Red blood cell transfusion-related eicosanoid profiles in intensive care patients—A prospective, observational feasibility study

**Background:** Transfusion of packed red blood cells (PRBCs) is associated with increased morbidity and mortality in ICU patients, potentially due to transfusion-related immunomodulation (TRIM). Eicosanoids—metabolic products of polyunsaturated fatty acids such as arachidonic acid (AA)—accumulate in PRBCs during storage (the "storage lesion") and may contribute to TRIM. However, no prospective clinical study had assessed eicosanoid dynamics after PRBC transfusion in postoperative ICU patients. This feasibility study aimed to determine whether such a study could be conducted and to generate preliminary data.

**Methods:** This single-center, prospective, observational feasibility study was conducted at the Medical University of Vienna (ethics approval #1595/2018). Three groups were defined: patients on acetylsalicylic acid (ASA) after aortic surgery (Aorta), patients on immunosuppressants after bilateral lung transplantation (LuTx), and postoperative ICU patients not on ASA or immunosuppressants (Comparison). Screening occurred from December 2018 to March 2020 (terminated early due to COVID-19). Adult patients transfused with one unit of PRBCs postoperatively were included. Exclusion criteria included use of NSAIDs, COX-2 inhibitors, or cysteinyl leukotriene receptor antagonists in all groups; ASA and protamine in LuTx and Comparison groups; and glucocorticoids within 24 hours in Aorta and Comparison groups. Blood was sampled from the PRBC unit immediately before transfusion. Patient plasma was collected at 60, 30, and 0 minutes before transfusion (T-2, T-1, T0) and at 60, 90, and 120 minutes after the start of transfusion (T1, T2, T3). Eicosanoid abundances were measured by liquid chromatography-tandem mass spectrometry and expressed as normalized area under the curve (nAUC). Spearman correlations, linear mixed models, and tobit regression were used for analysis. No correction for multiple testing was applied.

**Key Results:** Of 128 screened patients (66 Aorta, 62 LuTx), 85 provided consent, 26 required PRBC transfusion, and 21 were included in the final analysis (Aorta n=4, LuTx n=8, Comparison n=9). Mean PRBC storage duration ranged from 11 to 36 days. In PRBC supernatants, 20-HETE was not detected, 5-HETE was detected in all but one sample, and all other eicosanoids were ubiquitously detected. All measured eicosanoids showed significant positive correlations with storage duration: Spearman's rho ranged from 0.47 for 11-HETE (p=0.034) to 0.75 for 9-HETE (p<0.001). AA showed rho=0.49 (p=0.025), 5-HETE rho=0.70 (p<0.001), 12-HETE/8-HETE rho=0.52 (p=0.015), and 15-HETE rho=0.59 (p=0.005). In patient plasma, AA, 5-HETE, 12-HETE/8-HETE, and 20-HETE were detectable in all 125 samples; 15-HETE in all but one; 9-HETE in 83/137 (61%); and 11-HETE in 36/137 (26%). Baseline fluctuation was within the 0.5- to 2-fold range of the mean in 95.2% of measurements for AA, 11-HETE, and 20-HETE; 90.5% for 5-HETE and 15-HETE; 76.2% for 12-HETE/8-HETE; and 61.9% for 9-HETE. Contrary to the hypothesis, transfusion did not increase plasma eicosanoid levels. In the whole cohort, 12-HETE/8-HETE decreased by 11% per hour (slope 0.89, 95% CI 0.83–0.96, p<0.01) and 20-HETE decreased by 9% per hour (slope 0.91, 95% CI 0.85–0.98, p<0.01). In the LuTx group, 20-HETE decreased by 15% per hour (slope 0.85, 95% CI 0.77–0.94, p<0.01). No significant differences in post-transfusion slopes were found between groups. An untargeted analysis identified 47 additional analytes potentially involved in TRIM, of which 9 showed correlation with storage duration (r>0.4 or r<-0.4, p<0.05).

**Clinical Implications:** This study demonstrates that prospective investigation of eicosanoid dynamics after PRBC transfusion in postoperative ICU patients is feasible, with 9% of screened patients ultimately included. The confirmation that eicosanoids accumulate in stored PRBCs and are detectable in patient plasma supports the biological plausibility of their role in TRIM. The unexpected finding that plasma eicosanoids decreased after transfusion (rather than increasing) suggests that immediate post-transfusion dynamics may be complex, possibly involving rapid buffering, negative feedback regulation, or non-direct mechanisms. The limited baseline fluctuation of most eicosanoids justifies using a single pre-transfusion sample in future studies. Larger studies are warranted to clarify the role of PRBC-derived eicosanoids in TRIM and to explore personalized transfusion strategies based on storage duration and patient characteristics.