Should Steroids Be Used to Treat ARDS?

Acute respiratory distress syndrome (ARDS) has been associated with considerable morbidity and mortality, even before the COVID-19 pandemic hit. For the most part, ARDS management is limited to supportive care and prevention of complications. This means low tidal volumes, moderate to high positive end-expiratory pressure, and prone positioning when necessary. It may also mean corticosteroids (CS). I think. Maybe.

Twenty years ago, when I was a medicine resident interested in critical care, we used the "Meduri protocol" to treat ARDS. This was based on a small randomized controlled trial (RCT) published in JAMA by Umberto Meduri and colleagues in 1998. Patients enrolled had ARDS for at least 7 days without showing signs of improvement. Patients in the treatment arm received large amounts of CS (2 mg/kg/d initially) with a 32-day taper. The authors found a reduction in ICU and hospital-related mortality; the Meduri protocol was endorsed. So for a period of time, I gave CS to patients with ARDS who met the inclusion criteria from this study.

In 2006, a larger RCT, published in The New England Journal of Medicine by the National Heart, Lung, and Blood Institute ARDS Clinical Trials Network (ARDSNet), cast doubt on the Meduri protocol. The researchers found no differences in mortality rates in patients treated with CS vs those who received a placebo. There were differences in other outcomes (ventilator, ICU, and shock days within the first 4 weeks) that favored CS. There was also signal for an increase in reintubation due to neuromuscular weakness in the CS group. But what killed the Meduri protocol was an increase in mortality for patients started on CS more than 14 days after ARDS onset.

Given that the ARDSNet study questioned the utility of the protocol bearing his name, Meduri fought back. His smaller, unbalanced RCT (the percentage of patients with catecholamine-dependent shock in the placebo group was double that in the CS arm) was published a year later in the CHEST journal. His study found that CS, given within 48-72 hours of ARDS diagnosis, was associated with a reduction in ventilator days and ICU mortality.

At the time, I concluded that there was enough signal within the ARDSNet RCT to justify steroids for ARDS. Despite its flaws, Meduri's CHEST RCT influenced my choice of timing, as did physiology. ARDS follows a typical course, and the earlier proinflammatory phases should be more responsive to CS. My anecdotal experience, as always affecting my decisions more than it should, seemed to support using early CS.

Since 2007, publications on CS used for ARDS have generally shown improvements in softer outcomes with early low to moderate doses. They've also shown that higher doses increase mortality. Guidelines have tacitly endorsed steroids for ARDS or simply avoided comment. However, two recent articles have thrust CS for ARDS back into the spotlight.

The first was a multicenter RCT from Spain that evaluated dexamethasone as treatment for moderate to severe ARDS. The authors found impressive results: seven needed-to-treat (NNT) patients to prevent one death at 60 days. In the critical care world, this is a home run. It is rare that we find any intervention with that type of impact on mortality. Anytime we have found such an outcome, the results can't be replicated.

The second article was a COVID-19 guideline published by experts recruited as part of the Surviving Sepsis Campaign. The guideline makes a weak recommendation for using CS in COVID-19 patients with ARDS. The authors should be given some leeway here, because the pandemic evolved quickly and guidelines were rushed. Still, the Surviving Sepsis COVID-specific document reignited the CS in ARDS debate: if, when, and at what dose?

Where Does This Leave Us?

So where are we now? Personally, I was already comfortable using early CS for ARDS. Therefore, the Spanish RCT won't change my practice. It simply reinforces my current beliefs. It's the first RCT conducted in the era of low-tidal ventilation to show a reduction in mortality with CS for ARDS, and none of the authors were named Umberto Meduri. These two facts make it notable. An accompanying editorial points out a number of flaws to the study, and the NNT of seven almost sounds too good to be true, so I do take it with a grain of salt.

CS for ARDS related to COVID-19 is a little trickier because we know less about the disease. Investigators from the RECOVERY trial, a large, randomized study of treatments for hospitalized COVID patients, just reported a mortality benefit using dexamethasone. Unfortunately, we only have a preprint so we'll have to wait for the manuscript to officially pass judgment. Whether or not their data hold up against peer-reviewed judgment, I'm still in favor of using CS for COVID-19–related ARDS. Just like for non-COVID ARDS, I'd do it early (< 72 hours) and with a low dose (≤ 1 mg/kg/d of methylprednisolone equivalent). Let's hope that the RECOVERY dexamethasone results are published soon.

Aaron B. Holley, MD, is an associate professor of medicine at Uniformed Services University and program director of pulmonary and critical care medicine at Walter Reed National Military Medical Center. He covers a wide range of topics in pulmonary, critical care, and sleep medicine.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube