Home Cardiac Care Tuning up your heart failure patients at discharge

Tuning up your heart failure patients at discharge

Are you choosing the right therapies?

February 2019

DISCHARGING heart failure patients is pretty much a bread-and-butter job for hospitalists. But are you really optimizing patients before you send them home?

According to cardiologist Ronald Witteles, MD, a heart failure specialist at Stanford University Medical Center in Stanford, Calif., doctors can do a lot to improve these patients’ care, but they often don’t. Instead, many physicians rely on standard treatment strategies when evidence for better therapies exists.

Speaking at a hospital medicine conference last fall at the University of California, San Francisco, Dr. Witteles made the case for switching patients to more effective medications if they’re not already taking them and discontinuing some common prescriptions. “Don’t just diurese them and send them out the door,” he said.

“Which systolic heart failure patient should we not be switching to sacubitril-valsartan? Because probably most of them should.”

~ Ronald Witteles, MD Stanford University Medical Center

He presented this very common scenario: You admit a 78-year-old man with nonischemic NYHA class II-III heart failure, a left bundle branch block, a left ventricular ejection fraction of 25% and a 3+ mitral regurgitation. He’s gained seven kilos over the last three weeks and now has dyspnea on exertion.

The medications the patient comes into the hospital on include:
✓ Furosemide 40 mg bid.
✓ Lisinopril 10 mg qd.
✓ Metoprolol tartrate 25 mg bid.
✓ Digoxin 0.5 mg qd.
✓ Rosuvastatin 10 mg qhs.
✓ Potassium chloride extended release 20 mEq qd.

While you double his furosemide at discharge, the patient otherwise goes home on the same medications he came in with. Could you do better? According to Dr. Witteles, the answer is a definite “yes.”

SWITCH ACEI/ARB to sacubitril-valsartan
Dr. Witteles first zeroed in on the patient’s ACE inhibitor, lisinopril. In 2016, the ACC/AHA/HFSA issued guidelines that contained this class I recommendation: All NYHA class II-III patients with heart failure and reduced ejection fraction should be switched from an ACE inhibitor or ARB to sacubitril-valsartan.

That recommendation was based on results from the PARADIGM-HF trial, published in the Sept. 11, 2014, New England Journal of Medicine. Randomizing more than 8,400 patients to either sacubitril-valsartan or enalapril, researchers found that sacubitril-valsartan demonstrated an absolute risk reduction of 4.5%.

Even more impressively, said Dr. Witteles, “death from any cause was down 2.8%, with a number needed to treat of 35. Instead of asking which rare patient should we start this on, we should be asking this: Which systolic heart failure patient should we NOT be switching, because probably most of them should.”

But he also underscored two important caveats. One, sacubitril-valsartan causes a greater blood pressure drop than an ACE inhibitor or an ARB alone, so “a patient sitting there with a systolic of 80 is probably the wrong patient to switch.”

And patients taking an ACE inhibitor must be off that medication at least 36 hours before starting sacubitril-valsartan. “There’s an angioedema risk with ACE inhibitors that isn’t there with ARBs,” Dr. Witteles explained. While patients in the outpatient setting can just stop their ACE inhibitor and start sacubitril-valsartan 36 hours later, hospitalized patients aren’t as stable.

“You may want to switch them first to an ARB for 36 hours before starting sacubitril-valsartan,” he said. “The key is remembering that you can’t do a direct switch.”

SWITCH metoprolol to carvedilol
Dr. Witteles next turned to the patient’s beta-blocker, metoprolol 25 mg BID. He would instead recommend carvedilol 6.25 mg BID, based on the results of the COMET trial published in 2003. That trial randomized more than 3,000 patients with systolic heart failure to either carvedilol or metoprolol.

Researchers found “a pretty big treatment effect for carvedilol in terms of improved mortality,” he noted. “Unless they have a very good reason not to, the vast majority of patients with systolic heart failure should be on carvedilol, not metoprolol.” He added that both drugs are generic.

LOWER the digoxin dose
Then there’s the digoxin at 0.5 mg per day, which Dr. Witteles argued shouldn’t be discontinued. The DIG trial, published back in 1997, randomized about 6,800 patients whose ejection fraction was 45% or less to either digoxin or placebo. Researchers found “absolutely no difference” between the two in all-cause mortality, he pointed out.

But there was “a massive difference” between the two groups in heart failure hospitalizations. Those same results were borne out in a 2004 Cochrane review that looked at patients already on digitalis who were either continued or stopped.

“Digoxin doesn’t improve survival, but it does keep patients out of the hospital and it makes them feel better,” he said.

At the same time, Dr. Witteles recommended lowering the patient’s digoxin dose. Why? About 10% of the overall population have the bacteria eggerthella lenta in their GI tract, and “its sole purpose, as far as I can tell, is to break down digoxin.” Physicians targeting digoxin levels for heart failure or A fib patients “inevitably end up putting patients with this bacteria on an extremely high dose because the digoxin is being broken down.” When the patient gets antibiotics—”which all patients eventually do”—that patient gets “dig toxic.

“Nobody should require a maintenance dose higher than 0.25 mg a day, particularly when you’re using digoxin for heart failure and not rate control, he noted. “If patients do, that means they have this bacteria.”

That led Dr. Witteles to make this recommendation: Don’t target digoxin levels. “If you do, you risk running patients at a high level that seems normal, but then shoots into the toxic range once they get antibiotics.” There are only two reasons why doctors should be checking digoxin levels, he pointed out: “You clinically suspect toxicity, or you want to see if patients are adhering to their medication.”

Instead, estimate a patient’s digoxin dose according to two factors: whether or not the patient is on amiodarone, which raises digoxin levels by about 50%, and the patient’s GFR. “You shouldn’t use a dose higher than 0.125 mg per day for patients with a GFR less than 50,” he said. When patients’ GFR is less than 25 or 30, “it probably should be avoided all together.”

STOP the rosuvastatin
On the next medication considered—rosuvastatin 10 mg qhs—Dr. Witteles had this strong opinion: Just get rid of it in patients with nonischemic heart failure.

Two 10-year-old trials support ending its use. In the 2008 GISSI-HF trial, nearly 4,600 patients with either ischemic or nonischemic heart failure (and an ejection fraction of less than 40% or one greater than 40% but a heart failure hospitalization within the previous 12 months) were randomized to rosuvastatin 10 mg daily or placebo. Endpoints included survival, mortality and cardiovascular hospitalization.

The results? “Absolutely nothing” in terms of a rosuvastatin benefit, he said. “Not even a hint.”

The 2007 CORONA trial randomized patients with ischemic systolic heart failure to rosuvastatin 10 mg daily or placebo. But the catch in this industry-sponsored trial, Dr. Witteles pointed out, was that it looked at a vascular composite endpoint of cardiovascular death or nonfatal MI or stroke.

“That’s the really devious part, not using any heart failure endpoints and taking only patients with ischemic cardiac myopathy,” he said. Even with those very loaded odds, “they still couldn’t hit statistical significance with rosuvastatin.”

The take-away? Statins don’t benefit patients with nonischemic heart failure. “There’s absolutely no reason for these patients to be on them,” Dr. Witteles said. (That’s certainly not true, he added, for patients with ischemic heart failure and angina or symptomatic peripheral arterial disease.)

“Remember, polypharmacy issues are very real for these patients, so let’s focus on things that make a difference,” he noted. “I would actively take this one away.”

SWITCH potassium to spironolactone
You’ve doubled the patient’s furosemide and you expect him to have a metabolic panel checked within five days of leaving the hospital. So what should you do with his potassium at discharge: get rid of it, increase it to 40 mEq per day or switch it to spironolactone? Dr. Witteles recommended switching it.

The evidence for that switch comes from the 1999 RALES trial in which more than 1,660 “pretty sick” patients with NYHA class III-IV heart failure and an ejection fraction of 35% or less were randomized to spironolactone 25 mg daily or placebo. All had been hospitalized within six months of enrollment, and the primary endpoint was all-cause mortality.

“In terms of survival,” Dr. Witteles said, “there was a dramatic difference in favor of spironolactone.”

The 2011 EMPHASIS-HF trial tested another aldosterone antagonist, eplererone, against placebo. (Eplererone costs more than spironolactone, Dr. Witteles said, but it can be useful because “about 10% of men on spironolactone will get gynecomastia or nipple tenderness.”) More than 2,700 “slightly less sick” patients were enrolled in EMPHASIS-HF, with a primary endpoint of cardiovascular death or heart failure hospitalization. The trial was stopped after less than two years because of, again, “a really big treatment effect.”

And in 2004, Dr. Witteles said, Canadian researchers looked at “RALES in the real world.” In the wake of that 1999 trial, not surprisingly, “the number of spironolactone prescriptions really shot up.” But that spike, according to the 2004 analysis, didn’t appear to reduce the number of heart failure admissions. There were, however, increases in both the number of hospitalizations for hyperkalemia and of in-hospital hyperkalemic deaths.

“We have two very large trials—in slightly different patient populations—that both had really positive results in heart failure patients,” he said. “But you have to make sure patients have appropriate outpatient potassium monitoring.”

Further, he noted that he doesn’t know whether patients on spironolactone really run a higher risk of hyperkalemia than with supplemental potassium. Still, if you can assume that patients have good outpatient follow-up, “really think about making this switch for advanced heart failure patients.”

Phyllis Maguire is Executive Editor of Today’s Hospitalist.

The need for cardiac resynchronization
AT A RECENT hospital medicine conference, Ronald Witteles, MD, a heart failure specialist at Stanford University Medical Center in Stanford, Calif., presented the case of an admitted patient with nonischemic NYHA class II-III heart failure, a left bundle branch block and a left ventricular ejection fraction of 25%.

In addition to optimizing that patient’s medical therapies, Dr. Witteles said hospitalists can do something else in the hospital for such a patient: consult for biventricular pacemaker/implantable cardioverter defibrillator (ICD) placement.

He did note that such placement is “the outpatient cardiologist’s problem, not yours.” But he reminded the audience that being admitted “is a really important time in patients’ treatment course, and you can make a difference. Cardiac resynchronization can have as big an effect as any drug.” While he believes clinicians have an inflated expectation of what defibrillators do, “we often don’t have a high enough expectation of bi-V pacing in patients with left bundle branch block and systolic heart failure.”

He pointed to two trials. The 2004 COMPANION trial, which looked at about 1,500 patients with an ejection fraction of 35% or less and NYHA class III-IV heart failure, randomized patients to medical therapy alone, bi-V pacing alone or bi-V ICD, with a median follow-up of 16 months.

The results: Both bi-V pacing and bi-V ICD did much better in terms of survival without hospitalization than medical therapy alone. “What made the difference was the bi-V pacing,” he explained. In terms of just improved survival, “that’s where the defibrillator combined with the pacemaker arm did a little better.”

The 2005 CARE-HF trial enrolled more than 800 patients with similar entry criteria, randomizing them to either being resynchronized or not and following them for a median of more than 29 months. The study’s primary endpoint was all-cause mortality or CV hospitalization. In terms of survival without cardiovascular hospitalization, Dr. Witteles said the researchers found “a massive treatment effect” with resynchronization. “The same was true in overall survival.”

Any role for NT-proBNP levels?
CONSIDER THESE NT-proBNP levels for an admitted patient with nonischemic NYHA class II-III heart failure, a left bundle branch block and a left ventricular ejection fraction of 25%. The patient’s NT-proBNP was 3000 pg/ml four months before admission, 6500 pg/ml at admission and 5500 pg/ml when you’re preparing to discharge him. Given that the NT-proBNP level hasn’t returned to his outpatient baseline, should you discharge him as planned, push forward with further diuresis or further uptitrate neurohormonal antagonists?

Ronald Witteles, MD, a heart failure specialist at Stanford University Medical Center in Stanford, Calif., posed that question at a recent hospital medicine conference in San Francisco. The answer, said Dr. Witteles, is to discharge him as planned. While NT-proBNP levels can be useful for patients in the ED, “there really isn’t a role for using NT-proBNP levels to routinely guide therapy in the hospital.”

Good evidence supports that position, including from this most recent study: the PRIMA II trial published in the April 17, 2018, issue of Circulation. About 400 patients received usual care until clinical stability, then were randomized to either conventional therapy or NT-proBNP-guided treatment to try to target a greater than 30% NT-proBNP reduction between admission and discharge. Primary endpoints were all-cause mortality and heart failure readmissions within 180 days.

The result? “Absolutely nothing” in terms of any difference between the two study arms, Dr. Witteles pointed out. “Thirty-six percent to 36%. There’s no need for you to use repeated NT-proBNP checks.”

TTR amyloidosis: more common than you think
SPEAKING AT a recent hospital medicine conference, Ronald Witteles, MD, a heart failure specialist at Stanford University Medical Center, noted that he’s also co-director of the Stanford Amyloid Center. But he’s never spoken about amyloidosis to a general hospitalist audience because, he said, “it always seemed too esoteric.”

But three big changes have now occurred in the field of transthyretin (TTR) amyloidosis that hospitalists need to know, Dr. Witteles noted. For one, “we now have evidence that it’s common.” Two, “we have a much better way of diagnosing it, so we no longer have to send people for heart biopsies anymore.” And three, according to recent trials, “there’s a massive difference you can make for these patients” and a new treatment he expects to be approved soon.

While there are dozens of different types of amyloidosis, systemic forms that can involve the heart are light chains and TTR amyloidosis. In men, TTR amyloidosis “is actually quite common as they get older.” Patients can also be born with hereditary mutations in the transthyretin gene that make it more likely for them to begin having amyloid deposits at a younger age.

In terms of prevalence, researchers in a 2014 JACC: Heart Failure study reviewed the autopsies of more than 100 patients who’d been diagnosed with heart failure with preserved ejection fraction and no suspicion of amyloidosis. Those were compared to autopsies of age-matched controls with no heart failure diagnosis.

Among those with heart failure, “about a quarter of those in their 80s and two-thirds in their 90s had TTR amyloidosis,” he noted. And among controls, “about 15% had significant amyloid deposits in their heart at the time of their death.”

Another study, this one in a 2017 issue of the European Heart Journal, looked at patients referred for transcatheter aortic valve replacements (TAVRs). Everyone enrolled had technetium pyrophosphate (PYP) scans, and among the men, “almost one-quarter of the men referred for TAVRs actually had TTR amyloid, so it’s really common as people get older.”

It is those PYP scans, Dr. Witteles added, that have “completely transformed the field because we can now diagnose TTR amyloidosis easily.” The scans rely on an old bone-scanning agent that is taken up by hearts with TTR amyloid deposits; other hearts don’t.

If you suspect TTR amyloidosis—and “you should probably suspect it in most older men who come in with heart failure with preserved ejection fraction, particularly if they have left ventricular hypertrophy and their voltages aren’t high”—order a serum free light chains test. If that test comes back normal, “get a PYP scan and that will either tell you if the patient has TTR amyloidosis or not.”

As for treatment, the ATTR-ACT study, published in the Sept. 13, 2018, NEJM, randomized patients with TTR amyloidosis to one of two dosages of a drug called tafamidis or placebo.

“Tafamidis is a pill with no side effects,” Dr. Witteles explained. “Among patients alive at 30 months, the study found a more than 13% absolute risk reduction in mortality with tafamidis, and the number need to treat to prevent one death at 30 months was 7.5.” Further, “the number needed to treat to prevent one hospitalization per year was 4.5, so there was a massive treatment effect.” He expects the FDA to approve tafamidis this year.

Published in the February 2019 issue of Today’s Hospitalist
Notify of
oldest most voted
Inline Feedbacks
View all comments
Today's Hospitalist
March 2019 9:04 am

We wanted to share this question and Dr. Witteles’ response. A physician writes in: “This article suggests switching the patient’s metoprolol tartrate 25mg BID to carvedilol 6.25mg BID based on the COMET trial in 2003. “Unless they have a very good reason not to, the vast majority of patients with systolic heart failure should be on carvedilol, not metoprolol.” This is erroneous because metoprolol tartrate has never been shown to help in systolic heart failure! The only three beta blockers shown to help in systolic heart failure are metoprolol succinate, carvedilol, and bisoprolol. Metoprolol succinate is the long-acting version of… Read more »

Today's Hospitalist
March 2019 9:05 am

Dr. Witteles’ response: Thank you for your interest. Regarding your comments: 1) While it is true that metoprolol succinate (rather than metoprolol tartrate) is the approved form of metoprolol for systolic heart failure, that is only because the trial that led to its approval (MERIT-HF) chose metoprolol succinate to be tested. There is no plausible reason to believe that the succinate form is inherently superior to the tartrate form, and for better or for worse you will find many patients on metoprolol tartrate for this indication. 2) We have only the data that we have. The only trial comparing beta-blockers… Read more »

Rupesh Prasad
Rupesh Prasad
February 2019 12:16 pm


In addition to optimization medical therapy it is also essential to have a care transition plan in place for follow up (outpatient or home), along with proper patient and family education for compliance and to recognize early signs of decompensation.


Philip Chua, FB
Philip Chua, FB
February 2019 8:42 am

Our biggest bang for the buck is having a health navigator for our primary care clinics visiting the high probability readmission patients and working with the PCP and cardiology to implement the outpatient care plan. It has greatly reduced readmission.