INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Over the past decade, single-lung transplantation (SLT) has become an established therapeutic option for the treatment of end-stage emphysema (1). SLT increases FEV₁ to ~ 60% of the predicted normal value and substantially improves patients' exercise tolerance and functional status (2, 3). These changes result largely from an improvement in lung mechanics, but an improvement in inspiratory muscle function is also expected to occur as a result of SLT. We have previously shown that the hyperinflation induced by emphysema substantially reduces diaphragm length and surface area; thus, in nine patients with severe emphysema, we found that diaphragm surface area at FRC was decreased to 73% of normal values (4). Because it replaces the hyperinflated emphysematous lung with a graft having a normal FRC (5), SLT should increase diaphragm dimensions and curvature on the transplanted side, which might improve the pressure-generating capacity of the muscle. To investigate this question, we studied nine recipients of single-lung transplants for emphysema and nine matched normal controls, using computed tomography (CT), and we compared the dimensions and the configuration of the diaphragm on the side of the graft with those on the native side and on the ipsilateral side in the controls.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Between 1990 and 1997, SLT for emphysema was performed 14 times in 14 patients at our institution, and 10 patients were alive at the time of the present study. The study was done with nine of these patients and nine nonsmoking normal subjects matched for age, sex, height, and weight. Details of the subjects are given in Table 1. Seven patients had undergone a left SLT and two patients a right SLT. At the time of the studyan average of 573 ± 148 d after surgery (mean ± SE)the patients were 53.1 ± 17.7 yr of age; all were in clinically stable condition, with no evidence of infection or rejection, although one patient had developed a bronchiolitis obliterans syndrome (BOS) stage 1 (6). The patients were receiving triple immunosuppressive therapy with cyclosporine, azathioprine, and methylprednisolone. Their chest radiographs did not reveal any parenchymal abnormality on the side of the graft, but their pulmonary function tests (Table 2) showed an obstructive ventilatory defect with moderate hyperinflation, as usually seen after SLT for emphysema (2, 3). All subjects gave oral informed consent to the study, which was approved by the Human Studies Committee of the Erasme University Hospital.

Technique

Measurements of supine lung volumes and diaphragm dimensions were made with spiral computed tomography (CT) that extended from the lung apex to the lung base, and from 1 cm above the diaphragmatic dome to 1 cm below a metallic wire fixed around the costal margin, respectively. The techniques used for these measurements have been described in detail in our previous studies (4, 5, 7). Measurements were made at TLC, FRC, and FRC plus one-half inspiratory capacity (FRC+). To attain this last volume, the subjects were connected to a spirometer and instructed to inspire to TLC and slowly expire until FRC+ was reached. All measurements were made with the subjects wearing a noseclip and lying supine with their arms at their sides. At each volume the subjects were asked to hold their breath and relax against a closed airway. Before the actual data acquisitions, a few practice trials were always run to familiarize the subjects with the procedures and with the sensation of respiratory muscle relaxation. In addition, the data acquisitions were preceded by a 2- to 3-min period of increased tidal breathing, which made breathholding more comfortable and facilitated relaxation.

Data Analysis

Lung volumes. The individual volumes of the right and left lungs were calculated at each volume studied (5). The Pulmo CT option from Siemens (Somatom Plus S CT scanner; Siemens, AG, Erlangen, Germany) (8) was used to trace the lung contours on each scan and to measure lung area. Lung volume was calculated from values of lung area and slice interval.

Diaphragm surface area. Diaphragm contours were digitized on sagittal and coronal reconstructions and used to obtain a three-dimensional reconstruction of the muscle at each volume studied (Figure 1). We then created small triangular surfaces on the three-dimensional reconstructions and calculated the surface areas of the right and left diaphragm (4, 7). On each side the surface area of the muscle (A_di) was obtained as the sum of the area of the dome, defined as the portion of muscle apposed to the lung (A_do), and of the area of the zone of apposition, defined as the portion of muscle apposed to the rib cage (A_ap).

View larger version (40K): [in this window] [in a new window]   Figure 1.   Three-dimensional reconstruction of diaphragm at FRC in one emphysema patient before (A) and after (B) right lung transplantation. Reconstruction is shown from posterior-superior left lateral perspective. Units = cm.

