Lee Morrow, MD: Today I have been charged with going over the critical decisions that must be made when dealing with patients who have health-care-associated pneumonia, also known as HCAP.

There are my disclosures.

Because there are innumerable critical decisions that need to be made in HCAP patients, I have narrowed my discussion to the 4 decisions nodes I think are the most difficult, important, and/or controversial. The first critical decision will be, does my patient have HCAP? I think that this is the area where we see the most errors being made. Because we aren't rigorously looking for HCAP risk factors, we are not reliably identifying patients with HCAP. Second, we will ask, does HCAP really matter? I think that the answer there is a definitive yes, and it will matter more and more in the future for several reasons I will outline. Third, we will talk about treatment, specifically which pathogens merit empiric coverage. Finally, we will discuss the proper duration for a course of antibiotic therapy.

Although the definition of HCAP is unambiguous, it is also somewhat arbitrary. HCAP is diagnosed in essentially any patient with clinically diagnosed pneumonia who has any of the 5 risk factors shown here: they have been hospitalized for more than 48 hours in the last 3 months; they have lived in a nursing home or long-term care facility for any duration of time in the last 3 months; in the last month they received outpatient infusion therapy—and that can be antibiotics or chemotherapy—or they received home wound care; any patient who has attended a hospital-based clinic or chronic hemodialysis center in the last 30 days; and patients with a family member known to have a multidrug-resistant pathogen.

These definitions come straight out of the American Thoracic Society (ATS)/Infectious Diseases Society of America (IDSA) guidelines that were published in February of 2005. I am sure most everybody here has sat down and wrangled with this document in various levels of detail. It is a pretty incredible piece of work with a total of 28 pages divided among 7 major subcategories within hospital-acquired pneumonia (HAP), ventilator-associated pneumonia (VAP), and HCAP.

These subcategories address epidemiology and pathogenesis, prevention, strategies for diagnosis including diagnostic testing, empiric antibiotic treatment, and assessing the clinical response to therapy. Although I am going to say some things today that are somewhat critical of the ATS/IDSA guidelines, specifically as they relate to HCAP, I do not want you to think that I am trying to undermine the document. Quite the opposite, I think it is a phenomenal resource with multiple strengths. The guidelines are exhaustive, evidence-based, concisely written, and can be used at the bedside to guide patient care. They also represent collaboration between the most authoritative experts in the fields of pulmonary, critical care, and infectious disease medicine. I think it is particularly appealing that the guidelines also dissected the nosocomial pneumonias, specifically identifying HCAP patients as a distinct population. Finally, it is important to recognize that the new guidelines were very timely.

Why do I say they were timely? Well, if you look at the number of publications we had in 2005 and said these publications represented 100% of our knowledge base for nosocomial pneumonias, you can then go back and subtract out year-by-year the number of publications published in each calendar year. What you end up creating is this display of how our knowledge base has grown incrementally over time.

Does anybody remember when the previous nosocomial pneumonia guidelines were published by the ATS? It was in 1996. Looking at the graph we see that in 1996, our knowledge base was only about 17% of what it was when the updated guidelines came out in 2005. Because our knowledge base had grown 6-fold but we had not updated the guidelines, these revisions were clearly needed.

Whenever anybody provides something that is evidence based, it is important to assess the strength of the evidence. What I have done here is to stratify the evidence supporting the major recommendations made in each of the 7 major subtopics of the guidelines into level I, II, and III. Over here on the left, we have level I evidence consisting of multiple randomized control trials where you have a very low likelihood of drawing erroneous conclusions. On the other end of the spectrum, we have level III evidence consisting of uncontrolled studies, case reports, case series, expert opinion, etc. where the margin of error is significantly higher. If you look overall, you can see that 60% of the recommendations made by the guidelines are supported by Level II evidence. About 28% are supported by very strong (Level I) evidence, while about 13% is very weak (Level III) evidence. Given the limited supporting evidence for many of the recommendations, publication of the guidelines was swiftlyfollowed by innumerable letters to the editor. The bulk of the criticisms focused on the fact that although it intuitively made sense to carve out HCAP patients as a distinct entity, there was not much hard supporting data. There were also serious concerns regarding the heterogeneous groups lumped together under the HCAP umbrella and the potential implications of the guideline-recommended treatment for these patients.

