From January 2015 to December 2019, 1052 patients with biopsy-proven IMN from the First Affiliated Hospital of Zhengzhou University were included in this retrospective cohort analysis. The following inclusion criteria were applied: (1) complete baseline data and follow-up of ≥ 6 months; (2) age ≥ 18 years; (3) no previous history of thyroid disease; and (4) no glucocorticoids or immunosuppressants before the renal biopsy. Patients with other glomerular diseases, such as IgA nephropathy, diabetic nephropathy, minimal change disease, and/or secondary conditions, such as hepatitis, systemic lupus erythematosus, psoriasis, malignancy, serious infectious diseases, or severe cardiopulmonary diseases, were excluded. Blood was collected within 48 hours of admission to assess the free thyroxine (FT4), free triiodothyronine (FT3), and thyroid-stimulating hormone (TSH) levels. The patients were divided into euthyroid and thyroid dysfunction groups according to the thyroid hormone levels. Patients in the thyroid dysfunction group were further divided into three subgroups: hypothyroid, hyperthyroid, and non-thyroid illness syndrome (NTIS) groups.

This study followed the standards of the Helsinki Declaration and was supported by the Ethics Review Committee of the First Affiliated Hospital of Zhengzhou University (approval number:2022-KY-1187-002). Informed consent was waived owing to the retrospective nature of this study.

Demographic and clinical data were collected at the time of renal biopsy, including age, sex, blood pressure, blood urea nitrogen (BUN), serum creatinine (SCr), uric acid (UA), triglycerides (TG), total cholesterol (TC), blood white blood cells (WBC), hemoglobin, platelets, albumin, estimated glomerular filtration rate (eGFR), M-type phospholipase A2 receptor (PLA2R) antibody, D-dimer, proteinuria, FT3, FT4, and TSH. FT3 and FT4 levels were measured using a commercial radioimmunoassay (RIA) kit (Roche Diagnostics, Mannheim, Germany). TSH levels were measured using a commercial RIA kit (Immunotech, Marseille, France). Two experienced renal pathologists diagnosed the renal biopsy specimens using light microscopy, electron microscopy, and immunofluorescence. Pathological changes were classified into two grades based on the presence or absence of glomerulosclerosis, crescents, mesangial cell proliferation, renal tubular atrophy, renal interstitial fibrosis, inflammatory cell infiltration, and renal arteriolar lesions. Follow-up data included time, proteinuria, SCr, eGFR, albumin, and medications, such as renin-angiotensin-aldosterone system inhibitor (RAASi), corticosteroids, and immunosuppressants.

The endpoint event was a composite of SCr doubling, eGFR decrease of >40% from baseline, ESRD, or death from kidney failure (12–14). ESRD was defined as eGFR <15 mL/min/1.73 m² or the need for renal replacement therapy (dialysis or kidney transplantation). Follow-up time was defined as the interval from renal biopsy to the occurrence of the endpoint event or the last outpatient visit. Treatment response included clinical remission and relapse. Clinical remission included complete remission (CR) and partial remission (PR). CR was defined by proteinuria <0.3 g/d, serum albumin >35 g/L, and SCr stability. PR was defined as proteinuria <3.5 g/d with a >50% reduction from baseline. After clinical remission, the reappearance of proteinuria >3.5 g/d was defined as relapse. The reference ranges of FT3, FT4, and TSH levels were 3.28–6.47 pmol/L, 7.9–18.4 pmol/L, and 0.34–5.6 μIU/mL, respectively. NTIS, also known as euthyroid sick syndrome, includes low T3, low T4, low T3 and low T4, high T4, and other abnormalities (15). The time course of kidney disease was defined from discovery to renal biopsy. Hypertension was defined as systolic pressure ≥140 mmHg, diastolic pressure ≥90 mmHg, or the use of antihypertensive drugs. The mean arterial pressure was equal to one-third systolic pressure plus two-thirds diastolic pressure. Nephrotic syndrome was defined as proteinuria >3.5 g/d and serum albumin <30 g/L. The eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation (16).

Continuous variables with normal distribution were expressed as mean ± standard deviation, and the independent samples t-test, one-way analysis of variance, and Bonferroni method were used for comparisons between the groups. Data that did not follow the normal distribution were expressed as median and interquartile ranges (25%, 75%), and the differences were compared using the Mann-Whitney U test or Kruskal-Wallis test. Categorical variables were expressed as frequency (percentage) and compared between groups using the χ² test or Fisher exact test. Logistic regression was used to analyze the risk factors of thyroid dysfunction in patients with IMN. To reduce confounders and balance baseline variables, we applied propensity score matching (PSM) (17). Matching was performed in a 1:1 ratio using the nearest neighbor approach with no replacement and a matching tolerance of 0.02. The covariates entered into the propensity score model included age, sex, hemoglobin, platelets, albumin, BUN, SCr, UA, TC, TG, eGFR, WBC, PLA2R antibody, D-dimer, proteinuria, and treatment. To compare the rates of CR, relapse, and renal survival between the groups, the Kaplan-Meier curve and log-rank test were used, as well as multiple testing with a Bonferroni correction method. The association between thyroid dysfunction and IMN prognosis was investigated using Cox regression analysis. SPSS version 26.0 (IBM Corp, Armonk, NY, USA) and GraphPad Prism version 8.0.2 (GraphPad Software, San Diego, CA, USA) software were used for statistical analysis and figure creation. All tests were two-sided. A P value <0.05 indicated statistical significance.

