Optimizing red blood cell transfusion therapy in the 21st century: the power of data analysis for past understanding and future guidance

“It is a capital mistake to theorize before one has data.” Sir Arthur Conan Doyle, Author, 1859-1930 The optimal use of transfusion therapy is often the subject of ongoing debate in the scientific and lay media.1, 2 It has recently been stated that “Treatments used in transfusion medicine (TM) are among the most widely used in clinical practice, but are among the least well studied.”3 Some have suggested that 25% to 40% or even more of red blood cell (RBC) transfusions administered in the United States might be medically unnecessary,4, 5 and transfusion has been identified by the Healthcare Cost and Utilization Project (HCUP) of the Agency for Healthcare Research and Quality (AHRQ) as the most commonly performed procedure in hospitalized patients since at least 2003.6 In 2009, blood transfusion occurred in more than 10% of hospital stays that included a procedure.7 The potential pervasiveness of overtransfusion and/or wasteful transfusion was recently highlighted by the American Medical Association and The Joint Commission.8 Despite this, transfusion therapy is estimated by some to have saved over one billion lives worldwide since the middle of the 20th century and, by one taxonomy, is thought to represent the single greatest achievement in medical care throughout history.9 In the United States in 2009, a total of 10,877,000 blood donors provided 17,159,000 units of whole blood and RBCs resulting in 14,855,000 transfusions.10 Unfortunately, except for information provided in publications such as that by Tinegate and coworkers11 in this issue of TRANSFUSION and despite an emerging number of databases potentially harboring a plethora of information on the utilization of transfusion therapy (see Table 1), little information is readily available and accessible on how these transfusions are utilized in actual clinical practice over an extended period of time. While there appears to be, at least by the recent Health and Human Services snapshot, an ample supply of blood, changing population demographics,12, 13 especially a dramatic increase in the elderly as the baby boomers come of age,14, 15 and the use of increasingly intense treatment modalities16 threaten the adequacy of the blood supply to meet our future needs. This problem is further highlighted by information demonstrating marked variability in the use of RBC transfusion in patients with similar or even identical diagnoses again raising the concern that patients are unnecessarily and wastefully transfused.17-21 In this regard, the timely study by Tinegate and coworkers, together with two identically designed prior studies by this group concerning the demographics of clinical RBC utilization in Northern England between 1999 and 2009,22, 23 provide powerful data and raise important questions regarding the projected future supply and demand of our RBC resource. Specifically, Tinegate and coworkers, on behalf of the Northern Regional Transfusion Committee, studied regional blood use in the North of England by surveying 18 hospitals serving a total population of 2.8 million. During two separate 2-week periods in the spring and autumn of 2009, they sampled a total of 8025 RBC units from a total volume of 104,511 units issued, for recipient age, sex, and clinical indication. They indicate, based on separate audits throughout the United Kingdom, that this sample is likely representative of the national situation. Regarding age, predicted changes in population demographics in the United States, United Kingdom, and worldwide are impressive with the number of elderly (generally considered to be individuals 65 years of age and older) increasing more rapidly than at any other time in history and the oldest old (generally considered to be individuals 85 years of age and older) representing the fastest growing segment of the population.24, 25 In the United States between 2010 and 2030 those 65 and older are predicted to grow by 79%, with the 75 and older population increasing by 77%. By 2030 it is estimated that one in five Americans will be over 65, an increase from one in eight Americans in 2010. In the United Kingdom, by 2035, nearly one-quarter of the population is projected to be 65 and older. Despite these evolving changes, Tinegate and colleagues demonstrate only a slight increase in the mean age of hospitalized patients receiving RBC transfusions during their decade-long study (the mean age of the recipients was 63.2 years in this study compared to 62.7 and 63.7 years in the 2000 and 2004 studies). While this finding seems, on the surface, to be paradoxical, based on the aforementioned growth of the elderly population and the expectation there would be a commensurate increase in hospitalizations within this age group, it is consistent with data from HCUP (AHRQ), which shows, despite a stable rate of hospitalizations for all age groups combined from 1997 to 2009, a decrease by 8.2% in the number of hospitalizations in patients of 65 and older.7 One possible explanation is that as the “new” baby boomers are just now turning 65 the mean age of the over 65-year-old age group is for now decreasing. This factor combined with a greater emphasis on outpatient care may contribute to a temporary reduction in hospitalizations for the “new” elderly. However, HCUP data also demonstrate that despite this reduction in hospitalizations for the elderly overall, there has been and remains a disproportionately greater number of hospitalizations and use of transfusion therapy by those aged 65 and over. Thus, for example, while people aged 65 and over represented 13% of the US population in 2006, they comprised 34% of hospitalizations26 and accounted for