Published in the August 2014 issue of Today’s Hospitalist
TO FIGHT BACK AGAINST the ongoing overuse of antibiotics, many experts are rethinking treatment of health care-associated pneumonia (HCAP). Just because patients have had an IV in the last month does not mean their pneumonia needs the triple-antibiotic therapy that current guidelines recommend.
During this spring’s Society of Hospital Medicine meeting’s pneumonia update, Scott A. Flanders, MD, director of the hospitalist program at the University of Michigan in Ann Arbor, discussed recent research indicating that the decision about who needs a triple-drug regimen may be more nuanced. The study concluded that pneumonia patients with only one risk factor had no more risk of having a multidrug-resistant pathogen than patients with community-acquired pneumonia (CAP).
Only patients with several risk factors “prior antibiotics, gastric acid suppression therapy with PPIs and H2 blockers, tube feeding, nonambulatory status, immunosuppression, or a prior hospitalization “were truly at higher risk for multidrug-resistant organisms. That study was published in the Oct. 15, 2013, American Journal of Respiratory and Critical Care Medicine.
“I think the original HCAP criteria are facing a backlash now, and many people have said that triple therapy is too much too fast,” Dr. Flanders told hospitalists. “Maybe not all patients receiving hemodialysis or wound care need three drugs. You should probably treat them as you would regular CAP if they don’t have any other risk factors.”
“Strong” vs. “weak” risk factors
According to a 2006 definition, HCAP includes patients who have been hospitalized within 90 days of the infection, reside in a nursing home or long-term care facility, or received parenteral antimicrobial therapy, chemotherapy, or wound care within 30 days of pneumonia.
The category was created because some patients in the community develop pneumonia that resembles hospital-acquired (HAP) or ventilator-associated pneumonia (VAP). These patients are more likely to harbor resistant organisms.
To ensure that triple-antibiotic therapy is used only for patients who need it, Dr. Flanders recommended thinking about “strong” risk factors for resistant organisms vs. “weak or unclear” ones.
“Prior admission to the hospital probably warrants broad spectrum antibiotic treatment, but some of the weaker risk factors don’t necessarily,” he explained. Those include living in a nursing home as a sole risk factor, dialysis without other risk factors, wound care and home health. In addition to prior hospitalization, strong risk factors include being in a skilled nursing facility (especially if the patient was on antibiotics before or has poor functional status), critical illness, prior MRSA or pseudomonas, or having bronchiectasis or recurrent COPD exacerbations.
Dr. Flanders also pointed out that patients whose sputum and blood cultures fail to grow a concerning organism tend to be less sick than HCAP patients whose cultures come back positive for MRSA or pseudomonas. Those were the findings of a study published in the May 2010 issue of CHEST.
“What this says to me is that if you don’t find anything, patients are probably going to do just fine with run-of-the-mill CAP treatment,” he said.
The University of Michigan algorithm for patients at higher risk of having a drug-resistant organism calls for triple therapy of vancomycin, piperacillin-tazobactam (Zosyn) and tobramycin (or levofloxacin, if patients have renal insufficiency). The protocol also calls for treating patients for seven or eight days unless a culture isolates gram-negative bacteria that need longer treatment. The regular CAP treatment algorithm, on the other hand, specifies ceftriaxone plus azithromycin.
“Remember that treatment for more than seven days is no better than treatment for less than seven days in patients who are improving,” Dr. Flanders stressed.
“Seven days should be your cut point.” Moreover, when you reassess a patient you thought had pneumonia but who turns out to have heart failure, pneumonitis or something else altogether, stop all antibiotics. “Do not continue therapy,” he said.
Issues around azithromycin
Azithromycin presents another big challenge. “Does azithromycin kill patients?” Dr. Flanders asked. While that question may be an overstatement, “there is a potential downside” due to azithromycin prolonging QT intervals, “and two very large studies that came out relatively recently gave us some concern.”
The first study, published in the May 17, 2012, New England Journal of Medicine (NEJM), found an elevated risk of cardiovascular death in patients within five days of using azithromycin. The other study, published a year later in the May 2, 2013, NEJM, did not.
But instead of seeing these two studies as contradictory, experts see the results as possibly complementary. “When you have a sicker population” like the Medicaid population examined in the first study, “there is an increased risk of mortality with macrolides,” Dr. Flanders noted. That is probably not the case, “to as great an extent, in a population of young patients,” who were more the subject of the second paper.
