Less Is More: A 7-Day Course of Antibiotics Is the Evidence-Based Treatment for Pseudomonas aeruginosa Ventilator-Associated Pneumonia

(See the Viewpoints Article by Albin et al on pages 745–9.)

Recommendations for the length of antimicrobial therapy for ventilator-associated pneumonia (VAP) due to Pseudomonas aeruginosa (PsA) have changed over time. The PneumA study, published in 2003, found similar overall outcomes in patients with VAP treated with 8 versus 15 days, but a potentially increased risk of VAP recurrence with 8-day regimens in patients with VAP due to nonfermenting gram-negative bacilli (NF-GNB) [1].

The PneumA study triggered a change in treatment guidelines from 14–21 days [2] to 7 days for most organisms except for NF-GNB, which remained 14–21 days due to the perceived increased risk of recurrence among these patients [3]. In 2016, this recommendation was re-examined in the updated guidelines, in light of evidence that clinically meaningful patient-centered outcomes were similar in patients with PsA-VAP who received short- versus long-course therapy; 7 days of antibiotic therapy is now recommended for all bacteria as long as the patient is responding to treatment [4].

In this issue of Clinical Infectious Diseases, Albin et al [5] argue that the evidence in favor of short-course treatment is “murky,” and that further research should be done to address this question. While we agree that the existing data addressing the length of therapy for PsA-VAP are not perfect, we do have 2 high-quality randomized controlled trials (RCTs) to inform our practice, a luxury not available for many other important clinical decisions.

Albin et al base their argument against the recommendation for short-course therapy on 2 concerns. The first concerns the methodology used to create a meta-analysis on risk of recurrent pneumonia with short- versus long-course therapy for the 2016 guidelines. They noted that some patients who did not have PsA-VAP were mistakenly included in the recurrence meta-analysis and were thus concerned that the guideline panel's recommendation for short-course therapy was misinformed by this meta-analysis. The authors of this Editorial Commentary served as the Co-Chairs and a member of the 2016 Guideline panel. We have reviewed the documentation from this meta-analysis and agree that the meta-analysis results should not be used to support or refute the recommendation regarding length of therapy for PsA-VAP. Importantly, the same review of the guideline records confirmed that this meta-analysis was performed after the panel had reviewed and discussed the existing evidence identified by our systematic review, including previously published meta-analyses, and unanimously recommended short-course therapy. The meta-analysis in question was intended to provide supplemental information for readers alone. Indeed, the rationale for our recommendation stated the following:

“The panel considered whether a separate recommendation was indicated for patients with VAP due to non–glucose-fermenting gram-negative bacilli, in light of the previous evidence suggesting that recurrence may be increased in such patients [italics added] who receive a short course of antibiotics. The panel agreed that a different recommendation was not indicated because, even if there is a small increased recurrence rate, mortality and clinical cure do not appear to be affected” [4].

The argument against short-course therapy for PsA-VAP expressed by Albin et al centers on the potentially increased rate of recurrence in patients who receive short-course therapy. This is not a compelling argument because, when the appropriate statistical analysis is performed, accounting for time at risk for recurrence, there was no increase in the rate of recurrence. Because outcomes were measured at day 28 in the PneumA study, short-course patients were potentially at risk for recurrent VAP for 20 days after discontinuing antibiotics while long-course patients were at risk for 13 days. All told, the short-course patients in the PneumA study had 1.6 relapses per 100 patient-days at risk, while the long-course patients had 1.5 relapses per 100 patient-days at risk, while for any recurrent pulmonary infection, both had 2.0 per 100 patient-days. Furthermore, in the PneumA study, the patients with VAP due to NF-GNB who received short-course therapy had numerically lower mortality at 28 days (15/64; 23.4%) compared with long-course therapy (19/63; 30.2%), and a similar number of days on mechanical ventilation, organ-failure–free days, and intensive care unit (ICU) length of stay. Thus, the recurrences were most likely not clinically significant episodes of pneumonia; rather, they were more likely detection of persistent microorganism colonization.

