Hemovigilance studies report that from 6.7% to 15% of overall transfusion-related acute lung injury (TRALI) events occur among obstetrics-gynecological patients (Chapman et al, 2009, Transfusion, 49:440; Ozier et al, 2011, Transfusion, 51: 2102). Severe postpartum hemorrhage (sPPH, i.e. the blood loss in excess of 1000 ml), complicates from 1 to 5% of all deliveries and requires massive transfusion, which is a well acknowledged TRALI risk factor. In this retrospective study we evaluate the incidence and risk factors for TRALI among patients with sPPH. We identified in the blood bank database EmoNet those patients admitted to the delivery room of our hospital from January 2005 to December 2011, necessitating the urgent transfusion of a minimum of 3 red blood cell (RBC) units, with or without fresh frozen plasma (FFP) and platelet (PLT) concentrates. Clinical records of identified patients were then retrieved and demographics, clinical and obstetric data and laboratory and radiological findings were collected. Two anesthesiologists blinded to all information of transfused units independently examined clinical records. Suspected or possible TRALI were diagnosed according to the 2004 consensus criteria; (Kleinman et al. 2004; Transfusion 44: 1774; Toy et al. 2005; Crit Care Med 33:721). In total 71 patients received at least 3 RBC units for sPPH; suspected or possible TRALI was identified in 14 cases (overall incidence 21,6%). The 2 patients with possible TRALI had pneumonia and pulmonary embolism, respectively. On the whole, patients with TRALI were more frequently admitted to the intensive care unit and had a longer hospitalization (p=0.021 and p=0.001, respectively). At univariate analysis, patients with TRALI received a higher number of RBC (p=0,008), PLT (p =0,008,) and FFP units (p =0,034). No difference were found between TRALI and no TRALI groups according to the number of PLT and FFP units from female donors, the storage time of RBC and PLT units or the number of transfused RBC units with a storage time longer than 14 days. Relatively to patient-related factors, TRALI was not associated with age or smoke habit, or with the presence of co-morbidities pre-existing to pregnancy. In contrast, the presence of pregnancy-related hypertensive disorders (PR-HD, including gestational hypertension and preeclampsia /eclampsia) (8 patients, p=0.006) was an important risk factor for TRALI. The poor adverse role was confirmed also in gestational hypertension (6 patients, p=0.012 ) or preeclampsia/eclampsia (4 patients, p=0.022) separately evaluated. In a multivariate model combining both transfusion- and patient-related factors with significance level equal or inferior to 10% at univariate analysis, only PR-HD, gestational hypertension and preeclampsia/eclampsia maintained their significance. In particular, we found that the odds ratio for TRALI was 15.98 (95% IC 2.5-103.5, p=0.004) in PR-HD in total, 24.8 (95% IC 2.5 - 248.4, p=0.006) in gestational hypertension and 33.8 (95% IC 2.2-484.1, p=0.011) in preeclampsia/eclampsia. The strong association between PR-HD and TRALI has never been reported before and it is probably due to the widespread endothelial cell activation occurring in PR-HD. These results impose a careful observation when PR-HD patients receive transfusions.
The study was supported by the “Gruppo Donatori Sangue, Francesco Olgiati”
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.