In the normal subjects we used a midsagittal line to divide the dome of the diaphragm into its right and left portions. In recipients of single-lung transplants for emphysema, however, the mediastinum is displaced toward the transplanted side (5). Hence, using a midsagittal line (line A in Figure 2) would have overestimated A_do on the side of the graft. To overcome this problem, we assumed that the displacement of the dome toward the graft was similar to that of the anterior mediastinum, and we separated A_do into its right and left portions using a line (line B in Figure 2) joining the anterior aspect of the vertebral body and the anterior mediastinal line. The degree of mediastinal shift was determined by measuring the angle between lines A and B. In normal subjects, these two lines were almost superimposed, but in the patients the anterior mediastinal line was displaced toward the graft, with the angle between lines A and B increasing in magnitude with the degree of mediastinal shift.

View larger version (108K): [in this window] [in a new window]   Figure 2.   Representative axial CT slice in a patient with a left transplant. The arrow indicates location of the anterior mediastinal line. Note the shift of the mediastinum toward the graft. Line A is the midsagittal line, and line B was drawn from the vertebral body to the anterior mediastinal line and was used to compute right and left Ado. See text for details.

Diaphragm curvature. For determination of the mean curvature of the dome of the diaphragm, a second-order polynomial (z(x) = ax² + b x + c) was fitted to selected coronal and sagittal slices by means of a least-mean-square fit. The curvature of this one-dimensional model curve is given by

(1)

The mean curvature of each slice was then calculated over the range of points that defined the slice [x_min, x_max] by:

(2)

For each selected slice, the mean radius of curvature was given by  = 1/<\>. In the coronal plane, the radius of curvature of the dome was calculated as the average radius of curvature of four midcoronal slices. In the sagittal plane, three right and three left midsagittal slices were used.

Statistics

Statistical analysis was done with Wilcoxon's matched pairs test, analysis of variance (ANOVA), and analysis of covariance (ANCOVA), with lung volume as a covariant factor when appropriate. Data are expressed as mean ± SE throughout the text, tables, and figures. Values of p < 0.05 were considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Normal Subjects

In all subjects, the right lung had a greater volume than the left lung. The difference averaged 0.2 L at FRC and FRC+, and 0.3 L at TLC (p < 0.005). Figure 3A shows average values for A_di, A_do, and A_ap on the right and left sides. At FRC, A_di, A_do, and A_ap were significantly larger on the right side (p < 0.01); the difference averaged 12%. A similar observation was made for A_di and A_do at FRC+, and for A_do at TLC. Figure 3B shows the same data expressed as a function of the supine volume of each lung. On both sides, A_di and A_ap decreased linearly with lung volume, whereas A_do was unaffected. Data for the left diaphragm were displaced toward smaller values of surface area and volume such that, at a given absolute lung volume, diaphragm dimensions tended to be smaller on the left side; by ANCOVA, this difference reached statistical significance for A_di and A_ap (p < 0.05), indicating that differences in surface area between the two sides could not be explained by differences in volume.

View larger version (23K): [in this window] [in a new window]   Figure 3.   (A) Mean values (± SE) for Adi, Ado, and Aap at FRC, FRC+, and TLC on the right (black bars) and left (blank bars) side in nine normal subjects (**p < 0.01, ***p < 0.001). (B) Same data plotted as a function of the supine volumes of the right and left lungs. Right diaphragm: closed symbols; Left diaphragm: open symbols.

As expected, the radius of curvature increased in both the coronal and sagittal planes on going from FRC to TLC, but the rate of change with lung volume was greater on the left than on the right side. In the coronal plane, the rate of change averaged 1.4 ± 0.4 cm/L on the right side and 3.5 ± 0.6 cm/L on the left side (p = 0.01); corresponding values in the sagittal plane were 2.3 ± 0.7 cm/L and 4.5 ± 0.8 cm/L, respectively (p = 0.007).

