Development of a CALLY-based discrimination model for hyperemesis gravidarum and disease severity

Objective To investigate the association of the C-reactive protein-albumin-lymphocyte (CALLY) index with hyperemesis gravidarum (HG) and its severity, and to evaluate the discriminative performance of a CALLY-based model.Methods In this retrospective case-control study, 424 pregnant women at 6–16 weeks of gestation were enrolled between January 2023 and December 2025, including 204 patients with HG and 220 healthy pregnant controls. HG severity was classified using the modified Pregnancy-Unique Quantification of Emesis (PUQE) score. Clinical and laboratory data were collected, and the CALLY index was calculated. Independent factors associated with HG were identified using multivariate logistic regression. Receiver operating characteristic (ROC) curves and the DeLong test were used to assess and compare the discriminative performance of different models for HG and severe HG.Results Compared with healthy controls, patients with HG had significantly lower lymphocyte counts, albumin levels, and CALLY index values, but higher C-reactive protein levels (all p < .001). A lower log-transformed CALLY index (ln(CALLY)) was independently associated with HG (OR = 0.14, 95% CI: 0.09–0.22 p < .001). Among the evaluated models, the CALLY-based model showed the best discriminative performance for HG (AUC = 0.862, 95% CI: 0.826–0.897) and severe HG (AUC = 0.771, 95% CI: 0.692–0.850), with significantly higher AUCs than the models without CALLY (all p < .05).Conclusion The CALLY index was independently associated with HG, and a CALLY-based model demonstrated potential value for HG discrimination and severity assessment, warranting further validation in prospective multicenter studies.