Diaphragmatic Breathing Reduces Efficiency of ... - ATS Journals

Diaphragmatic Breathing Reduces Efficiency of Breathing in Patients with Chronic Obstructive Pulmonary Disease RIK A. A. M. GOSSELINK, ROBERT C. WAGENAAR, HANS RIJSWIJK, ANTHONY J. SARGEANT, and MARC L A. DECRAMER Respiratory Rehabilitation and Respiratory Division, University Hospita! Gasthuisberg, Leuven, Belgium; Faculty of Physical Education and Physiotherapy, Katholieke Universiteit, Leuven, Belgium; Department of Physiotherapy, Vrije Universiteit Hospita!, Amsterdam, the Netherlands; and Department of Muscle and Exercise Physiology, Vrije Universiteit, Amsterdam, the Netherlands

The effects of diaphragmatic breathing learning on chest wall motion, mechanical efficiency of the respiratory muscles, breathing pattern, and dyspnea sensation were studied in seven patients with severe chronic obstructive pulmonary disease (COPD) (FEVI 34 ± 7% of the predicted value) during loaded and unloaded breathing. Chest wall motion was studied focusing on amplitude and phase relation of rib cage and abdominal motion. Mechanical efficiency was defined as the ratio of added external power output and added oxygen consumption during inspiratory threshold loading (40% maximal inspiratory pressure [Pim]). After 2 wk run-in, all subjects participated in a diaphragmatic breathing program for 3 wk. Variables obtained during diaphragmatic breathing were compared with those obtained during natural breathing. During diaphragmatic breathing the ratio of rib cage to abdominal motion decreased significantly during unloaded (0.86 versus 0.37; p < 0.01) as well as during loaded breathing (0.97 versus 0.50; p < 0.01). Chest wall motion became more asynchronous during diaphragmatic breathing in the unloaded conditions (mean phase difference for natural breathing 3.5 versus 10.4% for diaphragmatic breathing; p < 0.02) and loaded conditions (mean phase difference for natura! breathing 6 versus 11.4% for diaphragmatic breathing; p < 0.02). Surprisingly, mechanical efficiency decreased significantly during diaphragmatic breathing (2.57 ± 0.76%) in comparison with natural breathing (3.35 ± 1.48%; p < 0.01). Tidal volume, respiratory frequency, and duty cycle did not change significantly during diaphragmatic breathing. Dyspnea sensation tended to increase during diaphragmatic breathing. We conclude that diaphragmatic breathing in patients with severe COPD detrimentally affected coordination of chest wall motion as well as mechanical efficiency, while dyspnea sensation tended to increase. Consequently, we question the usefulness of this technique in the rehabilitation of COPD patients. Gosselink RAAM, Wagenaar RC, Rijswijk H, Sargeant AJ, Decramer MLA. Diaphragmatic breathing reduces efficiency of breathing in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1995;151:1136-42.

Alterations of chest wall motion are common in patients with chronic obstructive pulmonary disease (COPD) (1-4). Several studies have described an increase in rib cage contribution to chest wall motion and asynchrony between rib cage and abdominal motion in these patients (1, 2, 5-7). The mechanisms underlying these alterations are not fully elucidated (8), but appear to be related to the degree of airflow obstruction (4, 5, 9), hyperinflation of the rib cage (4), changes in diaphragmatic function (1, 10), and increased contribution of accessory inspiratory muscles to chest wall motion (4, 11). Increased activity of accessory muscles is believed to enhance dyspnea sensation (12, 13). Consequently, diaphragmatic breathing is commonly applied in physiotherapy practice to correct abnormal chest wall motion, decrease the work of

breathing and dyspnea, increase the efficiency of breathing, and improve ventilation distribution (14). Although several studies investigated the effects of diaphragmatic breathing in COPD patients, to the best of our knowledge no controlled studies are available at present (15). Uncontrolled studies revealed a decrease in rib cage motion and an increase in abdominal motion during diaphragmatic breathing (3, 16), but pulmonary function (17), ventilation distribution (16), and exercise capacity (18) remained unaltered. These studies did not address the effects of diaphragmatic breathing on dyspnea or the mechanical efficiency of breathing. The present study was thus designed to investigate the effects of diaphragmatic breathing on mechanical efficiency of the respiratory muscles, chest wall motion, and dyspnea sensation in patients with severe COPD.

(Received in original form February 18, 1994 and in revised farm October 78, 1994) Correspondence and requests for reprints should be addressed to Dr. Rik Gosselink, Professor of Respiratory Rehabilitation, Division of Respiratory Rehabilitation, University Hospital Gasthuisberg, Herestraat 49, 3000 Leuven, Belgium. Supported by the Dutch Asthma Foundation and the Fonds voor Geneeskundig Wetenschappelijk Onderzoek. Am J Respir Crit Care Med Vol 151. pp 1136-1142, 1995

M ETHODS Study Population Seven patients with moderate to severe stable COPD were included in the study. The following criteria were used for patient selection: (/) FEV, of less than 40% of the predicted value; (2) age between 45 and 75 yr;

Gosselink, Wagenaar, Rijswijk, et al.: Diaphragmatic Breathing and Efficiency TABLE 1 ANTHROPOMETRIC AND PULMONARY FUNCTION DATA OF THE PATIENTS* Age, yr Gender, F/M Height, cm Weight, kg Body mess index, kg/rn , VC, % pred FEV 1 , % pred TLC, % pred RV, % pred FRC/TLC, % pred Pimax cm H 2 O Pi m ., % pred

4 65 2/5 169 ± 3 67 k 8 2.7 23.5 88 k 11 34 k 7 116 14 159 36 80 k 16 81 ± 10 88 k 16

Delinition of abbreviations: VC = vital capacity; RV = residual volume; Pimax = maximal inspiratory pressure. • Values are mean t SD.

