Predictors of Severe Respiratory Failure in Hospitalized Patients with SARS-CoV-2 Infection: Development and Validation of a Prediction Model (PREDI-CO Study)

Background: Severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2)-associated coronavirus disease-2019 (COVID-19) is spreading worldwide challenging healthcare resources. To provide the best possible care for patients, efficient prognosis is needed. Methods: We performed a multicentre cohort study to develop and validate a risk score for severe respiratory failure (SRF) among hospitalized patients diagnosed with COVID-19. Study population included all adult patients with RT-PCR on nasopharyngeal swab positive for SARS-CoV-2, hospitalized from February 22 to April 3 2020, at 11 hospitals from four Italian Regions. Exclusion criteria were early ( 30bpm, or respiratory distress. Findings: Of the 1265 included patients, 1044 patients were analyzed, (581 derivation, 463 validation cohort). Mean age was 65.7±15 years, 661 (63·3%) were male. SRF rates were 29% and 38% in derivation and validation cohort, respectively. At multivariate analysis, risk factors for SRF in the derivation cohort were age ≥70 years, obesity, body temperature ≥38°C, RR ≥22bpm, lymphocytes ≤900/mm3, creatinine ≥ 1 mg/dl, C-reactive protein ≥10mg/dl, and lactate dehydrogenase ≥350IU/L at hospitalization. Area under receiver-operator curve (AUROC) was 0·90 (95%CI 0.86-0.93). Assigning points to each variable an individual risk score (PREDICO score) was obtained. At risk score of >3, sensitivity, specificity, positive and negative predictive values were 71·6% (64-78), 89·1% (85-92), 72·9% (65-79), and 88·4% (85-91), respectively; positive and negative likelihood ratios were 6·56 (95% CI 4·9-8·77) and 0·31 (95%CI 0·25-0·40), respectively. PREDICO score showed similar prognostic ability in the validation cohort: AUROC 0·80 (0·74-0·88). Interpretation: Using eight variables, easy to obtain at hospitalization, patients with COVID-19 could be classified according to their probability of developing SRF during hospitalization. PREDICO score can be useful to allocate resources and prioritize treatments during COVID-19 pandemic. Trial Registration: NCT04316949. Funding Statement: None. Declaration of Interests: Authors state no conflict of interest related to the content of the present study. Ethics Approval Statement: The study was approved by the Ethic Committee of the promoting center (Comitato Etico Indipendente di Area Vasta Emilia Centro, n. 283/2020/Oss/AOUBo).