4, 20, 69
Use to structure referral discussions and to communicate prognosis. Never use as a threshold for offering or withholding ECMO - a low score means likely to die, not unlikely to benefit.
None - no untreated comparator exists in the cohort
DOI 10.1164/rccm.201311-2023OC; PMID 24693864
None - derivation of a prognostic score
Registry data 2000-2012; case mix no longer matches current practice. Derived only in patients who received ECMO, so it cannot estimate benefit. Discrimination is not calibration. External validation cohort was small (n=140).
1,338 of 2,355 patients (57%) discharged alive. RESP score derived from age, immunocompromised status, duration of mechanical ventilation before ECMO, diagnosis, CNS dysfunction, acute non-pulmonary infection, neuromuscular blockade, nitric oxide, bicarbonate infusion, cardiac arrest, PaCO2 and peak inspiratory pressure. Discrimination c = 0.74 (95% CI 0.72-0.76) in derivation; c = 0.92 (95% CI 0.89-0.97) in an external validation cohort of 140 patients.
2,355 adults supported with ECMO for severe acute respiratory failure, ELSO Registry 2000-2012