nearly 60% of the total blood transfusion episodes in hospital settings according to a slightly earlier survey.27 Most recent HCUP data demonstrate not only a disproportionately large use of transfusion by hospital inpatients aged 65 and over, but a substantial cumulative growth in the rate of transfusion for all age groups between 1997 and 2009.7 Not surprisingly then, in all three time periods studied by Tinegate's group, those aged 65 and older used more than 50% of the total RBCs transfused. Indeed, the 70- to 74-year-old age group in all periods used proportionally the largest number of RBC units (there was a virtual tie between the 70-74 and 75-79 age groups in the 2000 study). While it is tempting to compare head-on Tinegate's data with that of HCUP, since both are longitudinal and well suited to show trends, it must be noted that Tinegate et al. analyzes each blood transfusion used by inpatients and outpatients (confirmed by personal e-mail communication with J. Wallis, November 26, 2012) while the HCUP data analyze each discharged hospitalized patient receiving blood transfusions. Thus, Tinegate and colleagues are counting the total number of transfusions administered while the HCUP data are counting the total number of inpatients receiving transfusion therapy (transfusion episodes). This difference in data collection clearly limits comparisons. It would be helpful if HCUP was able to incorporate, at some point, a method to tally the actual number of transfusions administered to all patients (including outpatients) rather than just the number of patients transfused. Similarly, it would be helpful if Tinegate's group would also provide data on the number of transfusions administered to each patient—the number of transfusions per transfusion episode distinguishing inpatients from outpatients. Tinegate and coworkers also report from 1999 to 2009 a 20% decline in RBC use accounted for solely by a 40% decline in surgical (excluding obstetrics and gynecology [OB-GYN]) blood use in the 50- to 80-year-old age group. They attribute this decline to the following factors: 1) a nearly 50% decline in first-time coronary artery bypass (CABG) procedures from 1999 (n = 1101) to 2009 (n = 659); 2) a reduction in the number of patients undergoing first-time CABG procedures who are transfused—from 60% in 1999 to 25% in 2009; 3) a reduction in RBC use from 775 units to 608 units in total joint replacement surgery from 1999 to 2009 despite a 150% increase in total knee replacement surgery and a 50% increase in total hip replacement surgery; 4) improved surgical techniques to reduce surgical blood loss (but not otherwise specified); 5) increased use of surgical cell salvage; and 6) the appointment, beginning in 1999, of dedicated transfusion practitioners responsible for blood transfusion education and training focusing primarily on surgical blood use—that is, although further elaboration is not provided, the promulgation of what is certainly a blood management program. By 2009 the transfusion practitioner position was mandatory and 14 such individuals were employed. Some of the responsibilities of these practitioners appear to be delineated in a document issued by the UK Department of Health entitled, “Better Blood Transfusion—Safe and Appropriate Use of Blood.”28 Nonetheless, it seems apparent that this emphasis on surgical blood management was likely triggered, according to the authors, by a redefinition of the “transfusion trigger” to a more restrictive level (hemoglobin [Hb] level of 7-8 g depending on the clinical patient status) provided by the TRICC trial29 and subsequent trials such as FOCUS30 and was certainly responsible, at least in part, for the large reduction in surgical RBC transfusions documented during the study period. In this regard, it appears that, over the past few decades a paradigm shift may be taking place regarding RBC utilization. In 1988, in a consensus conference sponsored by the National Heart, Lung and Blood Institute (NHLBI) on the topic of perioperative RBC transfusion,31 it was estimated that two-thirds of RBC transfusions were utilized by surgical patients. In their first study period Tinegate's group reported that surgical blood use, including OB-GYN, accounted for 47% whereas medical blood use accounted for 52%. However, in the following two study periods medical blood use exceeded 60% of the total with the most recent data showing medical blood use of 64.2% and surgical blood use, including OB-GYN accounting for 35.8%—nearly the inverse of the numbers in the NHLBI Consensus Conference report. Thus, Tinegate and coworkers provide very good news indeed on the actualization of achieving a decrease in surgical blood use during the decade from 1999 to 2009 in the North of England. More data are necessary to ascertain just how widespread this change is throughout the United Kingdom and worldwide and the relationship to implementing a blood management program with its requisite elements.32 Is more reduction in surgical blood use possible? Almost certainly yes. Not mentioned, for example, is the importance of the assessment and treatment of anemia in patients scheduled for surgery. Anemia has been shown to be a risk factor for mortality in elderly patients in general33 and those undergoing surgical procedures.34 Attention to preoperative anemia is highlighted in the UK Department of Health Better Blood Transfusion document by its guideline of “ensuring that mechanisms are in place for the preoperative assessment of patients for planned surgical procedures to allow the identification, investigation, and treatment of anaemia and optimization of haemostasis.” Similarly, in the United States, The Joint Commission finalized a set of voluntary performance measures in late 2010 to monitor and improve patient