INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pleural fibrosis is the most common sequela of exposure to asbestos dust. In our experience with workers exposed to crocidolite at the Wittenoom mine and mill in Western Australia, the prevalence of asbestos-related pleural fibrosis (APF) among 384 applicants for pneumoconiosis compensation was about 70% (1). APF can take the form of circumscribed pleural plaques (PP) or diffuse pleural thickening (DPT). The former consist of largely acellular, interwoven bundles of collagen within the parietal pleura. In contrast, DPT involves both the parietal and visceral pleura, which may fuse; it often follows a benign asbestos effusion, and is usually associated with obliteration of the costophrenic (CP) sulcus (2).

The consequences of APF include reduced lung volumes (restriction) (3, 4), and exertional dyspnea (5, 6), and in severe cases, hypercapnic respiratory failure and death (7). The mechanisms by which APF cause pulmonary restriction have not been fully elucidated. Three possible contributors are reduced pulmonary distensibility caused by interstitial fibrosis not apparent on radiography ("occult" fibrosis), impaired rib-cage expansion, and impaired diaphragm shortening. Evidence from several studies suggests that when a chest radiograph shows APF without asbestosis, pulmonary restriction cannot be explained by occult interstitial fibrosis (5, 8). The effect of APF on rib-cage expansion and diaphragm shortening has not been studied.

DPT has been shown to be associated with a greater degree of restriction than PP (4, 5), and this could be due to obliteration of the CP sulcus. Adhesion of the parietal and diaphragmatic pleura in the zone of apposition of the diaphragm to the chest wall could restrict the ability of the diaphragm to shorten, and could, by limiting separation of the diaphragm and lower rib cage, limit the ability of the rib cage to expand. Reduced diaphragmatic excursion following chemical sclerosis of the pleural space has been observed on ultrasound examination by Loring and colleagues (11). Thus, although respiratory muscle strength appears to be preserved in APF (12), it is possible that diaphragmatic excursion is reduced, particularly when the CP sulcus is obliterated.

The aims of the present study were to examine the effect of APF on the ability of the diaphragm to shorten and contribute to lung-volume change, and on the ability of the lower rib cage to expand. We hypothesized that DPT with CP fibrosis would reduce diaphragm shortening and the contribution of the diaphragm to changes in lung volume, and that both DPT and PP would reduce lower rib-cage expansion. We confirmed that DPT was associated with reductions in diaphragm shortening, volume contribution of the diaphragm, and lower rib-cage expansion. The last-named effect was the main mechanism of restriction, because the volume contribution of the diaphragm was augmented by relative flattening of the dome of the diaphragm during inspiration. PP was associated with impaired diaphragm shortening and volume contribution, but with preservation of lower rib-cage expansion.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Subjects were recruited from a cohort of outpatients seen between October 1994 and September 1995 because of previous exposure to asbestos. From this cohort, subjects with clinical or plain radiographic evidence of asbestosis or other interstitial lung disease, asthma, emphysema, lung cancer, pleural effusion, or a neurologic or myopathic disorder likely to weaken respiratory muscles were excluded. Chest radiographs were read by one experienced reader (A.W.M.) using the 1980 International Labour Office classification of radiographs of pneumoconiosis (13). Twenty-six male subjects agreed to participate in the study. They were divided into three groups based on the radiologic appearance of their pleural space: (1) controls, who had no pleural disease (n = 7); (2) PP patients, who had costal and/or diaphragmatic plaques with no involvement of the CP angle (n = 12); and (3) DPT patients, who had CP angle obliteration and thickening with or without calcification of the costal and/or diaphragmatic pleura (n = 7). Ethical approval was granted by the Committee for Human Rights of the University of Western Australia.

