| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Medicine, Memorial Hospital of Rhode Island, Pawtucket, 02860; and Department of Medicine, Brown University Medical School, Providence, Rhode Island 02912.
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
FOOTNOTES
REFERENCES
ABSTRACT 
Cohn, David, Joshua O. Benditt, Scott Eveloff, and F. Dennis
McCool. Diaphragm thickening during inspiration.
J. Appl. Physiol. 83(1): 291-296, 1997.
Ultrasound has been used to measure diaphragm thickness
(Tdi) in the
area where the diaphragm abuts the rib cage (zone of apposition).
However, the degree of diaphragm thickening during inspiration reported
as obtained by one-dimensional M-mode ultrasound was greater than that
predicted by using other radiographic techniques. Because
two-dimensional (2-D) ultrasound provides greater anatomic definition
of the diaphragm and neighboring structures, we used this technique to
reevaluate the relationship between lung volume and
Tdi. We first
established the accuracy and reproducibility of 2-D ultrasound by
measuring Tdi
with a 7.5-MHz transducer in 26 cadavers. We found that
Tdi measured by
ultrasound correlated significantly with that measured by ruler (R2 = 0.89), with
the slope of this relationship approximating a line of identity
(y = 0.89x + 0.04 mm). The relationship between lung volume and
Tdi was then
studied in nine subjects by obtaining diaphragm images at the five
target lung volumes [25% increments from residual volume (RV) to
total lung capacity (TLC)]. Plots of
Tdi vs. lung
volume demonstrated that the diaphragm thickened as lung volume
increased, with a more rapid rate of thickening at the higher lung
volumes
[Tdi = 1.74 vital capacity (VC)2 + 0.26 VC + 2.7 mm] (R2 = 0.99; P < 0.001) where lung volume
is expressed as a fraction of VC. The mean increase in
Tdi between RV
and TLC for the group was 54% (range 42-78%). We conclude that
2-D ultrasound can accurately measure
Tdi and that the
average thickening of the diaphragm when a subject is inhaling from RV
to TLC using this technique is in the range of what would be predicted
from a 35% shortening of the diaphragm.
ultrasonography; diaphragm length; diaphragm thickness; diaphragm
mechanics
INTRODUCTION THE RELATIONSHIP BETWEEN diaphragm length
(Ldi)
and lung volume has been well described in animals. Using
sonomicrometers, previous investigators have found that the canine
diaphragm shortens by 20-42% as lung volume increases from
residual volume (RV) to total lung capacity (TLC) (5, 6, 8, 11,
14). Changes in
Ldi have not been
directly measured in humans. Instead, diaphragm shortening has been
indirectly assessed by using radiographic techniques. Between RV and
TLC, diaphragm shortening is estimated to be in the range of
25-35% (4, 9, 10, 17). It is likely that the diaphragm maintains
a constant volume and increases its circumference minimally as it
shortens; thus the diaphragm must thicken as it shortens. Measurements
of diaphragm thickening, then, may provide an alternate means of
assessing the relationship between
Ldi and lung
volume.
Measurements of diaphragm thickness
(Tdi) have been
made with the use of ultrasound to visualize the diaphragm in the area where it abuts the rib cage [zone of apposition (ZOA)] (12, 16). Wait et al. (16) used M-mode ultrasound to demonstrate inspiratory
thickening of the diaphragm in the area of the ZOA; however, the degree
of thickening and, therefore, the degree of shortening that they
measured was greater than the degree of diaphragm shortening previously
measured in animals (5, 6, 8, 11, 14) or estimated in humans (4, 9,
17). Discrepancies between the ultrasound measurements and the previous
human and animal studies may be related to the narrow range of volume
[from functional residual capacity (FRC) to one-half vital
capacity (VC)] over which the thickness measurements were made or
to limitations of M-mode ultrasound. Because two-dimensional (2-D)
B-mode ultrasound provides images for a broader expanse of tissue than
one-dimensional M-mode, it provides greater anatomic definition of the
diaphragm and neighboring structures. Thus it may be better suited to
assess changes in
Tdi. The purpose
of the present study was first to establish the accuracy and
reproducibility of 2-D ultrasound in measuring
Tdi and then to
reevaluate the relationship between lung volume and
Tdi by using 2-D
ultrasound over a range of lung volumes extending from RV
to TLC.
