Published in the January 2010 issue of Today’s Hospitalist
It’s been a busy year for physicians treating pneumonia.
While pneumonia remains America’s No. 1 discharge diagnosis and leading cause of infectious-disease death, new developments with H1N1 and MRSA are creating new challenges for hospitalists charged with treating the infection. At the same time, new data have been released that attempt to answer the longstanding debate about the importance of atypical therapies and the value of following clinical guidelines for treating pneumonia.
According to Scott Flanders, MD, a national expert on pneumonia care and director of the hospitalist program at University of Michigan in Ann Arbor, a number of recent studies have given hospitalists important clues about what causes pneumonia “and how to best treat the condition. But during an update on pneumonia at last fall’s hospital medicine conference at the University of California, San Francisco, Dr. Flanders also said that some of those studies “particularly those that seem to endorse close adherence to guidelines “raise as many questions as they answer.
Here’s a look at Dr. Flanders’ review of the latest research on pneumonia, and the strategies he recommends to contain both multidrug resistance and spiraling health care costs associated with the condition.
When thinking about pathogens that cause pneumonia, Dr. Flanders, who’s president of the Society of Hospital Medicine, said that a study published in the December 2008 Clinical Infectious Diseases found that between 30% and 40% of pneumonias are caused by typical bacteria, with S. pneumoniae the culprit in 45% of those cases. Between 10% and 30% are caused by atypical bacteria, the majority of which is mycoplasma and only rarely Legionella.
But in as many as 40% to 50% of pneumonia cases, the pathogen remains unknown. Another December 2008 study, this one in Chest, shed light on that group of unknown pathogens. “We’re finding that viral agents are increasingly common,” Dr. Flanders pointed out. With 15% to 20% of patients hospitalized with a pneumonia caused by a virus, he said, “influenza leads that pack.” (Human metapneumovirus is also frequently isolated.)
Patients with viral pneumonia tend to be older, more frail and less likely to have leukocytosis, Dr. Flanders explained, and the incidence of viral pneumonia has tended to be seasonal “at least until last year. Several studies over the past six months have detailed what he called “scary findings” on pneumonia caused by swine origin influenza or H1N1.
Studies have found, he said, that patients with severe H1N1 have bilaterial patchy basilar infiltrates; CPK abnormalities and lymphopenia; and a lack of comorbid disease, “which is,” he pointed out, “in sharp contrast to the usual severe flu cases.” Another major difference: 87% of deaths from the disease and 71% of severe cases were in the 5 to 59 year-old age group.
Patients with pneumonia caused by H1N1 have a “very rocky” clinical course, Dr. Flanders noted, with renal failure and little bacterial superinfection. Unfortunately, his institution got a lot of experience with pneumonia caused by H1N1 last spring, with findings from that experience detailed in the July 17, 2009, issue of the CDC’s Morbidity and Mortality Weekly Report.
Those patients required “very unusual ventilatory strategies,” he said, including oscillatory ventilation and extracorporeal membrane oxygenation. Seven of the 10 patients treated at the University of Michigan with severe H1N1 had body mass index readings of more than 40.
And patients were not only hypercoaguable, with 50% found to have pulmonary embolisms, but hemorrhagic. “These folks get a hemorrhagic pneumonitis, coughing up a lot of blood and then endotracheal bleeding, which makes them incredibly difficult to manage when they develop thrombotic events,” Dr. Flanders said. Treatment followed CDC recommendations, which “for severe cases ” include a doubled dose of oseltamivir (Tamiflu) at 150 mg BID po for 10 days.
When H1N1 is circulating along with seasonal flu, Dr. Flanders said his group prescribes rimantidine to patients hospitalized with influenza at 100 mg po BID for seven days.
While one-third of all pneumonias are caused by pneumococcus, Dr. Flanders said, hospitalists are getting much more experience with pneumonia caused by community-acquired MRSA (CA MRSA). Meta-analyses of the disease find that it “like H1N1 “presents with a flu prodrome, leucopenia, shock and multilobar disease.
In addition, “most of these organisms have a positive PVL [panton valentine leukocidin],” Dr. Flanders said. “It tends to be a necrotizing infection with average mortality of 40%.” While the incidence is 0.51-0.64 per 100,000, he explained, “this is a very serious disease.”