Patients

All patients had a displacement of the mediastinum toward the graft (Figure 2); the angle between lines A and B averaged 25° at FRC. As compared with the native lung, the TLC and the FRC of the graft were markedly reduced, on average to 56% and 50%, respectively (p < 0.0005). The TLC of the graft was also reduced as compared with the TLC of the ipsilateral lung in the controls, on average to 76% (p < 0.006). In contrast, the FRC of the graft averaged 100% of that of the control lung (p = NS). These results have been described and discussed in detail in our previous study (5).

Figure 4A shows average values for A_di, A_do, and A_ap on the side of the graft and on the ipsilateral side in the controls. A_do was significantly smaller in the patients than in the normal subjects, whatever the volume studied (p < 0.01); in contrast, A_ap at TLC was significantly greater in the patients (p < 0.05). Figure 4B, which displays the data as a function of the supine volume of the transplanted and control lungs, illustrates that the larger A_ap on the transplanted side at TLC was explained by the smaller TLC of the graft. The figure also shows that, as compared with control values, values for A_di and A_do in the patients were displaced toward smaller values of surface area (p < 0.02 for A_do by ANCOVA); thus, at a given absolute lung volume, diaphragm surface area tended to be smaller on the transplanted side than on the ipsilateral side in the controls because of a decreased A_do.

View larger version (23K): [in this window] [in a new window]   Figure 4.   (A and B) mean data (± SE) obtained from nine patients who had SLT and nine normal controls. Data are presented in the same format as in Figure 3 for the transplanted (LTx) side (black bars and closed symbols) and the ipsilateral side in the controls (blank bars and open symbols) (*p < 0.05, **p < 0.01).

Figure 5A shows average values for A_di, A_do, and A_ap on the side of the graft and on the native side. At the three volumes studied, A_do was significantly smaller on the transplanted than on the native side (p < 0.01), whereas the converse was true for A_ap (p < 0.05). Because the difference was more pronounced for A_do than for A_ap, A_di tended to be smaller on the transplanted side, but the difference only reached statistical significance at FRC+. Figure 5B shows the same data as a function of the volume of the transplanted and native lungs. As compared with values of A_di, A_do, and A_ap obtained on the native side, values obtained on the transplanted side were displaced toward lower values of surface area and/or lung volume; however, ANCOVA showed that the relations of A_di, A_do, and A_ap with lung volume were not significantly different on the two sides. In other words, differences in diaphragm dimensions between the native and the transplanted side could be explained by differences in lung volume.

View larger version (22K): [in this window] [in a new window]   Figure 5.   (A and B) Mean data (± SE) obtained from nine patients who had SLT. Data are presented in the same format as in Figure 3 for the transplanted (LTx) side (black bars and closed symbols) and the native side (blank bars and open symbols) (*p < 0.05, **p < 0.01, ***p < 0.001).

Table 3 shows average values of diaphragm curvature on the native and the transplanted side in the patients, and corresponding values on the ipsilateral side in the controls. The radius of curvature was significantly greater on the native side than on the control side, and this difference was more pronounced in the coronal than in the sagittal plane. On the other hand, no significant difference in radius of curvature was found between the transplanted side and the control side, indicating that the diaphragm had a normal configuration after SLT.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Normal Subjects

We observed that as compared with the left side, the diaphragm dome on the right side had a greater surface area at all lung volumes, and was located more cranially at FRC, making the right A_ap larger than the left one (Figure 3A). These differences cannot be explained by differences in volume between the two lungs. First, we have shown in both previous studies and the present study that the surface area of the diaphragm dome is relatively insensitive to the volume of the lung (4, 5, 7). Second, the dimensions of the zone of apposition should depend on intrathoracic rather than on lung volume. We found that the left lung was 0.2 L smaller than the right lung, a difference that corresponds roughly to the volume of the heart (9); it seems therefore that differences in intrathoracic volume cannot account for the observed differences in right versus left A_ap.

