Estimation of diaphragm length in patients with severe chronic obstructive pulmonary disease

Abstract

In patients with advanced chronic obstructive pulmonary disease (COPD) diaphragm function may be compromised because of reduced muscle fibre length. Diaphragm length (L_Di) can be estimated from measurements of transverse diameter of the rib cage (D_Rc) and the length of the zone of apposition (L_Zapp) in healthy subjects, but this method has not been validated in patients with COPD. Postero-anterior chest radiographs were obtained at total lung capacity (TLC), functional residual capacity (FRC) and residual volume (RV) in nine male patients with severe COPD (mean [S.D.]; FEV₁, 23 [6] %pred.; FRC, 199 [15] %pred.). Radiographs taken at TLC were used to identify the lateral costal insertions of the diaphragm (L_Zapp assumed to approach zero at TLC). L_Di was measured directly and also estimated from measurements of L_Zapp and D_Rc using a prediction equation derived from healthy subjects. The estimation of L_Di was highly accurate with an intraclass correlation coefficient of 0.93 and 95% CI of ∼±8% of the true value. L_Di decreased from 426 (64) mm at RV to 305 (31) mm at TLC. As there were only small and variable changes in D_Rc across the lung volume range, most of the L_Di changes occurred in the zone of apposition. Additional studies showed that measurements of L_Di from PA and lateral radiographs performed at different lung volumes were tightly correlated. These results suggest that non-invasive measurements of L_Zapp in the coronal plane (e.g. using ultrasonography) and D_Rc (e.g. using magnetometers) can be used to provide an accurate estimate of L_Di in COPD patients.

Introduction

The effectiveness of the diaphragm as an inspiratory muscle in the face of the large increase in functional residual capacity (FRC) that occurs in advanced chronic obstructive pulmonary disease (COPD) is controversial (De Troyer, 1997). While many studies of transdiaphragmatic pressure (P_Di) during tidal breathing have been made in COPD (Sharp et al., 1977, Druz and Sharp, 1982, Levine et al., 1988, Martinez et al., 1990) so far it has not been possible to estimate the accompanying dynamic changes in diaphragm length (L_Di). The nearest approach has been to make static measurements of L_Di at different lung volumes. Most measurements have been made using plain chest radiographs (Braun et al., 1982, Loring et al., 1985, Rochester and Braun, 1985, Sharp et al., 1986, Petroll et al., 1990). Scans using either computerised tomography (CT; Whitelaw, 1987, Cassart et al., 1997, Pettiaux et al., 1997) or magnetic resonance imaging (MRI; Gauthier et al., 1994) allow three-dimensional reconstruction of the whole diaphragm; unfortunately the time resolution of both techniques is poor (<1 Hz) and studies can only be made with the subject lying. Cine-radiography has better time resolution (around six frames/sec) and has been used to a limited extent (Wade, 1954, Gandevia et al., 1992) but entails significant radiation exposure.

To measure L_Di (muscle and central tendon) the lengths of the zones of apposition of the diaphragm against the chest wall on the right and left sides (L_Zapp,r and L_Zapp,l) respectively, and the cross-sectional lengths of the two domes are summed. On radiographs the costal insertions of the diaphragm in the coronal plane are assumed to attach to the rib at or just below the caudal limit of the costo-phrenic angle at total lung capacity (TLC). With CT and MRI the origin has been assumed from anatomical land marks or external markers have been applied to the skin. All three imaging techniques can be used to measure L_Di at TLC and at smaller lung volumes.

Loring et al. (1985) suggested that the length of the right and left domes in the mid-axillary coronal plane could be estimated from the internal diameter of the rib cage (D_Rc) in that plane plus an additional length which reflected the curvature of the domes. They obtained postero-anterior radiographs in three normal subjects at three different lung volumes. The effect of dome curvature could be expressed by multiplying D_Rc by a constant which was essentially similar in each subject and at each lung volume. Thus they were able to estimate changes in L_Di in normal subjects by measuring external diameter of the rib cage with a pair of magnetometers and using ultrasonography over the lateral rib cage to measure L_Zapp,r (L_Zapp,r equalled L_Zapp,l on the radiographs). Ultrasonography was also used to measure the thickness of the ribcage, which was subtracted from the diameter measured by the magnetometers to obtain D_Rc. The temporal resolution with ultrasonography is potentially higher (∼30 Hz) than that with other methods of imaging.

Petroll et al. (1990) extended these studies by measuring L_Di from postero-anterior radiographs taken at five different lung volumes between TLC and residual volume (RV) in 16 normal subjects. They found small differences in dome curvature at different lung volumes but showed that L_Di could be estimated accurately by the equation:

$$\text{L}{}_{\mathrm{<mtext>Di</mtext>}}\text{(}\text{mm}\text{)=2(0.98}\text{L}{}_{\mathrm{<mtext>Zapp</mtext>}}\text{,}\text{r}\text{+0.46}\text{D}{}_{\mathrm{<mtext>Rc</mtext>}}\text{+35),}\text{R}{}^2\text{=0.94}$$

If this equation could be applied to patients with COPD it would permit the use of surface measurements to measure dynamic L_Di during tidal breathing. However, the applicability of this equation to COPD cannot be assumed; obvious differences are that in COPD changes in D_Rc with volume may differ from control subjects (due to alterations in rib-cage shape and mechanics) and the diaphragm domes are often flatter and may exhibit more change in curvature with lung volume. In patients with COPD there are significant distortions in chest shape over the vital capacity range (Sharp et al., 1977) so that the relative changes in L_Di between the coronal and sagittal planes may be different from those in control subjects. Moreover, in contrast to the relative consistency of dome shape with volume in the coronal plane the diaphragm usually flattens completely in the sagittal plane at TLC (Gauthier et al., 1994).

In the present study we report measurements of L_Di made from postero-anterior radiographs at TLC, FRC and RV in nine men with severe COPD. Similar measurements were made from lateral chest radiographs taken at the same lung volumes in three patients to correlate measurements of L_Di in the coronal and sagittal planes. We had two distinct objectives: (i) to compare our measurements with the limited available literature using plain radiography or CT and (ii) to determine whether the equation of Petroll et al. (1990) derived in healthy subjects could predict static diaphragm length at different lung volumes in patients with severe COPD.