METHODS Autopsy studies. The diaphragm was
imaged through the ZOA (13) by B-mode ultrasound in 26 cadavers. This
site is ideal for ultrasonographic visualization, since the diaphragm
is bounded by soft tissue on either side and lies parallel to the skin
surface and, therefore, the transducer face. To image this area of the diaphragm, a 7.5-MHz transducer (Technicare) was placed on the skin in
the midaxillary line at the level of the right eighth or ninth
intercostal space. In this location, the diaphragm appears as a
three-layered structure comprising two parallel echogenic lines, the
diaphragmatic pleura and peritoneal membranes, sandwiching a relatively
nonechogenic layer (the muscle itself) (Fig.
1). Images were obtained with the chest
both closed and opened and were videotaped for subsequent analysis.
When the chest was open, diaphragm visualization was confirmed by
manual displacement of the diaphragm and simultaneous observation of
the ultrasound image. By introducing air into the pleural space, we
further identified the diaphragm as the most superficial structure to
be obliterated by the echogenic air-chest wall interface.
The image was either accepted for analysis or rejected if it was
indistinct or if the limiting membranes were not parallel. Once an
image appropriate for analysis was obtained, the intact diaphragm was
marked at the site being imaged. This portion of the diaphragm was then
excised and laid on a surface;
Tdi was then
measured by ruler at the marked site by at least one blinded observer.
Ultrasound images were converted to still photographs by a video
printer (Sony) and were also analyzed in a blinded fashion. For both
the ultrasound and gross anatomic measurements, Tdi was defined
as the distance between the outer edge of the limiting membranes. A
plot of Tdi
determined by ultrasound vs. Tdi determined by
ruler was constructed.
Fig. 1.
Ultrasound of right hemidiaphragm obtained in zone of apposition.
Diaphragm can be identified as a 3-layered structure consisting of
pleural and peritoneal membranes and the muscle itself.
[View Larger Version of this Image (116K GIF file)]
Table 1.
Subject characteristics
Subject No.
Sex
Age, yr
Tdi, mm
FVC, liters
FVC,
%Predicted
Height, cm
Weight, kg
1
F
29
2.3
3.35
95
158
52.3
2
M
30
2.9
5.47
105
175
70.5
3
F
42
3.3
3.50
100
159
54.1
4
M
31
2.8
5.16
100
174
76.4
5
M
30
2.7
5.05
102
171
61.4
6
M
32
2.4
4.98
105
170
60.0
7
M
35
2.8
6.12
116
178
76.4
8
M
33
3.3
5.10
106
171
66.4
9
M
41
3.0
6.23
127
175
77.3
Tdi, diaphragm thickness; FVC, forced vital
capacity; F, female; M, male.
![TF (V<SC>l</SC>) = [<IT>T</IT><SUB>di</SUB>(V<SC>l</SC>) − <IT>T</IT><SUB>di</SUB>(RV)]/ <IT>T</IT><SUB>di</SUB>(RV)](/content/vol83/issue1/fulltext/291/img001.gif)
|
(1) |
RESULTS
There was a wide range of
Tdi among the
cadavers (2-7 mm) and a significant correlation between the
ultrasound Tdi
and ruler Tdi
(R2 = 0.89;
P < 0.001) (Fig.
2). The mean ultrasound
Tdi in the
autopsy series was 3.4 ± 0.8 (SD) mm, and the mean ruler
Tdi was 3.2 ± 0.9 mm. These values are similar to those previously measured (16) or
calculated from autopsy studies (1, 2) (mean
Tdi of 3.5 mm for
a 70-kg adult). The slope of this relationship (0.89) approximated the
line of identity and the intercept (0.4 mm) approached the origin. The
mean value of Tdi
in the autopsy series was slightly greater than the value of
Tdi measured at
25% VC (2.8 ± 0.4 mm) in our subjects. The
slightly greater value of
Tdi in the
autopsy series may have been due to premortem hyperinflation or to a
larger body habitus of the autopsy subjects.
