IT’S ONE OF THE MOST popular offerings at each Society of Hospital Medicine annual meeting, and for good reason. The “things we do for no reason” sessions (and the Journal of Hospital Medicine columns with the same name) take aim at medical practices that, too often, abide because of inertia and familiarity, not evidence.
During this year’s SHM session, hospitalists Anthony Breu, MD, director of resident education for the VA Boston Healthcare System, and Leonard Feldman, MD, director of the combined internal medicine-pediatrics residency and the Osler Medical Residency Urban Health Track at Johns Hopkins Hospital in Baltimore, each tackled a common practice they believe needs to be scrapped. This year’s things you do for no reason: using furosemide (instead of torsemide) as the first-line diuretic in heart failure and ordering functional cardiac stress tests (instead of coronary CTAs) for patients with chest pain.
Torsemide vs. furosemide
When it comes to the diuretic that doctors prescribe for heart failure patients at discharge, a 2013 study in the Journal of the American College of Cardiology (JACC) found that furosemide accounted for a whopping 87% of those orders. Torsemide came in with less than 1%.
“There’s something about torsemide that leads to less loss of potassium.”
~ Anthony Breu, MD
VA Boston Healthcare System
“There are reasons why grabbing furosemide is the standard of care for many of us,” said Dr. Breu. For one, it has been around the longest, so “we have a comfort level with it.” Interestingly, the ACC/AHA doesn’t endorse any diuretic for monotherapy in heart failure, saying only that furosemide is the most commonly used.
But Dr. Breu marshaled three arguments for why physicians should switch. One, he believes the evidence favors torsemide—and that may be due to his argument No. 2, that torsemide’s “pharmacodynamics and kinetics are better.” And three, he prefers torsemide to furosemide because, he said, the drug “does more than just act as a loop diuretic.”
In presenting the evidence for torsemide, Dr. Breu cited three open-label studies. The first, which is 20 years old, randomized about 200 hospitalized patients with heart failure to either torsemide or furosemide, with results published in the American Journal of Medicine. The trial found that using torsemide resulted in fewer readmissions for both heart failure and cardiovascular causes.
The study also delved into patient symptoms and quality of life. “If you look at follow-up out to eight and 12 months,” Dr. Breu pointed out, “the patients who received torsemide were less fatigued.”
Another randomized study—published in 2003 by the European Journal of Heart Failure—looked at how well torsemide and furosemide improved patients’ New York Heart Association class. Torsemide performed better in terms of improvement by one or more class. Researchers also asked patients if they felt restricted by diuretics in their daily life.
“The torsemide arm reported feeling less restricted,” Dr. Breu noted, although it was unclear whether they were referring to less fatigue or less urination.
The third study, which was a nonrandomized, post-marketing surveillance study, came out in the European Journal of Heart Failure in 2002. It looked at both improvements in the New York Heart Association class and mortality.
With torsemide, “the authors found a mortality benefit overall and a benefit for cardiac mortality, or at least an association,” Dr. Breu pointed out. They also found better improvement of heart association class—and fewer potassium derangements. “There’s something about torsemide that leads to less loss of potassium,” he noted. “The decreased probability of hypokalemia may be why researchers saw a mortality benefit.”
As for why that may be, Dr. Breu discussed the drugs’ pharmacodynamics. Torsemide is “much more reliably absorbed” than furosemide, he explained, “because it’s not affected by food intake.” Torsemide also has less renal clearance, a benefit in heart failure patients who are being diuresed with changes in their GFR. And “the half-life of torsemide is longer, as is its duration of action.” Then there’s this big benefit: Patients need to take it only once a day, making it more convenient.
RAAS and antifibrotic effects
Dr. Breu drew attention to another potential benefit: Torsemide (not furosemide) might positively affect the renin-angiotensin-aldosterone system (RAAS). “There’s an interesting, almost spironolactone-type effect with torsemide that we don’t see with other loop diuretics,” he said. “There are some clinical data supporting that.”
