C37-01 Lung Function Phenotypes and 24-month Healthcare Utilization and Outcomes in Long COVID

Abstract Rationale Pulmonary function Test (PFT) abnormalities are frequent after SARS-CoV-2 infection, yet their prognostic value for long-term utilization of health resources and severe outcomes is unclear. We evaluated whether baseline PFT phenotypes (preserved ratio impaired spirometry (PRISm), obstructive, or normal) predict 24-month healthcare utilization and clinical outcomes among adults with long COVID. Methods We conducted a retrospective cohort study across the Mount Sinai Health System, including adults who met long COVID diagnostic criteria and completed baseline PFTs at diagnosis. Phenotypes were prespecified: obstructive (FEV1/FVC 0.70), PRISm (FEV1 80% predicted with FEV1/FVC ≥0.70), and normal. Primary outcomes were urgent care (UC) visit, emergency department (ED) visit and hospitalization counts by 24 months. Secondary outcomes were ICU admission and initiation of supplemental oxygen by 24 months. Negative binomial regression models were used to estimate incidence rate ratios (IRRs) for UC, ED and hospitalization counts by 24 months. Logistic regression models were used to estimate odds ratios (ORs) for ICU admission and initiation of supplemental oxygen by 24 months. Regression models adjusted for age, sex, race, ethnicity, obesity, smoking status, Charlson Comorbidity Index (CCI), COVID vaccination status, baseline dyspnea (mMRC scale) and insurance status. P-values were adjusted for multiple comparisons using the Benjamini-Hochberg procedure. Results Among 241 patients (median age 54 years; 61% female; 32.4% ever-smokers), phenotypes were normal, PRISm and obstructive in 161 (66.8%), 44 (18.3%) and 36 (14.9%) cases respectively. Median follow-up was 24 months. Compared with normal phenotype, PRISm phenotype was associated with higher counts of hospitalization at 24 months (IRR: 5.64; p=0.029) but unchanged counts of 24-month UC visits (IRR: 0.864; p=0.711) or 24-month ED visits (IRR: 1.090; p=0.837). The obstructive phenotype was not associated with 24-month counts of UC visits (IRR: 0.920; p=0.862), ED visits (IRR: 1.090; p=0.858) or hospitalization (IRR: 0.451; p=0.999). Additionally, dyspnea at baseline as measured by the mMRC scale was associated with higher counts of 24-month UC visits (IRR: 1.480; p=0.022) and 24-month ED visits (IRR: 1.436; p=0.020). Obstructive or PRISm phenotypes were not associated with odds of 24-month hospitalization, 24-month ICU admission or 24-month initiation of supplemental oxygen therapy. Conclusions In adults with long COVID, baseline PRISm and baseline dyspnea (as measured by mMRC scale) independently predicted greater healthcare utilization and clinically important 24-month outcomes. Simple PFT-based phenotyping may aid post-COVID risk stratification and guide follow-up intensity. This abstract is funded by: None