Thyroid Stimulating Hormone Levels and Risk of Hospital-Acquired Acute Kidney Injury: A Nationwide Cohort Study Authors: Caoxiang She, Yaduan Lin, Yanan Xie, Shiyu Zhou, Fan Luo, Ruixuan Chen, Luhua Jin, Lantian Li, Linjiang Chen, Xian Shao, Sheng Nie Published: 2026 ABSTRACT Rationale & ObjectiveThe thyroid–renal interaction remains underexplored in acute kidney injury pathophysiology. We aimed to investigate the relationship between thyroid status, defined by thyroid-stimulating hormone (TSH) levels, and incident hospital-acquired acute kidney injury (HA-AKI).Study DesignA retrospective cohort study.Setting & ParticipantsPatients with TSH quantification within 48 hours after admission from the China Renal Data System (CRDS) database.Exposure(s)Three TSH levels groups, including hyperthyroid (TSH ≤ 0.5 mIU/L), euthyroid (0.5 < TSH ≤ 5.0 mIU/L), and hypothyroid (TSH > 5.0 mIU/L) TSH levels.Outcome(s)The primary outcome was HA-AKI.Analytical ApproachMultivariable Cox hazard regression models and restricted cubic spline.ResultsA total of 20,317 (5.4%) patients developed HA-AKI, with incidences rates of 8.0%, 5.0%, and 6.4% in patients with hyperthyroid, euthyroid, and hypothyroid TSH levels, respectively. Compared with patients with euthyroid TSH levels, those with hyperthyroid or hypothyroid TSH levels were associated with an increased risk of HA-AKI (hyperthyroid TSH levels: adjusted hazard ratio [aHR], 1.28; 95% confidence interval [CI], 1.21-1.37; P < 0.001; hypothyroid TSH levels: aHR, 1.10; 95% CI, 1.01-1.21; P = 0.03). Additionally, the dose-response curve showed a U-shaped relationship between TSH levels and the risk of new-onset HA-AKI.LimitationsPossible residual confounding.ConclusionsTSH levels in the low (≤0.5 mIU/L) and high (>5.0 mIU/L) ranges were associated with increased risk of HA-AKI. These insights suggest that enhanced vigilance toward AKI may be justifiable in patients with thyroid dysfunction. Acute kidney injury (AKI), characterized by a rapid decline of kidney function, occurs in approximately 23% of patients admitted to hospitals, contributing to poor in-hospital outcomes and the potential development of chronic kidney disease (CKD) after discharge.1 Although significant advances in management have reduced in-hospital mortality of severe AKI patients, the prognosis remains discouraging.2,3 Given the challenges in detecting early stages of AKI and the lack of specific pharmacologic treatments up to now, it is crucial to identify risk factors associated with the episode of AKI.4,5 Accumulating evidence suggests an intricate interplay between kidney function and thyroid function, with alterations in the hypothalamic–pituitary–thyroid axis and thyroid metabolism potentially serving as mediators.6,7 The relationship between thyroid dysfunction and CKD has been extensively studied. These studies demonstrated that thyroid dysfunction was associated with the development or progression of CKD in various populations; however, some conflicting findings existing.6,8, 9, 10, 11, 12, 13 Compared with CKD, studies examining the association between thyroid dysfunction and AKI remains sparse and inconclusive. Existing studies have primarily focused on hypothyroidism in highly specific surgical or interventional settings, in which a low T3 level has been identified as an independent risk factor for postoperative/interventional-induced AKI.14, 15, 16, 17 However, the thyroid-AKI interplay in general populations is poorly characterized, particularly the role of hyperthyroidism. Critically, there is a compelling biological rationale to suspect that thyroid dysfunction may influence AKI risk. Thyroid-stimulating hormone (TSH)-mediated hormonal signaling exerts pansystemic effects, and thyroid hormones directly modulate kidney hemodynamics and tubular function.7,18, 19, 20 Although hypothyroidism may predispose to AKI via mechanisms like rhabdomyolysis, hyperthyroidism can contribute to AKI through high-output cardiac failure.20, 21, 22 As the central regulator of thyroid function, TSH may serve as a sensitive integrator of this thyroid–renal axis.23 Therefore, we conducted a nationwide cohort study to investigate the association between TSH levels and the risk of hospital-acquired acute kidney injury (HA-AKI), thereby providing evidence-based insights for identifying high-risk populations and implementing targeted prevention strategies. Materials and Methods Study Design The study population was drawn from the China Renal Data System (CRDS) database, a multicenter retrospective cohort comprising more than 8 million patients from 24 academic hospitals in China, from January 1, 2000, to December 31, 2022.24 The CRDS is a joint initiative of the National Clinical Research Center and the China Center for Disease Control and Prevention that contains comprehensive information on patients including demographic characteristics, vital signs, diagnoses, laboratory measurements, prescription, and medical records. This database has been used to conduct many