Because risk factors for CA MRSA include past infection, IV drug use or influenza, hospitalists should treat patients with these risk factors empirically. “We treat with vanco or linezolid, not daptomycin, which doesn’t work for this condition,” he explained. “Many of these isolates are resistant to things we commonly use for pneumonia, like macrolides and fluoroquinolones.”
Hospitalists should see higher vanco troughs of between 15 and 20 mcg/mL, Dr. Flanders added, and data do not support some therapies. Those include clindamycin; beta-lactams, because of their potential effect on toxins; and IVIG.
Dr. Flanders had one last word on etiologies: Because hospitalists are worried about pseudomonas, they should keep in mind that the big gram-negative predictors include aspiration, a prior admission or antibiotics, or pulmonary comorbidity. The latter, he pointed out, is “the biggest driver, particularly severe COPD that’s resulted in hospitalization or is actively being treated.” These patients are probably at risk for pseudomonas and warrant empiric coverage.
As for treating community-acquired pneumonia (CAP), recommendations for hospitalized patients call for a beta-lactam plus a macrolide or fluoroquinolone monotherapy. Those recommendations add atypical coverage, which is thought to be important for these patients.
But is that coverage really necessary? According to Dr. Flanders, a 2005 Cochrane review updated in 2008 looked at 24 randomized controlled trials and found that patients with atypical coverage vs. those without showed no difference in mortality or clinical failure. “It didn’t even make a difference when they had an atypical organism isolated,” he noted. While Legionella does need to be covered, it was isolated in only 43 out of thousands of cases.
Two studies published last September examined the issue again. One study, which covered 54,000 patients and is the largest study to date, compared patients who received guideline-based treatment to those who received “discordant” treatment, which deviated from the guidelines. The second study, which looked at older patients, likewise compared guideline-based treatment to undertreatment (omitting atypical coverage) and overtreatment (giving them all the guideline-recommended therapies, plus some).
Both studies found that patients did better “and hospitals saved money and bed days “with guideline-based care than with over- or undertreatment, delivering a relative risk reduction of between 30% and 40%. As an editorial in the same issue noted, given that both studies were adjusted for the usual severity indices, it would be hard to imagine what could be a possibly “unrecognized or uncontrolled confounder” that would have that big of an impact.
But Dr. Flanders pointed out that he could think of two subtle but real issues related to following guidelines for antibiotic CAP regimens. For one, hospitals in which physicians routinely follow guidelines probably do a lot of things better than hospitals with doctors who don’t, including maintaining hand hygiene and a culture of excellence.
Then there’s the issue of physician judgment. Physicians, himself included, are much more apt to give the broad, guideline-based therapy “including atypical coverage ” to patients who are sicker.
“I don’t think we’re able to adjust for all these things,” Dr. Flanders said, “and I’m still not convinced that there’s one regimen that’s better than another.” While guideline-based therapy remains a performance indicator, a randomized, controlled trial now taking place in Europe may provide an answer as to which patients don’t routinely need atypical coverage.
Weighing the risk of drug-resistant organisms
Dr. Flanders then turned to health care-associated pneumonia (HCAP), pneumonia cases in which patients have been “touched” by the health care system. That contact could come from home therapy for IV, wound or home-health agency care; from spending time in a hospital or dialysis unit within the past 30 days; from a hospitalization of at least two days within the past 90 days; or from living in a nursing home or long-term acute care facility.
Patients with these risk factors, Dr. Flanders explained, have more multidrug resistant pathogens, including pseudomonas and staph. The problem is that these risk factors have now been rolled into the larger body of risk factors for multidrug resistant organisms, he added, “and essentially given equal weight.” As a result, current guidelines recommend treating these patients with a big-gun regimen that includes vancomycin plus piperacillin/ tazobactam plus ciprofloxacin.
A study in the Jan. 6, 2009, Annals of Internal Medicine found that the dominant risk factor for HCAP was prior hospitalization within six months. Researchers found that when compared to CAP patients, HCAP patients were less likely to receive guideline-based treatment, which was associated with higher death rates.
Exactly how big is the gap between physician practice and guideline-based treatment? A survey published in the December 2009 Clinical Infectious Diseases of 1,300 physicians in the U.S., a group that included hospitalists, pulmonary/ critical care physicians and ED doctors, found that “only 10% of patients received guideline-based therapy,” Dr. Flanders said. Many doctors responding gave a single agent for the gram-negative rods but left out MRSA coverage and thus deviated from guideline recommendations.