I think that if you look at the evidence specifically relating to HCAP patients, you see that the data is really weak. Of the 294 references cited by the ATS/IDSA guidelines, only 7 had anything to do with HCAP. If you couple the various guidelines subtopics and the HCAP citations, you can see that the data are pretty thin in all of these subareas. Note that reference number 6 appears in multiple boxes in this graphic because it was essentially a review article. The most worrisome point here is that all 7 of the references cited regarding HCAP come from nursing home-associated pneumonia. So although HCAP includes patients with home wound care, home antibiotic therapy, and dialysis patients, there are really no data supporting these recommendations in those patient subsets.

If you are able to successfully identify HCAP patients, the next question becomes does HCAP really matter? Again, I think the answer is that it does. The logic behind pulling HCAP patients out from the overall pneumonia population really gets at the fact that these are the patients who are more likely to be colonized with resistant pathogens. As you all know, you cannot open up any major journal these days without seeing headlines like these screaming back at you. Historically, we have lulled ourselves into a false sense of security by saying, "Well, this resistance issue is really something that is an isolated ICU phenomenon. You intensivists take patients who are critically ill and marinate them in multiple antibiotics for a couple of weeks—of course you are going to see a lot of resistance. But we do not have to worry about that problem out on the general wards."

I think the Fridkin data displays that this line of thinking is simply not true. What you see in the general ward patients, these orange lines, is exactly parallel with resistance rates that you see in the ICU, the blue lines. Although these data show the trends for Gram-positive organisms, you could put up identical graphics for Gram-negative isolates.

Even more worrisome are the data from the NNIS, the National Nosocomial Infection Survey, where we see that the rates of antimicrobial resistance in the ICU are, indeed, very scary. Resistance rates on general wards are not that far behind, and when you look at isolates obtained in the outpatient setting there are very striking levels of resistant bugs there as well. The bottom line is that non-ICU resistance is increasing at a rate that is essentially in lockstep with what we see in the ICU.

The reason that antimicrobial resistance is so critically important is because the patients who harbor resistant pathogens are the ones who are most likely to receive inadequate empiric therapy and subsequently have bad outcomes. The studies listed on this slide highlight that if we do not get the empiric antibiotic prescription right, the mortality rate essentially doubles. You could pull up multiple other studies looking at outcomes, and although the magnitude of the bars will change based on the population being studied, they would show with remarkable consistency that the mortality rate is approximately doubled in patients infected by an organism not effectively treated by the empiric antibiotic regimen.

The next slide is here to remind me to comment that one area where we see a lot of misunderstanding is the frequent confusion between HCAP and severe community-acquired pneumonia (CAP). Frequently, I have gone to a physician who has consulted me on a pneumonia case and pointed out that the patient had risk factors for HCAP and the response has been, "they are not sick enough to have HCAP." Understand that according to the ATS/IDSA guidelines, severity of illness has nothing to do with HCAP. HCAP relates only to the risk of harboring multidrug-resistant pathogens. Because the CAP guidelines incorporate severity of illness in the empiric therapy choice, it is very important to make sure that everybody understands that illness severity is irrelevant in HCAP.

The figure on this slide comes from Don Craven's publication in Current Opinions in Infectious Diseases last year. It serves as a concise reminder that although HCAP is generally more similar to HAP and VAP than CAP, HCAP overlaps significantly with both CAP and HAP in terms of the risk of harboring multidrug-resistant pathogens, implications for morbidity, and mortality. As we will see shortly, this has significant implications for which pathogens we are going to empirically cover with our initial antibiotic therapy.

