Rethinking treatment approaches to organising pneumonia: more lessons to be learned from COVID-19

Extract

Organising pneumonia (OP) is an interstitial lung disease characterised by an abnormal repair response to lung injury [1]. Many forms of lung injury can lead to OP, including infections, radiation and drug exposures. OP is also associated with autoimmune conditions such as rheumatoid arthritis and polymyositis/dermatomyositis. In 15–65% of cases, no cause is identified and the phenomenon is termed cryptogenic organising pneumonia (COP) [2, 3]. Both COP and secondary forms of OP typically present with dyspnoea and cough, and are accompanied by radiological abnormalities consisting of ground-glass opacities and consolidation often with a peripheral predominance, although the spectrum of findings can be varied.

Shareable abstract

In this issue of ERJ Open Research, Espejo-Castellanos et al. present a noninferiority trial comparing two dosage regimens of prednisone for post-COVID organising pneumonia (OP) providing important insight into treatment duration of secondary OP https://bit.ly/422MWzj

Organising pneumonia (OP) is an interstitial lung disease characterised by an abnormal repair response to lung injury [1]. Many forms of lung injury can lead to OP, including infections, radiation and drug exposures. OP is also associated with autoimmune conditions such as rheumatoid arthritis and polymyositis/dermatomyositis. In 15–65% of cases, no cause is identified and the phenomenon is termed cryptogenic organising pneumonia (COP) [2, 3]. Both COP and secondary forms of OP typically present with dyspnoea and cough, and are accompanied by radiological abnormalities consisting of ground-glass opacities and consolidation often with a peripheral predominance, although the spectrum of findings can be varied. Due to overlapping clinical presentations with a myriad of other conditions, the diagnosis is usually one of exclusion and in some cases, biopsy may be necessary to exonerate alternative diagnoses [4].

Corticosteroids are considered the mainstay of therapy for all forms of OP; for COP, treatment courses of moderate dose prednisone (0.5–1.0 mg·kg⁻¹) with tapering over 6–12 months have been recommended [5]. These recommendations have never been studied in a randomised controlled trial, however, and are driven by observational studies and concern for relapse as steroids are weaned [6–8]. This approach of prolonged steroid treatment is not without risk due to side effects of insomnia, hyperglycaemia, weight gain, declining bone density and opportunistic infections. Secondary OP, that for which a trigger can be identified, is often treated similarly, and it is not known whether long courses of steroids are truly required in cases where the underlying insult has resolved (infection) or been withdrawn (medication).

Post-COVID-19 OP is increasingly recognised as a common complication of what is now a common disease. It has frequently been described in case series; one such series reported a prevalence of 12.5% among intensive care unit patients admitted with COVID-19 pneumonia [9]. Data supporting the use of corticosteroids for post-COVID-19 OP have come from observational studies with dosage and treatment courses varying among case series [10, 11]. Due to its ubiquity, post-COVID-19 OP provides a unique opportunity to test the efficacy of differing corticosteroid courses in a randomised fashion.

In this issue of ERJ Open Research, Espejo-Castellanos et al. [12] present a randomised, open-label noninferiority trial comparing two dosage regimens of prednisone for post-COVID OP and provide important insight into treatment duration of secondary OP. The trial enrolled 83 participants with post-COVID-19 OP from a post-COVID-19 recovery clinic in Spain. The study was terminated due to low recruitment before the planned 120 participants were enrolled. Participants were over the age of 18 years and had been hospitalised with COVID-19 pneumonia; 43% had required mechanical ventilation or extracorporeal membrane oxygenation while hospitalised. 55 (69%) had their OP diagnosis confirmed by biopsy; the remainder were diagnosed based on clinical and radiological evidence after discussion in a multidisciplinary conference. Mean age was 61 years and ∼70% were male.

