blood loss from intestinal ulcers²¹ lead to reduced intestinal absorption of dietary iron, and excessive blood loss results in low SI levels.

To further investigate the diagnostic value of SI, patients with IBD were divided into ID and non-ID groups according to ECCO guidelines⁴ for the diagnosis of ID. The median SI value (3.30 µmol/L) was lower in the ID group than in the non-ID group (7.20 µmol/L), and both groups exhibited median SI levels below the LLN. SI yielded the largest area under the ROC curve and Youden index for diagnosing ID in IBD, with a sensitivity of 81.60% and specificity of 74.20%. However, when the false-positive rate was set at 20%, the sensitivity of SI was relatively low. Despite known variability due to circadian rhythm and iron intake, our protocol involved morning fasting blood sampling at peak SI levels, minimizing confounding effects. These findings suggest that SI has meaningful clinical utility in diagnosing IDA in patients with IBD but lacks sufficient sensitivity for standalone diagnosis of ID.

Through retrospective analysis, we compared the diagnostic performance of SI and SF at different multiples of the LLN for detecting IDA. We found that 61.25% of patients with SF > 100 µg/L still met criteria for IDA, indicating that SI may be more sensitive than SF in detecting early IDA. Nevertheless, this study has several limitations. First, it was conducted at a single center with a relatively small sample size. The sample was drawn from patients attending a single center retrospectively, which introduced selection bias. Furthermore, the specific data for each patient relied on medical records, which may introduce information bias to a certain extent. Second, patients with IBD are more commonly characterized by moderate anemia, whereas patients with ordinary IDA predominantly present with severe anemia. The purposes of their medical visits differ: patients with IBD seek care due to gastrointestinal symptoms, while patients with IDA without inflammation typically present with reduced hemoglobin levels due to anemia-related symptoms, which may cause differences in their SI and SF values and introduce bias. In addition, patients with UC were predominantly in the severe active phase, while patients with CD were mostly in the moderate active phase, potentially introducing selection bias. All patients with IBD were in the moderate-to-severe active phase, which may also contribute to bias. Third, variations in laboratory assays across institutions may lead to differences in reference ranges for iron metabolism parameters. Moreover, we selected patients with a clear clinical diagnosis of IDA. However, some of these patients had SF levels between 100 and 300 µg/L, which may indicate concurrent anemia of chronic disease. Our hospital currently lacks the capability to measure TSAT for further differentiation.¹² In addition, in the ROC analysis performed on patients with ID and non-ID, considering clinical practical applicability, the sensitivity of SI was very low when the false-positive rate was controlled at 20%. Therefore, testing SI should be used as a supplementary method in the diagnosis of IDA. Fourth, pediatric and pregnant populations were excluded; given their unique physiological demands, iron assessment in these groups warrants further investigation. Future studies should expand sample sizes and validate these findings in diverse cohorts to better define the diagnostic role of SI in IBD complicated by IDA.

In conclusion, relying solely on SF levels is inadequate for diagnosing IDA in real-world management of patients with IBD. Current diagnostic thresholds may fail to capture functional or early-stage ID. SI is inversely correlated with inflammatory activity and appears more sensitive than SF in detecting IDA, potentially compensating for the limitations of SF in inflammatory contexts. Therefore, combining SI and SF measurements in clinical practice may enhance the detection rate of IDA in patients with IBD and provide more robust guidance for therapeutic intervention. In future studies, we plan to expand the sample size, incorporate additional iron metabolism indicators, further distinguish the effects of inflammation and ID on SI, and analyze these differences.

Footnotes

Acknowledgments

We thank the patients and clinical staff who provide care for patients.

Author contributions

JTT, XW, HC, and GBH were involved in the study design. JTT and XW contributed equally to this work and share first authorship. JTT and GBH wrote the main manuscript and prepared the tables and figures. XW and HC facilitated data collection. All authors reviewed the manuscript.

Availability of data and materials

All data generated or analyzed during this study are included in this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and publication of this article.

Ethical statement

In this retrospective study, the requirement for written informed consent from patients was waived. This study was reviewed and approved by the Medical Ethics Committee of the Affiliated Hospital of North Sichuan Medical College (file number: 2025ER487-1).

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

ORCID iD

Jing-Tong Tang

References

Auerbach

DeLoughery

Tirnauer

JS.

Iron deficiency in adults: a review. JAMA 2025; 333: 1813–1823.

Fiorino

Colombel

Katsanos

, et al. Iron deficiency anemia impacts disease progression and healthcare resource consumption in patients with inflammatory bowel disease: a real-world evidence study. Ther Adv Gastroenterol 2023; 16: 17562848231177153.

Iolascon

Andolfo

Russo

; from EHA‐SWG Red Cell and Ironet al. Recommendations for diagnosis, treatment, and prevention of iron deficiency and iron deficiency anemia. Hemasphere 2024; 8: e108.

Dignass

Gasche

Bettenworth

; European Crohn’s and Colitis Organisation [ECCO]et al. European consensus on the diagnosis and management of iron deficiency and anaemia in inflammatory bowel diseases. J Crohns Colitis 2015; 9: 211–222.

Loveikyte

Boer

Van der Meulen

; Dutch Initiative on Crohn and Colitis (ICC)et al. Anemia and iron deficiency in outpatients with inflammatory bowel disease: ubiquitous yet suboptimally managed. J Clin Med 2022; 11: 6843.