Ultrasound measurements of
Tdi were
reproducible over the range of volumes studied. When measured over
three sessions for each subject, the mean coefficient of variation for
Tdi at each lung
volume ranged from 0.09 to 0.14. The mean values of
Tdi at each lung
volume for all subjects are shown in Fig.
3. As lung volume increased, so did
Tdi. There was a
more rapid rate of thickening at the higher lung volumes. Thus the
relationship between lung volume and
Tdi could be
described by a polynomial equation
(Tdi = 1.74 VC2 + 0.26 VC + 2.7 mm), where
lung volume is expressed as a fraction of VC. The relationship between
Tdi and lung
volume (R2 = 0.99; P < 0.001) was highly
significant. The mean increase in
Tdi between RV
and TLC for the group was 54% (range 42-78%). When changes in
Tdi were
expressed as TF, a similar relationship with lung volume was noted. The
relationships between
Tdi, TF, and lung
volume are summarized in Table 2.
|
||||||||||||||||||||||||||||||||||
DISCUSSION
The major findings are as follows: 1) 2-D ultrasonography can accurately measure Tdi in the ZOA (autopsy study); 2) Tdi can be reproducibly measured and 2-D B-mode ultrasonography can be used to assess diaphragm thickening over a range of lung volumes from RV to TLC (in vivo studies); and 3) the degree of diaphragm thickening and, therefore, shortening during inspiration approximates the degree of diaphragm shortening previously measured with sonomicrometers in dogs (5, 6, 8, 11, 14) .
Accuracy and reproducibility of 2-D B-mode ultrasound. The depth into the body from which useful images can be recovered and the resolution of the image depend on the sound frequency used. Lower frequencies travel further, but, because of their longer wavelength, have less resolution than higher frequencies. For imaging the dome of the diaphragm, a 2.5- or 3.5-MHz transducer is usually used and can resolve structures on the order of 0.5 mm. For imaging a superficial structure such as the diaphragm in the area of apposition to the rib cage, higher frequency transducers in the range of 5-15 MHz are ideal as they provide better resolution (0.3-0.1 mm, respectively). Measurements of Tdi when these higher frequency transducers were used in the 2-D mode were accurate based on direct comparison at autopsy of ultrasonographic and direct measurements of Tdi (R2 = 0.89) (Fig. 2). The R2 for 2-D echo is similar to that reported for M-mode (R2 = 0.86; P < 0.001) (16). Furthermore, the slope between ultrasound measurements of thickness and direct measurements of thickness approximated a line of identity when 2-D ultrasound was used (0.89) and was also similar to that reported for M-mode ultrasound (0.97) (16).
As with any other diagnostic technique, diaphragm ultrasonography is operator dependent. However, with practice and adherence to the criteria for obtaining a suitable image, as described in METHODS, the measurement of Tdi should prove to be reproducible. When we repeated measurements of Tdi twice at each volume on three separate occasions, the coefficient of variation for Tdi was <15%, with the most variability seen at TLC. In the range of Tdi measured, this degree of variation approximates the limits of resolution of the ultrasound transducer (0.2 mm for a 7.5-MHz transducer).
Degree of diaphragm shortening. As one inhales from RV to TLC, the diaphragm is estimated to shorten by 25-35% (4, 9, 17). These values were obtained by imaging the dome of the diaphragm at different lung volumes. Ldi was calculated from the length of an arc that runs in a single plane from the insertion of the diaphragm on one side of the chest wall to the other. Although the estimates of diaphragm shortening made by using such techniques are in a reasonable range, this methodology is limited. First, the arc may not lie in the same plane as the diaphragm muscle fibers. Second, the arc includes the central tendon. Because the central tendon does not shorten, the actual shortening of the diaphragm muscle fibers is greater than the shortening of the arc.
