A systematic review of the effect of red blood cell transfusion on mortality: evidence from large-scale observational studies published between 2006 and 2010

Abstract

Objective To carry out a systematic review of recently published large-scale observational studies assessing the effects of red blood cell transfusion (RBCT) on mortality, with particular emphasis on the statistical methods used to adjust for confounding. Given the limited number of randomised trials of the efficacy of RBCT, clinicians often use evidence from observational studies. However, confounding factors, for example, individuals receiving blood generally being sicker than those who do not, make their interpretation challenging. Design Systematic review. Information sources We searched MEDLINE and EMBASE for studies published from 1 January 2006 to 31 December 2010. Eligibility criteria for included studies We included prospective cohort, case–control studies or retrospective analyses of databases or disease registers where the effect of risk factors for mortality or survival was examined. Studies must have included more than 1000 participants receiving RBCT for any cause. We assessed the effects of RBCT versus no RBCT and different volumes and age of RBCT. Results –32 studies were included in the review; 23 assessed the effects of RBCT versus no RBCT; 5 assessed different volumes and 4 older versus newer RBCT. There was a considerable variability in the patient populations, study designs and level of statistical adjustment. Overall, most studies showed a higher rate of mortality when comparing patients who received RBCT with those who did not, even when these rates were adjusted for confounding; the majority of these increases were statistically significant. The same pattern was observed in studies where protection from bias was likely to be greater, such as prospective studies. Conclusions Recent observational studies do show a consistently adverse effect of RBCT on mortality. Whether this is a true effect remains uncertain as it is possible that even the best conducted adjustments cannot completely eliminate the impact of confounding.