In Response

Our group read with great interest the letter by Mazzeffi et al,1 who deliver several thoughtful discussion points in response to our recent meta-analysis involving the use preoperative erythropoietin (pEPO) and its impact on perioperative allogeneic blood transfusion.2 The authors cite variation in institutional transfusion practice, inconsistent medication reimbursement and ubiquitous present-day use of antifibrinolytics and point-of-care testing as arguments against the reflexive adoption of pEPO in cardiac surgery. They suggest that patients and service lines instead should be scrutinized based on individual risk and fiscal responsibility before using the medication as a route transfusion prevention strategy. Taken in isolation, these arguments are both rational and pragmatic, introducing levity as it pertains to medical care costs and healthy skepticism that the available evidence reflects current practice. That is why the results of a recent article on the ultra-short-term treatment of patients with iron deficiency or anemia in cardiac surgery are so compelling.3 The randomized study found that an anemia bundle including pEPO, given the day before cardiac surgery, led to a significant reduction in allogeneic transfusions and a serial increase in postoperative hemoglobin (Hb). Based even on their modest definition of anemia (Hb < 120 g/L in women, Hb < 130 g/L in men), nearly one-third of people undergoing cardiac surgery may be candidates for the treatment. As evidence mounts for the bundled use of pEPO among transfusion prevention strategies, it becomes increasingly more difficult to establish the indications for individual constituent parts as opposed to the bundle as a whole. In example, Enhanced Recovery Programs (ERPs) have revolutionized perioperative practice through the bundled application of process measures based on the principle of aggregation of marginal gains, thereby improving intervention compliance and optimizing the value of perioperative care.4 Rather than determine if an individual patient should receive scheduled multimodal analgesia, lung protection ventilation, or deep vein thrombosis prophylaxis, efforts are instead focused on identifying those who should not. ERPs thus facilitate a less heralded, but equally important, benefit in not only reducing undesired interprovider variations but by selecting for desired ones. Predictable and consistent patterns of practice enhance one’s ability to identify contraindications, isolate and mitigate barriers to implementation, and realize new program and budget sustaining strategies. The coadministration of transfusion prevention strategies likely carries similar benefit. We contend that programs should be designed to provide patients with all reasonable prevention therapies unless otherwise contraindicated, not the other way around. This does not expressly contradict the motion by Mazzeffi et al,1 but it does represent an important difference in therapeutic philosophy. When systems are designed to automatically apply bundled best practice strategies, providers are no longer charged with selecting patients on which to intervene but instead are tasked with the identification of necessary deviations from that practice. The debate is less about whether pEPO and similar strategies should be provided, but in establishing circumstances where they should not. We believe, therefore, that pEPO is both an evolving standard of care and part of an aggregation of marginal gains. Michael C. Grant, MDLaeben Lester, MDBrian C. Cho, MDDepartment of Anesthesiology and Critical Care MedicineThe Johns Hopkins Medical InstitutionsBaltimore, Maryland[email protected]