Comparison of serial qualitative and quantitative assessments of caval index and left ventricular systolic function during early fluid resuscitation of hypotensive emergency department patients

Objectives: The objective was to determine whether serial bedside visual estimates of left ventricular systolic function (LVF) and respiratory variation of the inferior vena cava (IVC) diameter would agree with quantitative measurements of LVF and caval index in hypotensive emergency department (ED) patients during fluid challenges. The authors hypothesized that there would be moderate inter-rater agreement on the visual estimates.

Methods: This prospective observational study was performed at an urban, regional ED. Patients were eligible for enrollment if they were hypotensive in the ED as defined by a systolic blood pressure (sBP) of <100 mm Hg or mean arterial pressure of ≤65 mm Hg, exhibited signs or symptoms of shock, and the treating physician intended to administer intravenous (IV) fluid boluses for resuscitation. Sonologists performed a sequence of echocardiographic assessments at the beginning, during, and toward the end of fluid challenge. Both caval index and LVF were determined by the sonologist in qualitative then quantitative manners. Deidentified digital video clips of two-dimensional IVC and LVF assessments were later presented, in random order, to an ultrasound (US) fellowship-trained emergency physician using a standardized rating system for review. Statistical analysis included both descriptive statistics and correlation analysis.

Results: Twenty-four patients were enrolled and yielded 72 caval index and LVF videos that were scored at the bedside prior to any measurements and then reviewed later. Visual estimates of caval index compared to measured caval index yielded a correlation of 0.81 (p < 0.0001). Visual estimates of LVF compared to fractional shortening yielded a correlation of 0.84 (p < 0.0001). Inter-rater agreement of respiratory variation of IVC diameter and LVF scores had simple kappa values of 0.70 (95% confidence interval [CI] = 0.56 to 0.85) and 0.46 (95% CI = 0.29 to 0.63), respectively. Significant differences in mean values between time 0 and time 2 were found for caval index measurements, the visual scores of IVC diameter variation, and both maximum and minimum IVC diameters.

Conclusions: This study showed that serial visual estimations of the respiratory variation of IVC diameter and LVF agreed with bedside measurements of caval index and LVF during early fluid challenges to symptomatic hypotensive ED patients. There was moderate inter-rater agreement in both visual estimates. In addition, acute volume loading was associated with detectable acute changes in IVC measurements.