Jabłońska

Mrowiec

Nutritional status and its detection inpatients with inflammatory bowel diseases. Nutrients 2023; 15: 1991.

Rai

Role of iron deficiency anemia in inflammatory bowel disease. World J Clin Pediatr 2024; 13: 93697.

Loveikyte

Bourgonje

Van der Reijden

, et al. Hepcidin and iron status in patients with inflammatory bowel disease undergoing induction therapy with vedolizumab or infliximab. Inflamm Bowel Dis 2023; 29: 1272–1284.

Shah

Patel

Khan

Iron deficiency anemia in IBD: an overlooked comorbidity. Expert Rev Gastroenterol Hepatol 2021; 15: 771–781.

10.

Dignass

Farrag

Stein

Limitations of serum ferritin in diagnosing iron deficiency in inflammatory conditions. Int J Chronic Dis 2018; 2018: 9394060.

11.

Lin

Chiu

, et al. Low serum iron is associated with anemia in CKD stage 1-4 patients with normal transferrin saturations. Sci Rep 2021; 11: 8343.

12.

Elsayed

Sharif

Stack

AG.

Transferrin saturation: a body iron biomarker. Adv Clin Chem 2016; 75: 71–97.

13.

Gordon

Burisch

Ellul

, et al. ECCO guidelines on extraintestinal manifestations in inflammatory bowel disease. J Crohns Colitis 2024; 18: 1–37.

14.

Sninsky

Brooks

Sansgiry

, , et al. Anemia in inflammatory bowel disease: practice patterns and clinical outcomes in a large national veterans cohort. Crohns Colitis 2020; 7: otaf012.

15.

Belbellaj

Lona-Durazo

Bodano

, et al. The role of genetically predicted serum iron levels on neurodegenerative and cardiovascular traits. Sci Rep 2024; 14: 24588.

16.

World Health Organization. Guideline on haemoglobin cutoffs to define anaemia in individuals and populations, https://www.ncbi.nlm.nih.gov/books/NBK602198/ (2024, accessed 10 November 2025).

17.

Zhao

Huang

Dai

, et al. Serum iron level as a potential predictor of coronavirus disease 2019 severity and mortality: a retrospective study. Open Forum Infect Dis 2020; 7: ofaa250.

18.

Camaschella

Nai

Silvestri

Iron metabolism and iron disorders revisited in the hepcidin era. Haematologica 2020; 105: 260–272.

19.

Cuschieri

The STROBE guidelines. Saudi J Anaesth 2019; 13: S31–S34.

20.

Bergamaschi

Castiglione

D'Incà

, et al. Prevalence, pathogenesis and management of anemia in inflammatory bowel disease: an IG-IBD multicenter, prospective, and observational study. Inflamm Bowel Dis 2023; 29: 76–84.

21.

Broekaert

Assa

Borrelli

, et al. Approach to anaemia in gastrointestinal disease: a position paper by the ESPGHAN Gastroenterology Committee. J Pediatr Gastroenterol Nutr 2025; 80: 510–532.

22.

Kaitha

Bashir

Ali

Iron deficiency anemia in inflammatory bowel disease. World J Gastrointest Pathophysiol 2015; 6: 62–72.

23.

Sartain

Al-Ezairej

McDonnell

, et al. Iron deficiency and fatigue in inflammatory bowel disease: a systematic review. PLoS One 2025; 20: e0304293.

24.

Huguet

Cortés

Boscá-Watts

, et al. Ferric carboxymaltose improves the quality of life of patients with inflammatory bowel disease and iron deficiency without anaemia. J Clin Med 2022; 11: 2786.

25.

Peyrin-Biroulet

Bouguen

Laharie

; The CARENFER study groupet al. Iron deficiency in patients with inflammatory bowel diseases: a prospective multicenter cross-sectional study. Dig Dis Sci 2022; 67: 5637–5646.

26.

Huang

Feng

, et al. Inflammatory indexes and anemia in chronic kidney disease: correlation and survival analysis of the National Health and Nutrition Examination Survey 2005-2018. Ren Fail 2024; 46: 2399314.

27.

Xiong

, et al. Association between the immune-inflammation index and the severity and clinical outcomes of patients with inflammatory bowel disease: a systematic review and meta-analysis. BMC Gastroenterol 2025; 25: 414.

28.

Moreno Alfonso

Barbosa-Velásquez

Delgado-Miguel

, et al. Exploring the utility of cellular indices in the diagnosis of ulcerative colitis. Gastroenterol Hepatol 2025; 48: 502349.

29.

Bennett

Pettikiriarachchi

McLean

ARD

, et al. Serum iron and transferrin saturation variation are circadian regulated and linked to the harmonic circadian oscillations of erythropoiesis and hepatic Tfrc expression in mice. Am J Hematol 2024; 99: 2075–2083.

30.

Pasricha

Tye-Din

Muckenthaler

, et al. Iron deficiency. Lancet 2021; 397: 233–248.

31.

Kriel

Opie

Rusch

, et al. Old tests and new paradigms: how to interpret iron studies and related biomarkers for the diagnosis of iron deficiency in adult. Blood Rev 2025; 74: 101337.

32.

Grote Beverborg

Klip

Meijers

, et al. Definition of iron deficiency based on the gold standard of bone marrow iron staining in heart failure. Circ Heart Fail 2018; 11: e004519.