Ultrasound Versus Computed Tomography for Diaphragmatic Thickness and Skeletal Muscle Index during Mechanical Ventilation

Abstract

Background: Diaphragmatic alterations occurring during mechanical ventilation (MV) can be monitored using ultrasound (US). The performance of computed tomography (CT) to evaluate diaphragmatic thickness is limited. Further, the association between muscle mass and outcome is increasingly recognized. However, no data are available on its correlation with diaphragmatic thickness. We aimed to determine correlation and agreement of diaphragmatic thickness between CT and US; and its association with muscle mass and MV parameters. Methods: Prospective observational study. US measurements of the diaphragmatic thickness were collected in patients undergoing MV within 12 h before or after performing a CT scan of the thorax and/or upper abdomen. Data on skeletal muscle index (SMI), baseline, and ventilatory data were recorded and correlated with US and CT measures of diaphragmatic thickness. Agreement was explored between US and CT data. Results: Twenty-nine patients were enrolled and the diaphragm measured by CT resulted overall thicker than US-based measurement of the right hemidiaphragm. The US thickness showed the strongest correlation with the left posterior pillar at CT (r = 0.49, p = 0.008). The duration of the controlled MV was negatively correlated with US thickness (r = -0.45, p = 0.017), the thickness of the right anterior pillar (r = -0.41, p = 0.029), and splenic dome by CT (r = -0.43, p = 0.023). SMI was positively correlated with US diaphragmatic thickness (r = 0.50, p = 0.007) and inversely correlated with the duration of MV before enrollment (r = -0.426, p = 0.027). Conclusions: CT scan of the left posterior pillar can estimate diaphragmatic thickness and is moderately correlated with US measurements. Both techniques show that diaphragm thickness decreases with MV duration. The diaphragmatic thickness by US showed a good correlation with SMI.

Keywords: computed tomography; critical care; diaphragm; mechanical ventilation; skeletal muscle index; ultrasound.

Figure 1

Normal appearance of the diaphragm. The computed tomography scan clearly shows the crura in the direct coronal (A) and axial planes (B). The normal appearance of the diaphragm, both in the posterior region (crura) at the liver dome level and the left anterior diaphragm (arrows).

Figure 2

The mean diaphragmatic thickness measured using ultrasound (US right hemidiaphragm) and computed tomography (CT, at the level of different areas of the diaphragm). CT scan diaphragmatic thickness at the level of left anterior, right posterior, left posterior pillars, and splenic dome is significantly thicker than US-based thickness. * p < 0.05 (two-tailed) versus US-based thickness.

Figure 3

Bland–Altman plot exploring the agreement between the right hemidiaphragm thickness evaluated by US and the mean diaphragmatic thickness assessed by CT scan. Data are expressed in mm. UCL = upper confidence limit; LCL = lower confidence limit. Definition of abbreviation: US = ultrasound; CT scan = computerized tomographic scan.

Figure 4

Correlation between days of mechanical ventilation (MV) before imaging and Skeletal Muscle Index (SMI) (A); between SMI and US right diaphragmatic thickness (B); and days of MV before imaging and US right diaphragmatic thickness (C). Definition of abbreviation: SMI = skeletal muscle index; TMA = total muscle area; US = ultrasound.