NEUROMUSCULAR DISEASE can disproportionately affect

Transcription

1 123 in Lung Volumes in the Assessment of Diaphragmatic Weakness in Neuromuscular Disorders Claudine Fromageot, MD, Frédéric Lofaso, MD, PhD, Djillali Annane, MD, PhD, Line Falaize, Michèle Lejaille, Bernard Clair, MD, Philippe Gajdos, MD, Jean Claude Raphaël, MD ABSTRACT. Fromageot C, Lofaso F, Annane D, Falaize L, Lejaille M, Clair B, Gajdos P, Raphaël JC. Supine fall in lung volumes in the assessment of diaphragmatic weakness in neuromuscular disorders. Arch Phys Med Rehabil 2001;82: Objective: To determine whether diaphragmatic function can be determined by noninvasive respiratory indices in neuromuscular Design: Vital capacity (VC) and mouth pressure generated during a maximal static inspiratory effort (Pi max) were measured with patients in both sitting and supine positions. Setting: Rehabilitation hospital. Patients: Twenty-four patients with generalized neuromuscular Main Outcome Measures: Changes in indices from sitting to supine position were compared with invasive diaphragmatic function indices consisting of transdiaphragmatic pressures during maximal sniff (Pdi sniff) and the ratio of gastric pressure (Pga) increases over transdiaphragmatic pressure ( Pga/ Pdi) during quiet breathing. Results: The fall in VC in the supine position was greater in the 15 patients who had spontaneous paradoxical diaphragmatic motion ( Pga/ Pdi 0) than in the 9 patients who did not. Specificity and sensitivity of a greater than 25% supine fall in VC for the diagnosis of diaphragmatic weakness ( Pga/ Pdi 0 and/or Pdi sniff 30cmH 2 O) were 90% and 79%, respectively. Stepwise multiple regression analysis of Pdi sniff showed that both the supine fall in VC and Pi max were associated with diaphragmatic weakness (R 2.66; p.0001). These factors contributed 52% and 14% of the Pdi sniff variance, respectively. Conclusions: Simple VC measurement in the sitting and supine positions may be helpful in detecting severe or predominant diaphragmatic weakness. Key Words: Diaphragm; Hypercapnia; Forced expiratory volume; Functional residual capacity; Neuromuscular diseases; Rehabilitation; Vital capacity by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation From the Service de Physiologie Explorations Fonctionnelles (Fromageot, Lofaso, Falaize, Lejaille); Service de Réanimation Médicale (Annane, Clair, Gajdos, Raphaël), Hôpital Raymond Poincaré, Garches; and Institut National de la Santé etdela Recherche Médicale INSERM U 492, Hôpital Henri Mondor (Lofaso), Créteil, France. Accepted in revised form March 28, No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organization with which the authors are associated. Reprint requests to Dr. Frédéric Lofaso, Service de Physiologie-Explorations Fonctionnelles, Hôpital Raymond Poincaré, Garches, France, f.lofaso@rpc.ap-hop-paris.fr /01/ $35.00/0 doi: /apmr NEUROMUSCULAR DISEASE can disproportionately affect diaphragmatic function by predisposing patients to hypercapnic respiratory failure. 1-3 Accurate measures of diaphragmatic weakness include the transdiaphragmatic pressure (Pdi) during maximal inspiratory effort, and the ratio of gastric pressure increase over transdiaphragmatic pressure ( Pga/ Pdi) in the sitting position during quiet tidal breathing. 4-9 However, measuring the esophageal and gastric pressures simultaneously requires the insertion of a gastroesophageal catheter, which is not always clinically practical. Clearly, there is a need for a simple test that can detect severe or predominant diaphragmatic weakness. It is known that an increase in restrictive impairment when a person moves from an erect to a supine position indicates diaphragmatic weakness. 4,5 However, these studies by Newsom Davis et al 4 and Mier-Jedrzejowicz et al 5 were performed with patients who had isolated diaphragmatic weakness but no dysfunction of the other respiratory muscles. The researchers were, therefore, unable to determine whether a supine decrease in lung volumes was also a characteristic of predominant diaphragm weakness in patients with generalized neuromuscular Our study was designed to assess whether a decrease in lung volume resulting from a change from erect to supine position contributes to the respiratory weakness associated with neuromuscular METHODS Between 1990 and 1998, 24 patients with generalized neuromuscular disease from various causes, but who did not have chronic obstructive pulmonary disease, were referred to our laboratory for pulmonary function testing and transdiaphragmatic pressure measurement. All patients had experienced a recent increase in breathlessness and/or a clinical suspicion of diaphragmatic dysfunction and/or unexplained hypercapnia. The tests were performed after a 2-week period of stability for all subjects. Blood gas measurements were performed between 8:00 and 8:30 AM by using an arterial sample obtained with the patient resting in the sitting position and breathing room air. a Static lung volumes, dynamic lung volumes, forced ventilatory flows, and maximal inspiratory static mouth pressures (Pi max) were determined between 8:30 and 11:00 AM in triplicate in both the sitting and supine positions. During these measurements, the patients wore a noseclip and breathed through a flange-type mouthpiece. Slow inspiratory vital capacity (VC) and forced expiratory volume in 1 second (FEV 1 ) were obtained by using a spirometer. b Functional residual capacity (FRC) was determined by using the helium dilution technique. b All spirometry and lung volume techniques used met international standards. 10 Measured values were expressed as percentages of predicted values. 10 As in the study by Black and Hyatt, 11 Pi max was measured from the residual volume by using a differential pressure transducer c connected through the mouthpiece. A small leak in the mouthpiece prevented artifactual changes in pressure caused by use of the buccal muscles.

2 124 DIAPHRAGMATIC FUNCTION AND SUPINE FALL IN LUNG VOLUMES, Fromageot For each parameter and each position, the highest value of 3 technically satisfactory measurements was used. Diaphragmatic function was assessed on another day of the same week, again between 8:30 and 11:00 AM, in the sitting position. Data were recorded as follows. First, esophageal and gastric pressures (Peso, Pga) were measured by using a double balloon catheter c connected to 2 differential pressure transducers. b The esophageal balloon was positioned in the middle third of the esophagus. It contained 1mL of air and was used to measure pleural pressure. The gastric balloon contained 1mL of air and was used to measure abdominal pressure. Pga and Peso tracings differ in normal subjects, but can be similar in subjects with neuromuscular diseases. Consequently, 3 tests were performed to check that the catheters were correctly positioned: (1) an occlusion test to assess the validity of the Peso measurement 12 ; (2) a search for positive deflections on the Pga tracings recorded while 1 of the operators applied gentle pressure to the patient s stomach; and (3) a test to check that swallowing water resulted in a sharp rise in Peso caused by the muscular contraction of the esophagus, without any concomitant modifications in Pga. Pdi was measured by connecting the 2 catheters to a differential pressure transducer. Peso, Pga, and Pdi were recorded during quiet tidal breathing. The relative contribution of active motion of the diaphragm to quiet tidal breathing was assessed based on the index of Gilbert et al, 6 which is the Pga/ Pdi ratio where Pga is the difference between peak inspiratory Pga and end expiratory Pga and Pdi is the Pdi change during inspiration. This index is normally positive, but becomes negative if abdominal pressure decreases during inspiration, suggesting paradoxical diaphragmatic motion. 6-9 The mean value of Pga/ Pdi determined over 2 minutes of quiet breathing was used in the analysis. During this quiet breathing, the lack of an abdominal expiratory activity is confirmed by no increase in Pga if an expiratory decrease in abdominal motion is clinically observed. 13 Pdi was also measured during (1) 4 maximal inspiratory efforts with occlusion of the mouthpiece at the FRC (Pdi max) and (2) 10 single maximal sniffs (short sharp sniffs, as hard as possible, such that peak Pdi was not sustained) at the FRC, with at least 5 quiet breaths between 1 sniff and the next (Pdi sniff). For each condition of maximal effort, the highest recorded Pdi value was used for the analysis. STATISTICS Data were expressed as mean standard deviation. Comparison of groups with and without paradoxical diaphragmatic motion ( Pga/ Pdi 0) was performed by using the nonparametric Mann-Whitney test. In addition, we looked for correlations between noninvasive and invasive variables, by using least-square linear regression techniques. Univariate analysis was used to evaluate the independent contribution of each variable. A full model, stepwise, multiple, linear regression analysis was then performed to determine the influence of each variable. The level of significance was set at 5%. RESULTS A total of 18 men and 6 women were studied. Individual data are summarized in table 1. Paradoxical diaphragmatic motion ( Pga/ Pdi 0) was noted in 14 patients, of whom 7 were on nocturnal mechanical ventilation, versus none of the 10 patients without paradoxical diaphragmatic motion. No variation in expiratory abdominal activity, as judged by analysis of the shape of the Pga curve, was found in any of the patients. Age and height were similar in the 2 groups (table 1). Pdi sniff and Pdi max were abnormal in all patients. Mean Pdi max and Pdi sniff values were significantly lower in the group with paradoxical diaphragmatic motion than without (table 1). A noteworthy finding was that Pdi sniff was greater than 30cmH 2 Oin all 10 patients without paradoxical diaphragmatic motion and lower than 30cmH 2 O in the 14 patients with this abnormality (table 1). Total lung capacity (TLC), VC, and FEV 1 were higher in the group without paradoxical diaphragmatic motion (table 1). The supine falls in TLC, VC, and FEV 1 were significantly larger in the group with paradoxical diaphragmatic motion (table 1). No relationships were shown between daytime blood gas variables and other respiratory variables. The results of the univariate regression analysis of invasive and noninvasive variables are shown in table 2 and, as an example, the relation between Pdi sniff and supine fall in VC is shown in figure 1. Invasive variables were more closely correlated with the variables assessing the supine fall in lung volumes than with lung volumes measured in the sitting position. In contrast, invasive variables showed no correlation with the supine fall in Pi max but were correlated with Pi max (table 2). Noninvasive variables were more closely correlated with Pdi max and Pdi sniff than with Pga/ Pdi (table 2). In the stepwise multiple regression analysis (table 3), the closest correlations were between Pdi sniff and the noninvasive variables. The R 2 value in the stepwise multiple regression analysis was.66. Supine fall in VC contributed 52% of the Pdi sniff variance, whereas Pi max contributed 14%. None of the other variables contributed to Pdi sniff variance. DISCUSSION Almost all neuromuscular disorders causing limb weakness can affect the respiratory muscles, including the diaphragm. In some patients, the diaphragm is selectively or disproportionately affected. 1 Assessment of diaphragmatic function is important because several studies have found that the prognosis in neuromuscular disease was worse when diaphragmatic function was disproportionately impaired than when it was relatively spared. 2,3 There is a need for a simple test that can identify patients in whom invasive diaphragmatic function assessment is indicated. Simple observation of rib cage and abdominal motion is noninvasive and can provide information on use of the diaphragm compared with use of the intercostal muscles. However, the relation between diaphragm and intercostal muscle use is relatively loose, and observation is not a quantitative method. Lung volumes and maximal static respiratory pressure in the sitting position are the noninvasive parameters most widely used to assess global respiratory muscle function. However, they cannot detect disproportionate dysfunction of the diaphragm. For example, Laroche et al 14 showed that some patients with moderate diaphragmatic dysfunction had normal global respiratory muscle strength, presumably as a result of compensation by other respiratory muscles. Diaphragmatic weakness can result in elevation of the leaves of the diaphragm on the chest radiograph, though this is inconsistent. 14 Displacement of the diaphragm toward the chest cavity during inspiration is commonly observed on fluoroscopy 1 or ultrasonography, 15 but does not allow evaluation of the degree of diaphragmatic weakness. In addition, in patients who compensate partially for diaphragmatic weakness by contracting their abdominal muscles during expiration, relaxation of those muscles at the onset of inspiration can induce passive diaphragmatic displacement toward the abdominal cavity during inspiration. 1

3 DIAPHRAGMATIC FUNCTION AND SUPINE FALL IN LUNG VOLUMES, Fromageot 125 Table 1: Clinical Characteristics and Diaphragmatic and Pulmonary Function Studies in 24 Patients Patient Diagnosis Age (yr) Sex Height (cm) NMV Pga/ Pdi (%) Pdi max (cmh 2 O) Pdi sniff (cmh 2 O) Pi max (cmh 2 O) Pi max (%) TLC (% pred) TLC (% of seated) VC (% pred) VC (% of seated) FEV 1 (% pred) FEV 1 (% of seated) Arterial ph (UI) PaCO 2 (kpa) PaO 2 (kpa) Group with paradoxical diaphragmatic motion 1. Acid maltase deficiency 40 M 169 Yes Acid maltase deficiency 56 M 170 Yes Limb-girdle muscular 31 F 166 Yes Myotonic 49 M 180 Yes Acid maltase deficiency 54 F 168 Yes Myotonic 44 M 180 No Limb-girdle muscular 45 M 161 No Polyradiculoneuritis 53 M 170 No Myasthenia gravis thymoma 48 M 168 No Facioscapulohumeral muscular 48 F 168 No Myasthenia gravis 58 F 154 No Duchenne s muscular 40 M 173 No Becker s muscular 22 M 175 Yes Idiopathic muscular 64 M 179 Yes Mean / SD Group without paradoxical diaphragmatic motion 15. Myotonic 42 F 179 No Idiopathic muscular 58 M 170 No Polyradiculoneuritis 35 M 182 No Limb-girdle muscular 57 M 171 No Poliomyelitis 64 M 142 No Myasthenia gravis thymoma 71 F 156 No Myotonic 56 M 165 No Myotonic 42 M 169 No Myotonic 56 M 176 No Myotonic 42 M 175 No Mean /10 21* 48* 44* 55* 5 78* 8* 71* 3* 69* 8* SD Abbreviation: NMV, nocturnal mechanical ventilation. * Significant difference from value in the group with paradoxical diaphragmatic motion.

4 126 DIAPHRAGMATIC FUNCTION AND SUPINE FALL IN LUNG VOLUMES, Fromageot Table 2: Univariate Regression Analysis of Diaphragmatic Variables on Noninvasive Variables Pga/ Pdi Pdi max Pdi sniff Coefficient R 2 p Coefficient R 2 p Coefficient R 2 p TLC (% pred) VC (% pred) FEV 1 (% pred) Pi max (cmh 2 O) Supine fall in TLC (% of seated) Supine fall in VC (% of seated) Supine fall in FEV 1 (% of seated) Supine fall in Pi max (% of seated) NOTE. Regression analyses were performed in 24 patients. It is well known that the supine posture is associated with a further increase in restrictive ventilatory impairment in patients with diaphragmatic weakness 4,5 because of the effect of gravity on intra-abdominal organs. In the upright position, the hydrostatic pressure exerted by the abdominal contents tends to displace the diaphragm passively toward the abdominal cavity at the end of each expiration, thus, opposing the tendency of the rib cage inspiratory muscles to suck the diaphragm up into the thorax. In the supine position, in contrast, the hydrostatic forces displace the paralyzed diaphragm cranially, thus, reinforcing the sucking effect of the rib cage inspiratory muscles on the diaphragm. The result is that pulmonary volumes decrease in the supine position. Newsom Davis et al 4 studied 5 patients with total diaphragmatic paralysis and 3 with Pdi max values ranging from 2 to 6cmH 2 O; VC in the upright position was 65% to 30% of predicted value and fell by about half on changing to the supine position. In a population with various degrees of diaphragmatic weakness studied by Mier-Jedrzejowicz et al, 5 multiple regression analysis showed that Pdi sniff correlated significantly with both supine fall in VC and Pi max (R 2.64, p.0001). However, 73% of the patients in their study had isolated diaphragmatic dysfunction and only 8 patients had generalized neuromuscular disorders. We restricted our study to patients with generalized neuromuscular disorders, in whom respiratory dysfunction can result not only from diaphragmatic dysfunction but also from weakness of the other rib cage and abdominal muscles. Similar to Mier-Jedrzejowicz, 5 we found that Pdi sniff correlated significantly with both supine fall in VC and Pi max (R 2.66, p.0001). Thus, our study extends previous findings 4,5 to a population with generalized neuromuscular disorders. Limitations and Implications of the Study It can be considered that lack of homogeneity of the patients we tested is a major limitation. However, it is interesting to show that supine fall in VC remains a characteristic of diaphragmatic dysfunction in this heterogeneous population that suffered from generalized neuromuscular disease caused by various causes. Diaphragmatic weakness was found in 58% of our patients, indicating that our study group was not representative of the overall population of neuromuscular disease patients at our institution. This reflects the recruitment bias inherent in the fact that our study patients were referred to our laboratory by intensive care physicians. This bias may explain the absence of correlations between PaCO 2 and other respiratory variables because patients with unexplained hypercapnia were those most likely to be referred to us for full respiratory muscle assessment. Another explanation may be the well-known beneficial effect on PaCO 2 of nocturnal mechanical ventilation, 16 which was used in 7 of our 24 patients (table 1). However, no correlations between PaCO 2 and other respiratory variables emerged after exclusion of these 7 patients. We arbitrarily divided our patients into 2 groups, based on the value of the Pga/ Pdi ratio during quiet tidal breathing. Table 3: Stepwise Multiple Regression Analysis of Pdi sniff on Noninvasive Variables Fig. 1 Relation between Pdi sniff and supine fall in VC. Open symbols indicate patients with paradoxical diaphragmatic motion and closed symbols indicate patients without paradoxical diaphragmatic motion. Pdi sniff Coefficient R 2 p Supine fall in VC (% of seated) Pi max (cmh 2 O) NOTE. This stepwise multiple regression analysis tested the variables that correlated significantly with Pdi sniff in the univariate regression analysis of data from 24 patients.

5 DIAPHRAGMATIC FUNCTION AND SUPINE FALL IN LUNG VOLUMES, Fromageot 127 However, this parameter correlated less closely with spirometric data than Pdi max and Pdi sniff, perhaps because Pga/ Pdi is a reliable index only of paradoxical diaphragm motion and is not an accurate quantitative method for assessing diaphragm strength. Nevertheless, the significance of this index as an index of diaphragmatic dysfunction has been shown in healthy subjects, 6 in patients who have had abdominal surgery, 7,8 as well as in patients with phrenic nerve injury. 9 In addition, a significant positive linear correlation between this ratio and volume displacement of the abdomen within the tidal volume range has been established. 6 Because this ratio can be affected by active contractions of the abdominal muscles, we have systematically checked the absence of an abdominal expiratory activity as previously described. 13 Thus, a negative value of Pga/ Pdi (or Gilbert s index) 6 clearly indicates diaphragmatic dysfunction. In addition, this index has the advantage of being independent from patient cooperation. Pi max, Pdi max, and Pdi sniff are also reduced in patients with diaphragmatic dysfunction. However, all these parameters are dependent on the patient s ability and motivation to cooperate, a major limitation in the setting of neuromuscular disorder evaluation. Another noninvasive test is sniff nasal inspiratory pressure, which provides a reasonable estimate of inspiratory muscle strength. 17 This method uses the sniff maneuver that many patients find easier than static efforts to perform. Therefore, this measure provides probably a more reliable and reproducible method than Pi max; unfortunately, it was not described at the beginning of our study and consequently this test has not been systematically performed. It is possible to record Pdi after electric or magnetic stimulation of the phrenic nerves, 18 a test that has the advantage of being nonvolitional. However, whether this approach is more sensitive than Pdi max and Pdi sniff for detecting moderate diaphragmatic weakness has not been established. 18 Finally, in our stepwise multiple regression analysis, the closest correlations were between Pdi sniff and the noninvasive variables. This was to be expected because Pdi sniff is the most reliable volitional method for assessing diaphragmatic strength. 19 Our analysis showed clearly that Pi max was an independent factor of Pdi sniff, but that its influence was smaller than that of the supine fall in VC. These data are indirect evidence that the diaphragm is not affected in the same way as other respiratory muscle groups in patients with neuromuscular disorders. We observed a supine fall in VC in all but 3 patients. Those patients had no paradoxical diaphragmatic motion and had Pdi max and Pdi sniff values greater than 35cmH 2 O. However, a supine fall in VC can occur as a physiologic phenomenon from both the changes in respiratory mechanics related to gravitational unloading of the abdomen and to the increase in intrathoracic blood volume that is also related to the effects of gravity. Reductions of 8% in VC on changing from the standing to the supine position have been found in both unimpaired subjects and subjects with restrictive lung disease; the upper limits of the 95% confidence intervals in these 2 groups were 19% and 24%, respectively. 20 Allen et al 20 concluded that a 25% or greater supine fall in VC in a patient with restrictive lung disease should lead to further investigations of diaphragmatic function. Interestingly, in our study the supine fall in VC was greater than 25% in 11 of the 14 patients with paradoxical diaphragmatic motion (or Pdi sniff 30cmH 2 O), but was less than 25% in 9 of the 10 patients without paradoxical diaphragmatic motion (or Pdi sniff 30cmH 2 O) (fig 1, table 1). Thus, in our study, the specificity, sensitivity, positive predictive value, and negative predictive value of a supine fall in VC greater than 25% for the diagnosis of diaphragmatic weakness were 90%, 79%, 92%, and 75%, respectively. Therefore, it is clear that the diagnostic value of a supine fall in VC greater than 25% is less than that of abnormal Pdi measurements. However, there is agreement that Pdi measurement is required in only a small number of patients with neuromuscular disorders. 18 Because serial evaluations are required to monitor patients with progressive neuromuscular disorders, it should rest on easy-to-obtain parameters such as VC and Pi max. 18 Our data underline the importance of measuring VC in both the sitting and the supine positions because VC can decrease in the supine position in patients with neuromuscular disorders. Yet, VC measurement in these patients and in patients with acute respiratory failure is generally performed in the supine position and, therefore, cannot be compared with VC values obtained from the sitting position. Another important issue is the definition of criteria for long-term mechanical ventilation. The consensus conference of the American College of Chest Physicians 21 recommended that long-term mechanical ventilation be considered in patients with chronic hypercapnia (PaCO 2 8kPa) during the day, particularly if the cause is a neuromuscular disorder. In our study, of the 7 patients with paradoxical diaphragmatic motion who were not on mechanical ventilation, 6 did not meet the consensus conference criteria (table 1). Interestingly, during the next 2 years, 4 of the 6 patients developed acute respiratory failure and became dependent on long-term mechanical ventilation. During our study, 3 of the 4 patients were found to have a greater than 25% supine fall in VC (table 1). CONCLUSION Our data indicate clearly a need to perform lung volume measurements in both the sitting and the supine positions when respiratory function impairment caused by neuromuscular disorders is being evaluated. Patients with neuromuscular disorders are supine much of the time and detecting a supine fall in VC will help diagnose predominant diaphragmatic weakness and, therefore, predict the occurrence of hypercapnic respiratory failure. References 1. Gibson G. Diaphragmatic paresis: pathophysiology, clinical features and investigation. Thorax 1989;44: Newsom Davis J. The respiratory system in muscular. Br Med Bull 1980;36: Parhad I, Clark A, Barron K, Stauton S. Diaphragmatic paralysis in motor neuron Neurology 1978;28: Newsom Davis J, Goldman M, Loh L, Casson M. Diaphragm function and alveolar hypoventilation. Q J Med 1976;45: Mier-Jedrzejowicz A, Brophy C, Moxham J, Malcolm G. Assessment of diaphragm weakness. Am Rev Respir Dis 1988;137: Gilbert R, Auchincloss J, Peppi D. Relationship of rib cage and abdomen motion to diaphragm function during quiet breathing. Chest 1981;80: Simonneau G, Vivien A, Sartene R, Kunstlinger F, Samiio K, Noviant Y, et al. Diaphragm dysfunction induced by upper abdominal surgery. Am Rev Respir Dis 1983;128: Dureuil B, Viirès N, Cantineau J-P, Aubier M, Desmonts J-M. Diaphragmatic contractility after upper abdominal surgery. J Appl Physiol 1986;61: Diehl J, Lofaso F, Deleuze P, Similowski T, Lemaire F, Brochard L. Clinical relevant diaphragmatic dysfunction after open-heart surgery. J Thorac Cardiovasc Surg 1994;107: Quanjer P, Tammeling G, Cotes J, Pedersen O, Peslin R, Yernault J. Lung volumes and forced ventilatory flows. Report working

6 128 DIAPHRAGMATIC FUNCTION AND SUPINE FALL IN LUNG VOLUMES, Fromageot party. Standardization of lung function tests. European community for steel and coal. Eur Respir J 1993;6 Suppl: Black L, Hyatt R. Maximal respiratory pressures: normal values and relationship to age and sex. Am Rev Respir Dis 1969;138: Baydur A, Behrakis PK, Zin WA, Jaeger M, Milic-Emili J. A simple method for assessing the validity of the esophageal balloon technique. Am Rev Respir Dis 1982;126: Lessard M, Lofaso F, Brochard L. Expiratory muscle activity increases intrinsic positive end-expiratory pressure independently of dynamic hyperinflation in mechanically ventilated patients. Am J Respir Crit Care Med 1995;151: Laroche C, Mier A, Moxham J, Green M. The value of sniff esophageal pressures in the assessment of global inspiratory muscle strength. Ann Rev Respir Dis 1988;138: Houston J, Angus R, Cowan M, McMillan N, Thomson N. Ultrasound assessment of normal hemidiaphragmatic movement: relation to inspiratory volume. Thorax 1994;49: Annane D, Quera-Salva M, Lofaso F, Vercken J, Lesieur O, Fromageot C, et al. Mechanisms underlying effects of nocturnal ventilation on daytime blood gases in neuromuscular diseases. Eur Respir J 1999;13: Héritier F, Rahm F, Pasche P, Fitting J. Sniff nasal inspiratory pressure: a noninvasive assessment of inspiratory muscle strength. Am J Respir Crit Care Med 1994;150: Polkey M, Green M, Moxham J. Measurement of respiratory muscle strength. Thorax 1995;50: Miller J, Moxham J, Green M. The maximal sniff in the assessment of diaphragm function in man. Clin Sci 1985;69: Allen S, Hunt B, Green M. Fall in vital capacity with posture. Br J Dis Chest 1985;79: Make B, Hill N, Goldberg A, Bach J, Criner G, Dunne P, et al. Mechanical ventilation beyond the intensive care unit: report of a consensus conference of the American College of Chest Physicians. Chest 1998;113 Suppl:289S-344S. Suppliers a. ABL 330; Radiometer Medical A/S, DK 2700 Brønshøj, Denmark. b. Model DP45 Low Pressure Transducer; Validyne Engineering Corp, 8626 Wilbur Ave, Northridge, CA c. Marquat Gbm, BP11, Boissy-St. Leger Cedex, France.