PulmCrit Hot Take: Steroid for severe pneumonia (CAPE COD trial)

Steroid is worthless for pneumonia! Wait, a new study shows it's beneficial! Hang on, a fresh meta-analysis shows that steroid is useless! And wait, here's a fresh NEJM study showing mortality benefit!!

This feels like a roller coaster. What's going on here?

Several factors may explain this:

Steroid is beneficial, but only in a subset of pneumonia patients (the sickest patients, with the most systemic inflammation).
The benefit of steroid exists, but it's not huge.
Benefit probably depends on early treatment.
Demonstration of mortality benefit for any intervention in critical care is incredibly hard (discussed further here and here).

These factors create the following situation:

Impressive benefit from steroid in pneumonia may be detected in about a third of RCTs. Other studies will be neutral. For example, if the CAPE-COD trial were replicated in the United States, I fully expect that it would be a neutral trial in terms of mortality (patients in the United States tend to be less sick).
Meta-analyses will flip and flop all over the place, depending on inclusion criteria and how they are designed. Many meta-analyses are profoundly fragile (more on this here).

Overall, the CAPE COD trial won't affect my management very much. My take on the steroid data in pneumonia was explored in 2015 here. This hasn't fundamentally changed much since then. In essence: steroid is likely beneficial for sick ICU patients with pneumonia who don't have contraindications.

There is one golden nugget buried within the CAPE COD trial which is worth discussing. The benefit of steroid was stratified, based on the admission C-reactive protein (CRP) level:

Steroid was most beneficial among patients with a CRP level >15 mg/dL (i.e., >150 mg/L). This makes sense – steroid would be expected to be more beneficial in sicker patients with more systemic inflammation. Prior studies have likewise used CRP as a stratification tool to determine which patients benefit from steroid the most.(25688779) CRP is a cheap test with a rapid turn-around time. It situations where the benefit of steroid is clinically unclear, measuring a CRP would be a reasonable consideration.

This leads to a greater question of how we define pneumonia. In reality, our clinical definition of pneumonia is broad, vague, and prone to failure. Many patients who we diagnose with “pneumonia” probably have aspiration pneumonitis, atelectasis, heart failure, or a small pleural effusion. (In the desperate search for a source of sepsis, that nonspecific patchy abnormality on the chest X-ray can be awfully seductive.) Part of the magic of the CAPE-COD trial was that they were relatively judicious about excluding patients without true pneumonia (see highlights above). This raises the question of whether incorporating a biomarker into the definition of pneumonia (e.g., CRP or procalcitonin) could lead to a more homogeneous and meaningful clinical construct. So trials could stop flipping and flopping around like a dying fish. 🐟

Steroid is beneficial for patients with severe community-acquired pneumonia in the ICU who lack contraindications to steroid administration (benefits > risks). Please note that this study excluded influenza patients (that's a whole different debate for a different day).
Whether steroid actually causes a mortality benefit will remain debated until the heat death of the universe.
Elevated C-Reactive Protein may be used to determine which patients may benefit the most from steroid.

Full trial results at NEJM here.

Opening image credit: Photo by John Werner on Unsplash