Assuming that you now agree that HCAP matters, it is also important to appreciate that the incidence of HCAP is likely going to increase for a sustained period of time. In the United States, we have an ever-increasing number of ambulatory and non-hospitalized patients with multiple comorbidities who regularly come into contact with various aspects of the health-care system. As the baby boomer population ages, we are concurrently facing an explosion in the number of people who will be residing in nursing homes. Because the rates of pneumonia are much higher in these ambulatory patients with multiple comorbidities and in nursing home residents, HCAP is going to be a growing problem for the foreseeable future. A recent analysis by Scott Micek at Barnes-Jewish Hospital suggests that HCAP is already the largest subcategory of pneumonia. Micek and colleagues did a single center review of all patients admitted to their hospital with a diagnosis of pneumonia, teasing out the patients withHCAP criteria and comparing them to those who had CAP. What they showed was that two-thirds of pneumonia admissions met the HCAP criteria. Although I was initially surprised by that finding, we have performed a similar analysis at Creighton and found that 62% of our admissions last year were HCAP. I will share more data from the Micek study with you shortly.

When discussing whether HCAP matters, I think I would be terribly remiss if I did not address public opinion. Similar to what we see in the scientific literature, you can look anywhere in the lay press these days and see an endless stream of reports on antimicrobial resistance, the negligence of the medical system in terms of not addressing it, our failure to use routine prevention measures, etc. The bottom line is that this issue is now on the average person's radar screen, and it is clearly affecting their perception. As an example, this is Todd. I am using his picture today with his written consent. Todd is the husband of one of our former ICU patients at Creighton. His wife was hit by a delivery van while crossing the street and, given the extent of her injuries, went to the operating room almost immediately after arriving in the ED. Because it was late in the day, we suggested that Todd find her assigned room in the ICU and catch some sleep in a recliner rather than waiting upall night in the general surgical waiting room. As soon as Todd got to the ICU, he went to the isolation cart parked outside the room of the patient next door who had Clostridium difficile. He promptly stole a pile of blue gowns and gloves, set them outside his wife's room, and put multiple signs up instructing all health-care workers who went into her room to wear a gown and gloves and to wash their hands. When I asked Todd if she had a history of methicillin-resistant Staphylococcus aureus (MRSA) or vancomycin resistant Enterococcus (VRE) that we needed to know about he said, "No, but last week I saw a 60 Minutes story about how patients come from nursing homes with resistant bacteria and health-care systems spread them to the healthy patients, and I do not want her picking anything up." At least in the arena of public opinion, HCAP really does matter.

Does anybody have any idea why I am showing you the cover from the May 2007 Federal Register? I put it up here because the Centers for Medicare and Medicaid Services have decided that certain nosocomial infections will no longer be reimbursed. As of January 1, 2008, hospitals are not going to get paid for catheter-related blood stream infections and catheter-related urinary tract infections acquired in the hospital. There were significant discussions whether to also withhold reimbursement for VAP. That was struck down, apparently given the lack of an accepted gold standard for establishing the diagnosis. This issue will certainly be a major topic of discussion in 2009, and I think that it is just a matter of time until we find HCAP being scrutinized as well.

Let's shift gears and talk about treatment for the remaining time. First, let's talk about which pathogens need to be covered empirically. I think that the flow chart from the ATS/IDSA guidelines is a little bit confusing so I have created an edited version here. Essentially the guidelines say that all HAP, VAP, and HCAP patients need to be assessed for the presence of risk factors for colonization with multidrug-resistant pathogens (MDRP). However, by definition everybody with HCAP has a risk factor for MDRP. Accordingly, all HCAP patients should be treated with broad-spectrum therapy that covers conventional pneumonia pathogens as well as MDRP such as Pseudomonas, Acinetobacter, MRSA, and the extended-spectrum beta lactamase (ESBL)-producing organisms.