Pulmonary Function, Respiratory Muscle Strength, and Exercise Capacity

Respiratory function was assessed in all subjects as follows: lung volumes by plethysmography (Model 09103; Warren E. Collins, Inc., Braintree, MA), maximum expiratory flow-volume relationship and FEV₁ by pneumotachography (Model 400VR; Hewlett-Packard, Waltham, MA), transfer factor (TLCO) by the single-breath method (Model 1182; PK Morgan, Chatham, Kent, UK) with correction for hemoglobin and alveolar volume (TLCO/VA), and respiratory muscle strength by maximal inspiratory mouth pressure (MIP) at residual volume (RV) and maximal expiratory mouth pressure (MEP) at total lung capacity (TLC), using the technique of Black and Hyatt (14). Measured values were expressed as % predicted, using the following reference equations: TLC, Crapo and colleagues (15); vital capacity (VC), Kory and coworkers (16); RV, Goldman and Becklake (17); FEV₁, Cotes and associates (18); TLCO, Miller and coworkers (19); and mouth pressures, Black and Hyatt (14). All but one subject underwent an incremental exercise test on an electronically braked cycle ergometer (Model ER900; Jaeger, Würzburg, Germany). Workload was increased by 15 W each minute until the subject achieved his predicted maximal heart rate (HR) or was limited by symptoms. Breath-by-breath measurements of O₂ uptake (O₂), CO₂ output (CO₂), minute ventilation (E), tidal volume (VT), breathing frequency (f), inspiratory time (TI), total breath time (Ttot), and respiratory exchange ratio (R = CO₂/O₂) were measured continuously (Morgan Benchmark Exercise Test System; PK Morgan). Oxygen saturation (Sa_O2) was monitored by pulse oximetry (Biox 3700; Ohmeda, Boulder, CO), using an ear probe. HR was monitored with a Medeci M-1 cardiac monitor (PK Morgan). Maximal oxygen consumption (O_2max) and maximal HR were chosen as the highest O₂ or HR values recorded for any 30 s of exercise. Results were expressed as % predicted. Equations for predicted maximum workload and O_2max were obtained from Jones and colleagues (20) and Blackie and associates (21), and predicted maximum HR from Åstrand and coworkers (22). All equipment was calibrated before each measurement.

Grading of Pleural Disease

The severity of pleural disease was graded in each subject from PA chest radiographs, using a system proposed by Al Jarad and associates (12). This system ascribes a maximum score for each hemithorax of 9 for costal disease, depending on the length and thickness of pleural involvement; 2 for diaphragmatic plaques; and 1 for obliteration of the CP sulcus. Scores for each hemithorax are then summed, giving a total score.

Measurements from Chest Radiographs

Diaphragm length (Ldi), rib-cage dimensions, and subphrenic volume (Vdi), and their changes between RV and TLC were estimated radiographically from multiple PA and lateral chest radiographs taken at predetermined lung volumes during slow inspirations initiated from RV. Inspiratory flow was measured with a pneumotachograph, integrated to obtain inspired volume, and displayed against time on an oscilloscope (Model 5103N; Tektronix, Beaverton, OR). Rib-cage (Vrc) and abdominal (Vab) volume changes were measured with an inductance pneumograph (Model 10.9230; Respitrace, Ardsley, NY) calibrated with the isovolume maneuver. All signals were recorded continuously on a 12-channel direct writing polygraph (Grass Instruments, Quincy, MA).

PA and lateral chest radiographs were taken with a radiographic film auto changer (Puck; Siemens-Elema, Solna, Sweden) at RV and approximately 25%, 70%, and 100% of VC (Films 1 to 4, respectively). Exposure of each film was triggered manually by a radiographer guided by an oscillographic display of lung volume. To allow alignment of the PA and lateral films, radioopaque ball bearings were adhered to the midline of the chest wall as follows: anterior (single bearing) at the level of the xiphosternal junction, and posterior (two bearings) at the level of the 10th thoracic vertebra. The use of the radiograph autochanger imposed a limitation on film size, and in nine subjects (two control, six PP, and one DPT) the entire thorax could not be accommodated on the film. In these subjects the right hemithorax was imaged preferentially, and these subjects were excluded from analysis of left-sided measurements and estimation of Vdi. Radiation exposures for each PA and lateral radiograph were 85 kV and 96 kV, respectively, with a maximal cumulative dose to each subject of less than 0.2 mSv.

Measurement of diaphragm and rib-cage dimensions. Diaphragm and rib-cage dimensions were measured from radiographs, using methods adapted from Braun and coworkers (23). From each PA radiograph taken at TLC (Film 4), the costophrenic junction was identified and taken to represent the anatomic insertion of the diaphragm into the chest wall (Figure 1A). With the use of lateral ribs as landmarks, these insertions were identified on radiographs taken at lower lung volumes (Films 1 to 3). The outline of the diaphragm and inner surface of the rib cage was then traced on each radiograph, and a line was drawn through the midpoint of the vertebra at the level of the diaphragm to divide the thorax into its right and left halves. At each volume, the length of the right and of the left hemidiaphragm (Ldi) and the zone of apposition (Lzapp) of each were measured with flexible tape (Figure 1). A straight-edged rule was used to measure both the height of the most cephalad part of the dome of the diaphragm above the CP angle (Hdo) and the radius of the right and of the left rib cages (Rrc) at the level of the insertion of the diaphragm into each cage. Change in diaphragm shape was inferred from Hdo.