Of the 1052 patients with IMN, 736 (70%) had normal thyroid function and 316 (30.0%) had abnormal thyroid function. The most common type of thyroid dysfunction was hypothyroidism (n = 226; 21.5%) including 212 cases of subclinical hypothyroidism (20.2%) and 14 cases of clinical hypothyroidism (1.3%). This was followed by NTIS (n = 79; 7.5%). In contrast, hyperthyroidism occurred less frequently (n = 11; 1.0%) ( Figure 1A ). When eGFR (unit: mL/min/1.73 m2) was >90, 60–90, and <60, the incidence of thyroid dysfunction was 26.1% (224/857), 45.9% (78/170), and 56.0% (14/25), respectively ( Figure 1B ).

Patients in the thyroid dysfunction group had a higher prevalence of nephrotic syndrome than those in the euthyroid group (68.4% versus 46.9%, P < 0.001). However, no significant variations in age, sex, course, or blood pressure were seen between the two groups ( Table 1 ). Further subgroup analysis showed that, compared to the euthyroid group, the hypothyroid and NTIS groups had a higher incidence of nephrotic syndrome ( Table S1 ).

Patients in the thyroid dysfunction group exhibited higher levels of BUN, SCr, TG, TC, PLA2R antibody titer, D-dimer, and proteinuria and a greater percentage of PLA2R antibody titer > 50 RU/mL than those in the euthyroid group, but lower eGFR, albumin, and hemoglobin levels (all P < 0.05). Moreover, more patients in the thyroid dysfunction group had pathological changes in renal interstitial inflammatory cell infiltration (68.4% versus 61.5%, P = 0.035). Furthermore, the thyroid dysfunction group used a lower proportion of RAASi (58.9% versus 66.3%, P = 0.021) but a higher proportion of glucocorticoids combined with immunosuppressants (46.8% versus 35.2%, P < 0.001) ( Table 1 ). Further subgroup analysis revealed that compared to the euthyroid group, the hypothyroid and NTIS groups had higher levels of SCr, TC, D-dimer, and PLA2R antibody titer, proteinuria, and a higher proportion of PLA2R antibody titer >50 RU/mL, but lower albumin and eGFR levels. In addition, the NTIS group had higher BUN and lower hemoglobin levels, while the hypothyroid group had higher levels of TG, platelets, and a higher proportion of glucocorticoids combined with immunosuppressants ( Tables S1 , S2 ).

After PSM, 282 pairs were successfully matched, including 282 cases in the euthyroid group and 282 cases in the thyroid dysfunction group. The baseline data of the two groups reached a balance ( Table 1 ).

The univariate logistic regression analysis showed that BUN, SCr, albumin, TC, TG, eGFR, hemoglobin, platelets, D-dimer, PLA2R antibody, proteinuria, and renal interstitial inflammatory cell infiltration were predictors of thyroid dysfunction in patients with IMN. Age, sex, and significant variables in the univariate analysis were included in the multivariate logistic regression. The results of the multivariate logistic regression revealed that female sex, lower albumin and eGFR levels, higher D-dimer level, and more proteinuria were independent risk factors for thyroid dysfunction in patients with IMN ( Table 2 ).

The median follow-up durations in the euthyroid and thyroid dysfunction groups were 28 (range =17-40) months and 24 (range = 14-34) months, respectively. The incidence rates of treatment response and composite events in the two groups are shown in Table 1 . Results from the Kaplan-Meier curves revealed that the thyroid dysfunction group had a lower cumulative CR probability (P = 0.002, Figure 2A ), higher cumulative relapse probability (P < 0.001, Figure 2B ), and lower cumulative renal survival probability (P = 0.013, Figure 2C ) than the euthyroid group. In the subgroup comparisons ( Figures 3 , S1 ), the thyroid dysfunction subgroups were separately compared with the euthyroid group. Kaplan-Meier curves identified a lower cumulative CR probability in the hypothyroid (P = 0.023) and NTIS groups (P = 0.017), but the differences were not statistically significant. Additionally, the cumulative relapse probabilities were higher in the hypothyroid (P < 0.001) and NTIS groups (P = 0.002), while the cumulative renal survival probability was lower in the hypothyroid group (P = 0.011). In the unadjusted model, Cox regression analysis revealed that thyroid dysfunction was a significant predictor of CR [hazard ratio (HR) = 0.781, P = 0.003], relapse (HR = 1.699, P = 0.001), and composite endpoint event (HR = 1.703, P = 0.014) in patients with IMN. After adjusting for age, sex, hypertension, proteinuria, albumin, eGFR, PLA2R antibody, and treatment, thyroid dysfunction remained an independent risk factor for IMN relapse (HR = 1.726, P < 0.001) and composite endpoint event (HR = 1.576, P = 0.043), but not for CR (HR = 0.941, P = 0.486) ( Table 3 ).

In the matched cohort, the median follow-up time in the euthyroid and thyroid dysfunction groups was 34 (20, 53) months and 24 (14, 34) months, respectively. The incidence rates of the different outcomes are also presented in Table 1 . The thyroid dysfunction group had a lower cumulative CR probability (P = 0.044, Figure 2D ), a higher cumulative relapse probability (P < 0.001, Figure 2E ), and a poorer cumulative renal survival probability (P = 0.004, Figure 2F ) than the euthyroid group. Subgroup analysis revealed that compared with the euthyroid group, the hypothyroid group (P = 0.003) and NTIS group (P = 0.004) had a higher cumulative relapse probability; the hypothyroid group (P = 0.006) had a lower cumulative renal survival probability, but differences in cumulative CR probability between the groups were not significant ( Figures 3 , S2 ). Results from the multivariate Cox regression analysis showed that after adjusting for the confounding factors, thyroid dysfunction remained an independent risk factor for CR (HR = 0.810, P = 0.045), relapse (HR = 1.721, P = 0.001), and composite endpoint event (HR = 2.113, P = 0.014) in patients with IMN ( Table 4 ).