What constitutes broad-spectrum therapy? Well, it is actually very broad-spectrum therapy using 3 agents—one from each box in this figure. For our first agent, we pick one agent from the top box—an antipseudomonal cephalosporin, an antipseudomonal carbapenem, or a beta lactam/beta-lactamase inhibitor with antipseudomonal activity. To that agent, we add our second agent—an antipseudomonal fluoroquinolone or an aminoglycoside. Finally, our third agent is empiric MRSA coverage with vancomycin or linezolid. We should also stress the need to start the antibiotics in a timely fashion and to ensure adequate doses at appropriate intervals.

It seemed ironic to some people that at the same time the IDSA was recommending this antibiotic nuclear bomb, we were also seeing IDSA's "Bad Bugs, No Drugs" campaign. This educational effort reiterates the need for better antimicrobial stewardship given the problems with increasing resistance and the lack of new antibiotics coming to market. Accordingly, many individuals were skeptical of the need for such broad empiric therapy as there was a paucity of microbiologic data supporting the assertion that the pathogens causing HCAP were truly distinct from those causing CAP.

However, Marin Kollef came to our rescue in December of 2005 with this publication from Chest. Essentially, this study was a large retrospective database analysis where 4500 cases of pneumonia were identified using ICD-9 codes. The investigators then classified each patient as having CAP, HCAP, HAP, or VAP. Because all of the patients studied had positive cultures, they were able to correlate the specific pneumonia subtype with the causative pathogens. What they showed was that the pathogens causing HCAP were much more similar to those seen in HAP and VAP than in CAP. If you look at the rates of MRSA and Pseudomonas, they are much higher with HCAP, HAP, and VAP; whereas if you look at Streptococcus pneumoniae and Haemophilus, those were the big players in CAP but were insignificant factors in HCAP, HAP, and VAP.

This was corroborated by the study in Antimicrobial Agents and Chemotherapy by Scott Micek that we touched on earlier. These were data from a single institution with about 640 pneumonia admissions over the span of a year where patients were stratified into CAP and HCAP. Again, you see twice as much HCAP as you do HAP. If you look at the pathogens, you see the exact same thing as was demonstrated in Kollef's study; there is significantly more MRSA and Pseudomonas and much less S pneumoniae and Haemophilus in the HCAP patients when compared with CAP patients.

Another way of looking at this is to identify infections caused by a specific resistant pathogen and to then compare the incidence of various risk factors in those patients to a cohort of controls with infections caused by other pathogens. These are French data looking at risk factors for health-care-associated MRSA (HA-MRSA) acquisition, and surprisingly—or maybe not so surprisingly—the authors found that the risk factors associated with HA-MRSA acquisition were essentially the same risk factors for HCAP that the ATS/IDSA has put forth. Patients who were receiving home nursing care, including wound care and home IV antibiotics, had significantly more HA-MRSA. Patients who had previously been hospitalized or were transferred in from a nursing home also had more HA-MRSA. So what we have are data from a slightly different angle that again support the hypothesis that these subsets of patients are more likely to be colonized with MDRP.

One of the more controversial guideline recommendations is empirically double-covering Gram-negative rods. When you look at the objective data to support that practice, you see that there isn't any. A lot of people have erroneously extrapolated this as being a strategy for minimizing downstream Pseudomonas resistance; however, that is really not the goal behind this. The real goal is to increase the likelihood that patients are empirically put on at least one agent that their causative pathogen is susceptible to.

I think that ultimately the bottom line in terms of support for the ATS/IDSA's recommendations for such broad antibiotic therapy comes from the mortality data. Note that these mortality figures are eerily similar to what I showed you earlier with initially inadequate empiric antibiotic therapy. The mortality rates for patients hospitalized with HCAP are essentially double that from patients with CAP in both the Micek data and the Kollef paper from 2005.

That being said, I think that we should acknowledge the data suggesting that maybe the empiric therapy for HCAP does not have to be the same as that for HAP, despite all of the data we have reviewed. In the interest of time, we are not going to go through all of these points in detail. However, I would point out that while HAP therapy is almost always administered intravenously in the hospital, there are numerous studies showing that outpatient oral monotherapy is effective in various HCAP subsets, specifically end-stage renal disease patients with pneumonia and nursing home-acquired pneumonia patients.