The more cranial position of the right portion of the diaphragm dome at FRC is likely to be related to the anatomic asymmetry of the abdominal contents. On the left side, the diaphragm is in contact with organs (stomach, intestine) that are distortable and are likely to have an average density similar to that of water, with the effect that their behavior may be close to that of a liquid-filled system; as a result, under static conditions, the left diaphragm should be subjected to a surface pressure gradient that approximates a hydrostatic gradient (10). On the other hand, the right diaphragm is separated from the abdominal viscera by the liver, which is a relatively rigid structure; the surface pressure distribution along the muscle may thus differ substantially from a purely hydrostatic gradient. This factor, as well as the tendinous attachments of the liver with the inferior vena cava and the subhepatic veins, may explain why the right diaphragm dome was located more cranially than the left one at FRC, and hence why A_ap was greater on the right side. It may also explain why A_do was larger on this side. Our analysis of muscle curvature showed that although the radius of curvature of the dome increased in both the coronal and sagittal planes with lung volume, the change was substantially smaller for the right than for the left side. This observation thus supports the view that the liver has a significant influence on the shape (and presumably the surface) of the diaphragm dome.

In contrast to A_ap at FRC, A_ap at TLC was similar on the left and right sides, indicating that the right and left domes of the diaphragm had a similar position along the craniocaudal axis of the body. The lung has a much higher shear modulus at TLC than at FRC (11); it is therefore likely that at high lung volumes, the lung has a greater role than the abdominal contents in determining the position of the diaphragm muscle.

Patients

One of the most conspicuous findings of our study was that A_do was significantly smaller on the side of the graft than on either the ipsilateral side in the controls or on the native side in the patients, and that A_di showed a similar trend. Thus, although the hyperinflation present before transplantation was no longer present on the side the graft after SLT, the procedure did not allow diaphragm dimensions to return to normal because it produced a displacement of the mediastinum, and with it of the diaphragm dome, toward the graft.

The validity of these results depends critically on the adequacy of the method that was used to compute right and left A_do. In the patients, we used a line drawn from the vertebral body to the anterior mediastinal line, rather than a midsagittal line as we did in the normal controls. Inspection of diaphragms in cadavers indicated that the pericardium is so firmly attached to the central tendon that any displacement of the mediastinum and the heart (as occurred in the patients, see Figure 2) should be accompanied by a concomitant displacement of the diaphragm dome. We could not use CT to determine the position of the central tendon and the degree of displacement of the dome along the coronal axis because tendon and muscle have very similar thicknesses and densities. Therefore, we designed a method that computed right and left A_do with the assumption that the displacement of the dome toward the graft was similar to that of the anterior mediastinum. We acknowledge that this indirect method is an approximation, but it provides data that are as close as possible to the true in vivo circumstances.

The observation that A_di was smaller on the side of the graft than on the ipsilateral side in the controls suggests that the function of the diaphragm muscle on the transplanted side may not come back to normal after SLT. It does not imply, however, that diaphragm dimension and function did not improve with the procedure. In fact, it is very likely that with the reduction in lung volume, A_di increased after surgery (as illustrated for one patient in Figure 1), which should improve the pressure-generating capacity of the muscle. In addition, besides depending on fiber length, this capacity depends on the axially projected area of the diaphragm dome (A_thor) and configuration (i.e., all other things being equal, transdiaphragmatic pressure [P_di] increases as A_thor [12] and the radius of curvature of the muscle decrease). With the reduction in lung volume, A_thor presumably decreased, and we observed that the configuration of the muscle on the side of the graft was comparable to that found on the ipsilateral side in the normal subjects.

Altogether, these changes would be expected to improve diaphragm function; however there are no data of which we are aware comparing P_di before and after SLT. In a recent study, Wanke and colleagues (13) measured P_di during stimulation of the phrenic nerves at 1 Hz and during maximal sniff maneuvers in patients undergoing SLT for emphysema. They reported that as compared with a group of patients with chronic obstructive pulmonary disease who did not undergo transplantation, twitch and sniff P_di at FRC in the SLT patients were not significantly different, but this does not exclude that the procedure improved the diaphragm muscle pressure-generating capacity.

In conclusion, we have shown that after SLT for emphysema, diaphragm configuration comes back to normal, but the surface area of the dome, and with it the total surface area of the muscle, remain smaller than normal because the mediastinum is displaced toward the graft.