View larger version (18K): [in this window] [in a new window]   Figure 1.   Diaphragm silhouette as seen on PA and lateral chest radiographs at TLC (A) and RV (B). Measurements illustrated are diaphragm length (Ldi) (right hemidiaphragm [ac], left hemidiaphragm [ce], and lateral diaphragm [a'e']); length of zone of apposition (Lzapp) (ab, de, a'b', and d'e'); and dome height (Hdo).

Similarly, from lateral radiographs, the anterior and posterior insertions of the diaphragm were identified on radiographs taken at TLC and, using the sternum and vertebrae as landmarks, these insertions were identified on films at lower lung volumes. A line representing the midpoint between the silhouettes of the right and left hemidiaphragm was taken to represent the mean Ldi on the lateral film. This length and Lzapp were measured with flexible tape. A straight-edged rule was used to measure the diameter (Drc) of both the lower rib cage (at the level of the anatomic insertion of the diaphragm into the vertebrae) and of the mid-rib cage (at the level of the 7th vertebral body), the vertical distance between these points (Htx_lower), and the height of the most cephalad part of the dome of the diaphragm above the posterior CP angle (Hdo).

To correct for magnification caused by divergence of the X-ray beam, we determined an individual correction factor for each radiograph, using the distance between the X-ray source and film (180 cm), the distance between the subject and the radiographic film (1.5 cm), and the diameter and thickness of the rib cage as determined from radiographs, as described by Pierce and associates (24). The use of radiographs rather than calipers to measure rib-cage dimensions for correction of magnification introduces a small (< 1%) overcorrection.

Shortening of the diaphragm from its maximal length at RV, and expansion of the rib cage from its diameter at RV, were calculated at all higher lung volumes. To allow comparisons between subjects, all measurements were normalized by dividing by the subject's height in meters or by expressing dimensions as the fractional change from RV (e.g., fractional shortening of the diaphragm [FSdi], and fractional expansion of the rib cage [FErc]). Fractional shortening of the muscular component of the diaphragm (FSmus) was estimated in control subjects from the equation FSmus = Ldi/Ldimus_RV, where Ldi is the change in diaphragm length and Ldimus_RV is the length of the muscular component of the diaphragm at RV. Ldimus_RV was determined from the equation Ldimus_RV = Ldi_RV  0.25(Ldi_RV  0.6Ldi_RVTLC) by assuming that the central tendon forms 25% of diaphragm length at FRC, and that the diaphragm shortens by 60% of its maximal shortening capacity between RV and FRC (23). FSmus was then estimated in PP and DPT subjects by assuming in each subject a central tendon length equal to the mean central tendon length normalized for height in the control group (Lct/ht_{control mean}), and using the equation FSmus = Ldi/ht/(Ldi_RV/ht  Lct/ht_{control mean}).

In analyzing diaphragm and rib-cage dimensions, we gave greater emphasis to right hemidiaphragm length and right rib cage radius for two reasons. First, the outline of the right hemidiaphragm is less obscured by the cardiac silhouette than is that of the left hemidiaphragm. Second, measurements on the right side were available in all subjects, thus increasing the statistical power of comparisons.

Estimating volume displaced by the diaphragm. The volume displaced by the diaphragm was estimated from the change in subphrenic volume (Vdi) between RV (Film 1) and higher lung volumes (Films 2 to 4). The boundaries of the subphrenum were defined by the dome of the diaphragm cranially and by the diaphragm-apposed rib cage laterally. The lower limit of the subphrenum was defined as the most caudal point at which the diaphragm inserted into the chest wall on the RV film, and was identified through bony landmarks on films at higher lung volumes (Films 2 to 4). A horizontal line was drawn from this point to form the base of the subphrenum.

Vdi was measured by modification of the method of Pierce and colleagues (24), which uses PA and lateral chest radiographs to estimate lung volume. The PA and lateral radiographs were first aligned in the vertical axis, using the radioopaque balls and vertebrae as markers. The volume of the spinal mass within the subphrenum, not considered in the original method (24), was defined on the PA film by lines drawn on either side of the vertebral column, following the tips of the lateral processes of the vertebrae to take account of associated muscle masses. On the lateral film, the anterior limit of the spinal mass was drawn 1 cm in front of the vertebral bodies to allow for the great vessels and associated tissue. The combined subphrenic and spinal volume (Vdi+sp) was then divided into multiple horizontal slices, each 10 mm thick. The PA and lateral dimensions of each slice for Vdi+sp and for spinal volume alone (Vsp) were measured from the radiographs. Pierce and colleagues (24) determined from transverse sections at necropsy and from computed tomography of living subjects that the cross-sectional shapes of Vdi+sp and Vsp were best modeled as ellipses. This allows the volume of each slice to be estimated from the equation V = 0.25 ()(h)(a)(b), where a and b are the major and minor axes of the ellipse, respectively, and h is the height of the slice. The volumes of all Vdi+sp slices and all Vsp slices were summed, and Vdi was determined by subtracting Vsp from Vdi+sp. All dimensions were corrected for magnification. Vdi was corrected for differences caused by inspiratory expansion of the anterior abdominal wall, which formed the anteroinferior boundary of the defined subphrenum. To allow comparisons between subjects, Vdi was either expressed as a percent of inspired volume or was divided by the cube of the subject's height (Vdi/ht³); the latter procedure reflects normalization of each of the three measured dimensions used to calculate Vdi.