(3) no recent cardiac complaints; (4) absence of other pathologic conditions (e.g., cerebrovascular diseases, rheumatism, arthritis, and malignancy). Anthropometric and pulmonary function data are summarized in Table 1. All patients were outpatients on inhaled bronchodilators and steroids, and clinically stable during the period of the investigations. None of the patients used oral steroids on a regular basis and all had a normal body mass index. Two patients had a positive Hoover sign, one on deep inspiration and one during tidal breathing. The procedure was explained to all subjects and informed consent obtained. The study was approved by the Medical Ethical Board of the Vrije Universiteit Hospital. Study Design An A-B-A design was followed for the study. During a run-in period of 2 wk (Phase A), chest wall motion, mechanica] efficiency, and dyspnea were measured twice (Tests 1 and 2) during natural breathing (NB), defined as the naturally adopted pattern of breathing. The patients were at that time unaware of the type of breathing exercises they learned afterward. At least 1 wk was left in between these measurements. Subsequently, the subjects participated in a diaphragmatic breathing (DB) learning program for 3 wk (three treatment sessions a week; Phase B). The subjects were instructed to inspire predominantly with abdominal motion, while reducing upper rib cage motion. Initially, the subjects received instruction in the supine position; after two sessions DB was also practiced in the sitting position. Tactile feedback was given with one hand of the patient on the abdomen and the other hand on the upper rib cage. The three weekly sessions had a duration of 45 min and the patients practiced the exercises at home for 15 min twice a day. The physiotherapist was satis-

fied with the diaphragmatic breathing pattern if this pattern was associated with a doubling of abdominal tidal excursion, while rib cage excursion was greatly reduced. NB after the DB learning period was instructed as breathing without voluntary control of the coordination of breathing. Chest wall motion, mechanical efficiency, and dyspnea were measured during NB and DB, immediately after the 3-wk learning period and 2 wk later (second Phase A, Tests 3 and 4). The order of DB and NB was varied randomly. Materials The subjects inspired through a two-way Hans-Rudolph valve (Kansas City, MO), of which the inspiratory port was connected to a modified Threshold (Health Scan Products, Inc., Cedar Grove, NJ) with variable bad. Chest wall motion was monitored using flexible hollow tubes: one positioned around the entire rib cage at the level of the nipples, the other positioned around the abdomen at the level of the umbilicus. The change in length of the tubes during chest wall motions was continuously monitored using an acoustic technique. Each tube was attached to a loudspeaker at one end and to a microphone at the other. Transit time from loudspeaker to microphone and hence, length of the tubes was continuously monitored. The accuracy of this system was tested by imposing length changes

1137

at different initial lengths. At three different initial lengths of the coil (i.e., 60, 85, and 110 cm) 10 repeated measurements of a 2-cm change in circumference were made, showing a mean coefficient of variation of 4.1% (range 1.3 to 8.7 0/0). We expressed our measurements of circumferences of rib cage and abdomen as a percentage of the change in circumference of rib cage and abdomen during a maximal inspiration from FRC to TLC. This maximal inspiration was performed in each experiment. In fact this is analogous to a Konno-Mead analysis since during relaxation both rib cage and abdomen circumference will decrease from 100 to 0%, which yields a relaxation line with a slope of —1. Relaxation maneuvers are difficult to perform in our COPD patients. Then, the ratio of the amplitudes of rib cage to abdominal motion (RC/ABD ratio) was calculated, yielding mean values in healthy subjects of 0.44 ± 0.08 during NB and 0.22 ± 0.05 during DB. Inspiratory pressure (P10EZ; Datascope, Paramus, NJ), inspiratory flow (Spirolog; Dffiger, Germany), and chest wall motion were digitized (sample frequency 100 Hz) and stored on a computer (Olivetti M250). Data analysis was carried out using PC-MATLAB (The MathWorks, Inc., South Natick, MA). The expiratory side of the Hans-Rudolph valve was connected to a pneumotachograph. Flow was determined as the pressure drop across the pneumotachograph which was measured with a different pressure transducer (Validyne Co., Northridge, CA). This flow signal was used to calculate respiratory duty cycle (Ti/Ttot). It was integrated to volume to obtain tidal volume (VT). A capnograph was placed in series with the Rudolph valve (Sensornormocap Model CO-102; Datex Instrumentation, Helsinki, Finland) to measure end-tidal CO 2 . At the outlet of the pneumotachograph the Douglas bag system was arranged in series to measure minute ventilation (VE), oxygen consumption (Vo 2 ), and carbon dioxide production (Vco 2 ). During the last 5 min of the control and loaded runs, the expired gas was collected in the Douglas bag and subsequently analyzed (0M-11 oxygen analyzer and LB1 carbon dioxide analyzer; Beckman Co., Anaheim, CA; gas meter, Meterfabriek Dordrecht, the Netherlands; and LCD thermometer, Type 650-71g; RS Components Ltd., Hong Kong). Gas analyzers and pneumotachograph were calibrated before each experiment. Pulmonary function was assessed with a wet spirometer (Pulmonet III; Sensormedics, Bilthoven, the Netherlands). Measurements Prior to the initial experiments, pulmonary function was determined, using standard lung function tests, according to American Thoracic Society recommendations (19). The vital capacity (VC) and FEVI were determined from the tracing yielding the largest sum of VC and FEV I . VC and FEVI were related to normal values of Quanjer and associates (20). Maxima! inspiratory pressure (Pi max) was determined at residual volume using the technique and normal values of Black and Hyatt (21). Mechanical efficiency of the inspiratory muscles was studied according to the protocol of Collett and coworkers (22). Studies were carried out after the subjects had fasted for at least 4 h and they had taken their regular medication within 30 min before the start of the experiment. Prior to the experiments the subjects were seated in a comfortable chair for 30 min. Resting values were determined during the next 15 min. VE, endtidal CO 2 , Ti/Ttot, and mouth pressure (Pi) were measured continuously. VE, Vo 2 , and VCO 2 were measured over the last 5 min. Previous experiments showed coefficients of variation of the measurement of VE and Vo l of 8.6 and 7.6%, respectively, during the last 5-min period. After a period of at least 10 min rest, the patient started breathing against an inspiratory bad of 40% of Pi max either with DB or NB for 10 min. Feedback of endtidal CO 2 was given on digital display and the patient was instructed to keep end-tidal CO 2 at resting level. Measurements of VE, VCO 2 , and V02 were repeated during the last 5 min. The efficiency of the respiratory muscles in performing added work during natural and diaphragmatic breathing was calculated by dividing added mechanica! work by the added energy requirement. The amount of added mechanical work per minute during loaded breathing was calculated by multiplying the mean minute ventilation (L) by the mean mouth pressure (cm H 2 O) during the last 5 min of each run and expressed in kilogram-meters per minute. Mouth pressure during threshold loading is square-waved (23). So the peak pressure and the mean pressure are the same during this type of loading. Also the work (per minute) necessary for decompression of thoracic gas (Wdec) was taken into account. The following equation was used (24):

