Characteristics of transferrin saturation and anemia-related biomarkers in patients with uterine adenomyosis

Adenomyosis is a benign gynecological disorder characterized by the development of endometrial glandular epithelium and endometrial stromal tissue within the myometrium. The frequency of adenomyosis, as estimated from histopathological examinations after hysterectomy, varies widely, ranging from 8.8% to 61.5% [ 1 , 2 ]. The frequency is reported to be highest in patients in their early 40s [ 3 ]. Adenomyosis is associated with hypermenorrhea [ 4 ], dysmenorrhea [ 5 , 6 ], and obstetric complications such as preterm labor, preterm premature rupture of membranes, spontaneous abortion, gestational diabetes mellitus, small for gestational age, and preeclampsia [ 5 ]. The definitive treatment for adenomyosis is total hysterectomy; other treatments include endocrine therapies, such as low-dose estrogen-progestin, progestin, and gonadotropin-releasing hormone antagonist/agonist, and symptomatic treatments such as iron therapy and analgesics [ 7 ]. The management of anemia, a significant pathology of the disease, is particularly important. Anemia is typically identified based on hemoglobin (Hb) levels; however, a subtype of iron deficiency can occur without overt anemia. Iron is essential for erythropoiesis and oxygen transport, and iron deficiency can cause symptoms such as fatigue, impaired concentration, dizziness, and headache, regardless of anemia [ 8 , 9 ]. Iron deficiency affects many essential biological processes such as DNA synthesis and repair and enzyme activity [ 10 , 11 ]. In premenopausal women, blood loss due to regular menstruation causes iron deficiency. However, iron deficiency often remains unrecognized, even in women routinely screened for anemia. Iron deficiency presents as either absolute or functional deficiency. Absolute iron deficiency is characterized by reduced systemic iron levels, whereas functional iron deficiency is characterized by normal or higher systemic iron levels, uneven distribution of iron, and inadequate supply of iron to target tissues [ 2 , 12 , 13 ]. Absolute iron deficiency is diagnosed using low serum ferritin levels, which reflect a deficiency in iron levels of the body. Reference values for serum ferritin levels in patients with iron deficiency vary. The guidelines generally state a cut-off value of <15–30 ng/mL in apparently healthy individuals. Because ferritin is an acute-phase reactant, higher thresholds are recommended in the presence of infection or inflammation (e.g., < 70 ng/mL in adults), and cut-offs may vary by clinical context and guideline [ 14 , 15 ]. Other indicators of iron deficiency include a transferrin saturation (TSAT) of <20%, which reflects iron availability. TSAT is a measure of the proportion of iron-binding sites on transferrin that are occupied by iron; it is calculated as the ratio of serum iron to total iron-binding capacity (TIBC) [ 16 , 17 ]: A TSAT level <20% suggests iron deficiency, indicating that there is insufficient iron bound to transferrin, leading to inadequate iron supply to the bone marrow and other tissues [ 18 ]. Ferritin levels alone may not accurately reflect iron levels in the presence of chronic diseases or inflammation. Ferritin can be elevated in inflammatory conditions, masking an underlying iron deficiency [ 12 ]. The use of TSAT to diagnose iron deficiency is particularly valuable in cases of inflammation [ 12 ]. TSAT allows the assessment of iron availability for erythropoiesis and other iron-dependent processes, even when ferritin levels are elevated due to inflammation [ 19 ]. Several studies and guidelines recommend using both TSAT and ferritin to diagnose iron deficiency in chronic inflammatory conditions; a recent study proposed TSAT- and ferritin-based indices to support the interpretation of iron status in the context of inflammation [ 20 , 21 ]. Beyond gynecology, iron deficiency has clinically relevant effects on the kidney–cardiovascular axis. In chronic kidney disease and heart failure, functional iron deficiency is common and associated with poorer exercise capacity, quality of life (QoL), and outcomes even when Hb levels are preserved [ 22 , 23 ]. Because ferritin behaves as an acute-phase reactant in these inflammatory states [ 22 ], contemporary practice combines TSAT and ferritin findings to define iron deficiency [ 23 , 24 ]. TSAT is a useful indicator in inflammation, as it reflects bioavailable iron for erythropoiesis when ferritin may be spuriously elevated [ 22 , 24 ]. Although functional iron deficiency is associated with various diseases, it has not been extensively studied in benign gynecological conditions, including uterine adenomyosis. Therefore, the present study aimed to assess the frequency of functional iron deficiency in patients with benign gynecological diseases.

