Published in the December 2014 issue of Today’s Hospitalist
WHAT ANTIBIOTICS should you prescribe ICU patients with community- acquired pneumonia (CAP), and should you give thrombolytics to intermediate-risk patients with pulmonary embolism (PE)? Those are just two of the clinical questions that have been answered in the literature over the past year.
During the management of the hospitalized patient conference this fall at the University of California, San Francisco, hospitalists Bradley Sharpe, MD, UCSF’s associate division chief of hospital medicine, and Michelle Mourad, MD, the division’s director of quality and safety, highlighted what they found to be the most important studies of the last 12 months.
Those included “game-changers” in terms of hospitalist practice, as well as new evidence that bolstered what many doctors have been doing all along.
Case No. 1: Antibiotics in CAP
A 63-year old with COPD, coronary artery disease and diabetes presents with three days of fever, cough and shortness of breath. His oxygen saturation is 85% on room air, and his exam is notable for diffuse expiratory wheezes and crackles at the right base. His white cell count is 24,000, and his chest X-ray reveals a lower right lobe infiltrate.
You admit the patient to the ICU for COPD exacerbation, CAP and septic shock, and you need to decide the antibiotic regimen to prescribe for CAP. Is it levofloxacin monotherapy, ceftriaxone and doxycyline, ceftriaxone and azithromycin, or ceftriazone and levofloxacin?
While CAP is exceedingly common, Dr. Sharpe explained, “the evidence hasn’t been clear on which antibiotic strategy is better in the ICU: a beta-lactam plus a macrolide or a beta-lactam plus a fluoroquinolone.” One study, which appeared in the February 2014 issue of Critical Care Medicine, sought to answer that question.
The meta-analysis looked at 28 high-quality observational studies, which together covered close to 10,000 patients. The study compared macrolides vs. nonmacrolides as well as macrolides vs. fluoroquinolones. In addition, the meta-analysis reported on 30-day mortality for macrolide vs. nonmacrolide regimens from the studies that controlled for possible confounding factors.
Macrolides showed a statistically significant 3% absolute risk reduction (21% vs. 24%) vs. nonmacrolides. For macrolides vs. fluoroquinolones, macrolides again exhibited a trend toward lower mortality, although that did not reach statistical significance.
“Net benefit” for azithromycin
However, the adjusted risk ratio for macrolides vs. nonmacrolides was 0.75 in some of the better-designed trials included in the meta-analysis.
“That’s a 25% statistically significant decrease in mortality in ICU patients with CAP,” Dr. Sharpe pointed out. The same authors, he added, had looked at the effect of macrolides on mortality in CAP patients outside the ICU several years before. They had not found any statistical significance.
“Macrolides may be better in the ICU because of their anti-inflammatory properties,” he said. While guidelines for ICU patients with CAP recommend a third-generation cephalosporin with either azithromycin or a fluoroquinolone, his choice based on these study results would be ceftriaxone and azithromycin (with or without vancomycin).
But that may leave you wondering, Dr. Sharpe added: Isn’t azithromycin bad for the heart?
An analysis in the June 4, 2014, issue of the Journal of the American Medical Association (JAMA) of a large VA database found that azithromycin use among patients hospitalized with CAP who were older than 65 was associated with a small increase in the incidence of MIs (5.1% vs. 4.4%).
However, it was also linked to a more substantial mortality decrease (17% vs. 22%), leading the authors to conclude that azithromycin may confer “a net benefit” in older pneumonia patients admitted to the hospital.
Steroid dosing and COPD
While the patient starts dual antibiotic therapy, you also want to order steroids for his acute severe COPD exacerbation that has required ICU admission. What dose should you prescribe: 1 gram of methylprednisolone every six hours, 125 mgs of methylprednisolone every six hours, 60 mgs of prednisone twice a day or 60 mgs of prednisone only once a day?
As Dr. Sharpe pointed out, “a growing body of evidence has found that lower steroid doses are better for COPD patients who are not in the ICU.” But there hadn’t been much evidence on ICU dosing strategies until an observational study appeared in the May 1, 2014, issue of the American Journal of Respiratory and Critical Care Medicine.
