Improving Empiric Antibiotic Regimens for HABP/VABP in the ICU

Key Takeaways

• Institutions can implement rapid diagnostics, antibiograms, and prediction scores to improve the selection of effective empiric antibiotic treatment for .
• Although clinical guidelines provide recommendations for the antibiotic regimens and length of therapy for HABP/VABP, some treatment decisions are best personalized to each patient.

The following is a recap of the most commonly asked questions from a program discussing strategies to improve empiric antibiotic regimens for healthcare-acquired bacterial pneumonia (HABP)/ventilator-acquired bacterial pneumonia (VABP) in the ICU. 

How can institutions implement rapid diagnostics for lower respiratory tract infections? What are the limitations of this ?

Keith S. Kaye, MD, MPH:
Implementing rapid diagnostics requires an infrastructure and crossteam plan about what to do with the results. Obviously, you need buy‑in from your ICU, critical care team, respiratory therapy team, microbiology lab, and electronic medical records team.

Andrew Shorr, MD, MPH, MBA:
A problem with current rapid diagnostics for the respiratory tract is that the lungs are not a sterile site. So how do you differentiate pathogen from colonizer? Unfortunately, you can’t, and you need to consider what is going on with your patient. The current generation of rapid diagnostics for the respiratory tract are very genetically focused and do not give us quantitative information about the burden of culture or density of genome in the front vs the background of other pathogens. This is why I find bacteria identification by a rapid diagnostic not as in someone whom I think has HABP/VABP. This has hindered their adoption in respiratory infection. The next generation of diagnostics is addressing this, and when we get there, I think there will be rapid adoption.

For places that do not have ICU antibiograms, how would you suggest they select empiric antipseudomonal coverage?

Marion Elligsen, BScPhm, MSc, RPh, ACPR:
That is challenging. You really need to think about your local antibiogram, but in the absence of that, you can refer to regional antibiograms.

If a site wants to consider implementing advanced what steps are required? Are there any tips for creating institutional prediction scores?

Marion Elligsen, BScPhm, MSc, RPh, ACPR:
Getting your microbiology lab on board to give you full susceptibility panels is key. Once you have the data, you do not need a fancy software or statistician to manipulate them. For the escalation antibiograms, it was a little more challenging for us because we had to impute some gram‑negative resistances not tested by our microbiology lab.

The clinical prediction score comes next and requires a lot of patient‑level data. Our institution has a stewardship database that identifies patient‑specific factors hospital wide. We then derived the clinical prediction scores by comparing those factors with the microbiology results and validating the scores with another data set.

Are there any biomarkers that suggest whether an empiric antibiotic regimen is adequate?

Andrew Shorr, MD, MPH, MBA:
I think serial biomarkers are the best. Based on current data, increasing concentrations of procalcitonin is a bad prognostic sign. However, measuring procalcitonin is no better than looking at how the Clinical Pulmonary Infection Score changes over time. The biggest component of this measurement tool is whether the ratio of PaO₂ to FiO₂ improves within 1.5 days. If it is not improving within this timeframe, odds are that the patient is receiving inappropriate therapy.

How long do you typically treat patients who have VABP?

Marion Elligsen, BScPhm, MSc, RPh, ACPR:
We routinely treat patients for 7 days for VABP associated with Pseudomonasaeruginosa and Staphylococcus aureus. Whether you need to treat longer is controversial, so we typically individualize the length of treatment to the patient. That includes factors such as how sick the patient looks and how likely they are to survive a reoccurrence, should that happen.

Andrew Shorr, MD, MPH, MBA:
We do the same thing. The absolute “right” duration of therapy for a patient with documented P aeruginosa pneumonia is unknown. Although 14 days is probably too long, 7 days is probably too short for some subgroups of patients. The main thing to remember is that more than 14 days is actually worse for patients.

Keith S. Kaye, MD, MPH:
I agree. If a patient is still struggling and not improving, and only starting to turn the corner on Day 5 or 6, they should probably receive treatment a little longer. But patients who respond in the first 3 days and are almost ready to come off the ventilator probably only need 7 or 8 days.

For patients with HABP refractory to ceftazidime/avibactam or ceftolozane/tazobactam, what would be your next choice of therapy between cefiderocol and imipenem/relebactam/cilastatin? What considerations would you make in choosing the next therapy?

Andrew Shorr, MD, MPH, MBA:
I think the first thing to determine is whether this is drug failure or dosing failure. If you have a patient with augmented renal clearance, with a glomerular filtration rate of 150 or 200 mL/min/1.73 m², and you did not take that into consideration in the dosing, then you are likely underdosing. A lot of times it also can be host failure, where the patient is in shock and dying, so no drug is going to help.

Keith S. Kaye, MD, MPH:
Yes, and I also think susceptibility testing is important at this point. If you are using one of the novel combinations, you should tee up the other options for testing so that you already have susceptibility results ready. I would probably pick the drug that is more likely to be susceptible, and cefiderocol looks very good in vitro. We do have more experience with imipenem/relebactam/cilastatin, and there are some randomized data using it in drug-resistant P aeruginosa, but susceptibilities will be really important there.

You see extended infusion times built into a lot of the novel agents. When is the right time to do that?

Marion Elligsen, BScPhm, MSc, RPh, ACPR:
I think extended make a lot of sense, and we implement them in the setting of multidrug resistance, but not routinely. It is on our list of our stewardship priorities for the ICU to do more extended infusions.

Keith S. Kaye, MD, MPH:
I think the ICU and acute phase of illness are probably the most important, and we are implementing it for certain targeted antipseudomonals.

It is important to remember that the first dose should not be extended. You want to get that on board as quickly as possible and then extend. The individual patient level is where data have been most impressive in showing that it makes a difference for these borderline minimum inhibitory concentrations.

How do you feel about antibiotic cycling, and at what level (if at all) do you think it should be implemented?

Andrew Shorr, MD, MPH, MBA:
The original cycling studies that showed a benefit did so only because they cycled from antibiotics with low probabilities of hitting the 80% or 90% target to antibiotics more likely to hit the target. So, I think the point is you cycle in the individual patient. Cycling at the system level is not necessarily something you need to do.

Your Thoughts?
How do you reduce time to effective therapy in patients with HABP/VABP in the ICU? Join the discussion by posting a comment.