Blood Draws, Fluid Input, and Bleeding Events After Percutaneous Coronary Intervention

The American College of Cardiology (ACC) National Cardiovascular Data Registry (NCDR) CathPCI Registry reports bleeding events within 72-h of percutaneous coronary intervention (PCI) as a quality metric. It is plausible that NCDR post-PCI bleeding events may be influenced by markers of intensity of patient care not related to PCI quality. The relationship between iatrogenic phlebotomy, intravenous fluid input, and post-PCI bleeding events is unknown.

This is a patient-level, observational study of 13,553 PCI procedures at a single center between 6/11/2009 and 6/14/2017. Multivariable logistic regression determined associations between NCDR bleeding events, total oral and intravenous fluid volume input, and number of lab tubes drawn on day of PCI up to 72-h post PCI. Improvement in discriminatory ability of adding these variables to a bleeding risk model was assessed with C-statistic.

There were 767 (5.7%) bleeding events. Every 100 mL input was independently associated with 2.0% increased odds of bleeding (odds ratio [OR] 1.01-1.03, p < 0.001). Every 3 mL lab tube drawn was independently associated with 8.1% increased odds (OR 1.07-1.09, p < 0.001). A multivariable model including baseline patient characteristics and PCI variables achieved good discriminatory ability (C-statistic 0.858), but with addition of input, number of lab draws, and length of stay (LOS) achieved an excellent discriminatory ability (C-statistic 0.927, p < 0.001 comparing C-statistics).

Variables reflecting intensity of patient illness including frequency of lab draws and fluid intake within 72-h of PCI significantly improve the ability of models to predict NCDR post-PCI bleeding events. Consideration should be given to risk-adjusting for lab draws and fluid administration when nationally reporting bleeding event rates.