Transfusion with Washed vs. Unwashed Packed Red Cells in Coronary Artery Bypass Graft (CABG) Surgery: Major Outcome Differences

BACKGROUND: Blood transfusion (Tx) carries greater risks of adverse events (AEs) than previously appreciated. These adverse effects include higher incidence of post-surgical infections, longer hospital stay, higher mortality, more frequent serious adverse events (SAE’s), and generally poorer surgical outcomes. Accordingly, ameliorating these adverse effects constitutes an urgent challenge to medical science.

Factors responsible for Tx-related adverse events (AE’s) are not well understood. Many potentially toxic substances are released during blood storage, and many of them have been implicated or postulated as culprits. Washing of packed RBC remove these products and may ameliorate transfusion-related AE’s. Benefits of washed RBC are well established for pediatric surgical patients, chiefly by preventing hyperkalemia, but use of washed RBC in adult surgical patients has not heretofore been systematically investigated.

We here report results of a prospective randomized study directly comparing surgical outcomes, in terms of mortality and AE’s, between groups of adult CABG patients transfused with either washed or unwashed (conventional) RBC.

METHODS: A prospective randomized study of 148 patients undergoing coronary artery bypass graft (CABG) was conducted. Fifty-eight patients were randomized to receive unwashed (conventional) RBC (UW group) and 41 to washed RBC (W group). The remaining 49 did not require Tx. The main in-hospital outcomes recorded included mortality, serious adverse events (SAE’s), non-serious adverse events (AE’s), and SOFA scores pre- and post-surgery. A telephone interview was conducted at day 30 post-discharge, and mortality at one-year was also assessed. The statistical techniques used for the comparison of the UW and W RBC groups included: independent sample t-tests for variables with normal or approximately normal distribution; Mann-Whitney tests for variables with skewed distributions and for ordinal variables; chi-squared tests or Fisher’s exact tests for discrete variables; and logistic regression model for assessing different factors as predictors of the occurrence of each kind of event.

RESULTS: Between the 2 groups, demographic, clinical, and comorbidity data were similar and there was no statistically significant difference in number of serious AE’s (SAE’s). However, 4 of 6 patients died from SAE’s in the UW group but all 7of 7 with SAE in the W group survived. The in-hospital mortality was greater in the UW group (4 vs. 0, p = 0.149) but 1-year post-op mortality was significantly higher in UW group (7 vs. 0, p=0.036). Frequency of less serious AE’s was higher in UW group in every category. Negative binomial regression analyses showed that, after adjusting for comorbidities, UW-group are likely to experience 64% more AEs (p= 0.027). The 30-day follow-up showed similar trends of higher AE’s in UW-group, but only CNS-related AE’s were significant (30 vs. 5, p<0.01).

CONCLUSIONS / DISCUSSION: These data suggest major benefits to patient outcomes by use of washed RBC in CABG. Most important is significant reduction of mortality. Less serious AE’s were also lower in the W group in nearly every category, but only CNS-related AE’s were statistically significant in this comparatively small patient population.

To our knowledge, this is the first prospective randomized study in adults to assess possible benefits of washing RBC prior to cardiac surgery. At present, washed RBCs are seldom used in adults but the present study clearly demonstrates major advantages. It may be possible to reduce costs of washing by using on-site cell call-salvage equipment but this needs to be evaluated.

This study was undertaken with the hypothesis that cell-derived microparticles (MP) are major culprits in Tx-associated AE’s. Further study is needed to determine if that hypothesis is correct. Other evidence has led us to conjecture that MP are largely responsible for post-surgical adverse outcomes; the present study is consistent with that conjecture but does not prove it.

A major shortcoming of this study is the comparatively small patient population. A much larger study, including other types of surgery, is certainly warranted by these findings, and should be designed to include more quantitative evaluation of post-surgical cognitive impairment.