Early ionized calcium changes after postoperative blood product transfusion in adult cardiac surgical ICU patients: a retrospective database study

Ionized calcium (iCa) is frequently measured after adult cardiac surgery, but the clinical implications of early post-transfusion changes are uncertain. We evaluated early iCa change as a transfusion-associated measurement pattern rather than as proof of blood-product causality or calcium-repletion benefit. We conducted a retrospective database study using MIMIC-IV 3.1 as the primary data source, with eICU 2.0 used for exploratory analyses. Adult cardiac surgical ICU admissions with blood product transfusion within 48 h after ICU admission were eligible. The main MIMIC-IV analysis used transfusion episodes with paired pre- and post-transfusion iCa measurements within 6-hour windows. Product type was assigned from the index transfusion record, while co-administered products were measured separately. Additional analyses included non-transfusion sham windows, time-window analyses, recorded FFP amount analyses, pH and fluid-context sensitivity analyses, clinical scenario summaries, and inverse-probability weighting for paired iCa availability. Risk ratios (RRs) were estimated using Poisson regression with ICU-stay-clustered robust standard errors. In MIMIC-IV, 7,004 of 9,616 transfusion episodes from 3,157 ICU admissions had paired iCa measurements. First post-transfusion iCa decline of at least 0.10 mmol/L occurred in 1,764 episodes (25.2%). Compared with red blood cell (RBC) transfusion, fresh frozen plasma (FFP) was associated with this endpoint in the parsimonious model (RR 1.63, 95% CI 1.49–1.79), after pre-transfusion pH adjustment (RR 1.60, 95% CI 1.46–1.75), and in a broader clinical-context model that also included peri-transfusion fluid/solution variables (RR 1.39, 95% CI 1.26–1.54). First post-transfusion iCa was measured within 0–2 h in 4,819 paired episodes and within > 2–6 h in 2,185 episodes. Recorded positive FFP amount did not show a positive monotonic dose-response. In eICU, paired iCa episodes were sparse, so results were interpreted only as exploratory context. Among postoperative cardiac surgical transfusion episodes, FFP-index episodes were associated with a greater probability of early iCa decline than RBC-index episodes. These findings support selective early reassessment of iCa rather than any specific calcium-replacement strategy.