In a 2003 study published in Circulation, researchers looked at patients’ renin levels both before and after they took torsemide or furosemide. “With torsemide, they saw increased renin,” said Dr. Breu. “With furosemide, they did not,” suggesting aldosterone receptor blockade with torsemide.
He also highlighted torsemide’s antifibrotic effects. “To some extent, heart failure is a disease of myocardial fibrosis that occurs as a result of both increased collagen synthesis and turnover collagen.” Torsemide acts, again, “a bit like spironolactone,” which has been shown to mitigate fibrosis to some extent. “That could explain some of the longer-term benefits we see in studies.”
In one study published in JACC in 2004, researchers found that torsemide decreased patients’ serum carboxyl-terminal peptide of procollagen type 1 (a molecule correlated with myocardial fibrosis) while furosemide did not.
“I interpret these two impacts together—possibly positively affecting RAAS and having antifibrotic effects— as torsemide being something more than just a loop diuretic,” said Dr. Breu. “It has other effects that positively affect heart failure.”
He strongly recommends choosing torsemide when starting a new loop diuretic in hospitalized heart failure patients. As for switching patients to torsemide who are admitted on furosemide, “I’d certainly discuss that with their primary care physician or outpatient cardiologist, particularly when patients are having multiple readmissions for heart failure.” A definitive answer may come from the TRANSFORM-HF study, which is comparing torsemide to furosemide in about 6,000 heart failure patients.
Functional cardiac stress tests vs. coronary CTA
According to Dr. Feldman, the reason why physicians order functional cardiac stress tests—instead of coronary CTAs—for patients with stable chest pain is, again, familiarity.
“We’ve been using functional stress tests for a long time,” Dr. Feldman pointed out. “For some of us, when we trained in medicine, coronary CTA wasn’t even an option.” Every year in the U.S., noted a 2020 study in JAMA Network Open, more than 8.7 million patients undergo noninvasive diagnostic testing for coronary artery disease (CAD), at a cost of about $15 billion.
“About two-thirds of those are nuclear imaging, and one-third is stress echo,” said Dr. Feldman. “Maybe 5% is coronary CTA. I would say that coronary CTA probably should be done most of the time.”
“Changes in diagnosis and probabilities with coronary CTA led to changes in management.”
~ Leonard Feldman, MD
Johns Hopkins Hospital
Here’s his main argument against functional cardiac stress tests: They tell you only if a patient has obstructive disease—”and they may not even do that very well,” he pointed out. A 2010 study in the New England Journal of Medicine looked at 400,000 U.S. patients who had no known CAD and underwent elective catheterizations between 2004 and 2008.
“Most of them had some sort of noninvasive test, most of which were not coronary CTAs,” Dr. Feldman said. Researchers, who defined obstructive disease as 50% of the left main artery or 70% or more of the epicardial vessel, compared test results to what they actually found in the cath lab. They also looked at patients’ Framingham risk scores and their angina symptoms.
Their findings: Among patients with high-risk Framingham scores, 55% who had negative stress tests were found to have obstructive disease. And “even in patients with a low-risk Framingham score,” Dr. Feldman noted, “10% with a negative stress test still had obstructive coronary disease. That’s pretty worrisome.”
As for patients with classic angina symptoms and negative stress test results, 40% were found to have obstructive disease. That was also the case among 18% of patients who had atypical symptoms and a negative stress test. And researchers found that many patients with positive stress test results actually did not have obstructive disease.
A better diagnostic tool
Coronary CTAs, on the other hand, do a better job with diagnosis, “both of obstructive disease and nonobstructive atherosclerotic disease,” Dr. Feldman said. He discussed two big studies that compared coronary CTAs to functional stress tests.
The first is the PROMISE study, with results published in 2017 in Circulation. More than 9,000 patients with stable chest pain and intermediate risk for obstructive disease were randomized to either a functional stress test or a coronary CTA, with patients followed for more than two years.
Based on their test results, patients were categorized in one of four buckets: normal (no disease), or mildly, moderately, or severely abnormal.