Estimating the volume contributed to VC from expansion of the lower rib cage. The volume contributed to VC from expansion of the lower rib cage was estimated by modeling the lower thorax (between the superior aspect of the 7th thoracic vertebra and the vertebral insertion of the diaphragm) as a truncated elliptical cone in which displacement was limited to the walls of the cone. The volume of the model is:

(Vtx_lower) = ¹/₃  · Htx_lower · (R_PAR_LAT + R_PAr_LAT/2 + r_PAR_LAT/2 + r_PAr_LAT)

where Htx_lower is the height, R_PA and R_LAT are the semiaxes of the wider end, and r_PA and r_LAT are the semiaxes of the narrower end of the truncated cone. Values for Htx_lower, R_PA, R_LAT and r_LAT were the mean dimensions measured from lateral (Htx_lower, R_LAT and r_LAT) and PA (R_PA) chest radiographs in each group; r_PA was derived by assuming that the cross-sectional shape of the midthorax was the same as that of the lower thorax (i.e., r_LAT/r_PA = R_LAT/R_PA).

Protocol. Each subject was trained to perform a slow inspiratory VC maneuver through a pneumotachograph over a period of 10 s at a constant flow rate and in a constant posture. Actual and targeted inspired volume were displayed on an oscilloscope (Model 5103N; Tektronix) providing visual feedback of lung volume to the subject. During training maneuvers and when PA and lateral chest radiographs were obtained, the subject stood with his anterior or right chest wall, respectively, against the radiographic film changer, and with his arms elevated. For each set of chest radiographs, the subject exhaled to RV and a radiograph was obtained. The subject then inhaled slowly to TLC and a further three radiographs were obtained at approximately 25%, 70%, and 100% VC. No attempt was made to control the relative contributions of rib cage and abdomen. All subjects performed the maneuver satisfactorily, although more than one sequence was required in eight subjects.

Data Analysis and Statistics

All data were expressed as means ± SEM. Data from the PP and DPT groups were compared with those from the control group and the differences were tested for statistical significance using: (1) the unpaired t test corrected for multiple comparisons for measurements of pulmonary function, respiratory muscle strength, exercise response, and Ldi at RV and TLC; and (2) two-way analysis of variance (ANOVA) for Ldi, FSdi, FSmus, Hdo, Rrc, Drc, FErc, and Vdi measured on Films 2 to 4. Significance was defined as p < 0.05. Multiple linear regression analyses were performed on the entire sample using: (1) Ldi as the dependent variable and Lzapp and Drc as the independent variables, to verify their previously suggested significance (25); and (2) VC % predicted as the dependent variable and Ldi and Drc as independent variables, to quantify their relative importance. The correlation between VC % predicted and chest radiograph scores of pleural disease was analyzed with Pearson's product moment coefficient of correlation. Stepwise linear regression was used to determine the relative importance of costal, diaphragmatic, and CP involvement in VC and to allow modification of the chest radiograph scoring system.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Characteristics

The study groups were similar in age, height, body mass index, and smoking history (Table 1). Of the 12 PP subjects, six had bilateral costal plaques with no visible diaphragmatic involvement, and six had plaques involving both the costal and diaphragmatic pleura. Among the seven subjects with DPT, the CP angle was obliterated bilaterally in five, and on the right in the two remaining subjects. Radiographic scores for severity of pleural disease were similar for subjects with PP and DPT (Table 1).

Pulmonary Function, Respiratory Muscle Strength, and Exercise Test Peformance

Relative to controls, subjects with DPT had a lower TLC and VC, whereas these volumes were not reduced in subjects with PP (Table 2). In subjects with DPT, VC was 900 ml less than predicted. In the subgroup in whom Vdi could be measured, VC % predicted was 97.4 ± 4.6 for controls, 77.7 ± 5.7 for subjects with DPT, and 90.7 ± 7.0 for subjects with PP. TLCO and respiratory muscle strength were normal in all groups. Exercise capacity was similar in all groups, and no significant oxygen desaturation occurred during exercise.