“It may seem like a cop-out to say you need to balance risk” when deciding whether to prescribe azithromycin, he said. “But this is yet another reason for us to really consider whether our patient needs the antibiotic. If I had a high-risk patient taking other QT-prolonging medications “and there are a lot of them “I might pause before giving that patient azithromycin.”
Instead, he said, consider doxycycline, which is also recommended by pneumonia treatment guidelines. A study published online in May 2012 by Clinical Infectious Diseases established this added benefit: Doxycycline is associated with fewer C. difficile infections. And a more recent study in the June 4, 2014, JAMA specifically evaluated the safety of azithromycin use in an elderly hospitalized population. Results showed that overall, the benefits outweighed the risks.
Treating aspiration pneumonia
Aspiration pneumonia is another area where “run-of-the-mill CAP treatment” may be sufficient, while adding an extra antibiotic to fight anaerobic infection could be overkill.
“There are clearly patient populations where covering for anaerobes is important,” said Dr. Flanders. His list of aspiration pneumonia patients he tends to give antianaerobic treatment to include those with periodontal disease, putrid sputum, necrotizing cavitary lesions, large effusions, and elderly patients with poor functional status or recurrent aspirations.
While many elderly hospitalized patients with pneumonia have aspiration, that’s mostly “silent aspiration,” he said, caused by gram-negative rod organisms instead of anaerobes.
And most patients with anaerobic infections also have concomitant gram-negative rod infections. Studies have shown that 85% of these patients get better even without antibiotics to cover anaerobes. For Dr. Flanders, that means that treating anaerobes in this patient population is “not critically important. I think we over-treat anaerobes in aspiration patients.”
Hospitalists should also keep in mind that when a patient they thought had aspiration pneumonia gets well quickly, that patient probably has aspiration pneumonitis instead, an inflammatory response to gastric or oral content that can produce similar symptoms.
“If a patient resolves within 48 hours, consider stopping antibiotics,” Dr. Flanders said. That’s because aspiration pneumonitis is not due to a bacterial infection. And in the setting of rapid resolution, “treating for more than five days would not be indicated.”
Getting antibiotics right is also important for performance measurement. Although some pneumonia measures have been retired or suspended, the measure looking at appropriate initial antibiotic therapy continues to be collected, as well as those related to 30-day mortality and 30-day readmissions.
One study published in the Aug. 1, 2013, Clinical Infectious Diseases found that most pneumonia readmissions were of patients who met criteria for HCAP, not CAP. Although many of those criteria “like living in a nursing home “are beyond the ability of hospitalists and health systems to change, Dr. Flanders said that such findings indicate that readmission-reduction interventions should target specific patient subsets.
Research has also revealed some of the limitations of data now collected for performance measures. The authors of a study published in the April 4, 2012, JAMA, for instance, found that HCAP, CAP and HAP data were all lumped together “and that hospitals vary in how they code pneumonia. Some code more cases as “sepsis or respiratory failure with a secondary diagnosis of pneumonia,” while others code similar cases as “pneumonia.”
That affects how well your hospital performs on 30-day pneumonia mortality rates, Dr. Flanders pointed out. That was the conclusion of an important article published in the March 18, 2014, Annals of Internal Medicine.
“The CMS may change its methodology, using the results of some of these studies to improve on some measures,” Dr. Flanders said. “And coming soon are cost-based measures, which also include pneumonia.
These measures are not going away.”
Deborah Gesensway is a freelance writer who covers U.S. health care from Toronto.
A role for pneumonia risk scores
IN AN UPDATE ON PNEUMONIA presented at this spring’s Society of Hospital Medicine meeting, Scott A. Flanders, MD, director of the hospitalist program at the University of Michigan in Ann Arbor, urged hospitalists to become familiar with the various risk assessment scoring systems that pneumonia experts are developing. Those systems can help doctors decide whether pneumonia patients need to be hospitalized and, if they do, whether patients need ICU or step-down care.
According to Dr. Flanders, the best scoring systems to answer that first question include the pneumonia severity index (PSI) and the CURB-65. As for whether patients need to be in the ICU, he recommended using Infectious Diseases Society of America/American Thoracic Society severity criteria. “Guessing wrong in risk stratification worsens outcomes,” Dr. Flanders pointed out.
Risk-scoring systems should not trump clinical judgment. “But they should be added to clinical judgment and perhaps make you rethink when the score is discordant with what you would otherwise do for your patient,” he said.
Dr. Flanders also predicted that in the near future, clinicians will be able to use biomarkers like procalcitonin levels as well as risk scores to help decide how sick pneumonia patients are and whether patients are responding to treatment.