Objective: Severe community-acquired pneumonia (sCAP) is one of the major diseases within the ICU. We hypothesize that subtyping sCAP based on simple inflammatory markers, organ dysfunction, and clinical metagenomics results is feasible.Method: In this study, we retrospectively enrolled immunocompetent sCAP patients requiring invasive mechanical ventilation, who underwent clinical metagenomics from 17 medical centers. We collected data on potentially pathogenic species reported by clinical metagenomics and clinical information for all patients. Latent class analysis (LCA) was applied to routine clinical parameters such as gender, age, white blood cell (WBC), lymphocytes, C-reactive protein (CRP), and Procalcitonin (PCT), identifying two optimally fitting models.Results: A total of 569 patients were enrolled. Compared to class B, class A was characterized by a younger age, higher CRP and PCT levels, and a higher incidence of coagulation dysfunction, liver failure, circulatory failure, and renal failure. However, the mortality rates were similar between the two groups. In class A, more cases of Streptococcus spp. and fewer cases of HSV-1 and Candida spp. were detected. Among the patients in the two phenotypes, 48.7% and 57.5% received corticosteroid treatment, respectively. In the class A, corticosteroid treatment was not associated with patient mortality (unadjusted hazard ratio (HR)=0.988; 95% confidence interval (CI), 0.634–1.541; p=0.959). In contrast, in the class B group, the use of corticosteroids was associated with a reduced mortality rate (adjusted HR=0.719; 95% CI, 0.525–0.986; p=0.04). Additional analysis showed that in class B, methylprednisolone was associated with reduced mortality (adjusted HR=0.61; 95% CI, 0.44–0.86; p=0.005), while dexamethasone was not associated with mortality (adjusted HR=1.4; 95% CI, 0.89–2.22; p=0.148). In addition, after dose conversion, the results showed that higher doses of corticosteroids in class B were associated with increased mortality (adjusted HR=1.01; 95% CI, 1.00–1.01; p=0.005).Conclusion: We identified two classes based on clinical metagenomics and clinical features. Class B exhibited a better response to corticosteroid compared to class A. The rapid identification of these phenotypes could facilitate the screening of sCAP patients responsive to corticosteroid in future prospective clinical trials.