That study looked at more than 17,000 ICU patients with a COPD exacerbation. It compared a lower-dose strategy (defined as less than or equal to 240 mgs of methylprednisolone per day) to a high-dose one (more than 240 mgs of methylprednisolone per day) during patients’ first 48 hours. Researchers looked at hospital mortality.
While 64% of the COPD patients in the ICU received high-dose steroids, high doses were associated with worse outcomes and more adverse events. (Average doses in the two treatment arms were 100 mgs vs. 315 mgs per day.) Patients given lower doses had a trend toward lower mortality, but it was not significant, Dr. Sharpe said.
But, he noted, “they also had nearly a half day less in overall length of stay and more than $2,500 less in costs per patient.” Patients given lower steroid doses also had less need for insulin therapy and fewer fungal infections.
Despite that evidence, exactly what exact dose patients should receive remains unclear.
“Researchers conclude in the article, ‘We don’t know, but 60 mgs of prednisone once a day or the methylprednisolone equivalent is probably enough,’ ” said Dr. Sharpe. “You can try 80 if that makes you nervous, but the point is that the traditional 125 mgs q 6 that I used for years is going away.”
Case No. 2: PE and thrombolytics
You see a 56-year old woman with emphysema who had a knee replacement six weeks ago. Her leg has been swollen for several weeks, and she now presents with pleuritic chest pain, tachycardia and hypoxemia.
As her breathing becomes increasingly labored, she is intubated in the ED. You order a CT scan according to the pulmonary embolism protocol and find what Dr. Mourad called “a pretty large saddle embolus.”
Her heart rate is persistently in the 130s, she is on 100% FIO2 and a bedside echo demonstrates right ventricular dysfunction, which puts her at intermediate risk for mortality. “You know you’ll start anticoagulation, but this woman is very sick and you wish there was more that you could do,” said Dr. Mourad. “Is there a mortality benefit to using thrombolytics in PE patients at intermediate risk?”
A meta-analysis in the June 18, 2014, JAMA, provides some answers. That analysis, said Dr. Mourad, looked at 16 studies, some of which were randomized controlled trials, all covering patients with acute PE. The studies compared thrombolytics to anticoagulants in terms of all-cause mortality, major bleeding, recurrent PE and intracranial bleeds.
“The punchline,” Dr. Mourad said, “is that intermediate risk patients with right ventricular dysfunction are 50% less likely to die if they are given thrombolytics.” But they are also three times as likely to have major bleeding, defined as needing transfusion or a higher level of care “or having an intracranial hemorrhage.
However, she pointed out, the rates of patients receiving thrombolytics who have intracranial hemorrhage “are less than the mortality rates on anticoagulation alone. Now, thrombolytics are starting to seem a bit more appealing.”
Further, the meta-analysis stratified patients by age and found that patients under age 65 in both groups had no difference in their rates of major bleeding. This is a big change, Dr. Mourad explained, from a previous Cochrane Review, which found no benefit from thrombolytics in PE patients at intermediate risk. “So this is definitely a practice-changer.”
VAP and antibiotic duration
You administer tPA, and your patient’s blood pressure and tachycardia improve while her oxygen requirement falls. But because of her emphysema, she is difficult to extubate. And on day No. 4, she develops a fever with increased secretions, and a mini-bronchoalveolar lavage founds gram-positive cocci in clusters.
You start broad-spectrum antibiotics, then narrow them to nafcillin when the cultures are positive for MSSA. But how long do you want to continue that therapy?
A study that appeared in the December 2013 issue of Chest sought to gauge the optimal duration of antibiotics in ventilator-associated pneumonia. The meta-analysis looked at four randomized controlled trials that compared treatments of eight days or less to those of 10 days or longer. All VAP in the trials was diagnosed by culture, “and the largest study continued to do surveillance cultures after the treatment was concluded to look for what researchers called a ‘microbiological cure,’ ” Dr. Mourad said.
Shorter course durations of seven to eight days produced no increased mortality. But there was a trend, although it was nonsignificant, toward more relapses with shorter antibiotic courses.
However, that trend was due to the one study that continued to culture patients even after there were no more reported symptoms. In terms of secondary outcomes, patients receiving shorter courses had fewer antibiotic-free days for 28 days out, as well as no difference in ICU length of stay or days on a ventilator.