blood management activities.35 One of these, PBM-06, entitled “Preoperative Anemia Screening,” provides background information on the prevalence of undiagnosed anemia in preoperative elective surgical patients (estimates are highly variable and range from 5% to 75%), its relationship to increased morbidity and mortality, and the need for the “Development of formal protocols for preoperative testing of Hb for potential high-blood-loss elective surgeries (that) could be used to identify and intervene for optimal management of blood resources. Early recognition of anemia offers patients an opportunity to receive the most appropriate transfusion-sparing strategy and avoid the risk of a potential transfusion.” Of note is that the prevalence of anemia increases with age with a weighted mean prevalence of 17%, in 45 combined studies, for those 65 years and older thus prioritizing the need for diagnosis and treatment of anemia in the elderly—whether surgery is scheduled or not.36 The importance of diagnosing and treating anemia before it is severe is highlighted by a report demonstrating that 39.2% of 615 primarily elderly iron-deficient patients were treated with RBC transfusion rather than iron therapy because of the severity of the anemia and/or the presence of comorbidities.37 Also potentially applicable to a reduction in transfusion therapy is reduction in the often considerable phlebotomy associated blood loss38 through more limited blood draws, smaller sampling volumes, or most exciting, the emergence of noninvasive techniques for laboratory analysis.39, 40 Finally, changes in anesthetic and surgical techniques, hinted at by Tinegate and coworkers, are important in determining the need for transfusion. One particularly telling example is an analysis using the Premier database21 of bilateral total knee arthroplasty patients undergoing regional anesthesia compared to general anesthesia (or a combination of the two methods). Patients undergoing regional anesthesia or combined anesthesia had a lower risk of blood transfusion (28.5% vs. 38.0% compared to 44.7%). Perhaps not unexpectedly, contrary to the decline in surgical blood use reported by Tinegate's group over time, there was no change in the medical use of blood in the decade from 1999 to 2009 with the majority of RBC transfusions for medical use, in all three time periods, going to hematology-oncology patients and patients with gastrointestinal hemorrhage. In 2009, 46% of blood used by medical patients was for hematology and oncology diagnoses and 17% was for gastrointestinal hemorrhage. While changes in blood transfusion practices noted by Tinegate and coworkers are likely, in part, the result of application of patient blood management that has had success, less success appears evident to date with its application to the medical use of transfusion therapy. The reasons for this are uncertain. In addition, it is uncertain, over time, how the use of transfusions in inpatients versus outpatients may have changed. Inpatient transfusion utilization is likely, over time, to be increasingly utilized in sicker patients as a growing proportion of particularly elderly patients receive transfusions in the ambulatory setting.41 While Tinegate's group reports on changes in patient populations and transfusion practices, as noted above, they do not address how the blood supply may be increased to meet projected future demands in an increasing elderly population. In this regard, several investigators have suggested expanding blood donation into the geriatric-aged population to match those who are increasingly being transfused.42-45 While this is efficacious in a small percentage of donors—especially those who have been donors before becoming elderly—it may be, in an ecologic sense, relatively inefficient and costly compared to donations from more youthful donors. While safe and helpful and definitely to be encouraged, it is certainly not the final answer in providing for the expected needs of a growing elderly population. What is needed, as pointed out by Kolins and Herron46 is the need for more civic involvement in all aspects of our lives, including blood donation among the younger generations. Another suggestion for potentially increasing the supply of blood, as with other scarce medical resources, is a system of allocation and rationing based in part on age.47 According to the suggested model, age is proposed as one of the components with the optimal age being approximately 20 to 40 years. Although other factors, such as prognosis, are considered, infants and those above 60 years of age would receive the lowest priority. Patients with discordance between their physiologic and chronologic ages would be penalized under such a system raising ethical concerns.25 While blood, in general, has not yet reached a level of scarcity necessitating rationing, the possible shortfall between supply and demand raises this as a future possibility. As the studies by Tinegate and colleagues demonstrate, while there is reason for optimism in the use of our RBC resource—with a significant decline in surgical blood use in the 1999 to 2009 decade—there is much work still to be done in our efforts to further optimize the use of blood. This analysis and enhancements to our operational programs must be done carefully. Overuse and wastage of blood while widespread should be averted while the underuse of blood or its delayed administration, when lifesaving, must be avoided. Improved education for all health care professionals that is data driven is necessary. In this regard, accrediting agencies should consider a training requirement for all health care personal involved with transfusion therapy. Our success in better managing our blood resource is critical as we move forward in the 21st century. None.