Contrary to the Kollef and Micek data demonstrating a high frequency of MDRP in HCAP, older studies looking at nursing home patients and hemodialysis patients showed much lower frequencies of these agents. Whether this is center dependent or a consequence of obsolescence is unknown. There is also some uncertainty regarding possible interactions between the MDRP colonization risk factors in these various subsets of patients.

Additional sources of confounding information are the accumulated clinical trial data from pharmaceutical companies and independent investigators showing that antibiotic regimens that would not be compliant with the ATS/IDSA guidelines have often resulted in acceptable clinical cure rates. Specifically, there are multiple monotherapy studies that have been done in non-hospitalized nursing home-acquired pneumonia patients. There is also at least one study looking at patients who failed a previous course of antibiotic therapy who were successfully treated using an oral fluoroquinolone as monotherapy.

This next slide is a late addition. The reason it is here is because on October 12, 2007, the FDA granted approval to doripenem, a novel carbapenem with a very broad spectrum of activity. Although the FDA-approved indications are for complicated intra-abdominal infections and complicated urinary tract infections, there is also a head-to-head trial comparing doripenem with piperacillin/tazobactam in nosocomial pneumonia. Although the study was designed to meet a noninferiority endpoint, the increased susceptibility of Pseudomonas and Klebsiella strains to doripenem is particularly interesting. Given our concerns for empirically covering these pathogens in HCAP, I think it is going to be a very intriguing molecule to explore in these patients as well.

Last but not least, let's discuss the proper duration of a course of antibiotic therapy for a patient with HCAP. The answer here is essentially unknown. Because we don't have any good data in HCAP populations specifically, current recommendations are extrapolated from the HAP and VAP populations. Overall, it is now evident that shorter courses of antibiotics are just as effective as longer duration regimens if the patient was not initially toxic and was started on appropriate initial therapy. I suspect that everybody is familiar with the Chastre data from JAMA showing that the rates of death and recurrence of infection were not significantly increased by shortening the antibiotic duration from 15 days to 8 days. I personally believe that rather than arbitrarily picking a set duration of therapy, you have to consider measures of patient response to therapy.

I think most people are familiar with the data from Luna and colleagues looking at the clinical pulmonary infection score (CPIS) to identify patients who are responding to therapy. Realize that the CPIS is nothing more than a mathematical construct of several clinical factors we routinely assess at the bedside—fever, white blood cell count, sputum consistency, etc. Patients with a falling CPIS are those who are responding to the prescribed therapy and are more likely to survive their illness. An unchanging or rising CPIS identifies those patients who are failing therapy and are more likely to die in your ICU. The bottom line here is that there is no one-size fits-all answer to the duration of therapy in HCAP. As with other groups of pneumonia patients, you need to incorporate the clinical response to therapy as well.

In summary, I hope that I have highlighted the clinical importance of the various HCAP decisions we have talked about. I hope I have also conveyed that our current knowledge base regarding HCAP is really inadequate. Much of our data are extrapolated from other nosocomial infections. The data we do have come almost entirely from the nursing home population where there are multiple logistic problems in terms of obtaining microdata. When we do get microdata, we are never confident in its quality. We do not have a firm grasp on what to do with atypicals and Legionella because the bulk of the data from nursing home-associated pneumonia consist of clinical trials with rigorous inclusion and exclusion criteria. Most of the currently available microbiology studies look only at hospitalized patients. Last but not least, much of the data that we are basing our HCAP decision-making process on come from studies that are now 10 to 15 years old, which in absolute time value is notsignificant. If you look at what has happened in terms of antimicrobial resistance during that time span, one does have to wonder whether those studies are relevant. We simply do not know the effects of these cumulative limitations on our current guidelines.

Dr. Kollef: Okay, thank you, Lee. Our next speaker is Lena Napolitano, Professor of Surgery at the University of Michigan, and she will be discussing the topic of complicated skin and skin-structure infections (cSSSIs) and their management.