1138

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 151 1995

Wdec = ((FRC/2) + VT(.1212.f/(PB - PH,o - Pl), where FRC = functional residual capacity, VT = tidal volume, f = respiratory frequency, PB = barometric pressure, PH20 = water vapor pressure, and Pi = inspiratory mouth pressure. The amount of added energy per minute was calculated by determining the difference in n ./02 between the resting natural breathing and loaded breathing. Mean values of the last 5 min of each run were calculated and converted to its energy equivalent, using Joule's equivalent and a metabolic heat production of 4.825 cal/mlof oxygen consumed. Assuming a respiratory quotient of 0.82, 1 ml of oxygen consumption is equivalent to 2.1 kgm. Analysis of our data justified this assumption, as we found mean values for the respiratory quotient of 0.82 for unloaded breathing, and 0.83 and 0.81 during respectively loaded NB and loaded DB. The amplitude of rib cage and abdominal motion was calculated from the absolute changes in circumference of rib cage and abdomen, respectively. The amplitudes were expressed as the mean of the measurements of the sixth, eighth, and tenth minute, as the coefficient of variation of these values was only 7.8% for loaded as well as unloaded breathing. The phase relationship of rib cage and abdominal movements was analyzed by a cross-correlation function. The output of the cross-correlation function is a percentage. Zero percent (0 degrees) meant perfectly in phase, as 50% (180 degrees) meant that signals of rib cage and abdomen were completely out of phase with paradoxical movement during the entire breathing cycle. From 50 to 100% means between 180 and 360 degrees out of phase. One hundred percent means that signals are again completely in phase, but with a phase lag of one period. With breathing, this is of course not the case since all compartments move within one respiratory cycle. Dyspnea sensation was recorded in the sixth, eighth, and tenth minute with the modified Borg score (25). Statistics The amplitudes of rib cage to abdominal motion, the RC/ABD ratio, the phase relationship between rib cage and abdominal motion, mechanical efficiency, breathing pattern, and dyspnea sensation during Tests 1, 2, 3, and 4 were first analyzed for NB with the Friedman two-way analysis of variance. Then the comparison was made between DB and NB in Test 3 and Test 4, separately, using the Wilcoxon matched-pairs signed-ranks test. Stability of the pattern of thoracoabdominal motion was assessed with coefficients of variation of the amplitudes of rib cage and abdomen during loaded and unloaded breathing. Spearman correlation coefficients were calculated for repeated tests. All tests were performed with StatGraphics (Manugistics, Inc., Rockville, MD).

