Clinical Significance of Coagulation Studies in Predicting Response to Recombinant Factor VIIa in Cardiac Surgery Patients

While recombinant activated Factor VII (rFVIIa) is increasingly used for management of uncontrolled hemorrhage in a variety of clinical settings, its efficacy appears mixed. Few studies have examined the association between coagulation studies and rFVIIa efficacy in cardiac surgery.

We performed a retrospective study of all patients at Brigham and Women’s Hospital (BWH) from February 2005 through March 2011 who received rFVIIa for uncontrolled hemorrhage during or after cardiac surgery. rFVIIa-mediated hemostasis was determined based on clinical descriptions in the medical record. Laboratory studies, clinical parameters, blood product transfusions, and chest tube output were compiled from the medical record.

Forty-seven patients were identified for this study. The median age was 65 years (range, 16–85). Thirty-one patients were male. Eleven had a perioperative left ventricular ejection fraction < 35%. Of 47 surgeries performed, 30 were emergent or urgent and 16 elective (data was missing in 1 case). Twelve required an intraaortic balloon pump, and 12 placement or removal of a ventricular assist device. Median cardiopulmonary bypass (CPB) time was 281.5 min (range, 104–744; n = 45). Over a median follow-up period of 52 days (range, 0–1673), 22 patients (46.8%) died, 10 within one day of receiving rFVIIa. Median survival was 443 days (90% confidence interval (CI) [30, not-yet-reached]). The median dose of rFVIIa was 40.2 mcg/Kg. rFVIIa-mediated hemostasis was achieved in 23 patients (48.9%), while continued bleeding was observed in 24 (51.1%). Thromboembolic complications attributed to rFVIIa developed in 9 patients (arterial, 5; venous, 4). Median transfusion requirements decreased after rFVIIa administration for the entire population, although the differences were not statistically significant. Survival status could not be reliably incorporated into models of transfusion requirements in this retrospective study, prohibiting further analysis.

Mean coagulation studies before vs after rFVIIa were: PT, 24.9 sec +/− 18.3 (n = 44) vs 15.4 sec +/− 4.9 (n = 47); INR, 2.3 +/− 2.5 (n = 44) vs 1.2 +/− 0.5 (n = 47); and PTT, 57.3 sec +/− 27.2 (n = 40) vs 48.4 sec +/− 17.8 (n = 44), respectively. Plasma fibrinogen was excluded from the analysis, as levels were not measured in over one-third of cases. Univariate analyses identified several clinical or laboratory variables associated with rFVIIa-mediated hemostasis: PT after rFVIIa administration (p = 0.02 by Fisher’s exact test), INR before rFVIIa (p = 0.05), INR after rFVIIa (p = 0.05), PTT after rFVIIa (p = 0.008), pH after rFVIIa (p = 0.04), and survival status (p = 0.008). In stepwise multivariable logistic regression analyses, only normalization of PTT after rFVIIa was significantly associated with rFVIIa-mediated hemostasis (odds ratio (OR) = 0.169, 95% CI [0.041,0.694], p = 0.01). Stepwise multivariable regression models of PT, INR, and PTT after rFVIIa showed a strong association of PTT and only a marginal association of PT with rFVIIa-mediated hemostasis (PTT: OR = 0.181, 95% CI [0.047,0.722], p = 0.02; PT: OR = 0.10, 95% CI [0.01,0.98], p = 0.05).

Chest tube output was recorded for 14 patients who received rFVIIa in the CSICU rather than in an operating room. Of these, 8 (57.1%) had a reduction in chest tube output of ≥ 25% and 5 (21.4%) a reduction of ≥ 50%. While none of the variables examined showed a significant association with reduced chest tube output, concordance with rFVIIa-mediated hemostasis among this small subset of patients was poor (for ≥ 25% reduction: concordance 64.3%, 90% CI [39,84.7], p = 1.0 by McNemar’s test; for ≥ 50% reduction: concordance 57.1%, 90% CI [32.5,79.4]), p = 0.2).

In cardiac surgery patients who receive rFVIIa for uncontrolled hemorrhage, normalization of PTT after rFVIIa appears to be strongly correlated with rFVIIa-mediated hemostasis. This suggests that optimal hemostasis with rFVIIa occurs following concomitant repletion of other coagulation factors in both extrinsic and intrinsic pathways, although the possibility that normalization of coagulation studies in both pathways is a marker of rFVIIa-mediated hemostasis has not been ruled out. Prospective trials may help to determine whether correction of PTT and other coagulation parameters can improve rFVIIa efficacy in cardiac surgery.

No relevant conflicts of interest to declare.