During the study period, 1054 patients first visited the hospital, and their blood samples were collected. Based on the study criteria, 854 and 200 patients were initially excluded and included, respectively. The participants were initially categorized into the adenomyosis (42 patients), myoma (108 patients), and no uterine structural abnormality (50 patients) groups. We used two prespecified analysis populations: a correlation cohort (Hb and TSAT levels available at the first visit; summarized in Table 1 ) and a ferritin analytic cohort (first-visit ferritin levels available) for iron-deficiency frequencies. Median [range]; n (%) After further exclusion of patients without ferritin level measurements, 29 patients were ultimately included in the adenomyosis group, 55 in the myoma group, and 40 in the no uterine structural abnormality group. Patient demographics of the correlation cohort are presented in Table 1 . There were no significant differences in age, body mass index, or the frequency of self-reported hypermenorrhea between the three groups. However, a higher proportion of patients in the adenomyosis group reported fatigue compared to the other groups. The frequency of anemia and iron deficiency in each group, as diagnosed based on Hb for anemia and serum ferritin and TSAT levels for iron deficiency, is shown in Fig 2 . Anemia was defined as Hb < 11 g/dL. Iron deficiency frequencies were summarized using three marker-based criteria: ferritin-only (serum ferritin <20 ng/mL), TSAT-only (transferrin saturation <20%), and a combined criterion (ferritin <20 ng/mL and/or TSAT <20%). In the adenomyosis group, iron deficiency diagnosed using ferritin, and TSAT was significantly higher than that diagnosed using ferritin alone. Hb, hemoglobin; TSAT, transferrin saturation. Iron deficiency was more prevalent than anemia across all groups. Notably, in the adenomyosis group, iron deficiency diagnosed using ferritin and TSAT levels was significantly higher than that diagnosed using ferritin level alone. Agreement between ferritin-based and TSAT-based classifications was modest. Among patients with ferritin measured at the first visit (n = 124), 45 (36.3%) met both criteria (ferritin <20 ng/mL and TSAT <20%), 40 (32.3%) met none of the criteria, 14 (11.3%) met the ferritin-only criterion, and 25 (20.2%) met the TSAT-only criterion (TSAT <20% with ferritin ≥20 ng/mL). The agreement between ferritin-based and TSAT-based classifications was fair (Cohen’s kappa = 0.375; PABAK = 0.371). In multivariable logistic regression for TSAT-only classification (TSAT <20% with ferritin ≥20 ng/mL), adjusting for age and hypermenorrhea, adenomyosis showed higher odds compared with no uterine structural abnormality (OR 3.123, 95% CI 0.942–11.30; p = 0.0688), whereas myoma was not associated (OR 1.234, 95% CI 0.399–4.079; p = 0.7184). The model showed modest discrimination (AUC = 0.678, 95% CI 0.564–0.792) ( S1 Fig ). The correlation between log(TSAT) and Hb concentration in each group is shown in Fig 3 . Only a weak correlation trend was observed in the adenomyosis group (r = 0.2123, p = 0.1771), whereas a relatively strong and statistically significant correlation was found in the myoma (r = 0.5465, p < 0.0001) and no uterine structural abnormality groups (r = 0.6945, p < 0.0001). The correlation was weak in the adenomyosis group, compared to strong correlations in the myoma and no uterine structural abnormality groups. Hb, hemoglobin; TSAT, transferrin saturation. Using Spearman’s rank correlation on untransformed TSAT, we observed the same qualitative pattern: adenomyosis group ρ = 0.358, p = 0.020 (n = 42); myoma group ρ = 0.576, p < 0.0001 (n = 108); no uterine structural abnormality group ρ = 0.484, p = 0.0004 (n = 50). Stratified analyses similarly mirrored Pearson analysis results (adenomyosis with hypermenorrhea: ρ = 0.218, p = 0.342, n = 21; without hypermenorrhea: ρ = 0.563, p = 0.0079, n = 21). Furthermore, the patients were subdivided into six groups based on the presence or absence of hypermenorrhea. Approximately half of patients in each group experienced hypermenorrhea. The correlation between log(TSAT) and Hb levels was examined in these six subgroups. A notable absence of correlation was observed in the adenomyosis group with hypermenorrhea (r = 0.0526, p = 0.8209) ( Fig 4 ).

Our findings indicate that TSAT measurement in patients with adenomyosis is a critical diagnostic parameter for iron deficiency and could potentially inform diagnostic and management strategies. The incorporation of TSAT into the diagnostic repertoire enhances the detection of iron deficiency, particularly functional iron deficiency, which may not be identified using serum ferritin levels alone. Therefore, the assessment of TSAT may offer a more comprehensive evaluation of iron level in patients with adenomyosis, leading to improved therapeutic outcomes. However, further research is warranted to confirm these findings and explore the optimal incorporation of TSAT measurements in the clinical management of iron deficiency in uterine adenomyosis.

We conducted a retrospective cross-sectional observational study using electronic medical records from the first outpatient visit. Specifically, we performed a retrospective analysis of electronic medical records of patients who presented for their initial consultation at our hospital between April 2019 and January 2025. At our institution, a complete blood count is routinely performed as part of the initial assessment of new patients. Further iron studies, including serum iron, TIBC, and particularly serum ferritin, were performed at the discretion of the attending physician based on the clinical context. Non-menopausal women aged 30–55 years who underwent initial blood testing were included. The exclusion criteria were as follows: (1) pregnancy or the postpartum period; (2) malignant tumors; (3) postmenopausal status; (4) receipt of iron therapy prior to blood sampling; (5) autoimmune diseases; (6) renal diseases; and (7) insufficient clinical data for analysis. All analyses were based on this eligible first-visit cohort. For specific analyses involving serum ferritin, the dataset was restricted to data on patients for whom a ferritin value was available from their first visit ( Fig 1 ). To mitigate selection bias arising from non-protocolized ordering of iron studies, we enrolled all consecutive new patients who met the eligibility criteria, used only first-visit laboratory results, and applied only prespecified, non-discretionary exclusion criteria. Clinical data, including the presence of adenomyosis, uterine fibroids, menorrhagia, fatigue symptoms, Hb concentration, TSAT levels, and serum ferritin levels, were extracted from medical records. Uterine fibroids and adenomyosis were diagnosed using ultrasonography and magnetic resonance imaging. The participants were categorized into three groups: 1) those with adenomyosis (adenomyosis group); 2) those with uterine myoma (myoma group); and 3) those without structural uterine abnormalities (no uterine structural abnormality group). We defined anemia as Hb < 11 g/dL to capture clinically significant (moderate-to-severe) anemia. The WHO definition of anemia in non-pregnant women is Hb < 12 g/Dl; however, Hb < 11 g/dL corresponds to at least moderate anemia in severity classification, and it is commonly used in Japanese gynecologic practice for non-pregnant women with gynecologic conditions [ 25 , 26 ]. Absolute iron deficiency was defined as serum ferritin <20 ng/mL. Functional iron deficiency (FID) was defined as TSAT <20%, with ferritin ≥20 ng/mL. Overall iron deficiency was defined as ferritin < 20 ng/mL and/or TSAT < 20%. Continuous variables were summarized as medians (range) and compared across groups using the Kruskal–Wallis test. Categorical variables were summarized as n (%) and analyzed using the chi-squared or Fisher’s exact test, as appropriate. The McNemar’s test was used for paired binary comparisons within the same participants. Agreement between ferritin-based and TSAT-based classifications (ferritin <20 ng/mL vs TSAT <20%) was evaluated using Cohen’s kappa, with PABAK (2Po − 1) reported as a sensitivity analysis. Factors associated with discordant (“TSAT-only”) classification suggestive of functional iron deficiency (TSAT <20% with ferritin ≥20 ng/mL) were assessed using multivariable logistic regression, and model discrimination was summarized via ROC/AUC (ROC shown in S1 Fig ). Analyses requiring ferritin were restricted to patients with ferritin measured at the first visit. All tests were two-sided, with α = 0.05. Because transferrin TSAT level showed a right-skewed distribution, the primary association between Hb and TSAT was evaluated via Pearson’s correlation using log-transformed TSAT; Spearman’s rank correlation on untransformed TSAT data was performed as a sensitivity analysis. Coefficients were reported as r for Pearson and ρ for Spearman. Data were analyzed using EZR software and GraphPad Prism 10. This study was reviewed and approved by the Ethics Review Board (approval number: ERB-C-2981; approval date: 10/31/2023) of our institution. As per our institutional policy, patients were informed at the initial visit that their anonymized clinical data may be used for future research purposes. In this study, we included data exclusively from patients who had provided written broad consent for such secondary use. This consent procedure was fully approved by the Ethics Review Board as the ethical basis for this research. This study was conducted in accordance with the principles of the Declaration of Helsinki.

ROC curve of the multivariable logistic regression model for TSAT-only classification (TSAT <20% with ferritin ≥20 ng/mL) among patients with ferritin measured at the first visit (n = 124). Predictors included diagnosis group (adenomyosis and myoma, with no uterine structural abnormality as the reference), age (years), and hypermenorrhea. The area under the ROC curve (AUC) was 0.678 (95% CI 0.563–0.793). (PDF)