A new open-label noninferiority RCT comparing 8 versus 15 days of antibiotic therapy for PsA-VAP, the Impact of the Duration of Antibiotics on Clinical Events in Patients With Pseudomonas Aeruginosa Ventilator-associated Pneumonia (iDIAPASON) study, was published in July 2022 and, overall, it provides further support for the safety of short-course therapy [6]. The investigators used a composite endpoint of mortality and recurrent PsA-VAP within 90 days as their outcome and found an increased risk of the composite outcome in short-course patients. However, this trial lacked power to meet its prespecified noninferiority margins: only 31% of the planned sample size was enrolled (186 out of planned 600 patients). Just as an underpowered superiority trial cannot conclude “lack of superiority,” an underpowered noninferiority trial cannot conclude “lack of noninferiority.”

Importantly, closer review of the iDIAPASON results demonstrates remarkable similarity to the PneumA trial. The composite endpoint was driven almost exclusively by recurrent PsA-VAP in the short-course group, based on a post hoc analysis. But, as in the PneumA study, clinical outcomes, including mortality, ventilator days, ICU days, and acquisition of multidrug-resistant (MDR) pathogens, were the same for 8 versus 15 days, while antibiotic exposure was lower in the short-course patients.

The issue again was time at risk for recurrence. Albin et al contend that the imbalance in time at risk did not apply to the iDIAPASON study, as the outcomes were assessed at 90 days. However, this ignores the fact that the “recurrent” VAP was only sought while the patient remained in the ICU and starting at least 48 hours after antibiotics were stopped. The median ICU stay in the study was 34 days in both arms and the median time to enrollment was 14 days in the 15-day group and 11 days in the 8-day group. Hence, the median time at risk for recurrence for patients in the 15-day group was just 3 days (34 – [14 + 15 + 2]) versus 13 days for those in the 8-day group (34 – [11 + 8 + 2]), a 4-fold difference in time at risk. Thus, the iDIAPASON study suffers from the very same statistical pitfall as the PneumA study with respect to differential time at risk for recurrent VAP.

Even if one wrongly discounts the impact of the imbalance of time at risk on VAP recurrence rates, the results of these 2 studies beg the question as to why the higher recurrent VAP rates did not negatively impact the frequency of any patient-centered clinical outcomes. The PneumA trial was unblinded after day 8, while the iDIAPASON study was open-label. Thus, the decision to investigate patients for suspected recurrent VAP may have been substantially influenced by a greater concern for recurrence in the short-course patients by the unblinded frontline clinicians (the blinding of the post hoc panel ultimately adjudicating the diagnosis of recurrent pneumonia does not mitigate this concern because the adjudication panel had to rely on frontline clinicians' real-time decisions on whether to culture and image patients). Indeed, the potential for such over-investigation bias is a credible explanation for the seeming lack of clinical consequences of the “recurrent” cases of VAP. Additionally, in both the PneumA study and iDIAPASON study, there was a numerically higher risk of recurrent VAP caused by an organism other than PsA (labeled as superinfection) in short-course patients, a finding that is hard to explain other than by a greater bias to search for recurrent VAP, in addition to the increased time at risk in the short-course patients.

All told, there are 4 major flaws in the arguments of Albin et al:

1. Albin et al did not account for assessment/ascertainment bias (trial design): All RCTs were fully or partially open-label—that is, patients, healthcare providers, and research team members all knew which antibiotic treatment duration arm (short vs long) was being administered to the patient before the end of the trial. Compared with long–antibiotic-course patients, the short-course patients had an additional 7–10 days of observation during which fluctuation in cough, sputum, dyspnea, and laboratory results could prompt repeat culturing, imaging, and increase misdiagnosis. The positive-predictive value of a positive sputum or bronchoalveolar lavage (BAL) culture for histologically confirmed VAP is only 20–30% [7] given ventilated patients' high rates of microbial colonization, particularly among those with prolonged ventilation. Thus, the open-label RCTs are highly susceptible to recurrence assessment/ascertainment bias against the short course of antibiotics.