research studies.25,26 The study protocol was approved by the Medical Ethics Committee of Nanfang Hospital, Southern Medical University (NFEC-2019-213) and the China Office of Human Genetic Resources for Data Preservation Application (2021-BC0037). Study Population We selected 612,530 hospitalized adult patients from the CRDS database. These patients had at least one serum TSH measurement within 48 hours after admission for evaluating thyroid status, and at least 2 serum creatinine (SCr) tests within 7 days during the hospitalization to ensure adequate records for detecting AKI. Figure 1 and Figure S1 showed the study design and time frame definitions. Serum TSH levels were measured using standardized chemiluminescent immunoassay (CLIA) platforms following manufacturer protocols. For patients with multiple TSH tests, the baseline serum TSH levels was recorded as the earliest valid measurement obtained within the first 48 hours of admission. The date of the baseline serum TSH level was the index date. For patients with multiple qualifying hospitalizations, only the records from the first hospitalization were analyzed.Figure 1Flowchart of the study population selection. Abbreviations: TSH, thyroid stimulating hormone; CRDS, China Renal Data System; THs, thyroid Hormones; ATDs, antithyroid drugs; CA-AKI, community-acquired acute kidney injury; eGFR, estimated glomerular filtration rate; Scr, serum creatinine. Flowchart of the study population selection. Abbreviations: TSH, thyroid stimulating hormone; CRDS, China Renal Data System; THs, thyroid Hormones; ATDs, antithyroid drugs; CA-AKI, community-acquired acute kidney injury; eGFR, estimated glomerular filtration rate; Scr, serum creatinine. We excluded patients with any of the following conditions: (1) received thyroid hormone supplementation, antithyroid medication, and iodine medication, as these pharmaceutical agents may affect TSH levels; (2) diagnosed with community-acquired acute kidney injury (CA-AKI), which was defined as AKI events that occurred within the first 48 hours of hospitalization; (3) received kidney replacement therapy, including maintenance dialysis or kidney transplantation; (4) pregnant women; (5) missing estimated glomerular filtration rate (eGFR) at baseline; (6) patients with kidney failure or baseline eGFR < 30 mL/min/1.73 m2; (7) length of stay > 30 days; (8) the last SCr testing within 48 hours after admission; or (9) died within 48 hours after admission. Exposure and Outcome The exposure of interest was baseline serum TSH levels. TSH levels were categorized into 3 groups according to the TSH reference range used in the general population: hyperthyroid TSH levels (TSH ≤ 0.5 mIU/L), euthyroid TSH levels (0.5 < TSH ≤ 5.0 mIU/L), and hypothyroid TSH levels (TSH > 5.0 mIU/L).8,27,28 We used euthyroid TSH levels as the reference category. The primary outcome was incident HA-AKI, which was characterized by an increase in SCr of 26.5 μmol/L within 48 hours or a 50% increase in SCr within 7 days according to the Kidney Disease Improving Global Outcomes (KDIGO) criteria.29 Urine output criteria were not used because urine volume was not available in this study. The baseline SCr was recorded as the nearest measurement 7 days before the index date. The earliest day on which the change in SCr met the Kidney Disease: Improving Global Outcomes (KDIGO) criteria was established as the date of HA-AKI onset. Patients were considered censored at the earliest occurrence of any of the following events: (1) the last SCr test before discharge; (2) hospital mortality; or (3) HA-AKI. The secondary outcome was severe HA-AKI, which was defined as the composite of HA-AKI stage 2 or 3 (a 2.0-fold increase in SCr within 7 days or a peak SCr of ≥353.6 μmol/L within 48 hours) and HA-AKI requiring dialysis (the initiation for dialysis from the onset of AKI until discharge). Assessment of Covariates The study covariates included age, sex, length of stay, frequency of SCr measurement, body mass index (BMI), systolic blood pressure (SBP), diastolic blood pressure (DBP), comorbid conditions, laboratory values, concomitant medications, and operation procedures. The frequency of SCr measurement during the 180 days preceding the index date, which may affect AKI detection sensitivity, was documented for subsequent model adjustment. The comorbid conditions were defined by dichotomous covariates according to International Classification of Diseases, Tenth Revision (ICD-10) codes before the index date, including CKD, diabetes, hypertension, coronary heart disease (CHD), heart failure, sepsis, chronic obstructive pulmonary disease (COPD), moderate or severe liver disease, and previous AKI history. The Charlson Comorbidity Index (CCI) was calculated to assess baseline comorbidity. Laboratory tests were defined as the most recent measurement within 7 days before the index date and included TSH, SCr, eGFR, triiodothyronine (T3), thyroxine (T4), free triiodothyronine (fT3), free thyroxine (fT4), alanine aminotransferase (ALT), aspartate aminotransferase (AST), hemoglobin (Hb), white blood cell (WBC), platelets (Plt), low-density lipoprotein cholesterol, and total cholesterol. The categories of comedications analysis included statins, diabetes agents, renin–angiotensin system inhibitors (RASi), nonsteroidal anti-inflammatory drugs (NSAIDs), chemotherapy agents, contrast media, glucocorticoids, diuretics, aminoglycosides, other types of antibiotics, antiviral agents, antifungal agents, and proton pump inhibitors (PPI), which were determined based on any prescriptions from 7 days before admission to 48 hours of admission. Intensive care unit (ICU) admission was also included. The presence of procedures during hospitalization was identified using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD9-CM) codes. eGFR was calculated using the ethnicity-independent Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation. Statistical Analysis Categorical variables were described as frequencies with percentages (%), and continuous variables were presented as medians with interquartile ranges (IQR). Comparisons among different TSH level groups in the baseline characteristics were performed using analysis of variance (ANOVA) or Kruskal–Wallis tests for continuous variables and chi-squared test for categorical variables. Kaplan–Meier plots were generated to visualize the cumulative incidence of HA-AKI in each group, compared with the Log-rank test. Cox proportional hazards models were used to estimate hazard ratio (HR) with 95% confidence interval (CI). The multivariable Cox models were adjusted for prespecified covariates, including age, sex, frequency of SCr measurement, BMI, comorbid conditions (CKD, diabetes, hypertension, heart failure, CHD, sepsis, COPD, moderate or severe liver disease, and previous AKI history), CCI, laboratory tests (eGFR, Hb, WBC, ALT, and AST), concomitant use of drugs (RASi, NSAIDs, chemotherapy agents, contrast media, glucocorticoids, diuretics, aminoglycosides, other types of antibiotics, antiviral agents, antifungal agents, and PPI), ICU admission, mechanical ventilation, and operations. Under the assumption of missing at random, the multiple imputation method was used to impute the missing values. Five imputed datasets were generated, and the results of multivariable regression analyses in these datasets were pooled. The results of multivariable regression analyses from these datasets were pooled according to Rubin’s criterion. Additionally, we conducted restricted cubic spline (RCS) Cox regression to examine TSH levels as a continuous predictor of incident HA-AKI. All analyses were conducted using R software, version 4.1.2, with statistical significance defined as two-sided P value of < 0.05. Subgroup Analyses The possibility of effect modification of the exposure–outcome association was examined by introducing interactive terms into the regression model. Subgroup analyses were performed, with patients stratified by age, sex, CKD, diabetes, hypertension, heart failure, and CHD. Sensitivity Analyses We conducted sensitivity analyses to examine the robustness of the study findings. First, TSH levels were categorized based on more granular categorizations: overt hyperthyroid (≤0.1 mIU/L), subclinical hyperthyroid (0.1-0.5 mIU/L), low-normal (0.5-3.0 mIU/L), high-normal (3.0-5.0 mIU/L), subclinical hypothyroid (5.0-10.0 mIU/L), and overt hypothyroid TSH levels (>10.0 mIU/L). Additionally, we analyzed tertiles of TSH values, including tertile 1 (≤1.10 mIU/L), tertile 2 (1.10-2.14 mIU/L), and tertile 3 (>2.14 mIU/L), to explore the effect of different thresholds on the results. Second, to further analyze the relationship between TSH levels and the study outcomes, patients within euthyroid TSH ranges were identified and categorized into tertiles: low-normal (0.50-1.23 mIU/L), medium-normal (1.23-2.11 mIU/L), and high-normal TSH levels (2.11-5.00 mIU/L). Third, we excluded patients with a previous history of AKI to repeat the HA-AKI outcome analysis. Fourth, we excluded patients who developed HA-AKI within 4 days to reduce the potential impact of reversed causality. Last, we conducted a 1:1 propensity score matching (PSM) analysis using optimal matching to balance baseline characteristics and repeated the analysis in the matched cohorts. Results Baseline Characteristics A total of 373,204 eligible patients were identified (Fig 1). The characteristics of these patients, stratified by TSH levels, are summarized in Table 1. The median (interquartile range [IQR]) age of the patients was 60.8 (48.7, 71.5) years, with 54.5% being men. No significant differences were observed in the length of stay and frequency of SCr measurement among the 3 groups. Compared with group of euthyroid TSH levels, patients with hyperthyroid TSH levels had a higher prevalence of heart failure, sepsis, ICU admission, and mechanical ventilation. Additionally, they had a higher proportion of diabetes agents, NSAIDs, glucocorticoids, antibiotics, antifungal agents, antiviral agents, and PPI