Home Feature Medical practices that should be retired

Medical practices that should be retired

The case against routine blood cultures for SSTIs and doing away with dual therapy

September 2022

THE “THINGS WE DO FOR NO REASON” SESSION at the annual Society of Hospital Medicine conference always draws big crowds, and this year’s was no exception. During each presentation, speakers highlight common medical practices that persist long after evidence debunks their effectiveness.

That’s certainly true for the two practices discussed this year. In the first, Anthony C. Breu, MD, director of medical resident education at the VA Boston Healthcare System, traced how both the thinking and evidence have evolved against prescribing dual therapy (an anticoagulant plus an antiplatelet) for patients with stable coronary artery disease (CAD) and A fib.

And John Stephens, MD, a med-peds hospitalist at the University of North Carolina at Chapel Hill, honed in on a second practice that should be retired: ordering routine blood cultures for patients with skin and soft tissue infections (SSTIs). Such cultures, Dr. Stephens held, not only fail to offer any benefits, but they leave patients open to higher costs and potential harms.

Dual therapy in patients with stable CAD and A fib

Dr. Breu kicked off his portion of the session with this case: A 72-year-old man comes to the hospital with a GI bleed. The patient has a history of A fib (and is on apixaban) and of an acute MI and PCI with stenting two years ago, for which he’s taking aspirin. You initially hold both agents, although you restart the apixaban once the bleed subsides. But what should you do at discharge with the aspirin?

“Instead of dual therapy in patients with both A fib and CAD, what we should be doing is anticoagulant monotherapy.”

Anthony C. Breu, MD
VA Boston Healthcare System

“I’m going to argue that you should have no plan to restart,” Dr. Breu said. “Instead of dual therapy in patients with both A fib and CAD, what we should be doing is anticoagulant monotherapy.”

As Dr. Breu explained, heparin was introduced in the 1930s and 1940s as a treatment for coronary artery disease. Other anticoagulants also emerged and generated a lot of enthusiasm—as well as a lot of research, with more than 30 randomized trials between the 1940s and the 1970s.

Why so many? “For one, some of the trials weren’t very good,” he said. “But the main concern, and this should be no surprise, was bleeding.” Even the WARIS 1 trial, which was published in the New England Journal of Medicine (NEJM) in 1990 and was “strikingly positive” about the benefits of warfarin in reducing mortality and recurrent MI, didn’t lead to warfarin becoming the standard of care.

A big reason why, again, was the heightened risk of bleeding—and “no one likes to check INRs.” Another reason: In the 1970s and 1980s, “we begin to learn about the role of platelets in acute coronary syndromes.” That culminated in the massive ISIS-2 trial with more than 17,000 patients who received aspirin or placebo for five weeks. The study was published by The Lancet in August 1988.

An annual series, 2021 topics covered heart failure and the best test for chest pain. Read here.

“ISIS-2 is why we give all our patients aspirin,” said Dr. Breu, noting that the results were “resoundingly positive” in terms of aspirin reducing vascular and all-cause mortality as well as nonfatal MI. “With this study, aspirin supplanted warfarin because there was no increased bleeding risk.” In patients with stable CAD, aspirin—and later clopidogrel, another antiplatelet—emerged as the safer alternative.

Evolved thinking
“But now I’ve got to fold in A fib because the patient in our case has both conditions,” said Dr. Breu. For many years, the standard thinking was that CAD and A fib were separate problems that required separate therapies. He admitted to that thinking himself as a resident and an early career hospitalist.

Here’s how he described that conventional wisdom: “You had these conditions—CAD, AMI—that were platelet problems, so they required platelet therapy. You also had another silo of thrombotic conditions, A fib and DVT, that were issues of the coagulation cascade and required an anticoagulant.” It followed that patients with both conditions “needed both therapies.”

But in fact, “this is too simplistic” of an understanding of pathophysiology, and thinking has evolved with data now supporting a more nuanced view. “We now realize,” Dr. Breu said, “that all four conditions—AMI, CAD, A fib and VTE—are thrombotic conditions that all activate both platelets and the coagulation cascade. We ought to look at what the evidence tells us about whether or not we should use antiplatelet or anticoagulation therapy, or both.”