Diaphragm ultrasound provides an alternate approach for assessing changes in Ldi. Thickening of the diaphragm should reflect the shortening of the diaphragm as follows: first, the volume of the diaphragm muscle is the product of its Ldi, width (Wdi), and Tdi. With the assumption that the volume of the diaphragm muscle is constant as the diaphragm shortens, Ldi then will vary inversely with the product of Tdi and Wdi
|
(2) |
Wait et al. (16) measured thickening of the diaphragm in the ZOA and reported a TF of the diaphragm, defined as the difference between the thicknesses at end inspiration and end expiration divided by the thickness at end expiration. They found that TF increased linearly with inspired volume. The mean slope of the relationship between TF and inspired volume, where inspired volume is expressed as a fraction of inspiratory capacity, was 2.3 and varied widely among subjects (0.7-4.3). The slope of the TF-lung volume relationship is proportional to muscle shortening. With the assumption that Wdi, which is related to the perimeter of the thoracic cavity, remains constant during inspiration, and by applying Eq. 2, a slope of 2.3 implies that the thickening at end inspiration was 3.3 times that at end expiration. Accordingly, the muscle would have shortened to 30% of its original length. Such shortening is more than has been estimated for the diaphragm from direct measurements in animals (5, 6, 8, 11, 14).
We demonstrated less diaphragm thickening (54% over the VC range) than
measured by Wait et al. (16). Assuming once again that
Wdi remains
constant in the ZOA, we calculated a shortening to 64% of
the diaphragm rest length. The calculated diaphragm shortening at
intermediate lung volumes over the VC range is shown in Fig.
4. Because the optimal length of the
diaphragm for tension generation
(Lo) is at or
slightly below FRC (11), we used the value of
1/Tdi at 0.25 VC
as an approximation of
Lo and expressed our calculated lengths relative to this number. At lung volumes between
25 and 75% of VC, the relationship between diaphragm shortening and
lung volume is relatively linear
(R2 = 0.93)
(Table 2). The linear relationship over this range of volume may be
explained by the model of the diaphragm acting as a piston in a rigid
cylinder. In this context, diaphragm shortening in the ZOA may displace
the dome without changing its shape. Diaphragm shortening would then be
linearly related to the volume displaced by the dome of the diaphragm.
This model may be operant during tidal breathing (10). However, the
slope of this relationship increases as one approaches TLC and
decreases as one approaches RV. Accordingly, the data were better fit
by a model where
L/Lo = 
0.31 VC2 0.06 VC + 1.03, where lung volume is expressed as a fraction of VC
(R2 = 0.99).
The differences in results between the present study and that by Wait
et al. (16) may be related to study design. First, measurements of
Tdi at TLC and RV
were not obtained in their study, and TF was extrapolated over the VC
range. Our data show the slope of the relationship between
Tdi and lung
volume increasing over the VC range. Extrapolated data between 0.5 VC
and TLC then would overestimate TF, whereas a slope calculated from
data obtained between RV and 0.5 VC may underestimate TF. However, when
a slope for our data over a similar volume range (0.25-0.75 VC)
was calculated, our mean slope was 0.52, a value that is less than the
lowest slope measured by Wait et al. (16). It is
unlikely, then, that extrapolation of data to TLC and RV accounts for
the differences in slope between the two studies. There was, however,
another major methodological difference, namely the use of M-mode by
Wait et al. and of the B-mode in our study. With the M-mode, a single point on the diaphragm is recorded over time. In contrast, with B-mode,
the sound path is swept through an arc, and the diaphragm is then
displayed as a 2-D image recorded over time. We found that the dynamic
context of the real time 2-D B-mode sector scan was very helpful in
securely identifying and tracking the diaphragm. When directly
comparing images simultaneously obtained with the M-and B-mode
techniques in four subjects, we found that as the advancing edge of
inflating lung came into view the diaphragm peeled away and separated
from the chest wall in three individuals. The small echogenic area
wedged between the pleural surface of the diaphragm and the pleural
surface of the chest wall most likely represents a small amount of
pleural fluid (Fig. 5). With B-mode ultrasound, one can see the continuous curvature of the diaphragm and
the parietal pleura lining the chest wall; any fluid that may be
present between the two structures can be readily ascertained. However,
with M-mode, this space is not clearly defined, and one would be at
risk for mistaking the fluid as
Tdi.
