Serum Calcium Dynamics for Hydrocephalus Prediction After Subarachnoid Hemorrhage

An observational cohort of 302 spontaneous subarachnoid hemorrhage patients monitored for 90 days, with serial serum calcium measurements at admission, 72 hours, and 7 days post-ictus. The data was used to develop and validate a clinical nomogram for predicting delayed hydrocephalus, which occurred in 24.2% of patients.

Use Cases

• Train a logistic regression model to predict delayed hydrocephalus using admission serum calcium, hypertension, and Hunt-Hess grade as features.
• Analyze the predictive accuracy of serum calcium at 72 hours (cutoff ≤7.65 mg/dL) versus conventional severity scales using ROC analysis.
• Validate the performance of the admission-based nomogram, which incorporated variables like smoking history and dyslipidemia, in an external cohort from two hospitals.

Strengths

• Dataset includes 302 patient records with a clear outcome measure (24.2% delayed hydrocephalus incidence).
• Serial measurements of serum calcium at three specific time points (admission, 72h, 7 days) enable dynamic analysis.
• Prediction model demonstrated high performance with an AUC > 0.93 in internal validation and 0.898 in external validation.

Limitations

• Small sample size of 302 patients limits statistical power and generalizability of findings.
• Proposed serum calcium cutoff of ≤7.65 mg/dL may vary across different laboratory assays, requiring local calibration.
• Observational study design from two hospitals may introduce selection bias and limit geographic representation.

Provenance

Source

Di Wu via figshare.

Collection Method

Observational cohort study analyzing serial serum calcium measurements in SAH patients monitored for 90 days.

Time Range

Patients monitored for 90 days post-subarachnoid hemorrhage.

Freshness

Last updated March 2026.

Geography

Cohort from two hospitals; specific locations unknown.