usage. In contrast, patients with hypothyroid TSH levels tended to be older, were more likely to be women, and had a higher prevalence of CKD and heart failure. They also had lower eGFR levels and had higher proportion of glucocorticoid and diuretic usage.Table 1Comparison of Baseline Clinical Characteristics Stratified by TSH LevelsCharacteristicsOverallTSH≤0.50.55.0(N = 373,204)(N = 41,373)(N = 306,688)(N = 25,143)Age (y)60.8 [48.7, 71.5]60.6 [48.5, 71.8]60.7 [48.5, 71.3]62.8 [50.5, 73.2]Sex, male203,465 (54.5)23,862 (57.7)168,703 (55.0)10,900 (43.4)BMI21.9 [20.9, 24.7]21.8 [20.8, 24.1]21.9 [21.0, 24.8]21.8 [20.8, 24.4]SBP (mm Hg)126.0 [114.0, 141.0]126.0 [113.0, 142.0]126.0 [114.0, 141.0]125.0 [113.0, 140.0]DBP (mm Hg)77.0 [69.0, 85.0]76.0 [68.0, 85.0]77.0 [69.0, 85.0]78.0 [70.0, 85.0]Length of stay (d)10.0 [7.0, 14.8]10.1 [7.0, 15.0]10.0 [7.0, 14.1]10.5 [7.1, 15.0]Frequency of SCr measurement3.0 [2.0, 4.0]3.0 [2.0, 5.0]3.0 [2.0, 4.0]3.0 [2.0, 4.0]Comorbid conditions, N (%) CKD24,666 (6.6)1,644 (4.0)20,187 (6.6)2,835 (11.3) Diabetes54,437 (14.6)5,726 (13.8)45,574 (14.9)3,137 (12.5) Hypertension98,473 (26.4)10,304 (24.9)82,299 (26.8)5,870 (23.3) CHD83,947 (22.5)8,501 (20.5)70,485 (23.0)4,961 (19.7) Heart failure42,401 (11.4)5,074 (12.3)34,286 (11.2)3,041 (12.1) Sepsis2,135 (0.6)447 (1.1)1,587 (0.5)101 (0.4) COPD18,028 (4.8)2,780 (6.7)14,195 (4.6)1,053 (4.2) Moderate or severe liver disease5,587 (1.5)948 (2.3)4,311 (1.4)328 (1.3) Previous AKI history3,011 (0.8)346 (0.8)2,402 (0.8)263 (1.0) CCI4.0 [2.0, 5.0]4.0 [2.0, 5.0]4.0 [2.0, 5.0]4.0 [2.0, 5.0]Laboratory TSH (mIU/L)1.6 [0.9, 2.6]0.3 [0.2, 0.4]1.6 [1.1, 2.4]6.8 [5.7, 9.1] SCr, μmol/L70.0 [57.0, 87.0]68.1 [55.0, 85.6]70.0 [57.0, 87.0]72.0 [57.2, 92.0] eGFR (30 mL/min/1.73 m2)91.6 [73.4, 105.0]93.3 [75.6, 107.1]91.7 [73.7, 104.9]86.7 [65.5, 101.6] T3 (ng/mL)0.9 [0.7, 1.0]0.8 [0.6, 1.0]0.9 [0.7, 1.0]0.9 [0.7, 1.0] T4 (ug/dL)6.7 [5.6, 8.1]7.0 [5.6, 8.8]6.7 [5.7, 8.1]6.2 [5.2, 7.5] fT3 (pmol/L)3.8 [2.7, 4.5]3.5 [2.5, 4.4]3.8 [2.8, 4.5]3.8 [2.9, 4.4] fT4 (pmol/L)12.0 [1.3, 14.7]12.6 [8.9, 15.6]12.0 [1.3, 14.6]11.1 [1.2, 13.7] Hb (g/L)130.0 [115.0, 143.0]129.0 [113.0, 142.0]130.0 [116.0, 143.0]127.0 [111.0, 141.0] WBC (10ˆ9/L)6.8 [5.3, 8.9]8.4 [6.1, 11.7]6.7 [5.3, 8.7]6.3 [5.0, 8.0] Plt (10ˆ9/L)203.0 [155.0, 256.0]203.0 [153.0, 257.0]203.0 [156.0, 257.0]194.0 [143.0, 252.0] ALT (IU/L)20.3 [14.0, 35.0]22.4 [14.0, 42.0]20.0 [13.7, 34.0]20.0 [13.0, 33.0] AST (IU/L)23.1 [18.0, 36.0]25.0 [18.0, 46.0]23.0 [18.0, 35.0]25.0 [19.0, 36.3] LDL (mmol/L)2.5 [1.9, 3.2]2.4 [1.8, 3.1]2.5 [1.9, 3.1]2.5 [1.9, 3.3] Tcho (mmol/L)4.3 [3.5, 5.1]4.2 [3.4, 5.0]4.3 [3.5, 5.1]4.3 [3.5, 5.3]Medication, N (%) Statins125,876 (33.7)12,606 (30.5)105,528 (34.4)7,742 (30.8) Diabetes agents67,496 (18.1)8,838 (21.4)54,833 (17.9)3,825 (15.2) RASi83,581 (22.4)7,799 (18.9)70,125 (22.9)5,657 (22.5) NSAIDs108,866 (29.2)13,460 (32.5)89,235 (29.1)6,171 (24.5) Chemotherapy agents12,319 (3.3)1,426 (3.4)10,048 (3.3)845 (3.4) Contrast media32,686 (8.8)2,879 (7.0)27,999 (9.1)1,808 (7.2) Glucocorticoids39,405 (10.6)8,158 (19.7)28,680 (9.4)2,567 (10.2) Diuretics87,659 (23.5)10,339 (25.0)69,339 (22.6)7,981 (31.7) Aminoglycosides2,754 (0.7)528 (1.3)2,070 (0.7)156 (0.6) Other types of antibiotics18,267 (4.9)3,927 (9.5)13,393 (4.4)947 (3.8) Antifungal agents2,024 (0.5)526 (1.3)1,401 (0.5)97 (0.4) Antiviral agents12,666 (3.4)2,374 (5.7)9,758 (3.2)534 (2.1) PPI150,347 (40.3)22,428 (54.2)11,9223 (38.9)8,696 (34.6)ICU admission, N (%)30,386 (8.1)5,607 (13.6)23,304 (7.6)1,475 (5.9)Mechanical ventilation, N (%)1,904 (0.5)429 (1.0)1,387 (0.5)88 (0.3)Operations, N (%)131,734 (35.3)13,781 (33.3)11,0176 (35.9)7,777 (30.9)Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CCB, calcium channel blocker; CCI, Charlson Comorbidity Index; CHD, coronary heart disease; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; fT3, free triiodothyronine; fT4, free thyroxine; Hb, hemoglobin; ICU, intensive care unit; LDL, low-density lipoprotein cholesterol; NSAIDs, nonsteroidal anti-inflammatory drugs; Plt, platelets; PPI, proton pump inhibitors; RASi, renin-angiotensin system inhibitors; SBP, systolic blood pressure; SCr, serum creatinine; T3, triiodothyronine; T4, thyroxine; Tcho, total cholesterol; TSH, thyroid-stimulating hormone; WBC, white blood cell. Table 1Comparison of Baseline Clinical Characteristics Stratified by TSH LevelsCharacteristicsOverallTSH≤0.50.55.0(N = 373,204)(N = 41,373)(N = 306,688)(N = 25,143)Age (y)60.8 [48.7, 71.5]60.6 [48.5, 71.8]60.7 [48.5, 71.3]62.8 [50.5, 73.2]Sex, male203,465 (54.5)23,862 (57.7)168,703 (55.0)10,900 (43.4)BMI21.9 [20.9, 24.7]21.8 [20.8, 24.1]21.9 [21.0, 24.8]21.8 [20.8, 24.4]SBP (mm Hg)126.0 [114.0, 141.0]126.0 [113.0, 142.0]126.0 [114.0, 141.0]125.0 [113.0, 140.0]DBP (mm Hg)77.0 [69.0, 85.0]76.0 [68.0, 85.0]77.0 [69.0, 85.0]78.0 [70.0, 85.0]Length of stay (d)10.0 [7.0, 14.8]10.1 [7.0, 15.0]10.0 [7.0, 14.1]10.5 [7.1, 15.0]Frequency of SCr measurement3.0 [2.0, 4.0]3.0 [2.0, 5.0]3.0 [2.0, 4.0]3.0 [2.0, 4.0]Comorbid conditions, N (%) CKD24,666 (6.6)1,644 (4.0)20,187 (6.6)2,835 (11.3) Diabetes54,437 (14.6)5,726 (13.8)45,574 (14.9)3,137 (12.5) Hypertension98,473 (26.4)10,304 (24.9)82,299 (26.8)5,870 (23.3) CHD83,947 (22.5)8,501 (20.5)70,485 (23.0)4,961 (19.7) Heart failure42,401 (11.4)5,074 (12.3)34,286 (11.2)3,041 (12.1) Sepsis2,135 (0.6)447 (1.1)1,587 (0.5)101 (0.4) COPD18,028 (4.8)2,780 (6.7)14,195 (4.6)1,053 (4.2) Moderate or severe liver disease5,587 (1.5)948 (2.3)4,311 (1.4)328 (1.3) Previous AKI history3,011 (0.8)346 (0.8)2,402 (0.8)263 (1.0) CCI4.0 [2.0, 5.0]4.0 [2.0, 5.0]4.0 [2.0, 5.0]4.0 [2.0, 5.0]Laboratory TSH (mIU/L)1.6 [0.9, 2.6]0.3 [0.2, 0.4]1.6 [1.1, 2.4]6.8 [5.7, 9.1] SCr, μmol/L70.0 [57.0, 87.0]68.1 [55.0, 85.6]70.0 [57.0, 87.0]72.0 [57.2, 92.0] eGFR (30 mL/min/1.73 m2)91.6 [73.4, 105.0]93.3 [75.6, 107.1]91.7 [73.7, 104.9]86.7 [65.5, 101.6] T3 (ng/mL)0.9 [0.7, 1.0]0.8 [0.6, 1.0]0.9 [0.7, 1.0]0.9 [0.7, 1.0] T4 (ug/dL)6.7 [5.6, 8.1]7.0 [5.6, 8.8]6.7 [5.7, 8.1]6.2 [5.2, 7.5] fT3 (pmol/L)3.8 [2.7, 4.5]3.5 [2.5, 4.4]3.8 [2.8, 4.5]3.8 [2.9, 4.4] fT4 (pmol/L)12.0 [1.3, 14.7]12.6 [8.9, 15.6]12.0 [1.3, 14.6]11.1 [1.2, 13.7] Hb (g/L)130.0 [115.0, 143.0]129.0 [113.0, 142.0]130.0 [116.0, 143.0]127.0 [111.0, 141.0] WBC (10ˆ9/L)6.8 [5.3, 8.9]8.4 [6.1, 11.7]6.7 [5.3, 8.7]6.3 [5.0, 8.0] Plt (10ˆ9/L)203.0 [155.0, 256.0]203.0 [153.0, 257.0]203.0 [156.0, 257.0]194.0 [143.0, 252.0] ALT (IU/L)20.3 [14.0, 35.0]22.4 [14.0, 42.0]20.0 [13.7, 34.0]20.0 [13.0, 33.0] AST (IU/L)23.1 [18.0, 36.0]25.0 [18.0, 46.0]23.0 [18.0, 35.0]25.0 [19.0, 36.3] LDL (mmol/L)2.5 [1.9, 3.2]2.4 [1.8, 3.1]2.5 [1.9, 3.1]2.5 [1.9, 3.3] Tcho (mmol/L)4.3 [3.5, 5.1]4.2 [3.4, 5.0]4.3 [3.5, 5.1]4.3 [3.5, 5.3]Medication, N (%) Statins125,876 (33.7)12,606 (30.5)105,528 (34.4)7,742 (30.8) Diabetes agents67,496 (18.1)8,838 (21.4)54,833 (17.9)3,825 (15.2) RASi83,581 (22.4)7,799 (18.9)70,125 (22.9)5,657 (22.5) NSAIDs108,866 (29.2)13,460 (32.5)89,235 (29.1)6,171 (24.5) Chemotherapy agents12,319 (3.3)1,426 (3.4)10,048 (3.3)845 (3.4) Contrast media32,686 (8.8)2,879 (7.0)27,999 (9.1)1,808 (7.2) Glucocorticoids39,405 (10.6)8,158 (19.7)28,680 (9.4)2,567 (10.2) Diuretics87,659 (23.5)10,339 (25.0)69,339 (22.6)7,981 (31.7) Aminoglycosides2,754 (0.7)528 (1.3)2,070 (0.7)156 (0.6) Other types of antibiotics18,267 (4.9)3,927 (9.5)13,393 (4.4)947 (3.8) Antifungal agents2,024 (0.5)526 (1.3)1,401 (0.5)97 (0.4) Antiviral agents12,666 (3.4)2,374 (5.7)9,758 (3.2)534 (2.1) PPI150,347 (40.3)22,428 (54.2)11,9223 (38.9)8,696 (34.6)ICU admission, N (%)30,386 (8.1)5,607 (13.6)23,304 (7.6)1,475 (5.9)Mechanical ventilation, N (%)1,904 (0.5)429 (1.0)1,387 (0.5)88 (0.3)Operations, N (%)131,734 (35.3)13,781 (33.3)11,0176 (35.9)7,777 (30.9)Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CCB, calcium channel blocker; CCI, Charlson Comorbidity Index; CHD, coronary heart disease; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; fT3, free triiodothyronine; fT4, free thyroxine; Hb, hemoglobin; ICU, intensive care unit; LDL, low-density lipoprotein cholesterol; NSAIDs, nonsteroidal anti-inflammatory drugs; Plt, platelets; PPI, proton pump inhibitors; RASi, renin-angiotensin system inhibitors; SBP, systolic blood pressure; SCr, serum creatinine; T3, triiodothyronine; T4, thyroxine; Tcho, total cholesterol; TSH, thyroid-stimulating hormone; WBC, white blood cell. Comparison of Baseline Clinical Characteristics Stratified by TSH Levels Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CCB, calcium channel blocker; CCI, Charlson Comorbidity Index; CHD, coronary heart disease; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; fT3, free triiodothyronine; fT4, free thyroxine; Hb, hemoglobin; ICU, intensive care unit; LDL, low-density lipoprotein cholesterol; NSAIDs, nonsteroidal anti-inflammatory drugs; Plt, platelets; PPI, proton pump inhibitors; RASi, renin-angiotensin system inhibitors; SBP, systolic blood pressure; SCr, serum creatinine; T3, triiodothyronine; T4, thyroxine; Tcho, total cholesterol; TSH, thyroid-stimulating hormone; WBC, white blood cell. TSH Levels and Risk of HA-AKI During the hospital stay, the incidences of HA-AKI in patients with hyperthyroid TSH levels (8.0%) and hypothyroid TSH levels (6.4%) were significantly higher than those with euthyroid TSH levels (5.0%; Table 2). The Kaplan–Meier curves indicated that higher cumulative incidence rates of HA-AKI were observed in both hyperthyroid and hypothyroid TSH levels (Fig 2). Additionally, the RCS curves showed a U-shaped relationship between TSH levels and risk of HA-AKI (Fig 3). Compared with patients with euthyroid TSH levels, those with hyperthyroid or hypothyroid TSH levels were significantly associated with higher risk of HA-AKI (hyperthyroid TSH levels: adjusted hazard ratio [aHR], 1.28, 95% CI, 1.21-1.37, P < 0.001; hypothyroid TSH levels: aHR, 1.10, 95% CI, 1.01-1.21, P = 0.03; Table 2). Similar trend of effect was observed for severe HA-AKI outcome (hyperthyroid TSH levels: aHR, 1.48, 95% CI, 1.32-1.67, P < 0.001; hypothyroid TSH levels: aHR, 1.24, 95% CI, 1.03-1.47, P = 0.02; Table 2).Table 2The Association of TSH Levels With Incident HA-AKI Among Hospitalized Patients Stratified by Three GroupsTSH levels (mIU/L)TotalEvent (%)cHRa (95% CI)PaHRa (95% CI)PPrimary outcome Incident HA-AKI Hyperthyroid TSH levels (TSH≤0.5)41,3733,317 (8.0)1.50 (1.45-1.56)<0.0011.28 (1.21-1.37)<0.001 Euthyroid TSH levels (0.55.0)25,1431,613 (6.4)1.21 (1.15-1.28)<0.0011.10 (1.01-1.21)0.03Secondary outcome Incident Severe HA-AKIb Hyperthyroid TSH levels (TSH≤0.5)41,3731103 (2.7)1.91 (1.79-2.05)<0.0011.48 (1.32-1.67)<0.001 Euthyroid TSH levels (0.55.0)25,143438 (1.7)1.29 (1.17-1.43)<0.0011.24 (1.03-1.47)0.02Abbreviations: aHR, adjusted hazard ratio; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CCI, Charlson Comorbidity Index; CHD, coronary heart disease; CI, confidence interval; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; cHR, crude hazard ratio; eGFR, estimated glomerular filtration rate; HA-AKI, hospital-acquired acute kidney injury; Hb, hemoglobin; ICU, intensive care unit; NSAIDs, nonsteroidal anti-inflammatory drugs; PPI, proton pump inhibitors; RASi, renin–angiotensin system inhibitors; Ref, reference; SCr, serum creatinine; TSH, thyroid-stimulating hormone; WBC, white blood cell.aHazard ratios were adjusted for age, sex, frequency of SCr measurement, BMI, CKD, diabetes, hypertension, heart failure, CHD, sepsis, COPD, moderate or severe liver disease, previous AKI history, CCI, eGFR, Hb, WBC, ALT, AST, RASi, NSAIDs, chemotherapy agents, contrast media, glucocorticoids, diuretics, aminoglycosides, other types of antibiotics, antiviral agents, antifungal agents, PPI, ICU admission, mechanical ventilation, and operations.bSevere HA-AKI included the composite of HA-AKI stage 2/3 and HA-AKI requiring dialysis.Figure 2Cumulative incidence of HA-AKI among hospitalized patients in three TSH groups. Abbreviations: TSH, thyroid stimulating hormone; HA-AKI, hospital-acquired acute kidney injury.Figure 3Restricted cubic spline analysis of continuous TSH levels with risk of HA-AKI. Abbreviations: TSH, thyroid stimulating hormone; HA-AKI, hospital-acquired acute kidney injury. Table 2The Association of TSH Levels With Incident HA-AKI Among Hospitalized Patients Stratified by Three GroupsTSH levels (mIU/L)TotalEvent (%)cHRa (95% CI)PaHRa (95% CI)PPrimary outcome Incident HA-AKI Hyperthyroid TSH levels (TSH≤0.5)41,3733,317 (8.0)1.50 (1.45-1.56)<0.0011.28 (1.21-1.37)<0.001 Euthyroid TSH levels (0.55.0)25,1431,613 (6.4)1.21 (1.15-1.28)<0.0011.10 (1.01-1.21)0.03Secondary outcome Incident Severe HA-AKIb Hyperthyroid TSH levels (TSH≤0.5)41,3731103 (2.7)1.91 (1.79-2.05)<0.0011.48 (1.32-1.67)<0.001 Euthyroid TSH levels (0.55.0)25,143438 (1.7)1.29 (1.17-1.43)<0.0011.24 (1.03-1.47)0.02Abbreviations: aHR, adjusted hazard ratio; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CCI, Charlson Comorbidity Index; CHD, coronary heart disease; CI, confidence interval; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; cHR, crude hazard ratio; eGFR, estimated glomerular filtration rate; HA-AKI, hospital-acquired acute kidney injury; Hb, hemoglobin; ICU, intensive care unit; NSAIDs, nonsteroidal anti-inflammatory drugs; PPI, proton pump inhibitors; RASi, renin–angiotensin system inhibitors; Ref, reference; SCr, serum creatinine; TSH, thyroid-stimulating hormone; WBC, white blood cell.aHazard ratios were adjusted for age, sex, frequency of SCr measurement, BMI, CKD, diabetes, hypertension, heart failure, CHD, sepsis, COPD, moderate or severe liver disease, previous AKI history, CCI, eGFR, Hb, WBC, ALT, AST, RASi, NSAIDs, chemotherapy agents, contrast media, glucocorticoids, diuretics, aminoglycosides, other types of antibiotics, antiviral agents, antifungal agents, PPI, ICU admission, mechanical ventilation, and operations.bSevere HA-AKI included the composite of HA-AKI stage 2/3 and HA-AKI requiring dialysis. The Association of TSH Levels With Incident HA-AKI Among Hospitalized Patients Stratified by Three Groups Abbreviations: aHR, adjusted hazard ratio; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CCI, Charlson Comorbidity Index; CHD, coronary heart disease; CI, confidence interval; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; cHR, crude hazard ratio; eGFR, estimated glomerular filtration rate; HA-AKI, hospital-acquired acute kidney injury; Hb, hemoglobin; ICU, intensive care unit; NSAIDs, nonsteroidal anti-inflammatory drugs; PPI, proton pump inhibitors; RASi, renin–angiotensin system inhibitors; Ref, reference; SCr, serum creatinine; TSH, thyroid-stimulating hormone; WBC, white blood cell. Hazard ratios were adjusted for age, sex, frequency of SCr measurement, BMI, CKD, diabetes, hypertension, heart failure, CHD, sepsis, COPD, moderate or severe liver disease, previous AKI history, CCI, eGFR, Hb, WBC, ALT, AST, RASi, NSAIDs, chemotherapy agents, contrast media, glucocorticoids, diuretics, aminoglycosides, other types of antibiotics, antiviral agents, antifungal agents, PPI, ICU admission, mechanical ventilation, and operations. Severe