Alinearity of Tdi-VC relationship. The increase in Tdi as lung volume increased was alinear, with the slope of Tdi increasing more rapidly as lung volume increased (Fig. 3). One possible explanation of this disproportionate diaphragm thickening at high volumes is based on differences in length changes of the rib cage muscles and diaphragm as one approaches TLC. In the dog, the rib cage muscles shorten to a lesser degree than does the diaphragm (7). If they become less effective than the diaphragm at higher volumes, the diaphragm would then contribute proportionally more to the volume change at the higher volumes than the rib cage muscles, thereby shortening and thickening to a greater degree than at the lower lung volumes. However, a model of rib cage and diaphragm shortening based on a radiographic analysis of diaphragm volumes displaced between RV and TLC in humans (15) suggests that the diaphragm contribution to overall lung volume change is greatest between RV and FRC and not between FRC and TLC. A more likely explanation of the disproportionate thickening of the diaphragm at high lung volumes lies in the relationship between Tdi and Ldi (Eq. 2). Assuming that Wdi is constant, the product of Tdi and Ldi is a constant. A plot of Tdi vs. Ldi, then, would be a rectangular hyperbola. Changes in lung volume are, in turn, negatively related to changes in Ldi. By substituting lung volume for Ldi, one would then have a relationship between Tdi and lung volume that resembled a rectangular hyperbola with disproportionate thickening of the diaphragm at higher lung volumes. To summarize, we have demonstrated that 2-D ultrasonography can be used to accurately and reproducibly measure Tdi in the lateral region of the ZOA of the diaphragm. During inspiration, the diaphragm thickens as it shortens. When 2-D ultrasound is used, the average thickening of the diaphragm when inhaling from RV to TLC is in the range of what would be predicted from a 35% shortening of the diaphragm.
FOOTNOTES
Address for reprint requests: F. D. McCool, Pulmonary Division, Memorial Hospital of Rhode Island, Pawtucket, RI 02860.
Received 9 July 1996; accepted in final form 18 March 1997.
REFERENCES
| 1. |
Arora, N. S.,
and
D. F. Rochester.
Effect of body weight and muscularity on human diaphragm muscle mass, thickness, and area.
J. Appl. Physiol.
52:
64-70,
1982 |
| 2. | Arora, N. S., and D. F. Rochester. COPD and human diaphragm muscle dimensions. Chest 91: 719-724, 1987[Abstract]. |
| 3. |
Boriek, A. M.,
T. A. Wilson,
and
J. R. Rodarte.
Displacements and strains in the costal diaphragm of the dog.
J. Appl. Physiol.
76:
223-229,
1994 |
| 4. |
Braun, N. M. T.,
N. S. Arora,
and
D. F. Rochester.
Force-length relationship of the normal human diaphragm.
J. Appl. Physiol.
53:
405-412,
1982 |
| 5. |
Decramer, M.,
T.-X. Jiang,
and
M. Demedts.
Effects of acute hyperinflation on chest wall mechanics in dogs.
J. Appl. Physiol.
63:
1493-1498,
1987 |
| 6. |
Decramer, M.,
J. T. Xi,
M. B. Reid,
S. Kelly,
P. T. Macklem,
and
M. Demedts.
Relationship between diaphragm length and abdominal dimensions.
J. Appl. Physiol.
61:
1815-1820,
1986 |
| 7. |
Farkas, G. A.,
and
D. F. Rochester.
Contractile characteristics and operating lengths of canine neck inspiratory muscles.
J. Appl. Physiol.
61:
220-226,
1986 |
| 8. |
Farkas, G. A.,
and
D. F. Rochester.
Functional characteristics of canine costal and crural diaphragm.
J. Appl. Physiol.
65:
2253-2260,
1988 |
| 9. |
Gauthier, A. P.,
S. Verbanck,
M. Estenne,
C. Segebarth,
P. T. Macklem,
and
M. Paiva.
Three-dimensional reconstruction of the in vivo human diaphragm shape at different lung volumes.