Inspired Volume during Collection of Radiographs

In all groups, VC during collection of radiographs was systematically lower than VC measured during assessment of lung function (mean reduction of 15.2 ± 1.6% and 16.4 ± 1.9% for PA and lateral films, respectively). Inspired volume at Films 2 and 3 was 24 ± 1% and 71 ± 2% of VC, respectively, with no difference between groups. There was close concordance between inspired volumes for matched PA and lateral radiographs.

Diaphragm Length and Shape

In subjects with DPT, Ldi at RV (Ldi_RV) was reduced relative to that of controls (Table 3), as was Ldi at TLC (Ldi_TLC) of the right hemidiaphragm on the PA film (p < 0.01). In subjects with PP, Ldi_RV appeared shorter and Ldi_TLC appeared longer than in controls, but these differences were not statistically significant. FSdi and FSmus between RV and TLC in control subjects were approximately 0.31 and 0.39, respectively (Table 3 and Figure 2). Relative to controls, DPT and PP subjects had reduced Ldi, FSdi, and FSmus. Lzapp was reduced in subjects with DPT, and was reduced on the right side (PA film) in subjects with PP. There was no difference in Ldi between PP subjects with and without visible diaphragmatic plaques. Between RV and TLC, Hdo increased in control and PP subjects and decreased in DPT, implying relative flattening of the diaphragm dome (Table 3 and Figure 3).

View larger version (13K): [in this window] [in a new window]   Figure 2.   Fractional shortening of the right hemidiaphragm (FSdi). Values are mean ± SEM (*p < 0.0001 relative to DPT and PP, by ANOVA).

View larger version (14K): [in this window] [in a new window]   Figure 3.   Change in height of right hemidiaphragm dome normalized for subject's height (Hdo). Values are mean ± SEM (*p < 0.01 relative to control, by ANOVA).

Volume Displaced by the Diaphragm

Figure 4 illustrates the Vdi normalized for each subject's height (Vdi/ht³) between RV and TLC. Relative to that of controls, Vdi/ht³ was reduced in subjects with PP and to a lesser extent in those with DPT; over the range of VC, the reduction in Vdi/ht³ was 93 ml/m³ (457 ml) in subjects with PP and 70 ml/m³ (356 ml) in those with DPT. The percent contributions of Vdi to VC were 51.7 ± 6.0% in controls, 55.1 ± 4.8% in subjects with DPT, and 38.1 ± 6.5% in subjects with PP; in controls and subjects with DPT the relative contribution of the diaphragm was greater at lower lung volumes.

View larger version (14K): [in this window] [in a new window]   Figure 4.   Volume displaced by the diaphragm, normalized for subject's height (Vdi/ht3). Values are mean ± SEM (*p < 0.01 relative to DPT, by ANOVA, and p < 0.001 relative to PP, by ANOVA; p < 0.05 relative to DPT, by ANOVA).

Rib Cage and Abdominal Displacement

Volume changes of the rib cage (Vrc) and abdomen (Vab) were measured with inductance pneumography and normalized for each subject's height. Relative to that of controls, Vrc/ht³ between RV and TLC was reduced in subjects with DPT (controls: 0.45 ± 0.04; subjects with DPT: 0.34 ± 0.03; p < 0.01), but was unchanged in subjects with PP (0.45 ± 0.04). Vab/ht³ was similar in all groups (controls: 0.15 ± 0.03; subjects with DPT: 0.14 ± 0.02; subjects with PP: 0.21 ± 0.02).

Figure 5 and Table 3 show the effect of APF on expansion of the rib cage measured radiographically. In control subjects, FErc at the level of the diaphragm was about 0.08 and 0.18 on the PA and lateral radiographs, respectively. In subjects with DPT, lower-rib-cage expansion was reduced on the lateral film and on the right side in the PA film. PP caused no significant change in lower-rib-cage expansion. The ratio of coronal to sagittal thoracic diameter at the level of the diaphragm decreased by 8.5% in the control group, from 1.53 ± 0.04 at RV to 1.40 ± 0.03 at TLC, and similar changes were noted in subjects with PP (5.6% decrease) and DPT (6.3% decrease). Rib-cage expansion at the level of the 7th thoracic vertebra on the lateral film was not significantly different among the study groups.

View larger version (15K): [in this window] [in a new window]   Figure 5.   Fractional expansion of the right hemithorax at the level of the insertion of the diaphragm (FErc lower), measured radiographically. Values are mean ± SEM (*p < 0.01 relative to control, by ANOVA).