HA-AKI included the composite of HA-AKI stage 2/3 and HA-AKI requiring dialysis. Cumulative incidence of HA-AKI among hospitalized patients in three TSH groups. Abbreviations: TSH, thyroid stimulating hormone; HA-AKI, hospital-acquired acute kidney injury. Restricted cubic spline analysis of continuous TSH levels with risk of HA-AKI. Abbreviations: TSH, thyroid stimulating hormone; HA-AKI, hospital-acquired acute kidney injury. Subgroup Analyses and Sensitivity Analyses The associations remained consistent across various subgroups (Fig S2). We found that age, sex, heart failure, and CHD were modified factor in this association (all P for interaction < 0.05). The findings were similar to the primary results after applying novel threshold values based on more detailed TSH categorizations or tertiles of TSH levels (Table S1). Among patients within euthyroid TSH ranges, both higher and lower TSH levels were associated with an increased risk of HA-AKI (Table S2). The results did not materially change when excluding patients with a previous history of AKI as well as those who experienced outcome within 4 days (Tables S3 and S4). Results from the PSM dataset also mirrored those of the primary outcome analysis (Table S5). Discussion To our knowledge, this is the largest study to analyze the association between TSH levels and HA-AKI. Our findings indicated that TSH in both the low (≤0.5 mIU/L) and high (>5.0 mIU/L) levels were independently associated with an increased risk of incident HA-AKI, compared with normal TSH levels. This association was consistent across subgroup and sensitivity analyses. The dose-response curve showed a U-shaped relationship between TSH concentrations and HA-AKI susceptibility. This study suggests that patients with thyroid dysfunction may be at a higher risk of developing AKI, highlighting the need for close monitoring of kidney function in these patients. Recently, a multivariable mendelian randomization analysis has found that elevated TSH levels are causally associated with reduced kidney function, highlighting the potential importance of thyroid function in kidney health.30 However, the pivotal role of thyroid hormonal disorder in kidney pathophysiology remains a frequently neglected field of clinical research, particularly in the context of AKI.19 According to a prospective cohort study, alterations in thyroid function indices are prevalent among patients with AKI (82.9%).31 However, the causal-temporal relationship between this clinicopathological phenomenon and AKI remains enigmatic. Scientific insights emerge from cardiovascular surgical cohorts in which a preoperative low level of T3 has been identified as an independent risk factor for postoperative AKI among patients undergoing complex interventions, including acute type A aortic dissection repair and valvular/coronary procedures.14,15 Similar results were reported in 2 other percutaneous coronary intervention cohort studies.16,17 These results were consistent with our observations that elevated TSH levels were associated with an increased risk of HA-AKI, as elevated TSH levels often coincide with low levels of T3 mediated by the hypothalamic–pituitary–thyroid axis. Of note, another observational study reported no significant association between preoperative thyroid function and AKI after thyroidectomy.32 In summary, the aforementioned studies have been limited by small-sample and single-center designs, specific population like surgical/interventional setting, and incomprehensive thyroid dysfunction assessment excluding hyperthyroidism. Our nationwide cohort study demonstrated that both high and low TSH levels are risk factors for incident HA-AKI in hospitalized patients, with low levels exhibiting a particularly pronounced risk elevation compared with high counterparts. To some extent, the results aligned with previous studies in terms of hypothyroid associated with AKI, while fundamentally advancing the field by establishing the first robust epidemiological association between hyperthyroid (characterized by low levels of TSH) and AKI.14, 15, 16, 17 The observed thyroid axis perturbations may serve to predict kidney prognosis in clinical settings. Therefore, prompt evaluation of thyroid function may be considered to assist in detecting AKI in high-risk populations, particularly in critically ill patients who often experience euthyroid sick syndrome (ESS). The potential mechanisms of TSH–AKI interactions may be attributed to systemic disturbances in hypothalamic–pituitary–thyroid axis homeostasis and thyroid hormones metabolism.6,7 As the central regulator of thyroid function, TSH-mediated hormonal signaling exerts pansystemic regulatory effects, including kidney pathophysiology. Thyroid hormones directly modulate nephron function through various mechanisms, including effects on kidney development, changes in kidney hemodynamics, activation of tubular ion transporters, modulation of the tubuloglomerular feedback system, and creatinine metabolism.7,18, 19, 20 