“So we now have very good evidence that we can give patients a seven- or eight-day course, see how they’re doing and then continue their course if they need it,” said Dr. Mourad. “That saves on costs and resistance and avoids the harms “like C. diff “of prescribing too many antibiotics.”
After several days of antibiotics, the patient is extubated, but then develops worsening renal failure. You’re considering all the possible reasons for that including prerenal azotemia, pyelonephritis and acute interstitial nephritis (AIN). Would ordering urine eosinophils help with the differential?
It would not, according to a study published in the Nov. 7, 2013, issue of the Clinical Journal of the American Society of Nephrology. In a case control study of more than 560 patients with suspected AIN, researchers compared the presence of urine eosinophils with the final diagnosis.
“They found that urine eosinophils were most prevalent in acute tubular necrosis, not AIN, for which they were neither sensitive nor specific,” said Dr. Mourad. “It turns out that you have to use the patient’s clinical history and presentation as well as AIN risk factors, so we can stop ordering them.”
Case No. 3: Albumin for fluid resuscitation?
A 79-year old man with a history of lung cancer presents with respiratory failure. His chest X-ray shows an enlarging mass with a post-obstructive pneumonia.
He is placed on 100% FIO2 with a non-rebreather and is transferred to the ICU where, despite receiving antibiotics, he is progressively hypotensive and his urine output is falling.
His WBC is 20, his creatinine is 2.8 and his albumin is 12.8, and he has leukocytosis and acute kidney injury with a very low albumin. You want to write for more fluids. But should you also give him albumin?
Researchers writing in April 10, 2014, issue of the New England Journal of Medicine looked at adding albumin to fluid resuscitation of patients with severe sepsis. They randomized 1,800 patients. “They set the target at an albumin 3 and gave those in the intervention group boluses of albumin if they fell below that level,” Dr. Mourad noted. “Crystalloids were then administered if patients needed fluid.” Researchers were specifically looking at both 28- and 90-day mortality and organ dysfunction.
Both sets of mortality outcomes were the same in those receiving albumin plus crystalloids and those receiving crystalloids alone. Nor was there any difference in the number of patients with organ dysfunction.
The study’s conclusion? “Albumin has no significant benefit over other fluids in severe sepsis,” Dr. Mourad said.
Estimating in-hospital mortality
Despite his resuscitation, the patient worsens and you feel he should be intubated. You also want to speak to the family and provide some estimate of the patient’s inhospital mortality risk. Based on the information you have, how high is that risk?
Researchers conducting an observational trial in 28 Brazilian hospitals looked at more than 260 patients who represented many different types of cancers and had been intubated for less than 24 hours. Results were published in the August 2014 issue of Chest.
While patients’ overall hospital mortality was 67%, Dr. Sharpe explained, predictors of higher mortality included active or progressive cancer, poor performance, and organ dysfunction. Patients with controlled cancer, good performance status and other organ failure ran a 52% risk of in-hospital mortality; patients with metastatic cancer, poor performance status and other organ failure had an in-hospital mortality risk of 84% “the category the patient in the case presentation would fall in.
“The overall mortality is quite high, with very high mortality for some cancer patients who need to be intubated,” Dr. Sharpe pointed out. “That’s information we can use for patient and family discussions.”
Phyllis Maguire is Executive Editor of Today’s Hospitalist.
Fistbumps vs. handshakes
DO YOU HAVE a favorite study this year? During a year in review discussion at a conference this fall at the University of California, San Francisco, hospitalists Bradley Sharpe, MD, UCSF’s associate division chief of hospital medicine, and Michelle Mourad, MD, the division’s director of quality, presented their favorite study of the year: the now famous fistbump research, published in the August 2014 American Journal of Infection Control.
One researcher put a sterile-gloved hand in an E. coli culture, then exchanged different greetings with a recipient, also wearing a sterile glove.
“Nearly twice as many bacteria were transferred during a handshake compared to a high five,” said Dr. Sharpe. “But the fistbump far and away consistently led to the lowest bacteria transmission.”
The short take, he added, is that doctors need to be aware that they may be transmitting “incredible amounts of bacteria. Consider using the fistbump or high five instead of a handshake.”