Corticosteroids for Community-Acquired Pneumonia
A 66 year old man presents to the emergency department with sudden-onset fevers, chills, and scant hemoptysis. He is hypoxemic, tachypneic and noted to have egophony with focal crackles and wheezing across his right, anterior chest. Further, he is found to be in acute kidney injury and his chest x-ray reveals a bulging fissure sign with dense right middle lobe infiltrate. The emergency department resident contacts the patient’s pulmonologist who confirms that 6 weeks ago the patient had a chest CT for malignancy screening which demonstrated only bi-apical emphysema. Ceftriaxone, azithromycin and methylprednisolone are administered.
Earlier this year, Wan et al. published a systematic review and meta-analysis on the efficacy and safety of corticosteroids for community-acquired pneumonia [CAP]. Reports of corticosteroids for the treatment of pneumonia date back at least 6 decades; as with a plethora of other pathologies, the pendulum for corticosteroids in community-acquired pneumonia continues to swing.
The mechanism of action of glucocorticoids is manifold. There are immediate and delayed effects on leukocyte trafficking, as well as upon the innate and adaptive immune responses. Of clinical relevance, perhaps the most immediate effect of glucocorticoids is to prevent the transudation of neutrophils and monocytes to the site of inflammation; this is partially responsible for the nearly universal rise in peripheral WBC count in response to corticosteroids.
As varied as the biological workings of glucocorticoids is their reported clinical benefit in community acquired pneumonia. A number of well-designed randomized-controlled trials [RCTs] have found differing benefits and degrees thereof. Further, a previous meta-analysis reported a decrement in mortality when patients with severe CAP received corticosteroids.
Overview of methods and results
Wan and colleagues located 9 RCTs comprising nearly 1700 patients and an additional 6 cohort studies with nearly 4100 patients. 4 of the 9 RCTs included patients with severe CAP. The mean duration of steroid duration was 7 days at an average dose of 30 mg/day of methylprednisolone. [Recall that relative to hydrocortisone, prednisone and prednisolone have 4 times, methylprednisolone 5 times and dexamethasone 30 times anti-inflammatory effect]. Importantly, the dose and type of corticosteroid used between trials differed considerably.
The primary outcome of their meta-analysis was mortality in patients with severe CAP and in the trials which included such patients, there was no statistically significant difference in mortality. The authors noted that the sample size required to reliably detect a difference in mortality in severe CAP would be over 2500 patients! This lack of effect on mortality held even when the authors considered trials which assayed inflammatory mediators [e.g. C-reactive protein] and in patients who were treated with long versus short-duration corticosteroids.
With regards to their secondary outcomes, corticosteroids did show a consistent reduction in length of hospital and ICU stay, IV antibiotic duration and time to clinical stability. Also of significance, glucocorticoids were not associated with hyperglycemia, GI bleeding, superinfection or pleural space infection.
The authors made mention that their results remained stable when considering only multicenter trials, low risk of bias trials and those with large sample sizes. Moreover, there was no evidence of publication bias.
The benefit of corticosteroids in CAP – including severe CAP – are reminiscent of their known benefits in COPD exacerbations [AECOPD]. The landmark VA-Cooperative trial in 1999 demonstrated that, compared to placebo, corticosteroids got patients with AECOPD out of the hospital earlier by about 1 day. Additionally, there was a small improvement in FEV1 and patients receiving steroids were less likely to have ‘treatment failure’ [interestingly, mostly driven by the need for open-label steroids].
As Wan et al. note, a fairly large trial would be required to confidently detect a mortality benefit for corticosteroids in CAP. Until then [if ever], we will likely squabble and make use of post-hoc, subgroup analyses to justify or condemn their use. What is, perhaps, most important is that a short course of steroids is unlikely to cause clinical deterioration in patients with community-acquired pneumonia. Also of note is the need for keen recognition of inclusion and exclusion criteria. Many studies utilized by Wan et al. in their meta-analysis specifically excluded patients with nosocomial pneumonia, immunosuppression, uncontrolled hyperglycemia, pregnancy, adrenal insufficiency, recent gastrointestinal hemorrhage, etc. One even excluded patients en route to the ICU.
Regardless, the calls to add corticosteroids to the general management of community-acquired pneumonia have begun; I do not foresee them getting softer, especially with exquisite emphasis placed on length of stay and cost-containment.
Have you changed your practice?
*For a 2019 update on community acquired pneumonia, please see this post.