Volume Contribution of the Lower Rib Cage

Table 4 lists measured rib-cage dimensions used to calculate Vtx_lower at RV and TLC. Estimated volume contributions of the lower rib cage over the range of VC (Vtx_lower) were 255 ml/m³ for controls, 242 ml/m³ for subjects with PP, and 103 ml/ m³ for subjects with DPT. The difference in Vtx_lower between the control and DPT groups was 152 ml/m³, or 773 ml, of which 534 ml could be attributed to differences in chest wall expansion and 239 ml to differences in Htx_lower.

Predictors of Ldi and VC

Changes in right Lzapp accounted for most of the variability in Ldi on the PA film, with the latter measure expressed by the equation Ldi = 0.49 + 0.82 Lzapp (r ² = 0.86, standard error of the estimate [SEE] = 0.92); Drc failed to account for the unexplained variability. VC % predicted was most accurately obtained through a linear combination of total (right + left hemidiaphragm) Ldi on the PA film and Drc of the lower rib cage on the lateral film, as VC % predicted = 54.04 + 2.99Ldi + 9.7Drc (r ² = 0.69, SEE 8.74). VC % predicted could not be determined from radiographic scores for severity of pleural disease. A multiple regression analysis was done, using VC % predicted as the dependent variable and the scores for costal involvement, diaphragmatic involvement, and CP angle obliteration as independent variables. The resulting equation suggested that the presence of CP obliteration was undervalued, and modification of the scoring system by ascribing a score of 15 for each CP angle obliterated significantly improved the correlation between chest radiograph score and VC % predicted (r ² = 0.36, p < 0.01).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The findings in our study provide insights into the mechanisms of restriction in APF. Subjects with diffuse pleural thickening had a shorter diaphragm at RV and reduced shortening of the diaphragm and expansion of the lower rib cage during inspiration. The volume displaced by the diaphragm was reduced, but this was partly compensated by flattening of the dome during inspiration. As a consequence, the observed reduction in VC was mainly attributable to a reduced change in volume of the lower thorax; this was due to: (1) reduced expansion of the lower rib cage; and (2) a reduced axial height of the lung- apposed rib cage, resulting from obliteration of the CP sulcus. Figure 6 illustrates these changes schematically. In contrast, subjects with PP had normal lower-rib-cage expansion. In this group, diaphragm length appeared shorter at RV and longer at TLC, and although these differences were not statistically significant, diaphragm shortening during inspiration was significantly reduced. This was associated with a reduced contribution of the diaphragm to lung volume change, but VC remained normal. For the entire group, VC % predicted was best obtained with an equation that included both lower-rib-cage expansion and diaphragm shortening. These results suggest that APF can restrict lung expansion through adhesion of the parietal and diaphragmatic pleura within the area of apposition of the diaphragm to the rib cage, which can in turn reduce: (1) expansion of the lower rib cage; (2) pulmonary recruitment of the area of apposition; and (3) the capacity of the diaphragm to shorten and displace volume. Moreover, diaphragmatic plaques can restrict lung expansion by limiting elongation of the diaphragm at RV, thereby reducing maximal shortening. The results of the study support the conclusions of other studies that APF alone, and particularly with diffuse pleural thickening, can reduce lung volumes (3), and emphasize that full expansion of the lung requires separation of the diaphragm from the rib cage in the zone of apposition during inspiration (26).

View larger version (24K): [in this window] [in a new window]   Figure 6.   Schematic illustration of observed changes in the diaphragm silhouette and lower-rib-cage margin in control subjects and subjects with diffuse pleural thickening during an inspiration from RV (black lines) to TLC (grey lines). In subjects with diffuse pleural thickening, CP fibrosis was associated with: (1) a reduced volume contribution of the diaphragm, partly compensated by relative flattening of the dome; (2) reduced lower-rib-cage expansion; and (3) reduced axial height of the lung-apposed rib cage.

Limitations

The conclusions of this study are based on estimates of diaphragm length, rib-cage dimensions, and cross-sectional area and volume of the subphrenic space made with measurements obtained from PA and lateral chest radiographs and the methods described by Braun and colleagues (23) and Pierce coworkers (24). The major assumptions in calculating diaphragm length and shortening are that: (1) skeletal structures adjacent to CP angles at TLC give a reasonable approximation of the anatomic insertions of the diaphragm; (2) the coronal and sagittal planes determining the diaphragm silhouette on PA and lateral chest radiographs in a particular subject remain fairly constant at different lung volumes; and (3) changes in length of the diaphragm silhouettes on PA and lateral chest radiographs are representative of overall change in length of the diaphragm. Using magnetic resonance imaging, Gauthier and associates (27) found that in healthy subjects the zone of apposition was not eliminated when subjects relaxed against a closed glottis after reaching TLC. A persistent zone of apposition at TLC would result in an underestimation of diaphragm length at RV and an overestimation of fractional shortening. However in our study, radiographs were obtained during active expiration (RV) and inspiration, making it likely that diaphragm length in the control subjects was systematically longer at RV and shorter at TLC than in the subjects of Gauthier and associates' study. Despite these methodologic differences, estimated shortening of the diaphragm and of diaphragm muscle in our control subjects was similar to that estimated by Gauthier and associates (27) and by others, who used chest radiography to evaluate diaphragm and rib-cage dynamics (23, 25, 28).

In healthy subjects, the diaphragmatic silhouette on a PA radiograph is produced by the contour of the diaphragm at a near-mid coronal plane, and this plane moves ventrally with increasing lung volume (27, 29). This movement could lead to underestimation of fractional shortening of the diaphragm. In subjects with CP fibrosis, in which adhesion of the pleural surfaces may be irregularly distributed around the circumference of the rib cage, the planes determining diaphragm silhouettes between RV and TLC could differ systematically from those in control subjects. In particular, as lung volume increases, the diaphragmatic silhouette is likely to be determined by the most cephalad part of the diaphragm that has adhered to the rib cage. For these reasons, measurements of length and fractional shortening of the diaphragm may not be representative of the entire diaphragm, and would tend to be underestimates. Estimates of fractional shortening of diaphragm muscle may also be complicated by unpredictable changes in the proportion of the diaphragmatic silhouette occupied by the central tendon at each lung volume. PP involving the diaphragmatic pleura could limit elongation of the diaphragm near RV and shortening near TLC and, depending on the sites of such plaques, could cause inhomogeneous diaphragmatic shortening and change the planes determining the silhouette at each lung volume. Again a consequence could be underestimation of diaphragm shortening. Despite these uncertainties, the finding of reduced shortening and fractional shortening of the diaphragm in both PA and lateral projections supports the likelihood that shortening was indeed decreased in subjects with diffuse pleural thickening and subjects with PP.

The contribution of the diaphragm to change in lung volume (Vdi) was estimated with the method of Pierce and coworkers (24), which was modified to allow for the volume occupied by the vertebral bodies, spinal muscles, and great vessels. We chose this method because when applied to the thorax it gave precise estimates of TLC, and alternate methods (30, 31) were considered likely to overestimate Vdi because of their shape assumptions and failure to account for spinal volume. The method of Pierce and coworkers (24) assumes that the subphrenum and spinal structures have elliptical cross-sections, and we assumed the same for our model of the lower thorax. However, our data and that of Gauthier and colleagues (27) show that as lung volume increases in healthy subjects, the thorax expands more in a sagittal than in a coronal plane, and this assumption could result in underestimation of the volume contribution of the diaphragm and lower thorax. Nevertheless in control subjects, the relative contribution of the diaphragm to VC was 51.7 ± 6.0%, a figure similar to previous estimates made with radiographic techniques and measurements of surface displacements (31, 32). Although lower-rib-cage expansion was more severely impaired in the coronal than in the sagittal plane in subjects with diffuse pleural thickening, the ratio of coronal to sagittal diameter of the subphrenum and its variance between RV and TLC, were similar to the values in the control group, and impaired lower rib cage expansion in these subjects was thus unlikely to be a significant source of error.

Inspired volume measured during the radiographic sequence was lower than VC measured during formal assessment of lung function. This is likely to relate to the posture adopted for radiographs, and particularly to arm elevation (24, 33), and the inspired volume during the radiographic maneuver was reduced to the same extent in all groups. There was also no difference between groups in the volume increments at which chest radiographs were obtained in the PA and lateral sequences. These factors are therefore unlikely to have influenced the results.

Implications

We attribute the restriction observed in subjects with diffuse pleural thickening to CP fibrosis which, in effect, attaches the diaphragm to the chest wall at a position more cephalad than the anatomic insertions of the diaphragm, thus reducing its maximum length at RV and the area of apposition available for recruitment during inspiration. As lung volume increases, increasing neural drive to and tension within the diaphragm must result in a transverse force vector at sites of adhesion, which opposes outward displacement of the lower rib cage resulting from the action of inspiratory intercostal muscles and increased abdominal pressure. In addition, a more cephalad insertion of the diaphragm into the rib cage would decrease its mechanical advantage in displacing the costal margin outward and upward.