But again, Dr. Flanders noted, maybe one-size-fits-all isn’t really the answer and not every patient with risk factors needs such broad therapy. One 2008 study in Archives of Internal Medicine assigned points to different risk factors in HCAP patients and created a graded algorithm.
Among patients who scored between 0 and 2 points, “four out of five didn’t have resistant organisms, so we’re probably overtreating a lot of them,” Dr. Flanders said. At the high end of the scale, on the other hand, most patients did have multidrug resistant organisms, and “we absolutely should be giving them broad therapy.”
Then there are subsets of patients who do quite well with narrow therapy, Dr. Flanders added, including some with nursing home-acquired pneumonia. Studies have found that many nursing home patients with pneumonia who are relatively functional can be managed with less aggressive therapy at a much lower cost.
That was borne out in a 2009 study of patients with nursing home-acquired pneumonia. That study found that 77% of these patients were treated with what were essentially recommendations from 2003 guidelines, which was ceftriaxone plus a macrolide. The other 23% received the full complement of three antibiotics, which was called for in 2005 guideline recommendations.
What researchers found was essentially no difference in outcomes, although patients getting broader regimens had higher costs and greater lengths of stay.
“We’re beginning to see with the nursing home population that it really depends on the patient,” Dr. Flanders said. The algorithm for treating HCAP is evolving, and a new iteration of guidelines “should be coming out in another year.”
In the meantime, Dr. Flanders urged, patients should be stratified by risk, and those with severe pneumonia who need ICU ventilation should have their risk factors added up for multidrug resistance. “If it’s severe, they’re going to get broad therapy,” he said.
Risk factors should include any indwelling devices “urinary catheters, PICCs, feeding tubes “as well as advanced respiratory disease and bronchiectasis. Hopefully, Dr. Flanders said, such an approach will be prospectively validated because it leads to “more targeted therapy.”
Phyllis Maguire is Executive Editor of Today’s Hospitalist.
IN LOOKING AT THE EVIDENCE published in the past year on preventing pneumonia, one area receiving new attention, according to national pneumonia expert Scott Flanders, MD, is the importance of oral hygiene.
One recent study found that, along with practicing good hand hygiene and elevating the head of the bed 45 degrees, brushing patients’ teeth every eight hours in the hospital or nursing home was as effective as using chlorhexidine mouthwashes in terms of reducing the incidence of hospital-acquired pneumonia.
In his hospitalist program at University of Michigan in Ann Arbor, where Dr. Flanders is the director, that research has led residents to write orders for nurses to brush patients’ teeth every several hours ” an approach, Dr. Flanders said, that just won’t work.
Instead, “I try to engage the patient’s family if they’re around” to brush patients’ teeth, Dr. Flanders said last fall at a hospital medicine conference at the University of California, San Francisco. “I’m convinced this works.”
PHYSICIANS HAVE LONG SOUGHT what Scott Flanders, MD, a national expert on pneumonia, called “the Holy Grail of blood tests, like cardiologists who get their BNP” to see if patients have pneumonia. Such a sure-fire test would let physicians know if a pneumonia is viral vs. bacterial, and whether or not patients need antibiotics. “Right now,” said Dr. Flanders, giving an update on pneumonia care last fall at the University of California, San Francisco, “procalcitonin is bubbling to the top of that list.”
That may be frustrating news to physicians in the U.S., where procalcitonin testing isn’t even available. The biomarker tends to go up as a reaction to a bacterial infection, Dr. Flanders explained, but not with viral infections. It correlates well to disease severity, and “its response tends to predict outcomes,” he pointed out. “We’ve had a lot of preliminary work that suggests its usefulness.”
That usefulness was borne out in a study in the Sept. 9, 2009, Journal of the American Medical Association. Some of the more than 1,300 patients, most with pneumonia, received antibiotics (or did not) according to a procalcitonin algorithm. That enabled 10% of patients with pneumonia to receive no antibiotics, while those given antibiotics based on serial procalcitonin readings were treated seven days vs. 11 days for the control group treated according to guidelines. The procalcitonin group, while it had no difference in outcomes, had fewer adverse effects due to its shorter antibiotic course.
The study was criticized for treating patients too long with antibiotics, which Dr. Flanders said was a valid criticism. And “we don’t know the costs of procalcitonin testing yet, so we’ll see whether this will catch on,” he noted. That certainly seems likely, he concluded, “given what we’ve seen with other biomarkers.”