41 participants were randomised to receive a 3-month course of prednisone with an initial starting dose of 0.5 mg·kg⁻¹, while 41 were randomised to the control group to receive a 6-month course of prednisone with 0.75 mg·kg⁻¹ as a starting dose. Both groups received prophylaxis with trimethoprim–sulfamethoxazole, omeprazole, calcium and alendronate. Adherence rates were good, with >90% completing their assigned regimens. The primary outcome was noninferiority of the experimental (3-month) group, with a margin set at 10% difference in diffusing capacity of the lung for carbon monoxide (D_LCO) at 6 months. Baseline D_LCO was 59.3% predicted in the experimental group and improved by 11.7%, while baseline D_LCO in the control group was 55.5% predicted and improved by 12.8% (absolute difference 1.16%, 95% CI −5.1–7.4%). Thus, the inferiority margin was not met, and the experimental (3-month) regimen was deemed noninferior to the control (6-month) regimen. The absolute difference in improvement widened at 12 months to 6.75% but did not reach statistical significance. Three participants in the control group and two in the experimental group had relapses after completing the treatment regimen. Improvements in high-resolution computed tomography findings and 6-min walk test distance were similar between groups; however, the control group had a larger improvement in forced expiratory volume in 1 s (although not forced vital capacity (FVC)) than the experimental group at 6 months (absolute increase 9.5% versus 2.8%, p=0.03). Although overall rates of adverse events such as adrenal insufficiency, hyperglycaemia and insomnia were low, importantly, these events were reduced by >50% in the 3-month regimen, occurring in only 23% of the experimental group compared to 56% of the control group.

This study is the first randomised controlled trial evaluating corticosteroid duration in OP and as such, sets an important standard for the field. While FVC is often the outcome of choice in other studies of interstitial lung diseases, D_LCO can be more sensitive for mild lung disease [13]. Radiological changes are often used clinically and experimentally to assess treatment response; these were appropriately included as a secondary endpoint, but they did not differ between groups. Although OP (particularly secondary OP) is often diagnosed clinically, one strength of this trial was that most of the participants had histological confirmation of their diagnosis. Furthermore, adherence rates were very high. One challenge in the interpretation of this study is that spontaneous remission rates in mild OP can be as high as 40% [14]. All participants had severe COVID-19 pneumonia during their hospitalisations; however, upon enrolment in this study, while 96% of subjects reported dyspnoea, none reported ongoing fevers, fewer than one in six had a cough and <10% had modified Medical Research Council dyspnoea scores ≥3. In the absence of an established post-hospitalisation COVID clinic, it is unclear how many of these participants would have sought medical attention in other circumstances. The addition of an untreated control group would have made it easier to assess expected outcomes in each group regardless of treatment assignment, though equipoise may not exist for this depending on OP disease severity. Data on whether participants had received corticosteroids while hospitalised with COVID-19, which is likely given guideline recommendations for severe COVID-19, would also be useful for understanding outcomes, as they may be considered “partially treated” prior to enrolment in the trial [15]. Thus, while the selection of treatment regimens for this trial mimic widely observed clinical practice, it does not resolve the question of whether steroids are indicated in mild forms of post-infectious OP.

Secondary OP is an increasingly common phenomenon, particularly in oncology, where immunotherapies and targeted therapies such tyrosine kinase inhibitors can lead to pneumonitis and OP [16]. As use of these drugs continues to increase, we can expect the incidence of OP to increase as well. This trial provides important guidance on duration of treatment of post-infectious OP that may have important bearing on the management of other forms of secondary OP in which the primary insult is expected to be time limited (infection, drug exposure, radiation, etc.). How well these results can be extrapolated to either forms of secondary OP where the insult is ongoing, such as in OP due to systemic autoimmune disease, or those where the insult is unknown, such as COP, remain important unanswered questions. This study shines light on the need for more clinical trials to define the gold standard of therapy for different forms of OP and to determine if even shorter courses of corticosteroid treatment for OP are equally efficacious. Recent trials in the field of infectious disease have found noninferiority of shorter courses of antibiotics; perhaps it is time for the field of pulmonology to reconsider similar dogmas [17]. This trial is an excellent first step.