Serially measured diaphragmatic thickening fraction, but not diaphragmatic excursion, was associated with subsequent intubation during high-flow nasal cannula therapy: a prospective repeated-measures cohort study.

OBJECTIVE: To evaluate whether serial diaphragmatic excursion (DE) and diaphragmatic thickening fraction (DTF) are associated with subsequent tracheal intubation during high-flow nasal cannula (HFNC) therapy and whether they add information beyond the ROX index. DESIGN: Prospective, single-center, repeated-measures cohort study. SETTING: Intensive care unit. PATIENTS: Consecutive adults with acute hypoxemic respiratory failure receiving HFNC. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Right-sided diaphragm ultrasound was performed at HFNC initiation and 1, 3, 6, 9, and 12 h. The primary outcome was intubation within 48 h of HFNC initiation. At T1-T5 landmarks, current DE, DTF, and ROX were used to estimate remaining intubation risk through 48 h using pooled logistic regression with patient-clustered robust standard errors. Next-window prediction was a sensitivity analysis. Among 270 patients, 64 (23.7%) were intubated; 1491 ultrasound assessments yielded 1221 person-period records. After adjustment for age and APACHE II, each 1-SD increase in DTF (35.7 percentage points) was associated with higher intubation odds (OR, 1.50; 95% CI, 1.29-1.74; P < 0.001), whereas DE was not. ROX remained the stronger bedside marker. Adding DTF to ROX improved apparent AUC from 0.707 to 0.736 (delta AUC, 0.029; cluster-bootstrap 95% CI, 0.009-0.055); adding DE provided little improvement. Next-window results were directionally consistent. CONCLUSIONS: Serially measured DTF, but not DE, was associated with subsequent intubation risk. DTF may complement, but should not replace, ROX and clinical assessment.