Rocket Science and the Infectious Diseases Society of America/American Thoracic Society (IDSA/ATS) Guidelines for Severe Community‐Acquired Pneumonia

To the Editor—We have some concerns about the clinical usefulness of the findings by Liapikou et al. [1] in their recent article. The fact that patients who need mechanical ventilation or require vasopressor support almost always need admission to an intensive care unit (ICU) is self-evident. This fact is not much use to a clinician who is trying to decide which patients presenting with community-acquired pneumonia have severe disease and conditions likely to deteriorate. Thus, what is the point of these guidelines? Is their role to allow the accurate characterization of patients in large databases, or are they intended to be clinically useful for predicting an outcome in an individual? We believe there are at least 2 major methodology issues that raise important concerns about the findings of Liapikou et al. [1]. First, the definition of what constitutes an ICU was not stated, and the criteria for ICU admission in each of the participating centers were not defined. On the basis of previous experience, we believe that these criteria are likely to vary substantially from site to site and to be impacted by random issues, such as bed availability [2, 3]. Thus, the primary end point definition of “ICU admission” may not have been stable in the study. Second, the fact that “do not intubate” orders were available for only 41% of patients assessed is concerning. Although this limitation was cited by the authors, the impact of this possible confounder is difficult to quantify. In our studies in Australia [4], “do not resuscitate” orders were documented in 17.9% of all community-acquired pneumonia admissions, which is ∼2–3 times the rate described by Liapikou et al. [1]. Although cultural issues may explain these potential differences, the absence of firm data on an issue that clearly impacts whether a patient is admitted to an ICU is a problem that may affect much of the data analysis, especially with regard to the study's minor criteria. We have recently described the SMART-COP (systolic blood pressure, multilobar chest radiography involvement, albumin level, respiratory rate, tachycardia, confusion, oxygenation, and pH) approach to predicting the severity of community-acquired pneumonia on the basis of whether patients are likely to require intensive respiratory or vasopressor support. Although not cited by Liapikou et al. [1] or in the associated editorial [5], the SMART-COP algorithm identified 85%–92% of patients who required intensive respiratory or vasopressor support [2, 6]. For managing clinicians, the likely need for intensive respiratory or vasopressor support would appear to be a more clinically useful end point against which to measure the severity of community-acquired pneumonia than ICU admission. SMART-COP appears to be clinically useful for predicting this outcome in an individual. Potential conflicts of interest. All authors: no conflicts.