J. Appl. Physiol.
76:
495-506,
1994 |
| 10. |
Loring, S. H.,
J. Mead,
and
N. T. Griscom.
Dependence of diaphragmatic length on lung volume and thoracoabdominal configuration.
J. Appl. Physiol.
59:
1961-1970,
1985 |
| 11. |
Margulies, S. S.,
G. A. Farkas,
and
J. R. Rodarte.
Effects of body position and lung volume on in situ operating length of canine diaphragm.
J. Appl. Physiol.
69:
1702-1708,
1990 |
| 12. | McCool, F. D., J. O. Benditt, P. Conomos, L. Anderson, C. B. Sherman, and F. G. Hoppin, Jr. Variability of diaphragm structure among healthy individuals. Am. J. Respir. Crit. Care Med. 155: 1323-1328, 1997[Abstract]. |
| 13. | Mead, J. Functional significance of the area of apposition of diaphragm to rib cage. Am. Rev. Respir. Dis. 119, pt. 2, Suppl.: 31-52, 1979. |
| 14. |
Newman, S.,
J. Road,
F. Bellmare,
J. P. Clozel,
C. M. Lavigne,
and
A. Grassino.
Respiratory muscle length measured by sonomicrometry.
J. Appl. Physiol.
56:
753-764,
1984 |
| 15. | Rochester, D. F., G. A. Farkas, and J. Lu. Contractility of the in situ diaphragm: assessment based on dimensional analysis. In: Respiratory Muscles and Their Neuromotor Control. New York: Liss, 1987, p. 327-336. |
| 16. |
Wait, J. L.,
P. A. Nahormek,
W. T. Yost,
and
D. F. Rochester.
Diaphragmatic thickness-lung volume relationship in vivo.
J. Appl. Physiol.
67:
1560-1568,
1989 |
| 17. |
Whitelaw, W. A.
Shape and size of the human diaphragm in vivo.
J. Appl. Physiol.
62:
180-186,
1987 |
This article has been cited by other articles:
|
|
 |
|
  E. M. Summerhill, Y. A. El-Sameed, T. J. Glidden, and F. D. McCool Monitoring Recovery From Diaphragm Paralysis With Ultrasound Chest, March 1, 2008; 133(3): 737 - 743. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Singh, J. A. Panizza, and K. E. Finucane Diaphragm electromyogram root mean square response to hypercapnia and its intersubject and day-to-day variation J Appl Physiol, January 1, 2005; 98(1): 274 - 281. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. A. DePalo, A. L. Parker, F. Al-Bilbeisi, and F. D. McCool Respiratory muscle strength training with nonrespiratory maneuvers J Appl Physiol, February 1, 2004; 96(2): 731 - 734. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Pinet, M. Cassart, P. Scillia, M. Lamotte, C. Knoop, G. Casimir, C. Melot, and M. Estenne Function and Bulk of Respiratory and Limb Muscles in Patients with Cystic Fibrosis Am. J. Respir. Crit. Care Med., October 15, 2003; 168(8): 989 - 994. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. H. Mills, J. M. Wild, B. Eberle, and E. J. R. Van Beek Functional magnetic resonance imaging of the lung Br. J. Anaesth., July 1, 2003; 91(1): 16 - 30. [Full Text] [PDF]
|
|
|
|
 |
|
 G Dimitriou, A Greenough, L Pink, A McGhee, A Hickey, and G F Rafferty Effect of posture on oxygenation and respiratory muscle strength in convalescent infants Arch. Dis. Child. Fetal Neonatal Ed., May 1, 2002; 86(3): F147 - F150. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. K. Rehan, J. Laiprasert, M. Wallach, L. P. Rubin, and F. D. McCool Diaphragm Dimensions of the Healthy Preterm Infant Pediatrics, November 1, 2001; 108(5): e91 - 91. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. K Rehan, J. M Nakashima, A. Gutman, L. P Rubin, and F D. McCool Effects of the supine and prone position on diaphragm thickness in healthy term infants Arch. Dis. Child., September 1, 2000; 83(3): 234 - 238. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
