There's something in the water . . . Could it be Legionella?

Legionnaire's cases have been increasing, so hospitalists should learn to spot them and curb potential outbreaks.


More than 40 years after the first outbreak of Legionnaires' disease in Philadelphia, Legionella continues to wreak deadly havoc on vulnerable populations despite the availability of effective treatments.

Image by Thinkstock
Image by Thinkstock

Patients who are already sick are especially susceptible to serious infection from the common water organism (which also causes the more mild Pontiac fever, a self-limiting flu-like illness). Those most at risk for Legionnaires' disease include people ages 50 and older, current or former smokers, and people with chronic diseases or weakened immune systems, according to the CDC. In 2015, about 6,000 Legionnaires' cases were reported in the U.S., although this is likely an underestimate, the agency said.

Since 2000, the number of cases has more than quadrupled, according to CDC data, although it's unclear whether this is due to increased awareness and testing, increased presence of Legionella, increased susceptibility of the population, or a combination of factors.

About 20% of Legionnaires' cases reported in 2015 were definitely or possibly health care-associated, according to the June 9 Morbidity and Mortality Weekly Report. People typically acquire the disease by inhaling aerosolized droplets of water containing Legionella, which can grow in health care facilities' complex water systems and be transmitted through decorative fountains, faucets, and showerheads.

“This is one of the few pneumonias that we see routinely that comes from the environment rather than being spread person-to-person,” said John A. Sellick, DO, MS, FACP, an associate professor of medicine in the division of infectious diseases at the University at Buffalo.

While hospitalists can do little to stop the environmental spread of Legionella, they can rapidly identify cases of Legionnaires' disease and curb future outbreaks, if they know where to look, experts said.

The right test

In the summer of 2015, the largest outbreak of Legionnaires' disease in New York City's history sickened 138 people and killed 16 in the South Bronx, according to an investigation published in March by Public Health Reports. The outbreak, which was later traced to a hotel cooling tower, prompted authorities to enact a new city law on the operation and maintenance of cooling towers.

Most often, clinicians in the U.S. diagnose Legionnaires' by using urinary antigen (UA) testing for Legionella pneumophila. But UA has its limitations (e.g., it detects only one serogroup of one of the 58 Legionella species), and clinicians should also obtain a sputum culture, noted Tara N. Palmore, MD, FACP, the hospital epidemiologist at the National Institutes of Health Clinical Center in Bethesda, Md.

Although 90% of human Legionella infections are thought to be caused by L. pneumophila serogroup 1, “Those stats are skewed very much by the fact that 97% of cases are diagnosed by Legionella UA, and only 5% of cases are confirmed by culture, according to the CDC. . . . Many hospitals almost certainly are missing cases that are due to other serogroups of L. pneumophila or even other species of Legionella, and they just don't know it because they're not doing enough cultures,” Dr. Palmore said.

Case in point: the Bronx outbreak. Of the 138 cases, only 26 patients (19%) were confirmed by culture as having L. pneumophila serogroup 1.

“Even though most strains are L. pneumophila 1, we see plenty of the other ones,” said Burke Cunha, MD, MACP, chief of the infectious disease division at NYU Winthrop Hospital in Mineola, N.Y., and professor of medicine at the State University of New York School of Medicine in Stony Brook, N.Y. “And in compromised hosts, non-L. pneumophila strains are more common.”

Guidelines on community-acquired pneumonia (CAP) published by the Infectious Diseases Society of America and the American Thoracic Society (IDSA/ATS) in 2007 (scheduled to be updated in summer 2018) recommend obtaining a sputum culture, a blood culture, and UA tests for L. pneumophila and Streptococcus pneumoniae in cases of severe pneumonia or when Legionella is suspected.

Dr. Sellick said he doesn't restrict Legionella testing to patients with severe pneumonia. “I encourage our housestaff to always be thinking about Legionella. I'm never going to yell at you for testing a CAP patient for Legionella, even if they're at low risk,” he said, noting that the broader testing indication helps physicians remember. “I especially work on this with the intensivists because someone making it into the ICU with pneumonia is much more likely to be a very sick person.”

In addition to ordering UA testing, Dr. Sellick also encourages his clinicians to obtain sputum for culture if Legionella is suspected. “But you have to specify to the lab that you're looking for Legionella because labs don't routinely culture all sputums for Legionella,” he noted.

There are other reasons to get a sputum culture. “Often, a clue to Legionella is a sputum Gram stain that doesn't reveal a bacterial source because Legionella doesn't show up well on a sputum Gram stain,” said Dr. Palmore.

She also recommended ordering a sputum culture for Legionella any time the UA test is positive for the infection. “In those cases, it's prudent to get a sputum culture not to confirm the UA test, although the UA test is actually more sensitive than a culture, but actually for epidemiological reasons,” she said. “Public health authorities cannot connect cases to the source without having an actual isolate of the bacteria, and the only way to get that is by culture.”

Since the incubation period for Legionnaires' ranges from two to 14 days, Dr. Palmore noted, the hospital environment may be the source if a hospitalized patient becomes infected.