2. Albin et al did not adjust for the differential time-at-risk bias (statistical analysis): The simple fact that patients in the short-course groups are observed for recurrent pneumonia for at least 7–10 days more than patients in long-course groups makes them not only more prone to be misdiagnosed with recurrence if they develop one of the many hospital-acquired conditions that mimic recurrence (fluid overload, thromboembolic disease, atelectasis, acute respiratory distress syndrome [ARDS], hypersensitivity, etc) but also prone to have persistently positive cultures due to colonization that support a misdiagnosis of recurrent infection. When calculated as recurrence of PsA-VAP per day at risk, the recurrence rates for short- and long-course treatments are nearly identical.

3. Albin et al did not evaluate clinically meaningful patient-centered outcomes (morbidity and mortality): Powerful evidence that most “recurrences” are wrongly defined and do not represent another episode of VAP is the fact that the higher “recurrence” rate is not associated with more prolonged hospital or ICU stay, mechanical ventilator duration, or mortality. This important and consistent lack of association between “recurrence” and any clinically meaningful outcomes demonstrates that “recurrent VAP,” as defined in the published RCTs, is an unreliable outcome measure on which to base recommendations for the duration of antibiotics for PsA-VAP.

4. Albin et al did not consider patient harms introduced by a longer course of antibiotics: Patient harm is a universally accepted consequence of unnecessarily prolonged antibiotic treatment, with complications including serious drug side effects, Clostridioides difficile colitis, and development of antimicrobial resistance. In the PneumA trial, long-course therapy resulted in 4.5 days more antibiotic exposure and more frequent subsequent isolation of MDR pathogens among patients with recurrent VAP (P = .04), further evidence of the harm associated with long-course treatment.

Albin et al conclude their Viewpoint urging further study of this issue and suggest some aspects of study design. We would not argue against the development of new data, if clinically necessary and ethically sound. However, it is unlikely that more conclusive data will be generated on this issue for several reasons. First, the 2 existing high-quality RCTs demonstrated remarkably concordant results, providing little imperative for further study. Second, the similarity in all clinically meaningful patient-centered outcomes between short and long antibiotic courses has been consistent. Third, patient harm from antibiotics side effects is directly and proportionally related to the duration of antibiotic courses. Given the lack of equipoise (no additional clinical benefit and increased risk of harm with a long course), it could be argued that it would not be clinically appropriate to conduct another trial, or ethical to expose future trial patients to unnecessary and unsafe prolonged courses of antibiotics.

Thus, clinicians will have to decide on short- versus long-course therapy for PsA-VAP based on currently available data. Fortunately, the evidence is more robust and clinically applicable than Albin et al believe. The 2 RCTs demonstrated equivalent patient-centered clinical outcomes and lower antibiotic exposure in patients with PsA-VAP treated with short-course therapy. The potentially increased frequency of recurrence in short-course patients is falsely elevated due to the bias of over-investigation and, most importantly, differential time at risk. And most notably, even if the “recurrences” were increased, there were no differences in clinically meaningful outcomes such as hospital and ICU stay, duration of mechanical ventilation, and mortality among patients with PsA-VAP treated with short versus long courses. We stand by our recommendation from the 2016 guidelines. Less is more, a 7-day course of antibiotics is the evidence-based treatment for PsA-VAP.

Author notes

Potential conflicts of interest. Each of the co-authors served on the panel of the guideline, entitled “Management of Adults With Hospital-acquired and Ventilator-associated Pneumonia: 2016 Clinical Practice Guidelines by the Infectious Diseases Society of America and the American Thoracic Society.” M. K. reports grants or contracts to his institution from the Centers for Disease Control and Prevention and the Agency for Healthcare Research and Quality; royalties or licenses to the author from UpToDate; and other financial or nonfinancial interests as a member of the Infectious Diseases Society of America/American Thoracic Society (IDSA/ATS) guideline panel on management of adults with hospital-acquired pneumonia and ventilator-associated pneumonia. A. C. K. reports other financial or nonfinancial interests as Co-Chair of the IDSA/ATS guideline panel on management of adults with hospital-acquired pneumonia and ventilator-associated pneumonia. M. L. M. reports other financial or nonfinancial interests as Co-Chair of the 2016 IDSA/ATS guideline panel on management of adults with hospital-acquired pneumonia and ventilator-associated pneumonia Guideline panel.