The first hint that adding aspirin to an anticoagulant provides no benefit came in the WARIS II trial, published in NEJM in September 2002. “Until then, warfarin and aspirin were each up against placebo,” he said. “Here, they compared warfarin and aspirin head-to-head in post MI patients and also combined them to see what happens.” The results: Warfarin beat both aspirin alone and the combination of both agents together.

No additional benefits
Dr. Breu then mentioned three studies that looked at patients with both stable CAD and A fib. The first, a cohort study published by Circulation in 2014, found that both clopidogrel and aspirin monotherapy were inferior to vitamin K antagonist monotherapy. “When you add antiplatelets to VKA, you get no additional benefit for mortality,” he pointed out. “You also get no additional benefit for thromboembolism, MI or coronary death. This is kind of the theme.”

The next two studies were randomized trials—although Dr. Breu said that the first, published by Circulation in January 2019, was “flawed” due to being “wildly underpowered.”

“They wanted to enroll 2,000 patients but enrolled fewer than 700—because nobody wanted to take their patients off aspirin.” Because the study was underpowered, the results didn’t establish the noninferiority of anticoagulant monotherapy vs. dual therapy on efficacy alone, although it did meet noninferiority when safety was factored in as well.

Interestingly, 25% of the patients in that study were taking a DOAC for anticoagulation. “I’ve been talking a lot about warfarin because that’s where the data are,” said Dr. Breu. “But this is 2022, and we use DOACs for A fib.” The final study he highlighted, the AFIRE study published in NEJM in September 2019, looked at patients taking either rivaroxaban alone or rivaroxaban plus an antiplatelet.

Researchers actually showed superiority of rivaroxaban monotherapy in terms of a primary endpoint that included fewer deaths, with no increased risk of MI. “They also showed reductions in major bleeding,” Dr. Breu pointed out. “These are the best data we have in favor of monotherapy in this exact population, and it was a pretty darn positive study.”

Waiting on updates
Dr. Breu noted that while guidelines support monotherapy, including CHEST guidelines issued in 2018, the ACC/AHA guidelines still have not been updated. He suspects a similar evolution in thinking is taking place with VTE, but “I just don’t have those data to show you, and I’m not sure they exist yet.”

His recommendation: For patients who come in on dual therapy with a bleed, “stop the antiplatelet and don’t restart it.” He appreciates that many hospitalists will remain cautious until updated ACC/AHA guidelines provide some backing, “but I anticipate those are coming.” In the meantime, patients in whom dual therapy is “absolutely necessary” include those less than a year out from PCI or acute MI, as well as patients with mechanical valves.

Routine blood cultures in SSTIs

In his portion of the presentation, Dr. Stephens pointed out how common SSTIs are, with more than 14 million cases of cellulitis in the U.S. every year leading to about 650,000 hospitalizations. Studies indicate that blood cultures are run in anywhere between one-third and three-quarters of those patients.

Current IDSA guidelines recommend against routine blood cultures for nonpurulent SSTIs like cellulitis with some exceptions; those include unusual mechanisms of injury, such as infected animal bites, and cases of immunosuppression, among others. “But the guidelines are less clear about what to do with purulent infections and with severe nonpurulent ones in terms of blood cultures,” Dr. Stephens said.

“Guideline uncertainty tends to breed practice variation.”


John Stephens, MD
University of North Carolina at Chapel Hill


The guidelines define severe and moderate infections “largely by SIRS criteria, so signs and symptoms of systemic infection.” And while the guidelines recommend “culture and sensitivity” for severe and moderate purulent infections, “it’s not clear if that means a blood culture or culturing what someone just debrided,” Dr. Stephens said.

Severe nonpurulent infections have the same “C&S” notation in the IDSA guideline. “Again,” he said, “it’s not entirely clear what they’re recommending be cultured. That kind of guideline uncertainty tends to breed practice variation.”

Low culture yields
To bolster his case against routine cultures, Dr. Stephens discussed four studies. The first, which was the largest, looked at nonpurulent infections in one facility over a three-year period. Published in 1999 by Clinical Infectious Diseases, the study included about 750 patients with cellulitis, three-fourths of whom had cultures drawn. Out of those, only 11 patients (2%) were found to have pathogens—while 20 cases (3.6%) had false positives due to contaminants.

Three studies that are more recent “inform the era in which we practice, where MRSA is everywhere,” he noted. One, which was published in 2017 in the American Journal of Emergency Medicine, looked at both purulent and nonpurulent infections among about 250 patients presenting to a single center in San Francisco. “One-third had blood cultures drawn,” Dr. Stephens said. “One caveat of this study is that about 40% of the patients in it used IV drugs.” Six patients (7%) had positive cultures, all Staph aureus. Four of those six injected drugs, and none of the positive cultures resulted in a change in antibiotics.

IF YOU TEND to order stress tests or DVT prophylaxis for low-risk patients just to be safe, you were the target of our last article on Things we do for no reason.

The next study, also a single-site trial, was published in 2018 in JAMA Internal Medicine. The authors did an analysis of both routine blood cultures and imaging in cellulitis patients. They also generated some cost estimates for imaging and blood cultures in the U.S. for SSTIs by extrapolating from their data.

Among the 180 patients in the study, “researchers also looked at how often they met IDSA guidelines for culture, based on their history and vital signs,” Dr. Stephens pointed out. “While only 10% met guideline criteria, about one-third had cultures drawn—and only one was positive.”

No guideline indications
The last trial that Dr. Stephens highlighted was published in 2021 by the American Journal of Emergency Medicine and included about 140 observation patients with SSTIs, both purulent and nonpurulent. “The authors looked at patients’ SIRS criteria to see if they met the CMS sepsis guidelines for cultures,” he noted. “They also looked at whether patients met IDSA indications for culture.”

Again, only 10% met IDSA criteria, but more than half had cultures drawn, Dr. Stephens said. “So meeting IDSA guideline indications did not predict if patients were cultured or not,” he explained. Interestingly, “the number of SIRS criteria didn’t predict culture either. In terms of who got cultures, the authors concluded that it appeared to be based on practice patterns and variation.”

Only 3% of those with cultures drawn had true pathogens, while another 3% uncovered contaminants. The upshot of all four studies, said Dr. Stephens, is that the yield from drawn cultures is always in the single digits—and in the mostly low single digits at that. The other big reveal: Cultures result in at least the same number of contaminants as true positives.

The downsides
“So what’s the big deal?” Dr. Stephens asked. “We do other things that yield only single-digit outcomes, so what are the downsides?” Cost is an obvious one. In the 1999 study, for instance, the authors estimated that the 700 cultures drawn cost $36,000. “That’s in 1999 dollars,” he said, “so it would be more now.”

The 2021 study had a smaller cohort and put the cost of cultures drawn against guidelines at more than $26,000. As for the 2018 study that estimated the annualized cost of both cultures and imaging for cellulitis in the U.S., the total came to $226.9 million.

Then there are downsides and potential harms of contaminant false positives. A study published in 2011, for example, found that contaminant cultures led to patients staying in the hospital more than five days longer and incurring additional costs of more than $10,000.

Another study, also published in 2011, looked at 85 patients with contaminant cultures to see what happened to them in the hospital. “Not surprisingly,” Dr. Stephens said, “60 had another blood culture drawn.” More than one-quarter ended up with a course of antibiotics, while close to one-quarter had serum vancomycin levels collected. And two patients were actually readmitted after discharge.

As for children, who Dr. Stephens also treats, “the data are more robust.” He was the lead author of a study in Hospital Pediatrics in April 2020 that looked at how often blood cultures are drawn for SSTI patients in more than 40 children’s hospitals. Those figures ranged from a low of about 30% to 80%.

“Hopefully, I’ve convinced you that even with the murkiness of the IDSA guidelines, most patients don’t meet indications for culture,” Dr. Stephens said. And even in the MRSA era, “there’s really no increased evidence that cultures help management, compared to the non-MRSA era.”

His conclusions: “The yield of blood cultures for nonpurulent—and I’d argue for purulent as well—typically do not change management,” he said. “And contaminant cultures, which are found at least as frequently as true positives, have associated costs and harms.”

Phyllis Maguire is Executive Editor of Today’s Hospitalist.

Published in the September/October 2022 issue of Today’s Hospitalist

Notify of
oldest most voted
Inline Feedbacks
View all comments
Alexandre Andrianov MD, MBA
November 2022 8:32 am

You still need cultures in SSTI if
1. Innunosupression
2. Animal bite
3. Exposure to 💧

Jensa Morris - LinkedIn
Jensa Morris - LinkedIn
November 2022 8:27 am

Time to evolve our thinking on concurrent anticoagulant and antiplatelet therapy in patients with CAD 💔 and 💓 AF. Really well explained here.