Relative to predicted values, VC measured plethysmographically in subjects with diffuse pleural thickening was reduced by 900 ml. To partition the relative contributions of abnormalities in diaphragm and lower-rib-cage motion to this restriction, we estimated the volume change of the lower thorax between RV and TLC, using measured dimensions and modeling of the lower thorax as a truncated elliptical cone (Table 4). We assumed that the volume contribution of the upper rib cage was the same in both the control and DPT groups, and the available data support this (Table 3). When adjusted for differences in mean height between the study groups, the model estimated that the volume change of the lower thorax between RV and TLC was 773 ml less in subjects with diffuse pleural thickening than in controls, because of differences in rib-cage expansion (534 ml) and axial height of the lower rib cage (239 ml). This difference, together with the reduction in volume displaced by the diaphragm in the group with pleural thickening (356 ml), yields an estimated reduction of VC close to that measured plethysmographically. These estimates suggest that impaired expansion of the lower rib cage and inability of the diaphragm to separate from it are principally responsible for volume restriction in CP fibrosis. Despite the substantial decrease in shortening within the zone of apposition, the volume displaced by the diaphragm was decreased by only a small amount because the diaphragm flattened during inspiration, thereby helping to maintain its volume contribution.

Our finding that the diaphragm flattens during inspiration in subjects with CP fibrosis but not in control subjects or subjects with PP presumably reflects a difference in the balance of forces determining diaphragm shape that results from the more cephalad attachment of the diaphragm and rib cage in CP fibrosis. In subjects with diffuse pleural thickening, the reduced area of apposition available for recruitment is likely to result in elimination of its availability at lower lung volumes than in controls; diaphragm flattening with inspiration is then likely to occur because transdiaphragmatic pressure, which opposes flattening, is lower, and because the axial forces that can be generated by muscle fibers within the dome are higher, since the average length of these muscle fibers is longer. Volume change achieved by flattening of the dome of the diaphragm rather than diaphragmatic shortening in the area of apposition could have implications for the efficiency of diaphragm muscle, since comparable axial displacement would require disproportionate shortening of muscle fibers in the dome.

It is important to recognize that change in subphrenic volume during the VC maneuver is not purely a function of diaphragm action, but also of accessory muscles (intercostal muscles, abdominal muscles) and of rib-cage and abdominal elastances. In our study, maximum mouth pressures suggested similar respiratory muscle strength in the three study groups, and systematic intergroup differences in the elastance of the abdomen and rib cage appear unlikely to explain the observed differences in behavior of the chest wall.

In subjects with PP, diaphragm shortening and the contribution of the diaphragm to lung volume change was reduced. Possible explanations for this include occult pulmonary fibrosis, occult CP fibrosis, and increased diaphragm elastance. Reduced pulmonary compliance from occult pulmonary fibrosis could increase the load on the diaphragm and limit its excursion. We believe that this is an unlikely explanation for the reduced diaphragmatic contribution to lung volume change in PP for the following reasons: (1) any reduction in pulmonary compliance is likely to also limit chest wall excursion, and this was not observed; (2) physiologically significant asbestosis in the absence of clinical and radiologic changes is rare (8, 10); and (3) measurements of respiratory function showed no evidence of interstitial lung disease in these subjects. The presence of minor degrees of CP fibrosis was also unlikely, given that at TLC, diaphragm length was not reduced, and dome flattening was not observed. The reduced Vdi with inspiration in subjects with PP could be explained by an increased elastance of the diaphragm resulting from the presence of diaphragmatic plaques, thereby limiting diaphragmatic elongation at RV and diaphragmatic shortening, particularly during early inspiration, when intraabdominal pressure decreased with relaxation of expiratory muscles. Although half the subjects with PP had no radiographic evidence of diaphragmatic plaques, radiographs are a relatively insensitive measure of plaques, to the extent that radiologically apparent plaques constitute only a fraction of those found at necropsy (34). In the six subjects with PP in whom a volume contribution of the diaphragm could be measured, Vdi between RV and TLC was significantly reduced relative to that in control subjects (mean reduction 457 ml), but VC (which was 90.7 ± 7.0% predicted, or 369 ml below predicted) was not. This apparent anomaly is attributable to the relatively greater variability in VC.

The failure of chest radiograph scores of the severity of pleural disease to correlate with VC in our sample was due to the low value ascribed to the most physiologically important abnormality: obliteration of the CP sulcus. Our results suggest that if a classification is to reflect functional impairment, the presence of CP fibrosis should be given greater emphasis.

The results of our study provide novel information about the effect of asbestos-related pleural disease on the behavior of the diaphragm and rib cage; they suggest that CP fibrosis is more likely to cause restriction than are PP because it impairs the interaction of the diaphragm and the lower rib cage in expanding the thorax. In addition, both CP fibrosis and PP limit the volume contribution of the diaphragm to inspiration.