In hospital-acquired cases, Dr. Sellick said concerns always focus on the sink in the patient's room and occasionally the shower. “In fact, when we used to have our heart transplant program here at Buffalo General [Medical Center], we had it all set up that the patients did not drink tap water and they did not shower, at least in the earliest postoperative phase,” he said. “So they got bed baths and they drank only sterile water to avoid this.”

Dr. Palmore said hospitalists can be the best line of defense against undetected nosocomial clusters. “One way that hospitalists can be sure to detect clusters before they become big outbreaks with high mortality rates is to be more aggressive in obtaining diagnostic testing, because you won't detect a Legionella cluster if you don't test for Legionnaires' disease,” she said.

At the bedside

Legionnaires' disease will remain hidden if physicians don't know what findings suggest the disease, agreed Dr. Cunha. “Legionella is a headliner now. It used to be just some unusual pneumonia,” he said. “Now, there's an increased awareness, better testing. But it's not a matter of better testing; it's a matter of knowing when to suspect Legionella to order the appropriate tests.”

In the days between a patient's first symptoms and the return of Legionella test results, Dr. Cunha said, the savvy hospitalist should be able to determine low or high probability of the infection and start patients on an empiric drug for Legionella if necessary.

“The misconception is that clinically, you can't tell an atypical from a typical pneumonia, or you can't tell Legionella from S. pneumoniae, which is not the case,” Dr. Cunha said. “The problem is, few know what they're looking for, so therefore others don't see anything suggesting Legionnaires' disease.”

Decades ago, Dr. Cunha (considered a Legionella expert) began describing certain weighted clinical and lab characteristics that increase or decrease the probability of Legionella. He's now honed them down to a few key diagnostic characteristic findings and eliminators, which were published in the March issue of Infectious Disease Clinics of North America, which focused on Legionnaires' disease.

Dr. Cunha's key clinical predictor of the disease is fever greater than 38.9 °C (102 °F) with relative bradycardia, and lab predictors include hypophosphatemia at admission or soon after, microscopic hematuria at admission, and highly elevated erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), or creatine phosphokinase (CPK). Clinical eliminators include sore throat, hoarseness, splenomegaly, or severe myalgia. Lab results that can eliminate Legionella include leukopenia, lymphocytosis, and thrombocytopenia.

Hyponatremia can be a “soft clinical clue” or a reminder to test for Legionnaires', but it can also be found in any severe pneumonia, Dr. Palmore said. And Dr. Sellick noted that neurologic symptoms and gastrointestinal symptoms, while sometimes present in Legionella infection, are also not sufficient to help make the diagnosis.

But Dr. Cunha said his core predictors and eliminators of the disease are more valuable in diagnosis than hyponatremia or GI symptoms because they're more specific. Considered together, they equal more than the sum of their parts.

“You see only what you know,” Dr. Cunha said, paraphrasing Johann Wolfgang von Goethe, the German philosopher. “If you don't know about relative bradycardia and you don't know about low phosphorus [as signs of Legionella], you will miss the boat.”

However, not every case of Legionnaires' will feature all of Dr. Cunha's clues. “Not everybody has a low phosphorus, not everybody has a high CPK, but they all have otherwise unexplained relative bradycardia,” he said. “So there's enough to come up with a weighted point score that would suggest Legionnaires' disease is likely. Not perfect, but it's the basis of clinical syndromic diagnosis that warrants further testing.”

Another key indicator for Legionella risk is recent travel history, which the IDSA/ATS guidelines recommend obtaining from the patient. “There tends to be a huge amount of reliance on the technology when the majority of the time, the answer is at the bedside,” said Dr. Sellick. “So you go talk to the patient, you find out what their habits are, what their underlying medical problems are, where they've traveled, where they work, what their hobbies are.”

Hotels, cruise ships, and other buildings with manmade water supplies are what lead to Legionnaires' disease, so it doesn't matter whether patients have traveled in the U.S. or internationally, said Dr. Palmore. “I think hospitalists are conscientious about getting social histories and medical histories, and travel history often is something the patient doesn't bring up unless you specifically ask, so I think that it's good to be mindful to ask that specific question,” she said.

For treatment, there are limited choices, but they are all effective, Dr. Cunha noted. Among the macrolides, azithromycin is used; among the tetracyclines, doxycycline is used; and among the fluoroquinolones, levofloxacin is most often used, he said. “They all are equally effective, with perhaps a slight edge to levofloxacin,” Dr. Cunha said.

Although there's never been a randomized controlled trial to determine the best treatment for Legionnaires' disease, “Most of us tend to prefer fluoroquinolones . . . Yes, there are side effects, but in general, many people feel that they're more active,” agreed Dr. Sellick.

For both typical and atypical pathogen CAP coverage in admitted adults, either doxycycline or levofloxacin works, Dr. Cunha said. “Empiric monotherapy is preferable and has the advantage of IV to PO switch,” he said.

Although New York and certain other locations bear much of the country's Legionella burden, infection with the organism may occur virtually anywhere, said Dr. Cunha.

“This is a day-to-day problem across the world, not just the United States, and people keep finding Legionella in the strangest places: windshield wiper fluid, dental irrigation [systems], misters in supermarkets, cooling towers, hot tubs, towel warmers in hotels,” he said. “The common denominator, of course, is aerosolized water containing a sufficient inoculum of Legionella species pathogenic for humans.”