Early versus deferred intravenous calcium in patients with sepsis and ionized hypocalcemia: a target-trial emulation

BACKGROUND: Intravenous calcium is frequently administered during sepsis, yet whether early treatment improves 24-h hemodynamic recovery remains uncertain. We compared 24-h hemodynamic recovery between calcium initiation within 2 h of a non-emergency low-ionized calcium (iCa) result and no calcium initiation during the first 6 h after the result. METHODS: We conducted a retrospective target-trial emulation of 2016-2025 high-resolution electronic health records from Zigong Fourth People's Hospital. Adults with sepsis, final or corrected iCa of 0.90-1.10 mmol/L, and no emergency indication were eligible at result availability. Strategies were executed calcium within 2 h versus no calcium through 6 h; care after 6 h was unrestricted. Clone-censor weighting and overlap tilting addressed grace periods, measured time-varying confounding, and clinical equipoise. The primary outcome was being alive and vasopressor-free at 24 h. Estimates and two-sided 95% percentile intervals came from 2,000 patient-level bootstrap samples with model refitting. RESULTS: Of 5,993 patients, 1,538 initiated calcium within 2 h, 147 initiated at 2-6 h, and 4,308 received none through 6 h; none of the latter initiated calcium during 6-24 h. The standardized probability of 24-h hemodynamic recovery was 69.9% with early calcium and 61.1% with deferral, corresponding to a risk difference of + 8.69 percentage points (95% percentile interval, + 4.66 to + 12.84) and a risk ratio of 1.142 (1.074-1.214). The estimated association varied with baseline norepinephrine-equivalent dose (Holm-adjusted P = 2.00 × 10⁻⁶), whereas the iCa interaction did not remain statistically supported after adjustment (P = 0.097). Early calcium was associated with a 1.29-percentage-point higher observed risk of hypercalcemia (95% percentile interval, + 0.64 to + 2.05). Intervals for severe arrhythmia, new renal replacement therapy, and hospital mortality included the null. CONCLUSIONS: Early intravenous calcium was associated with a higher probability of 24-h hemodynamic recovery than the deferred strategy, whose observed pathway involved no calcium through 24 h. Estimated effects were larger at higher vasopressor doses. Early calcium was associated with a modestly higher observed risk of hypercalcemia. These findings support randomized evaluation in vasopressor-treated patients. Baseline iCa modification remains exploratory.