Objective To explore the dynamic trajectories of heart rate variability (HRV) in the acute phase of intracerebral haemorrhage (ICH) based on a latent class growth model (LCGM), and to analyze their realtionship with short-term prognosis.

Methods A total of 173 patients with ICH were enrolled. Heart rate data were continuously collected for 7 days after admission using electrocardiographic monitoring equipment. The standard deviation of hourly heart rate (SDHR) over each 24-hour period was used as the HRV assessment indicator. LCGM was applied to identify distinct HRV trajectory patterns, and multivariate logistic regression was performed to determine independent predictors for poor prognosis in ICH patients.

Results The HRV (SDHR) values within 7 days after onset of ICH patients were (12.47 ± 4.67), (12.01 ± 4.89), (13.16 ± 5.04), (12.16 ± 4.72), (12.03 ± 4.53), (12.18 ± 4.94), and (12.65 ± 4.74), respectively, with statistically significant differences across time points (P ＜ 0.01). LCGM identified three HRV trajectory subtypes: a low-level slowly decreasing group (45 cases, 26.0%), a moderate-level slowly increasing group (61 cases, 35.3%), and a high-level slowly decreasing group (67 cases, 38.7%). Multivariate logistic regression revealed that HRV trajectory was an independent strong predictor of poor prognosis (P ＜ 0.01): compared with the low-level group, the risk of poor prognosis increased to 95.724 times (95%CI: 11.107–824.973, P ＜ 0.01) in the moderate-level group and 26.264 times (95%CI: 3.219–214.267, P ＜ 0.01) in the high-level group.

Conclusions HRV in the acute phase of ICH exhibits three distinct dynamic trajectories that are closely associated with the long-term prognosis of patients. Identification of HRV trajectories may facilitate precise prognostic risk stratification and provide a critical foundation for early clinical intervention.