Pseudomonas Infections: Selecting Upfront Antibiotics

Peter L. Salgo, MD; Andrew Shorr, MD; Jason Pogue, PharmD, BCPS-AQID; Yoav Golan, MD; and Marin Hristos Kollef, MD, reflect on their personal upfront approaches to treating Pseudomonas infections before lab results return, considering monotherapy versus treatment cocktails and the inherent risks that need to be accounted for.

Peter L. Salgo, MD: You’ve got a sick patient with pneumonia: an older patient, perhaps immunosuppressed or immunocompromised at the very least because of surgery. What do you start with before you get the cultures back?

Andrew Shorr, MD: In our hospital, if you’re septic, that’s our criteria. If you’re on pressors and you have an elevated lactate, we’re going to call you septic and we know getting it wrong is increasing your risk of death substantially. For pneumonia, playing the odds based on our antibiogram, our best combination is piperacillin, tazobactam, amikacin, and vancomycin. That’s our best start.

Do I feel comfortable doing that routinely? Absolutely not, especially with the emerging literature that piperacillin/tazobactam plus vancomycin equals acute kidney injury in a way that we never thought would exist, and it’s a classic case of medical hubris. Could these 2 drugs that we use like water be harmful? Well, they are, because drugs do a lot of things together that we never conceive of. But playing the odds and knowing that hopefully 2 of those drugs will be off within 48 hours, if not all 3, and that the patient will be on something else, that’s what we start with based on what we know about the current level of data for drugs—even including the new drugs—in terms of actual patients with treated pneumonia in phase III clinical trials; what we know about the cost of these agents; and what I know about susceptibility.

Peter L. Salgo, MD: So, what are you trying to cover on that waterfront? What’s on that waterfront?

Andrew Shorr, MD: It’s Pseudomonas, it’s protease, it’s Enterobacteriaceae, and it’s MRSA [methicillin-resistant Staphylococcus aureus]. Now, MRSA rates in my ICU [intensive care unit]—and I think most ICUs in the United States are not as good as they are in the United Kingdom, which is de minimis—are going down, especially among pneumonia patients in this country, which is why our gram-negative problem is going up. Some of our prevention strategies like chlorhexidine work better on gram positives than gram negatives, so as we squeeze one end of the balloon the other side is counter pulsing.

Jason Pogue, PharmD, BCPS-AQID: Just to reiterate things that have already been discussed, that’s going to be specific to a given institution in what that cocktail ends up looking like. I will, again, emphasize the need to take patients’ specific factors.

Peter L. Salgo, MD: You know what he starts with.

Jason Pogue, PharmD, BCPS-AQID: I know what he does.

Peter L. Salgo, MD: What do you do?

Jason Pogue, PharmD, BCPS-AQID: What I was going to say is that I know he takes into account patient-specific factors, but I bet that not every hospital, not every physician at your institution practicing in the ICU, does that. They go to the cocktail quickly, overnight, from the standpoint that the patient might have a history of an ESBL [extended-spectrum beta-lactamase] that requires a different empiric therapy. What we do is very similar. It’s vancomycin/cefepime and tobramycin, and that’s taking into account the organisms that we have in our unit, the vancomycin /piperacillin/tazobactam thing that’s already been discussed, and just the ability to give the best likelihood of coverage.

Yoav Golan, MD: I think another factor that you have, too, is if you have some clue as to what the source of the infection is. You should definitely take it into account. If we think someone has urosepsis, we would not use a third-generation cephalosporin: We’d use a carbapenem because our ESBL break is very, very high. But with pneumonia we are more interested in covering the more resistant Pseudomonas, and we use, like in Jason’s hospital, vancomycin and cefepime. Again, you have to tailor that to the patient.

Peter L. Salgo, MD: But again, this is all institution-specific. You’ve got to know your institutional antibiogram, right?

Marin Hristos Kollef, MD: I think that’s absolutely correct. Each institution is different, and sometimes within an institution you can see areas that are different. Our BMT [bone marrow transplant] ICU is very different than some of the other ICUs in terms of the pathogens that they see. It’s never that simple, and that’s the problem. That’s why I echo Yoav’s comments about rapid diagnostics. If we have a septic patient, sure, you want to take into account where you think the source of the infection is. But sometimes you just don’t know. Many times, you don’t know because they may have infiltrates or they may have some white cells. Could it be lung? Could it be urinary tract?

We know for example that if we err on 1 side, if I pick an aminoglycoside—we’ll throw that out for a minute—if I go with carbapenem, piperacillin, tazobactam, and cefepime, I’m going to run about a 20% chance of not covering the gram negative with any of those in our unit. You’re right. If I knew it was urinary tract and I went with a carbapenem, I’d be more likely to cover the ESBL. If I really knew it was going to be Pseudomonas in the lung and use cefepime, for example, I’d do a little better, but there’s still a 15% likelihood that I’m not going to cover the bug with a single agent. That’s where we throw in the aminoglycoside. Unfortunately with aminoglycoside, particularly in someone who’s in septic shock, we’re looking at 1 or 2 doses, sometimes more. There is a little penalty to be paid for that from the standpoint of nephrotoxicity.

Peter L. Salgo, MD: What is the morbidity of renal failure in this setting? It’s high, isn’t it? If you start off with a creatinine of 0.9 mg/dL and you add an aminoglycoside in the setting of sepsis, how often do you see creatinine levels of 2, 3, or 4 mg/dL? A lot?

Andrew Shorr, MD: Right, but it’s always hard to tease out cause and effect because these patients are under-resuscitated and there’s a whole host of issues that may be interacting with them. The nephrotoxin that is the most abused thing in my hospital is the CAT scanner. In addition to getting an antibiotic cocktail at night, what happened was that they actually had a hypoxemic event because someone sent them to the CAT scanner.

Peter L. Salgo, MD: With contrast.

Andrew Shorr, MD: Of course, and the pickup rate for Pes [pulmonary embolisms] in this country on CAT scans is 3%, which tells you that we’re overtesting the vast majority of people and exposing them to harm only for the sake of us to sleep better, which is absurd. The point is that there are a lot of nephrotoxins floating around in the ICU and you have to manage all of them, not just the antimicrobials.