Abstract Rationale Long COVID is a prevalent and often disabling condition. Respiratory symptoms affect over 50% of patients, yet the pathobiology, diagnostic evaluation, and treatment strategies for Respiratory Long COVID (RLC) remain undefined. Here, we sought to delineate RLC phenotypes to enable mechanism-based diagnosis and therapy. Methods We conducted a retrospective analysis of two cohorts of patients evaluated at the Yale Center for Lung Diseases between 11/1/2020 and 11/1/2025 for RLC, defined as 12 weeks of respiratory symptoms following SARS-CoV-2 infection. Data were abstracted via standardized manual chart review. Cohort 1 included 102 individuals who underwent comprehensive pulmonary evaluation, including chest imaging for parenchymal injury and bronchoprovocation testing (BPT) via mannitol (PD15) and/or methacholine challenge (PD20) per ATS/ERS standards. Cohort 2 comprised 146 patients treated with asthma biologics (49/146 received 1 therapy, yielding 200 treatment instances). The primary outcome was clinical response, defined by the treating provider’s assessment after 12 weeks of therapy. Results In Cohort 1, parenchymal injury was observed in 23/102 patients (22.5%), predominantly among those hospitalized for acute COVID (15/23, 65.2%). Persistent fibrotic post-COVID interstitial lung disease (PC-ILD) was rare (7/102, 6.9%). Patients without parenchymal injury showed a high prevalence of airway hyperreactivity (AHR) on BPT (51/79, 64.6%). Notably, AHR was common in PC-ILD as well (11/23, 47.8%). Patients with BPT+ post-COVID small airway disease (PC-SAD) demonstrated low rates of bronchodilator reversibility (BDR, 4.8%), peripheral eosinophilia 300/microliter (6.6%), IgE 150 IU/mL (20.0%), FeNO 25 ppb (22.9%), pre-existing asthma (30.6%), and exacerbations (8.1% with ≥1 systemic corticosteroid-requiring exacerbation/year). Onset of PC-SAD was associated with extrapulmonary symptoms such as fatigue (72.6%) and cognitive dysfunction (48.4%). In Cohort 2, patients with PC-SAD showed limited response to anti-IL5 and anti-IgE biologics (22/59, 37.3%; 3/8, 37.5%), intermediate response to anti-IL4R (44/73, 60.3%), and a superior response to anti-TSLP therapy (56/60, 93%, Fisher’s exact, two-sided, p 0.001 vs each other class). Conclusions In the Yale RLC cohort, fibrotic PC-ILD was uncommon. The predominant phenotype was a T2-low, BDR-negative small airway disease associated with multisystem complaints and infrequent exacerbations. We propose to call this condition PC-SAD. AHR on BPT was a key diagnostic biomarker and anti-TSLP biologics yielded a significant therapeutic response in our cohort. To our knowledge, this is the first report to link a quantitative Long COVID biomarker to a mechanistically targeted therapy with demonstrated clinical benefit. This pathophysiology-informed ‘treatable trait’ framework provides a rational strategy for managing RLC and warrants prospective evaluation. This abstract is funded by: NIH K08 HL159422
Gautam et al. (Fri,) studied this question.