“We see really no difference in performance between functional stress tests and coronary CTA in identifying patients with severely abnormal disease,” said Dr. Feldman. “Where we see big differences is in the mildly abnormal and normal groups.” Within the coronary CTA group, 55% of the patients tested ended up being classified as mildly abnormal, a classification given to only 10% of the patients in the functional stress test group.
Many more patients—80%—who received functional stress tests were instead classified as normal vs. only 33% in those given coronary CTA. But the event rate (defined as death, MI or unstable angina hospitalization) among the normal group as determined by functional stress testing was more than 2%—the same event rate found among patients considered mildly abnormal per functional testing.
Among those given coronary CTA, however, less than 1% of the normal patients had an event, while events occurred among 3% of those classified as mildly abnormal. “When a functional stress test classifies your patient as normal,” said Dr. Feldman, “that patient’s event rates are still much higher than when classified as normal as per a coronary CTA.” When coronary CTA classifies patients as normal, he pointed out, they actually are—which is not true for functional stress tests.
Targeted treatments, fewer tests
The second study is the SCOT-HEART study, which took place in the U.K. and was published in 2015 by The Lancet. About 4,000 patients with suspected stable coronary heart disease were randomized to either standard care or standard care plus a coronary CTA. As Dr. Feldman explained, the patients’ physicians were also asked when patients were first enrolled how likely they thought patients were to have coronary heart disease (CHD) and angina.
Six weeks later, doctors were again asked about patients’ likelihood of CHD and angina. “What you see is that having a coronary CTA done in those first six weeks really helped doctors reclassify patients,” he pointed out. “Instead of only 10% of the patients with a pretest probability of ‘yes’ for coronary heart disease with a coronary CTA, physicians the second time around thought 23% of them did.” By contrast, among the standard care group, the pre- and posttest probabilities for CHD were exactly the same.
As for the no disease group, doctors of patients who underwent coronary CTA predicted the first time around that only 6% would fall into that group; the post-test probability at six weeks for that group had risen to 15%. Similar post-test reclassifications were also found in terms of angina with coronary CTA.
“Here’s the really important part,” said Dr. Feldman. “These changes in diagnosis and probabilities with coronary CTA led to changes in management.” Forty percent more of the patients who received coronary CTA were started on preventive therapy compared to those receiving only standard care. And among the group found with coronary CTA to have a post-test probability of no disease, “therapies and tests were cancelled.” Patients were followed for five years for primary endpoints of death from CHD or MI; the event rate among the standard care group was 3.9% vs. 2.3% among those with coronary CTA.
For Dr. Feldman, using coronary CTAs means that more patients are prescribed the medications they need. And “that makes a difference in their outcomes.”
The bottom line: Coronary CTA “is finding a lot of the atherosclerotic disease that we didn’t know was there.” While doctors certainly care about obstructive disease, he added, “we need to care even more about nonobstructive disease because those are the patients who actually have events. Most cardiovascular death events occur among individuals without angina or evidence of ischemia.”
Guidelines in the U.K. hold that coronary CTA is the first-line tool for patients presenting with new onset chest pain due to suspected coronary artery disease. “The U.K. folks also argue that coronary CTA is cost effective because it helps prevent downstream tests, like MRIs and PET tests, that patients don’t need if you start them on the right medications.”
While radiation is a downside of coronary CTA, “it’s probably not higher than background radiation we all get over the course of a year when performed at an academic center.” But coronary CTA isn’t for everyone. “This is not a test for patients with acute coronary syndromes, or those who’ve had previous cardiac interventions like stents,” he said. You also shouldn’t use coronary CTA for patients with tachycardia or those for whom you can’t get good imaging due to extensive coronary calcifications, irregular heartbeats or significant obesity.
“It’s not a perfect test,” Dr. Feldman concluded. “But it’s a much better test overall than functional stress tests.”
Phyllis Maguire is Executive Editor of Today’s Hospitalist.
Published in the July/August 2021 issue of Today’s Hospitalist