Additionally, these hormones play a role in regulating water–electrolyte balance and cardiovascular performance.7,18,33 Collectively, these effects lead to significant alterations in kidney function associated with thyroid dysfunction. Several studies have reported that hypothyroidism can induce AKI through rhabdomyolysis and changes in kidney hemodynamics.20,21 Additionally, there is a biological rationale suggesting that hyperthyroidism may negatively affect kidney function and contribute to AKI because of thyrotoxicosis-induced cardiomyopathy, such as high-output cardiac failure.22 Taken together, TSH may mediate AKI risk through downstream hormonal perturbations affecting multiorgan networks. Further studies are required to elucidate the underlying pathways by which thyroid dysfunction contributes to AKI. The subgroup analyses enhanced the generalizability of our findings. Notably, we found that age, sex, heart failure, and CHD emerged as significant effect modifiers (P interaction <0.05) in TSH-AKI associations. Stronger point estimates of the association between hyperthyroid TSH levels and HA-AKI were observed among those who were younger, without heart failure, and without CHD. This may be explained by reduced responsiveness of the hypothalamic-pituitary feedback system or less bioactivity of thyrotropin receptor antibody in the elderly.34,35 Additionally, both heart failure and CHD are risk factors for AKI, which may mitigate some of the risk associated with hyperthyroid TSH levels.36 Nevertheless, the results showed no relationship between hypothyroid TSH levels and HA-AKI among female populations, potentially attributable to alterations in the TSH–fT4 relationship or serum thyroxine-binding globulin concentration under the mechanisms of female hormones.34,37,38 In addition, therapeutic paradoxes emerge in thyroid modulation for AKI management. In a randomized trial of AKI patients with ESS, thyroid hormone therapy had no beneficial effects on the severity or duration of AKI, while incurring a higher risk of adverse clinical outcomes because of TSH suppression.39 Conversely, several studies have indicated that appropriate treatment with thyroxine can reverse kidney injury in cases of hypothyroidism-induced AKI.21,40 These diametric outcomes underscore critical pathophysiological distinctions between primary thyroid disorders and ESS-mediated hormone perturbations. Thus, despite the considerable challenges involved, clinicians should endeavor to determine the causal or temporal relationships between thyroid dysfunction and AKI for thyroid-modulating pharmacotherapies. This study could provide novel insights and facilitate clinical practice to address this issue. This study possesses several notable strengths. To our knowledge, it is the first large-sample and multicenter cohort study to comprehensively evaluate the association between TSH and incident AKI, thereby addressing the evidence gap regarding the association between hyperthyroid and AKI, which has not been previously reported. During the enrollment of the cohort, patients diagnosed with CA-AKI at admission were excluded to mitigate potential reverse causality, as CA-AKI may be accompanied by abnormal TSH levels.18,19 Moreover, AKI was diagnosed based on changes in SCr levels rather than relying on the ICD-10 code, which is more objective and reliable.41 However, several limitations of our study are worth mentioning. First, the study relied on isolated TSH measurements at baseline without concurrent assessment of thyroid hormones (T3/T4). This precludes a precise etiological diagnosis of thyroid dysfunction and limits our ability to distinguish between overt and subclinical disease states. Second, the absence of daily SCr measurements coupled with unrecorded urine output may attenuate AKI detection sensitivity. Third, the inability to investigate the underlying causes of AKI hindered our capacity to explore the mechanisms by which thyroid dysfunction contributes to AKI. Fourth, given that reverse T3 (rT3) was not routinely measured in clinical practice, we could not directly identify patients with ESS to analyze the effect of ESS on AKI or adjust for this variable in our multivariable models. Lastly, despite adjustment for potential confounders, we could not fully eliminate residual confounding from unmeasured factors, such as lifestyle behaviors, hospital characteristics disparities, and primary reasons for hospitalization. In summary, our study confirmed that both higher TSH levels (>5.0 mIU/L) and lower TSH levels (≤0.5 mIU/L) were associated with a higher risk of incident HA-AKI, showing a U-shaped relationship. These findings suggest that clinicians should consider incorporating TSH measurement as a potential predictor for evaluating the risk of AKI, particularly in specific clinical settings. Furthermore, there is an urgent need for further research to explore the potential of thyroid-modulating pharmacotherapies in ameliorating established AKI among patients with thyroid dysfunction.