RESU LTS Natura! Breathing

Chest wall motion. The ratios of the amplitudes of rib cage to abdomen during loaded and unloaded NB did not change significantly. The ratios for unloaded breathing were 0.83 ± 0.50 (Test 1) and 0.75 ± 0.39 (Test 2) before the DB. After the learning period the ratios did not change significantly in Test 3(0.86 ± 0.42) and Test 4(0.70 ± 0.26, p = 0.77) (Figure 1, upper part). In addition, no significant differences were observed during loaded breathing. The values were respectively 1.23 ± 0.47 (Test 1), 1.07 ± 0.58 (Test 2), 0.97 ± 0.10 (Test 3), and 0.93 ± 0.36 (Test 4: p = 0.60) (Figure 1, lower part). The phase relationship of the movements of rib cage and abdomen during NB were not significantly different (p = 0.58) in the unloaded conditions (mean phase difference for all tests was 4.6%; Figure 2, upper part) and loaded conditions (mean phase difference for all tests was 4.8%; Figure 2, lower part). Respiratory muscle efficiency Mechanical efficiency during NB did not change significantly (p = 0.54) over all tests (Figure 3). Total body oxygen consumption (Vo2) during unloaded NB was not significantly different (p = 0.23). During loaded breathing V02

1.5

3

unloaded natural breathing unloaded diaphragmatic breathing

c E -0 -g2 1. o a> cs) ea -o

0.5

cC

0 1 test test 2 test 3 test 4

Assessments 2 -

loaded naturel breathing loaded diaphragmatic breathing

c E 1.5 o -Ct CD C» .0 CO 0 -Cl

C-

0.5 co

CC

test 1

test 2 test 3 test 4

Assessments Figure 1. Ratio of the excursion of rib cage and abdomen during naturel and diaphragmatic breathing during unloaded (upper) and loaded (lwei-) conditions before (Tests 1 and 2) and after (Tests 3 and 4) DB learning. Asterisk indicates significant difference between NB and DB. The error bars represent 1 SD.

significantly increased in the successive measurements during NB (p < 0.03). Breathing pattem. During loaded and unloaded breathing minute ventilation, tidal volume, TifTtot, and respiratory frequency were not significantly different between the tests during NB. Dyspnea. No significant changes were observed in dyspnea sensation measured with a Borg score (p = 0.94) in the successive measurements. Diaphragmatic Versus Natura! Breathing

Chest wall motion. After DB learning abdominal excursion increased significantly during DB both during unloaded (p < 0.001) and loaded conditions (p < 0.001) and rib cage paradox was frequently present. Figure 4 shows a representative tracing of chest wall motion during NB and during DB in the same patient. After DB learning (Tests 3 and 4, Figure 1) a significant alteration of

1139

Gosselink, Wagenaar, Rijswijk, et al.: Diaphragmatic Breathing and Efficiency

25

E

_ natural breathing

unloaded naturel breathing

diaphragmatic breathing —

unloaded diaphragmatic breathing 20

5

15

10 2

test 2 test 1

test 3 test

4

1 test test test 2 3 4 test

Assessments 25

Assessments Figure 3. Mechanical efficiency during natural and diaphragmatic breathing before (Tests 1 and 2) and after (Tests 3 and 4) DB learning. Asterisk indicates significant difference between DB and NB. The error bars represent 1 SD.

loaded naturel breathing loaded diaphragmatic breathing

ao

and abdomen during NB and DB in Test 3 showed significant differences for both unloaded conditions (phase difference was 4.1% during NB and 11.0% during DB; p < 0.02; Figure 2, upper part) and loaded conditions (phase difference was 3.6% during NB and 12.0% during DB; p < 0.03; Figure 2, lower part). These differences were also evident in Test 4. Respiratory muscle efficiency. Mechanical efficiency was significantly different between NB and DB in Test 3 (respectively 3.35 ± 1.48% and 2.57 ± 0.76%, p < 0.01) (Figure 3). Also the

15

10

test 2 test 1

test 3 test

4 90

Assessments Figure 2. Phase relation of rib cage and abdominal motion during unloaded (upper) and loaded (lower) breathing before (Tests 1 and 2) and after (Tests 3 and 4) DB learning. Asterisk indicates significant difference between NB and DB. The error bars represent 1 SD.

C.5

< o

0

60-

30 the ratios of the amplitudes of rib cage to abdomen during unloaded DB (0.37 ± 0.16 and 0.32 ± 0.16, respectively) and loaded DB (0.50 ± 0.12 and 0.41 ± 0.08, respectively) was found in comparison to NB (p < 0.01), indicating that the patients were able to perform diaphragmatic breathing correctly. During loaded NB the coefficient of variation of the excursions of rib cage was 15.1% and for abdominal movements 24.5%. With unloaded breathing the coefficients of variation for rib cage and abdomen were 132 and 15.5%, respectively. During unloaded DB a small, not significant decrease of the coefficients of variation for both rib cage and abdomen was present during DB (11.5 and 11.3%, respectively, p = 0.96) compared with NB (13.8 and 15.5%, respectively, p = 0.68). During loaded breathing the coefficient of variation of the excursions of rib cage tended to increase during DB (24.5%) in comparison to NB (15.1%) (p = 0.09), while the coefficient of variation decreased significantly for abdominal movements (NB 24.5% and DB 15.0%, p < 0.03). The phase relationship between the movements of rib cage

-

FRC

0

1 1 90 120 30 60 0

ABDOMEN (% inspiratory capacity) Figure 4. Representative tracing of rib cage and abdominal excursions during NB (solid line) and DB (dashed line) in a COPD patient. Closed circle represents functional residual capacity; open circles, end-inspiratory lung volume. During DB there is a doubling of abdominal motion, whereas the rib cage is moving paradoxically during inspiration.

1140


♦

6

TABLE 2 BREATHING PATTERN (Vr, RR, TifTtot), MINUTE VENTILATION, AND TOTAL BODY OXYGEN CONSUMPTION DURING DIAPHRAGMATIC AND NATURAL BREATHING AFTER THE TRAINING PERIOD*

0

5

• ♦ 0

4

Unloaded Breathing Loaded Breathing NB DB NB DB Vo l , ml/min 215 ± 23 232 ± 13* 288 ± 28 305 ± 22 VT, ml 797 ± 229 936 ± 221 744 ± 283 677 ± 325 Ti/Ttot 0.37 ± 0.04 0.32 ± 0.05 0.37 ± 0.06 0.37 ± 0.08 VE, Umin 7.7 ± 1 8.2 ± 1.6 10.5 ± 1.7 11.4 ± 2.6 RR, breaths/min 10.4 ± 3.2 9.6 ± 3.6 17.1 ± 7.6 20.9 ± 8.7

0

qj

3

• 2

♦

o TEST 3 ♦

Definition of abbreviations: NB natura! breathing; DB = diaphragmatic breathing; Vo, = oxygen consumption; VT = tidal volume; TifTtot = respiratory duty cycle; VE = minute ventilation; RR = respiratory rato. * Values are meen ± SD. Significant difference, p < 0.05.

TEST 4

3 4 5 2 6

Mechanical efficiency DB (%) Figure 5. Identity plot of the mechanical efficiency in COPD patients during natura! and diaphragmatic breathing after DB learning (Tests 3 and 4). Almost all data points fel) left from the identity line indicating that mechanical efficiency was higher during NB.

differences between NB and DB in Test 4 were significantly different (p < 0.01). Figure 5 compares the mechanica! efficiency during NB to the mechanical efficiency during DB. As can be seen, virtually all data points fall left of the identity line, demonstrating a greater efficiency during NB compared with DB. Total body oxygen consumption during unloaded NB and DB in Test 3 (respectively 215 ± 23 and 232 ± 13 ml/min) showed significant differences (p < 0.05) (see Table 2). During loaded breathing Vo l of NB and DB in Test 3 (respectively 288 ± 28 and 305 ± 21 ml/min) showed no significant differences (p = 0.27). Similar results were obtained in Test 4. Breathing pattern. Du ring loaded and unloaded breathing, VE, VT, f, and Ti/Ttot were not significantly affected by DB (Table 2). Dyspnea. In Test 3 the Borg score was significantly different between NB (median score 5, range 3 to 7) and DB (median score 6, range 4 to 9, p < 0.05). No significant differences were found in Test 4 (p = 0.37). DISCUSSION The present study demonstrates that during diaphragmatic breathing, patients were able to increase abdominal motion. Meanwhile the phase relationship between rib cage and abdominal motion deteriorated as well as the mechanical efficiency. There was a tendency for an increase in dyspnea sensation during diaphragmatic breathing. The increase in abdominal motion during diaphragmatic breathing is in agreement with the findings of others (3, 7, 26). Concomitantly, the increase in abdominal motion was accompanied by a significant deterioration of the phase relationship during diaphragmatic breathing. This was also observed ~lier (7, 26). This implies that, contrary to what is commonly believed, chest wall motion became more abnormal with diaphragmatic breathing. We also found that diaphragmatic breathing is associated with a de-

crease in the mechanical efficiency of breathing. This has not been studied before, but the increased oxygen cost of breathing in our patients during unloaded breathing was previously also suggested in an uncontrolled study by Weitzenblum and colleagues (27). Dyspnea sensation tended to increase during (loaded) diaphragmatic breathing compared with natural breathing in our patients. In a previously published study (28), a significant increase in dyspnea sensation was demonstrated in six healthy subjects during diaphragmatic breathing in comparison to rib cage breathing at different inspiratory resistive loads. Moreover, as others, we did not observe significant differences in respiratory variables between the natural and diaphragmatic breathing pattern during loaded and unloaded breathing. There was a tendency for an increase in VE and a decrease of TifTtot during diaphragmatic breathing, but as in another study (3), this tendency did not reach statistical significance. The previously mentioned detrimental effects were, however, only present during diaphragmatic breathing; no permanent changes in NB were observed. These results suggest that the diaphragmatic breathing pattern was not adopted as the natural pattern. Critique of the Methods

The term "diaphragmatic breathing" requires further specification. Diaphragmatic breathing in the present study was applied as it is routinely used in clinical physiotherapy. This means facilitating outward motion of the abdominal wall, while reducing upper rib cage motion during inspiration. Physiologically, it does not necessarily imply a greater diaphragmatic contribution to chest walt motion for several reasons. First, the lower rib cage constitutes a movable part of the boundary of the abdominal container. Consequently, diaphragmatic displacement is not only accommodated by moving the abdominal wall, but also by lower rib cage expansion (29). Nevertheless, variations in abdominal dimensions are likely to remain proportional to diaphragmatic displacement (30). Second, relaxation of the abdominal muscles at the onset of inspiration produces outward motion of the abdominal wall, which is not caused by active contraction of the diaphragm. Ou r patients were instructed to exhale without abdominal muscie contraction. However, recently Ninane and coworkers (31) observed abdominal contraction during expiration at rest in a majority of their COPD patients. Because we did not control for electromyogram (EMG) activity of the abdominal muscles, we cannot substantiate this suggestion. Although the diaphragmatic contribution to chest wall motion was not measured directly, and no attempt to record diaphragm

Gosselink, Wagenaar, Rijswijk, et al.: Diaphragmatic Breathing and Efficiency

1141

EMG was made in the present study, it appears likely that breathing with greater abdominal excursion will require a greater diaphragmatic displacement and hence, a greater diaphragmatic contribution to chest wal! motion. Measurement of mechanical efficiency may be associated with errors in measurement of the work of breathing as well as in the measurement of the energy expenditure during loaded breathing (32). However, the test-retest results during NB from Tests 1 and 2, as well as Tests 3 and 4, showed a high Spearman correlation coefficient (respectively 0.82, p < 0.04 and 0.94, p < 0.01) and no significant differences (p = 0.94), which means that the measurement performed in the present study was reproducible. Measurement of chest wall motion from changes in circumference does not provide pertinent information on shape (anteroposterior or lateral diameter) of the rib cage or abdomen. The cross-correlation analysis was performed on both inspiratory and expiratory signals, such that the phase relationships were not analyzed separately over inspiration and expiration. The information on chest wall motion obtained in the present study was thus approximate. Nevertheless, it provided sufficient control of the diaphragmatic breathing maneuver.

breathing, thereby improving mechanica! efficiency. However, contrary to what is intuitively expected, a significant decrease in efficiency was observed. Several factors may be related to the decrease in respiratory muscle efficiency. An increase in energy expenditure during diaphragmatic breathing could be due to recruitment of additional muscles during diaphragmatic breathing. It was shown that during diaphragmatic breathing, at least in healthy subjects, the EMG activity of the parasternal intercostal muscles increased, while upper rib cage motion was reduced (34, 35). Probably, this additional nonagonistic work is not measured as work, but it increases the energy expenditure, causing a reduction in mechanical efficiency (32). As we found an increase of paradoxical movements, compensatory movement of the chest wall was necessary to maintain a constant VT. This is likely to have increased the oxygen cost of breathing. If paradoxical chest wall movements are present, more inspiratory muscle activity is necessary to generate the same inspiratory pressure and the same global chest wall expansion. In addition, the mechanical coupling between inspiratory muscles and rib cage and abdomen may deteriorate during diaphragmatic breathing, such that less pleural pressure is reached for a similar degree of muscle contraction.

Mechanisms of the Deterioration of Chest Wall Motion and Mechanical Efficiency

Practical Implications

The deterioration of the phase relationship indicates a negative effect of diaphragmatic breathing on the timing of chest wall motion; rib cage and abdomen were moving more asynchronously. Contrary to our findings, in other studies, especially the abdominal wall was moving paradoxically during inspiratory loading (9, 33). We observed from the recordings of rib cage and abdominal motion that paradoxical movement was most apparent at the beginning of inspiration (Figure 4). As the deterioration in phase relationship was found during both loaded and unloaded breathing, it seems unlikely that the lower pleural pressure during loaded breathing could be entirely responsible for the deterioration of the phase relation. Several factors could be responsible for this deterioration allowing for an increase in paradoxical rib cage motion. First, paradoxical rib cage motion could be related to reduced activity of upper rib cage muscles. Experiments in healthy subjects support this contention, since De Troyer and Estenne (34) observed a reduction or suppression of scalene EMG activity and paradoxical upper rib cage motion during diaphragmatic breathing in healthy subjects. By way of contrast, parasternal intercostal EMG activity clearly increased during diaphragmatic breathing (34, 35). Second, an increase of the paradoxical movement of the rib cage could also be the result of flattening of the diaphragm, since contraction of the flattened diaphragm may cause lower rib cage paradox (Hoover sign). Third, relaxation of the abdominal wall prior to the onset of diaphragmatic contraction could be responsible for rib cage paradox. Indeed, relaxation of the abdominal wall facilitates diaphragmatic descent because the abdominal contents no longer act as a fulcrum for the diaphragmatic dome. Consequently, the insertional force of the diaphragm is acting more radially than axially. In the presence of a reduction of the area of apposition due to severe hyperinflation on, and hence, a clear reduction in appositional force, the insertial force may cause inward motion of the lower rib cage (10). The effect of diaphragmatic breathing on respiratory muscle efficiency was studied for the first time in the present study. It is commonly thought that in patients with COPD the increased contribution of rib cage muscles to breathing enhances the oxygen cost of breathing. Consequently, increasing the diaphragmatic contribution would then be expected to decrease the oxygen cost of

No positive effects were observed after a learning period of diaphragmatic breathing in patients with severe COPD. At first sight diaphragmatic breathing during loaded breathing appears as different from diaphragmatic breathing in clinical practice, because no external loading is applied. Conversely, diaphragmatic breathing is often advocated in periods of increased demands on the ventilatory pump, like exercise or exacerbations of COPD. These also represent loading conditions and thus the difference between the experimental conditions and the conditions of application in clinical practice may be less pronounced than it appears at first sight. It may be argued that in our patients, the changes in the ventilatory pump due to hyperinflation of the rib cage and alterations in diaphragm geometry were virtually irreversible. This is consistent with the absence of changes in NB after the learning period of diaphragmatic breathing. Chest wall motion in COPD patients is characterized by less abdominal and increased rib cage movement (1, 4), resulting from the reduced mechanical effectiveness of diaphragmatic contraction, and hence, greater parasternal and accessory muscle contribution. Diaphragmatic breathing learning is in this situation apparently not the treatment of choice. It should also be noted that our patients had relatively high, almost normal, values of Pim.. This is probably the result of adequate adaptations of the inspiratory muscles to chronic hyperinflation (36). In addition, none of the patients used oral steroids on a regular basis (37), all had a normal body mass index and all were outpatients not frequently admitted to the hospita! (38). Different effects may be obtained in patients with more apparent inspiratory muscle weakness. In patients with reversible or less irreversible bronchial obstruction, hyperinflation is less marked, resulting in less dysfunction of the diaphragm. Diaphragmatic breathing may possibly have positive effects in these patients (4). Consequently, the present study does not allow definite conclusions to be drawn on the effects of diaphragmatic breathing in patients with reversible airflow obstruction. Several studies suggest that breathing with upper rib cage motion may be advantageous in patients with severe airflow obstruction. First, because of hyperinflation of the rib cage the diaphragm is placed at a shorter length in these patients. The diaphragm probably adapts to this situation, by loss of sarcomeres in series,

1142


but the effective length range over which the muscle fibers operate will be reduced (39). Second, in dogs, acute hyperinflation puts the parasternal intercostal muscles at a more favorable part of the length-tension relationship (40). Third, fatigability of the rib cage muscles in normal subjects was found to be less than the diaphragm (41). If similar concepts were to apply in COPD patients, rib cage muscles might be more apt to deal with the increased work of breathing. The present study appears to confirm this speculation. These factors probably account for the relatively high, almost normal, values of Pi m ,, in these patients with severe airflow obstruction. In conciusion, the present study demonstrated that diaphragmatic breathing in patients with severe COPD and moderate hyperinflation resulted in a decrease in mechanical efficiency of the respiratory muscles. Chest wall motion became more asynchronous and paradoxical. Dyspnea sensation tended to increase. Acknowledgment: The writers thank R. Heslinga, P. Schepers, G. Tiesinga, and B. van Tol for their excellent technical assistance during the experiments. References 1. Gilmartin, J. J., and G. J. Gibson. 1984. Abnormalities of chest wall motion in patients with chronic airflow obstruction. Thorax 39:264-271. 2. Ashutosh, K., R. Gilbert, J. Hj. Auchinloss, and D. Peppi. 1975. Asynchronous breathing movements in patients with COPD. Chest 67: 553-557. 3. Sackner, M. A., H. F. Gonzalez, G. Jenouri, and M. Rodriguez. 1984. Effects of abdominal and thoracic breathing on breathing pattern components in normal subjects and in patients with COPD. Am. Rev. Respir. Dis. 130:584-587. 4. Martinez, F. J., J. I. Couser, and B. R. Celli. 1990. Factors influencing ventilatory muscle recruitment in patients with chronic airflow obstructien. Am. Rev. Respir. Dis. 142:276-282. 5. Delgado, H. R., S. R. Braun, J. B. Skatrud, W. G. Reddan, and D. F. Pegelow. 1982. Chest wall and abdominal motion during exercise in patients with COPD. Am. Rev. Respir. Dis. 126:200-205. 6. Sharp, J. T., N. M. Goldberg, W. S. Druz, H. Fishman, and J. Danon. 1977. Thoracoabdominal motion in COPD. Am. Rev. Respir. Dis. 115:47-56. 7. Sackner, M. A., H. F. Gonzalez, M. Rodriguez, A. Belisto, D. R. Sackner, and S. Grenvik. 1984. Assessment of asynchronous and paradoxic motion between rib cage and abdomen in normal subjects and patients with COPD. Am. Rev. Respir. Dis. 130:588-593. 8. Jubran, A., and M. J. Tobin. 1992. The effect of hyperinflation on rib cageabdominal motion. Am. Rev. Respir. Dis. 146:1378-1382. 9. Tobin, M. J., W. Perez, S. M. Guenther, R. F. Lodato, and D. R. Dantzker. 1987. Does rib cage-abdominal paradox signify respiratory muscle fatigue? J. AppL Physiol. 63:851-860. 10. Gilmartin, J. J., and G. J. Gibson. 1986. Mechanisms of paradoxical rib cage motion in patients with COPD. Am. Rev. Respir. Dis. 134:683-687. 11. Levine, S., M. Gillen, P. Weiser, G. Feiss, M. Goldman, and D. Henson. 1988. Inspiratory pressure generation: comparison of subjects with COPD and age-matched normals. J. Appl. Physiol. 65:888-899. 12. Ward, M. E., D. Eidelman, D. G. Stubbing, F. Bellemare, and P. T. MackIem. 1988. Respiratory sensation and pattern of respiratory muscle activation during diaphragm fatigue. J. AppL Physiol. 65:2181-2189. 13. Breslin, G. H., B. C. Garoutte, V. Kohlmann-Carrieri, and B. R. Celli. 1990. Correlations between dyspnea, diaphragm, and sternomastoid recruitment during inspiratory resistance breathing. Chest 98:298-302. 14. Miller, W. F. 1954. A physiologic evaluation of the effects of diaphragmatic breathing training in patients with chronic pulmonary emphysema. Am. J. Med. 17:471-477. 15. Gosselink, H. A. A. M., and R. C. Wagenaar. 1993. Efficacy of breathing exercises in chronic obstructive pulmonary disease and asthma: a metaanalysis of the literature. J. Rehab. Sci. 6:66-87. 16. Grimby, G., H. Oxhoj, and B. Bake. 1975. Effects of abdominal breathing on distribution of ventilation in obstructive lung disease. Clin. Sci. Mol. Med. 48:193-199. 17. Cole, M. B., C. Stansky, F. E. Roberts, and S. M. Hargan. 1962. Studies in emphysema: long-term results of training diaphragmatic breathing

on the course of obstructive emphysema. Arch. Phys. Med. Rehabil. 43:561-564. 18. Williams, I. P., C. M. Smith, and C. R. McGavin. 1982. Diaphragmatic breathing training and walking performance in chronic airways obstruction. Br. J. Dis. Chest 76:164-166. 19. American Thoracic Society. 1987. Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease (COPD) and asthma. Am. Rev. Respir. Dis. 136:225-244. 20. Quanjer, P. H. 1983. Standardized Lung Function Testing (Report Working Party Standardization of Lung Function Tests of the European Community for Coal and Steel). Bull. Eur. Physiopathol. Respir. 19 (Suppl. 5):1-95. 21. Black, L. F., and R. E. Hyatt. 1969. Maximal respiratory pressures: normal values and relationship to age and sex. Am. Rev. Respir. Dis. 99:696-702. 22. Collett, P. W., C. Perry, and L. A. Engel. 1985. Pressure-time product, flow, and oxygen cost of resistive breathing in humans. J. Appl. Physiol. 58:1263-1272. 23. Larson, J. L., M. J. Kim, J. T. Sharp, and D. A. Larson. 1988. Inspiratory muscle training with a pressure threshold breathing device in patients with chronic obstructive pulmonary disease. Am. Rev. Respir. Dis. 138:689-696. 24. Collett, P. W., and L. A. Engel. 1986. Influence of lung volume on oxygen cost of resistive breathing. J. AppL Physiol. 61:16-24. 25. Borg, G. A. V. 1982. Psychophysical basis of perceived exertion. Med. Sci. Sports Exerc. 14:377-381. 26. Willeput, R., J. P. Vachaudez, D. Lenders, A. Nys, T. Knoops, and R. Sergysels. 1983. Thoracoabdominal motion during chest physiotherapy in patients affected by chronic obstructive lung disease. Respiration 44:204-214. 27. Weitzenblum, E., B. Moyes, A. Vandevenne, C. Hirth, and G. Methlin. 1980. Regional ventilation and perfusion during diaphragmatic breathing in patients with severe chronic bronchitis. (Abstract). Clin. Respir. Physiol. 16:263. 28. Stubbing, D. G., E. H. Ramsdale, K. J. Killian, and E. J. M. Campbell. 1983. Psychophysics of inspiratory muscle force. J. Appl. Physiol 54: 1216-1221. 29. Mead, J., and S. H. Loring. 1982. Analysis of volume displacement and length changes of the diaphragm during breathing. J. AppL Physiol. 53:750-755. 30. Decramer, M., X. I. Jiang, M. Reid, S. Kelly, P. T. Macklem, and M. Demedts. 1986. Relationship between diaphragm length and abdominal dimensions. J. Appl. Physiol. 61:1815-1820. 31. Ninane, V., F. Rypens, J. C. Yernault, and A. De Troyer. 1992. Abdominal muscle use during breathing in patients with chronic airflow obstruction. Am. Rev. Respir. Dis. 146:16-21. 32. Roussos, C. S., and E. J. M. Campbell. 1986. Respiratory muscle energetics. In A. Fishman, P. T. Macklem, J. Mead, and S. R. Geiger, editors. Handbook of Physiology, The Respiratory System III. American Physiological Society, Washington, DC. 509-527. 33. Sampson, M. G., and A. De Troyer. 1982. Role of intercostal muscles in the rib cage distortions produced by inspiratory loads. J. AppL Physiol. 52:517-523. 34. De Troyer, A., and M. Estenne. 1984. Coordination between rib cage muscles and diaphragm during quiet breathing in humans. J. Appl. Physiol. 57:899-906. 35. De Troyer, A., and M. G. Sampson. 1982. Activation of the parasternal intercostals during breathing efforts in human subjects. J. AppL Physiol. 52:524-529. 36. Similowski, T., S. Yan, A. P. Gauthier, P. T. Macklem, and F. Bellemare. 1991. Contractile properties of the human diaphragm during chronic hyperinflation. N. EngL J. Med. 325:917-923. 37. Decramer, M., L. M. Lacquet, R. Fagard, and P. Rogiers. 1994. Corticosteroids contribute to muscle weakness in chronic airflow obstruction. Am. J. Respir. Crit. Care Med. 150:11-16. 38. Decramer, M., H. Gosselink, M. Verschueren, K. Demuynck, and G. Evers. 1994. Medical consumption is related to muscle weakness in COPD patients. (Abstract). Am. J. Respir. Crit. Care Med. 149:A140. 39. Rochester, D. F. 1991. The diaphragm in COPD: better than expected, but not good enough. N. EngL J. Med. 325:961-962. 40. Decramer, M., T. X. Jiang, and M. Demedts. 1987. Effects of acute hyperinflation on inspiratory muscle use. J. AppL Physiol. 63:1493-1498. 41. Zocchi, L., J. W. Fitting, U. Majani, C. Fracchia, C. Rampulla, and A. Grassino. 1993. Effect of pressure and timing of contraction on human rib cage muscle fatigue. Am. Rev. Respir. Dis. 147:857-864.