Different definitions of ultrasonographic diaphragmatic dysfunction can improve its correlation with in-hospital death

<bold>Background:</bold> Ultrasound(US)diaphragmatic dysfunction(DD)is defined as an excursion(Ex)<10 mm and/or a thickening fraction (TF)<20%. However, these cut-offs do not correlate with clinical outcomes as were validated in invasively ventilated(IV)patients. <bold>Aim:</bold> to test different thresholds of Ex and TF in non-IV patients with acute respiratory failure (ARF) for the DD diagnosis and their correlation with in-hospital death. <bold>Methods:</bold> we prospectively enrolled 43 in non-IV patients with ARF of any severity and diagnosis. All patients underwent diaphragmatic US upon admission (T0, n= 43) and discharge (TD, n=33). Each assessment included: diaphragmatic thickness, TF, and Ex during maximal volitional inspiratory maneuver. The prevalence of DD was considered for a TF<30%, 20% and 10% and an Ex of 30, 20 and 10 mm, and their combinations. <bold>Results:</bold> modifying the threshold of Ex from ≤10 to ≤30 mm and of TF from ≤10 to ≤30% a higher prevalence was observed- from 5% to 46% and from 5% to 30% respectively. During hospitalization 7 patients died(16%). An Ex≤20 mm had a sensitivity of 77% and a specificity of 71% in predicting in-hospital death. Combining different Ex and TF values increased sensitivity to 92-100%, reducing specificity to 14%, but increased the negative and positive predictive value up to 100% and >86%, respectively, when extreme levels of Ex impairment were considered, regardless of the TF. <bold>Conclusion:</bold> in non-IV patients, the prevalence of DD varies significantly when different Ex and TF are considered. Combining different Ex and TF thresholds may increase the accuracy of DD in predicting in-hospital mortality. A larger sample size is required for further investigations.