Acute kidney injury (AKI) is defined by a rapid loss of kidney function, as evidenced by elevated blood creatinine and decreased urine output (1, 2). Severe AKI may lead to death, and even with proper fluid management and renal replacement therapy, a proportion of patients may still develop chronic kidney disease (CKD), resulting in decreased quality of life and a significant healthcare burden on society (1–3). As the blood creatinine and urine output indicate the loss of kidney filtration function, direct assessment of structural damage to the kidney is not available at the moment. Searching for kidney damage biomarkers is an endeavor of many researchers (4, 5).

Serum ferritin is often used clinically as an indicator of iron store, and it has been found to be elevated in inflammatory diseases (6, 7), malignancies (7, 8), coronary heart disease (9, 10), and other conditions (11). In clinical work, many researchers including us have noticed that serum ferritin is also elevated in AKI patients. However, there are arguments as to whether elevated serum ferritin is associated with the risk of adverse events (12–14). In this study, we collected data on critical ill patients with AKI to explore whether serum ferritin elevation is associated with adverse prognosis in patients with AKI.

Acute kidney injury (AKI) is the sudden loss of excretory function of the kidneys and has a high prevalence in critically ill patients. As a global disease, the causes of AKI vary in different regions. Infection and trauma are the most common causes of AKI. In less developed regions, infection, and hypovolemic shock are the main causes of AKI. In developed countries, AKI occurs mostly in hospitalized elderly patients and is associated with sepsis, nephrotoxic drugs, or invasive procedures. The diverse etiology of AKI implies diverse pathophysiological mechanisms. AKI that occurs in ICU mostly belongs to one of the multiorgan failure categories and is often accompanied by hepatic impairment, respiratory failure, heart failure, and shock, which carries a higher risk of death (1, 16).

As shown in Table 1, Our study population was included AKI patients admitted to the ICU, most of whom had multi-organ dysfunction and required RRT (17%) and respiratory ventilation support (88%). 70.7% of these patients met the diagnostic criteria for Sepsis 3 (Third International Consensus Definitions for Sepsis and Septic Shock), 34.2% of them had comorbid diabetes mellitus, and some of them had comorbid cardiovascular and other diseases, whereas previous studies have shown that elevated serum ferritin may occur in these disease states. Previous studies have found that ferritin is not only an indicator of iron stores, but is also abnormally elevated in inflammatory responses, tumors, and other conditions. Studies of Sepsis (23), acute cardiac infarction (9), cerebrovascular disease (24), and rheumatoid disorders have suggested a correlation between elevated ferritin and poor disease prognosis (25, 26). Therefore, our study does not imply that elevated serum ferritin is a specific manifestation of AKI, but rather the result of a combination of systemic diseases in critically ill patients.

Most critically ill patients are in a state of high depletion and negative nitrogen balance, with impaired gastrointestinal function and low early energy and nutrient intake (27–29), resulting in the fact that traditional markers of kidney injury, such as serum creatinine and blood urea nitrogen, may be insufficient to comprehensively assess critically ill AKI patients. As an indicator of iron stores and acute phase response, serum ferritin can be a good addition to the assessment of the condition of patients with severe AKI. We suggest that there may be several reasons for elevated SF in AKI patients: (1) Infections and other conditions lead to an inflammatory response, which promotes ferritin synthesis (12, 30). (2) Tissue and cellular destruction lead to the release of intracellular iron into the bloodstream, resulting in the formation of more ferritin (7). (3) Disturbances in the internal environment might lead to impaired iron utilization and lower ferritin consumption (14, 31). All three of these causes are a result of the seriousness of the disease, and therefore the more serious the disease the higher SF is likely to be, which is accompanied by a worse prognosis. This is consistent with our findings. Most basic medical research on AKI has focused on renal tubular injury. Ferroptosis is a type of programmed cell death driven by iron-dependent phospholipid peroxidation. Recent studies have shown that ferroptosis is elevated in tubular cell ferroptosis (32, 33). However, ferroptosis as an intracellular phenomenon in relation to circulating iron has not been correlated and deserves continued attention in this direction.

In our study, we found that AKI patients in the ICU with elevated SF had a higher length of hospitalization, length of ICU stay and higher risk of death (28-day mortality, 90-day mortality, hospital mortality and ICU mortality) than other patients, and this high risk was still significant after applying fully adjustment. Grouping by tertiles showed that SF was generally in the normal range in group Q1, highest in group Q3, and HR for each mortality event was significantly higher in group Q3 compared to group Q1, whereas in group Q2, the increase in HR was not significant. Curve fitting showed that the HR for the occurrence of each mortality event was increasing as the SF was higher, and the value of SF was positively and curvilinearly correlated with the risk of death.

The predictive value of serum ferritin for all-cause mortality in critically ill and septic patients has been previously reported (22, 23). Ferritin levels were higher in septic patients than in patients with other diagnoses in the intensive care unit, and were higher in patients who developed septic shock (22). Our subgroup analysis revealed that AKI patients in the Q3 group (SF∈ (680, 200,205) ng/ml) had an increased risk of death within 28 days regardless of the presence of sepsis (sepsis: HR = 1.285 95% CI (1.206–1.370); without sepsis: HR = 1.261,95% CI (1.075–1.479); p for interaction = 0.730). We subgrouped age, gender, and common comorbidities, the HR for 28-day mortality was significantly higher in the Q3 group in all subgroups. This suggests that high SF (>680 ng/mL) is an independent risk factor for the occurrence of death within 28 days in patients with AKI.

