Clinical Laboratory Strategies: October 13, 2011
Strategies logo

Exploring the Link between Diagnostic Blood Loss and Hospital-Acquired Anemia

In Acute MI Patients, Risk Rose with Increasing Volume of Blood Drawn
By Genna Rollins

Reducing complications from care provided in the hospital has been a major goal of patient safety groups and other organizations. One potential problem associated with hospitalization that has not been studied extensively is hospital-acquired anemia, which may be due in part to blood loss from lab testing. Now, however, researchers have explored the role of blood loss in patients hospitalized with acute myocardial infarction, and their findings are presented in this issue of Strategies.

Hospital-acquired anemia (HAA), which develops in patients who have normal hemoglobin levels upon admission to the hospital, is a multifactorial complication of treatment. Although gastrointestinal bleeding or bleeding from procedure sites may contribute to HAA, several other factors could be at play, including blood loss from phlebotomy and impaired hematopoietic response. Based upon their observations in clinical practice, researchers at Saint Luke’s Mid America Heart and Vascular Institute in Kansas City, Mo. became interested in exploring the role of diagnostic blood loss as a risk factor in HAA among patients hospitalized with acute myocardial infarction (AMI). 

“Our preliminary work in the past showed that it’s common for patients who come into the hospital with acute MI and have normal hemoglobin at the time they’re admitted to develop new-onset anemia,” explained senior author, Mikhail Kosiborod, MD, a cardiologist at Saint Luke’s and an associate professor of medicine at the University of Missouri-Kansas City. “We demonstrated in our original paper that not only was hospital-acquired anemia common, but it was associated with adverse outcomes like persistent anemia, increased mortality, and worsened health status compared to patients who didn’t develop anemia in the hospital. Interestingly, the majority of patients who developed hospital-acquired anemia did not have a documented bleeding event.”

Building on their original findings, the investigators recently conducted a study that looked at the association between diagnostic blood loss and HAA in patients hospitalized with AMI (Arch Intern Med 2011; doi:10.1001/archinternmed.2011.361). Kosiborod and his colleagues accessed a database with de-identified data from 17,676 patients hospitalized with AMI in 57 hospitals. None of the patients had anemia at the time of admission. The researchers defined anemia as the nadir hemoglobin level during hospitalization, with moderate-to-severe anemia having a nadir <11.0 g/dL. The database included the time and date of every phlebotomy event, along with results from lab tests collected with each blood draw. The researchers made several assumptions about the volume of blood drawn from each phlebotomy event, including that all tests that could be processed from a particular type of specimen tube were run from a single tube, that only the minimum amount of blood needed to run each test was drawn, and that no blood was wasted at the time of blood draws. The investigators assumed that 5 mL of blood would be drawn for hematology tubes, 4.5 mL for coagulation tubes, 5 mL for chemistry and miscellaneous tubes, 2 mL for arterial blood gas tubes, and 10 mL for blood cultures.

The authors found that 20% of AMI patients developed moderate-to-severe HAA, and that the estimated mean blood loss from phlebotomy was nearly 100 mL higher among these patients in comparison to those who did not develop HAA. Particularly large blood loss from phlebotomy >300 mL was observed in 12.5% of patients who developed moderate-to-severe anemia versus just 0.8% of those who did not. The relationship between greater diagnostic blood loss and increased risk for moderate-to-severe HAA persisted after multivariate adjustment, with an 18% higher risk of developing HAA for every 50 mL of blood drawn. The researchers also found significant variation in estimated mean diagnostic blood loss between hospitals, particularly in patients who developed moderate-to-severe HAA, with a range from 119.1 mL to 246.0 mL.

The findings add to the body of evidence around HAA and should be a call-to-arms among laboratorians and clinicians, according to Timothy Hamill, MD, professor and vice chair of the University of California, San Francisco (UCSF) department of laboratory medicine and director of UCSF Clinical Laboratories. “Laboratorians have long recognized that iatrogenic anemia in hospitalized patients is a real thing, but this is probably the first study I’ve seen in a homogenous group across a large number of hospitals. The study shows that it’s highly likely that phlebotomy and lab testing are contributing to the development of anemia in these patients.” Hamill was not involved in the study.

Kosiborod noted that some observers have questioned how the average estimated phlebotomy draw of 173.8 mL could cause HAA. “If you look at the estimate of average diagnostic blood loss in our study, it’s a little less than half a pint. So I’ve been asked, if people who donate blood give a pint or even more and they don’t become anemic, why was this an issue in our study population? My answer is that these are not healthy volunteers. They’re people with an acute illness that can impair their bone marrow’s ability to produce red blood cells. So a much smaller volume of diagnostic phlebotomy may contribute to the development of anemia in these patients versus healthy individuals.”

Hamill added that the authors’ assumptions about blood loss were very conservative and that in most patients, the amount of blood actually drawn probably was much higher. “They assumed for each blood draw only the minimum amount of blood requested for the test was drawn. That probably underestimates the amount drawn since we teach our phlebotomists to fill the tubes,” he observed.

Both Kosiborod and Hamill agreed that the findings present an opportunity to identify practice changes that could lead to less diagnostic blood loss. “I think up until our study came out, there was very little awareness, at least among cardiologists, about how much the amount of blood we draw in the hospital actually matters,” said Kosiborod. “If clinicians have awareness that this is an issue, then that alone can have an impact. It may make one pause and say, well maybe we don’t need to draw blood on patients every day of their hospitalizations. When they’re critically ill, no question, but maybe by the time they get better and are about to go home we don’t need to keep drawing a CBC and chemistry panel every single day.”

Hamill and his colleagues at UCSF have taken steps to cut back on diagnostic blood loss by, for example, presenting on the topic at ground rounds. “We’ve been working with various physicians, like our anesthesiologists and hospitalists, so that if a patient’s labs have been normal for a couple of days, they stop ordering those tests day-after-day,” he said. “With busy clinicians, it’s easy to write lab orders for multiple days and then forget about discontinuing them. But in stable patients, we really need to think about whether a test is needed or if it will just be another data point in the chart that doesn’t impact care.”

In addition to giving more thought about the frequency of lab tests, Kosiborod suggested that labs might consider using smaller volume specimen tubes. A sensitivity analysis he and his coauthors performed found that if low-volume pediatric tubes had been used instead of standard tubes, the estimated diagnostic blood loss volume would have declined dramatically to a mean of just 65.3 mL among patients with moderate-to-severe HAA. However, Hamill cautioned that this would need to be a longer-term solution. “There’s virtually no automated testing system that can handle those tubes. They virtually all have to be handled manually. You can stick them in a plastic sleeve that makes it taller and seem like an adult sample, but you can’t put them through a cap piercer or spin or aliquot them in any of the major automation systems,” he cautioned.

Hamill also pointed out that while point-of-care testing typically requires much less blood volumes than lab-based testing, its comparatively high cost makes it impractical to implement for the entire hospital stay of a broad class of patients like those with AMI. He suggested that a more promising avenue would be to learn about practices at hospitals with lower average diagnostic blood loss volumes.

Kosiborod emphasized that the best solutions would come from close laboratorian-clinician collaboration. “We’ve worked really closely with our laboratory staff to try to figure out where we go from here, and I can tell you the entire staff, from technicians to the lab director, have been extraordinarily supportive, and understand that this is an issue we need to do something about,” he said. “So the best advice I can give lab professionals is to try and find clinical champions that understand that this is an important issue that needs to be addressed and try to come up with some solutions on how to minimize phlebotomy volumes.”

Join Us On LinkedIn Join us for a discussion of this article, and interact with your peers, on LinkedIn

Rate this page:       
Page Access: