They read the literature so you don't have to

COPD, stroke and drug choice were among the hot topics in hospital medicine research last year.

Bradley A. Sharpe, FACP, knows about your secret stash. He has his own—a pile of articles and journals that need to be read. “That stack for me seems to always just get bigger and bigger,” he told attendees at Hospital Medicine 2011.

Dr. Sharpe, a hospitalist at the University of California San Francisco, and Daniel Steinberg, FACP, of Beth Israel Medical Center in New York, braved the ever-growing stack, or at least the 2010 portion of it, to present an “Update in Hospital Medicine,” collecting 15 studies published last year that have potential to change, modify or confirm hospitalist practice.

Bradley A Sharpe FACP Photo courtesy of the Society of Hospital Medicine
Bradley A. Sharpe, FACP. Photo courtesy of the Society of Hospital Medicine.

They began with a few publications about chronic obstructive pulmonary disease (COPD). The first was an observational study of the effect of long-term beta-blocker use on COPD exacerbations and mortality (Rutten FH, et al. Arch Intern Med. 2010;170:880-7). “There's plenty of research showing that beta-blockers are safe in COPD, but interestingly, most of those studies were either pulmonary function test studies or they only followed patients for four or six weeks,” said Dr. Sharpe. This seven-year trial found a 36% decrease in mortality and exacerbations in patients taking the drugs.

“What we can say from this is that beta-blockers are certainly not harmful in patients with COPD,” said Dr. Sharpe. “In general, long-term beta-blocker use should be continued in patients with COPD as there may be real benefits.”

Another study (Lindenauer PK, et al. JAMA. 2010;303:2359-67) also filled a key gap in the evidence, according to Dr. Sharpe. “You may be aware that there are lots of studies showing clinical benefit to systemic steroids [but] very little good evidence on what is the right dose and route,” he said. The observational cohort study of nearly 80,000 admitted COPD patients compared 20 mg to 80 mg of oral prednisone a day versus 120 mg to 800 mg of intravenous (IV) equivalents as treatments for acute exacerbations. Although most patients got the high-dose IV drugs, the low-dose oral ones actually led to slightly better outcomes—a trend toward less treatment failure and shorter length of stay.

“This does confirm guidelines,” said Dr. Sharpe. “Most patients should get low-dose oral steroids for COPD exacerbations… This may not be true for those who are admitted to the ICU. In those patients, it may be your clinical judgment and you may choose to give higher-dose intravenous steroids.”

Patients in a third study (Sharma G, et al. Arch Intern Med. 2010;170:1664-70) had also been admitted for an acute COPD exacerbation and then discharged to the care of their known primary care physician or pulmonologist. But only 67% of them had a follow-up appointment with either of those physicians within 30 days. “You can reflect if that's a high or a low number at your institution,” said Dr. Sharpe. The patients who got follow-up were less likely to be readmitted or go to the emergency department within 30 days, and it didn't appear to matter whether they saw a pulmonologist or a primary care doctor. “This just seems like the right thing to do,” said Dr. Sharpe. “Just make sure they have the appointment.”

Stroke was the next major clinical topic, with two studies about the time window for administration of tissue plasminogen activator (tPA). The first (Lees KR, et al. Lancet. 2010;375:1695-703) found that while tPA's benefits decline and risks increase over time, it's still more likely to help than harm patients any time before four and a half hours after symptom onset. “This is useful if you find yourself in this situation, but it's really not a license to wait,” said Dr. Steinberg. “Time is brain.”

The second study (Ahmed N, et al. Lancet Neurology. 2010;9:866-74) revealed that physicians do find themselves in such situations. Using a 24,000-patient registry, researchers assessed use of the expanded window for thrombolysis before and after publication of trials supporting its safety. “Three times more patients were treated in the expanded time window after the publication of the evidence in the fall of 2008,” said Dr. Steinberg. “A nice collateral effect was that they found that increased numbers of patients were treated within the three-hour time window.”

Nice effect sizes were seen in the next study, which addressed audible gurgling (Vazquez R, et al. Chest. 2010;138:2848). It was a small trial of 20 patients plus 60 controls, but it indicated that gurgling noises are a good predictor of pneumonia and transfer to the ICU (with odds ratios of 140 and 35, respectively).

In addition to listening to chests, hospitalists might want to listen to infectious disease (ID) specialists, according to another cohort study, this one of patients with Staphylococcus aureus bacteremia (Honda H, et al. Am J Med. 2010;123:631-7). An ID consult significantly increased the rate of correct selection and duration of antibiotics and reduced mortality by 56%. “It's definitely going to lower my threshold for calling ID, particularly in complex cases,” said Dr. Steinberg.

Drug choice was the subject of several covered studies. A randomized controlled trial of patients with cirrhosis and a history of hepatic encephalopathy (HE) looked at the effect of rifaximin for preventing episodes of HE (Bass NM, et al. N Engl J Med. 2010;362:1071-81). Patients who took rifaximin had 60% fewer episodes of HE and half as many admissions for the condition. “I'm sure many of us have considered rifaximin as an alternative to lactulose, but this new data is actually rifaximin on top of lactulose,” Dr. Steinberg noted.

To help select drugs for some even sicker patients, another randomized trial compared dopamine and norepinephrine for treatment of shock (De Backer D, et al. N Engl J Med. 2010;362: 779-89). “In terms of positive outcomes, they were very similar, but there was a huge difference in arrhythmias: 12% of patients on norepinephrine had an arrhythmia versus 24% with dopamine. Most of those were atrial fibrillation, although there were a handful of ventricular arrhythmias,” said Dr. Sharpe. In a subgroup analysis of patients with cardiogenic shock, dopamine was associated with higher mortality.

The results indicate that hospitalists should not only choose norepinephrine, but spread the word. “If you have a relationship with your emergency department, you should share that you prefer them to start this. We often have trouble where dopamine is started in the ED and then you have to decide, is the transition worth it?” said Dr. Sharpe.

The word on metformin has long been that it can cause lactic acidosis. That turns out to be wrong, according to a meta-analysis of 70,000 patient-years (Salpeter SR, et al. Cochrane Database Syst Rev. 2010;(1):CD002967). “In all of this data, not a single case of fatal or non-fatal lactic acidosis was found,” said Dr. Steinberg. “To me in practice, what this means is, if you have a patient who comes in with lactic acidosis and they're on metformin, absolutely stop it. That's certainly an easy prudent thing to do. But you should really take a careful look elsewhere as to the cause because it's unlikely metformin is going to be the cause.”

Another piece of conventional wisdom was corrected by a study of venous thromboembolism (VTE) in patients with liver disease (Dabbagh O, et al. Chest. 2010;137:1145-9). Coagulopathy in liver disease does not protect against VTE: More than 6% of the 190 admitted patients had a VTE, and an even higher rate (8%) was seen in the patients with the most advanced liver disease. “Whether you give them heparin, Lovenox [enoxaparin] or compression stockings, that's going to be a clinical decision,” said Dr. Steinberg. “The point here is that the incidence is too high to ignore and patients aren't protected just because they have liver disease.”

Moving on from meds to their delivery, Dr. Sharpe reported on a study of whether to replace peripheral IVs after three days (Webster J, et al. Cochrane Database Syst Rev. 2010;(3):CD007798). “There was no difference in bacteremia between routine replacement and replacing based on symptoms,” he said, noting that the thrombophlebitis rates were also the same between groups. “It probably is OK to extend peripheral IVs beyond the routine placement if asked by your nursing staff.”

You can skip the IV, on the other hand, if you're trying to feed a patient with acute pancreatitis, according to the next meta-analysis (Al-Omran M, et al. Cochrane Database Syst Rev. 2010;(1):CD002837). “Use of enteral nutrition instead of total parenteral nutrition reduced mortality with some nice numbers needed to treat: for mortality 13, reduced multi-organ failure 7, reduced operative intervention 5, reduced systemic infection 5,” reported Dr. Steinberg. He noted, however, that the review authors had judged the quality of the evidence to be low to moderate. “We rarely have perfect evidence, but we always have to take care of patients and make decisions. This study is the best available evidence on this topic and is enough to act on in practice.”

The final two studies dealt with determining prognosis. A retrospective analysis of the National Registry of Cardiopulmonary Resuscitation used data from 2000 to 2008 to assess outcomes from cardiac arrest in the ICU (Tian J, et al. Am J Respir Crit Care Med. 2010;182:501-6). “The big picture is that one in six patients survives to discharge,” said Dr. Steinberg. But subgroup data from the study may actually be more useful for prognosis: Patients whose initial rhythm was pulseless electrical activity/asystole were much less likely to survive than those in ventricular tachycardia or fibrillation (11% vs. 33%). Other factors, including previous use of pressors, nonwhite race, older age and mechanical ventilation, all brought the likelihood of CPR success down even further. “Just having been on pressors alone predicted a poor prognosis,” said Dr. Steinberg. In patients who were on pressors before a cardiac arrest in the ICU, all outcomes—survival to discharge, discharge to home, and discharge to home and ability to perform activities of daily living—were reduced by over 50%.

Even when physicians can offer precise calculations, family members may base their prognosis assumptions on other factors, according to the last study covered (Boyd EA, et al. Crit Care Med. 2010;38:1270-5). Surrogate decision makers were interviewed about what sources of information they used to predict patients' survival. “Only 2% relied exclusively on what the physician said. Only 47% said that at least part of their estimate was based on what we said,” said Dr. Sharpe.

Other considerations for the surrogates were: 1) the patient's will to live, 2) his or her physical appearance and 3) prior history of survival, and 4) the surrogate's own faith or intuition. Awareness of these factors may be useful to physicians communicating with surrogates, according to Dr. Sharpe. “It tells us that determining prognosis for surrogates is complicated and our input is only part of the picture,” he said.