A few previous studies have suggested that high ferritin is associated with a promising prognosis. A small sample (N = 112) prospective cohort study among AKI patients at the Clinical Center Nis, Serbia suggested that serum ferritin levels on admission were higher in AKI recovery group compared with the non-recovery group (34). However, a prospective study evaluated 120 patients who underwent procedures at another tertiary referral center, serum ferritin levels did not differ in the groups with and without ARF in those patients (35).

OASIS (Oxford Acute Severity of Illness Score) (36), SAPS II (Simplified Acute Physiology Score ii) (37) and SOFA (Sequential Organ Failure Assessment Score) (38, 39) are common scores for disease assessment in ICU. APACHE ii is the most commonly used illness scoring system in ICUs today (40), but due to the wide variety of assessments, accurate data cannot be extracted from the MIMIC-IV database currently. These scoring systems are designed to focus on the patient’s vital signs, state of consciousness, and important organ function indicators, with higher scores suggesting that the patient’s condition is more severe and the likelihood of a worse prognosis (41). A prospective, multicenter, observational study included 2,954 critically ill patients with AKI in 30 intensive care units of 28 tertiary hospitals in Beijing, China, intending to evaluate the performance of four scoring systems in predicting 28-day mortality. The study showed that OASIS, APACHE II, and SAPS II had better predictive accuracy than SOFA, but due to the computational complexity of APACHE II and SAPS II, OASIS is a good alternative (42). In our study, as the value of SF increased, the OASIS, SAPSii and SOFA scores increased accordingly in both Q2 and Q3 groups, indicating that the severity of the disease was higher in these patients. The results suggest that when dealing with critically ill AKI patients, we should perform illness assessment scores such as OASIS, SAPS ii, etc., meanwhile we should also be concerned about hyperferritinaemia, which is strongly associated with a poorer prognosis, regardless of the presence of sepsis in this AKI patient.

In chronic kidney disease (CKD) patients, serum iron abnormalities were commonly occurred because of lack of erythropoietin (EPO), uremic toxins or dialysis-induced erythrocyte destruction, immune disorders and acid–base balance disturbances. These abnormalities may lead to an increased risk of death (24, 43). Many studies have attempted to find an appropriate ferritin threshold to distinguish between iron deficiency or iron overload, or to predict disease progression, but a single serum ferritin threshold cannot be appropriate for all diseases. The baseline level of SF may change if the patient has a comorbid condition such as liver disease, iron-deficiency anemia, or malignancy, but this is also a sign of a complex and serious medical condition that may lead to a worse prognosis. Therefore, it is necessary to develop corresponding thresholds for each specific condition. In this study, patients comorbid hematologic diseases, tumor, carcinoma, hemorrhage diseases, or undergoing maternity events was excluded, minimized effects of blood loss and consumptive diseases on ferritin. In our study, the Q3 group had a significantly higher HR in all death events in 90 days, which was not observed in the Q2 group, and the range of SF values in the Q3 group was >680 ng/mL. These results suggest that serum ferritin >680 ng/mL is associated with all-cause mortality in AKI patients.

Our study demonstrated serum ferritin elevation (SF > 680 ng/mL) was significantly associated with 28-day mortality in AKI patients, with or without sepsis, which may further facilitate the clinical application of serum ferritin as a biomarker in prognosis of AKI.

This is a retrospective cohort study based on MIMIC-IV 2.2, a single-center clinical database. Whether the same phenomenon exists in clinical medical centers in other countries and regions requires further study. Due to the lack of SF results, quite a number of AKI patients were excluded. Erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) are well recognized acute-phase response proteins, but they were not included in this study because they had more than 25% missing values in the study population. Cystatin C, a recognized biomarker of kidney injury, were not included in the MIMIC-IV database. Further studies need to incorporate more factors to validate this result.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found at: This is a retrospective cohort study. All data were obtained from the Medical Information Mart for Intensive Care (latest version MIMIC-IV 2.2) database (https://mimic.mit.edu/). MIMIC is a public database and contains the clinical information of critical ill patients from 2008 to 2019 who admitted in intensive care unit (ICU) of Beth Israel Deaconess Medical Center (Boston, Massachusetts). The author has passed the Protecting Human Research Participants exam and obtained the certification (certification number 38100618) for utilizing the MIMIC-IV database.

The studies were conducted in accordance with the local legislation and institutional requirements. The human samples used in this study were acquired from Establishment of the MIMIC-IV database was approved by the Massachusetts Institute of Technology (Cambridge, MA) and Beth Israel Deaconess Medical Center (Boston, MA), and consent was obtained for the original data collection. Therefore, the ethical approval statement and the need for informed consent were waived for the studies on this database. Written informed consent for participation was not required from the participants or the participants’ legal guardians/next of kin in accordance with the national legislation and institutional requirements.

XR: Data curation, Investigation, Resources, Writing – original draft. ZJ: Data curation, Formal analysis, Software, Writing – review & editing. FL: Methodology, Visualization, Writing – original draft. QW: Data curation, Validation, Visualization, Writing – original draft. HC: Methodology, Visualization, Writing – review & editing. LY: Investigation, Software, Writing – review & editing. CM: Visualization, Writing – review & editing. RW: Funding acquisition, Methodology, Resources, Supervision, Writing – review & editing.