Peripheral blood inflammatory indices across thyroid functional states in Graves’ disease: associations with liver biochemical abnormalities and hyperthyroid relapse

Abstract

Graves’ disease is an organ-specific autoimmune disorder driven by thyrotropin receptor antibody (TRAb) and accompanied by persistent immune and inflammatory activation. Peripheral blood inflammatory indices are readily available and have been increasingly used to assess inflammation in autoimmune diseases. However, their clinical significance in Graves’ disease across different thyroid functional states is unknown. In this retrospective study, 687 patients with Graves’ disease were enrolled. Five peripheral blood inflammatory indices were calculated, including the neutrophil-to-lymphocyte ratio (NLR), monocyte-to-lymphocyte ratio (MLR), systemic immune-inflammation index (SII), systemic inflammation response index (SIRI), and pan-immune-inflammation value (PIV). Patients were stratified and compared according to thyroid functional status. Because sex distribution differed among the groups, sex-adjusted median regression models were additionally fitted, and group-by-sex interactions were examined. Correlation analysis, logistic regression, and receiver operating characteristic (ROC) curve analysis were used to evaluate associations of these indices with clinical characteristics, liver biochemical abnormalities, and hyperthyroid relapse. The outcome models were internally validated using bootstrap resampling and repeated 10-fold cross-validation. Sex distribution differed significantly among the thyroid functional groups. After adjustment for sex, MLR, SIRI, and PIV remained significantly higher in the overt hyperthyroid group than in the subclinical hyperthyroid and euthyroid groups. In patients with overt hyperthyroidism, MLR was positively correlated with age, FT3, FT4, ALT, and AST (r s = 0.152–0.206, all P  < 0.05). Multivariable logistic regression identified TRAb, hemoglobin, and standardized MLR as independent factors associated with liver biochemical abnormalities, and the resulting model showed limited-to-moderate discrimination (apparent AUC, 0.703; optimism-corrected AUC, 0.692; cross-validated AUC, 0.680). Among the five inflammatory indices, PIV showed the highest individual discrimination for hyperthyroid relapse (AUC, 0.734). The TRAb–PIV model showed relatively good internal discrimination (apparent AUC, 0.815; optimism-corrected AUC, 0.805; cross-validated AUC, 0.788). Compared with NLR and SII, the monocyte-containing indices MLR, SIRI, and PIV showed clearer differences across thyroid functional states in Graves’ disease and may provide complementary inflammation-related information. Among them, MLR was associated with liver biochemical abnormalities, whereas PIV showed potential adjunctive value for hyperthyroid relapse risk assessment.