Accuracy of noninvasive estimates of respiratory muscle effort during spontaneous breathing in restrictive diseases

J Appl Physiol 95: 1542 1549, 2003. First published June 13, 2003; 10.1152/japplphysiol.01010.2002. Accuracy of noninvasive estimates of respiratory muscle effort during spontaneous breathing in restrictive diseases Francisco García-Río, José M. Pino, Angeles Ruiz, Salvador Díaz, Concepción Prados, and José Villamor Servicio de Neumología, Hospital Universitario La Paz, 28034 Madrid, Spain Submitted 4 November 2002; accepted in final form 6 June 2003 García-Río, Francisco, José M. Pino, Angeles Ruiz, Salvador Díaz, Concepción Prados, and José Villamor. Accuracy of noninvasive estimates of respiratory muscle effort during spontaneous breathing in restrictive diseases. J Appl Physiol 95: 1542 1549, 2003. First published June 13, 2003; 10.1152/japplphysiol.01010.2002. Mean inspiratory pressure (PI), estimated from the occlusion pressure at the mouth and the inspiratory time, is useful as a noninvasive estimate of respiratory muscle effort during spontaneous breathing in normal subjects and patients with chronic obstructive pulmonary disease. The aim of this study was to compare the PI with respect to mean esophageal pressure (Pes) in patients with restrictive disorders. Eleven healthy volunteers, 12 patients with chest wall disease, 14 patients with usual interstitial pneumonia, and 17 patients with neuromuscular diseases were studied. PI, Pes, and mean transdiaphragmatic pressure were simultaneously measured. Tension-time indexes of diaphragm (TT di) and inspiratory muscles (TT mu) were also determined. In neuromuscular patients, significant correlations were found between PI and Pes, PI and transdiaphragmatic pressure, and TT mu and TT di. A moderate agreement between PI and Pes and between TT mu and TT di was found. No significant correlation between these parameters was found in the other patient groups. These findings suggest that PI is a good surrogate for the invasive measurement of respiratory muscle effort during spontaneous breathing in neuromuscular patients. respiratory muscles; neuromuscular diseases; chest wall diseases; usual interstitial pneumonia RESTRICTIVE LUNG DISORDERS increase the load on the inspiratory pump while at the same time decreasing the pressure-generating capacity of the respiratory muscles (25). Although inspiratory muscle fatigue has been extensively studied in acute respiratory failure and in patients with chronic obstructive pulmonary disease (COPD) (8, 18, 30, 36), little is known about the role of the load imposed on the inspiratory muscles in the pathogenesis of hypercapnic ventilatory failure, exercise intolerance, and dyspnea experienced by patients with restrictive disorders (2, 31). Mean esophageal pressure (Pes) is the most frequently used parameter for estimating the inspiratory muscle effort during spontaneous breathing (6). However, measurement of Pes is an invasive procedure that Address for reprint requests and other correspondence: F. García- Río, Alfredo Marqueríe 11, izqda, 1 A, 28034 Madrid, Spain (E-mail: fgr01m@jazzfree.com). requires the placing of an esophageal balloon, making it difficult to use for the routine assessment of respiratory muscles (7, 20). Mean inspiratory pressure (PI) is estimated from the mouth occlusion pressure [at 0.1 s after beginning of inspiration (P 0.1 )] and the inspiratory time (TI) with the assumption that the actual intrathoracic pressure increases linearly over the entire inspiratory duty cycle. This method has been proposed as a noninvasive measurement of respiratory muscle effort during spontaneous breathing (16, 29). This parameter is a good surrogate for Pes in normal subjects and COPD patients (28). Moreover, in these patients, the respiratory muscle effort relative to strength [PI/maximum PI (PI max )] and the tension-time index of inspiratory muscles (TT mu ) are related to the diaphragmatic effort and the tension-time index of diaphragm (TT di ), respectively (28). Nevertheless, the relationships PI-Pes and TT mu -TT di may change, depending on the different recruitment of inspiratory muscles. These results must not be extrapolated to disease conditions. In this study, a noninvasive method to measure PI force is compared with pressure measurements via esophageal balloons in patients with restrictive disorders. MATERIALS AND METHODS Study subjects. Forty-three patients with restrictive lung disease, referred to the lung function laboratory for routine assessment of respiratory muscle function, were studied in their stable clinical state. The general selection criteria were the ratio of forced expiratory volume in 1 s to forced vital capacity ratio 80% and total lung capacity 80% of predicted. The causes of restrictive disorders were neuromuscular disease (NMD), chest wall disease (CWD), or usual interstitial pneumonia (UIP). Twelve patients were diagnosed with CWD, due to thoracoplasty in four subjects and to idiopathic kyphoscoliosis, with a Cobb s angle of kyphosis 25, in eight patients. Fourteen patients were pathologically identified as UIP, the most common idiopathic interstitial pneumonia, which is characterized by the presence of patchy, nonuniform, and alternating zones of interstitial fibrosis, inflammation, and honeycomb changes, in accordance with the definitions of Katzenstein (21). Among 17 patients diagnosed with NMD, six had myasthenia gravis, five amyotrophic lateral sclerosis, The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1542 8750-7587/03 $5.00 Copyright 2003 the American Physiological Society http://www.jap.org

NONINVASIVE RESPIRATORY MUSCLE EFFORT 1543 two Duchenne muscular dystrophy, two congenital myopathy, one brachial plexus paralysis, and one myotonic dystrophy. In all patients, the diagnosis of NMD was based on clinical and laboratory examinations, including nerve conduction studies, electromyograms, and muscle biopsies. In addition, these patients with restrictive lung disease were compared with a control group of 11 healthy subjects. Control subjects were judged healthy based on their clinical history, findings from physical examination, and the results of ECG, basal spirometry, and chest radiography. None of the study subjects had a history of asthma or chronic bronchitis, and none was accustomed to performing the maneuvers. All patients gave informed consent for the respiratory muscle assessment. Methods. Testing was begun at the same time in the morning in similar conditions for all subjects. They were asked not to eat for 4 h before the exploration and to refrain from drinking coffee, tea, and alcohol for at least 12 h, and from using tobacco for at least 2 h before each study. The subjects were studied in the sitting position without prior training. An arterial blood sample was obtained while patients breathed room air for the measurement of blood gases (ABL520, Radiometer, Copenhagen, Denmark). Spirometry was performed by means of a pneumotachograph, and static lung volumes were measured with the multibreath helium dilution technique (MasterLab Body 4.0, Erich Jaeger, Würzburg, Germany), according to European Respiratory Society standardization (27). The normal values for lung volumes were those proposed by the European Community for Coal and Steel (12). PI max was measured by using a standard flanged mouthpiece, which could be manually occluded at the distal end, leaving a small leak to prevent glottic closure during inspiratory and expiratory maneuvers. The mouthpiece was connected to a pressure transducer (M-163, Sibelmed, Barcelona, Spain) by a polyethylene tube 2 mm in internal diameter. A 1-mm leak was incorporated to prevent participation of orofacial muscles. The transducer was calibrated before each study by using a U-tube water manometer. PI max was measured from functional residual capacity (FRC), with a nose clip. Maximum inspiratory efforts were encouraged verbally with simultaneous visual feedback from a monitor. Maneuvers were separated by 1 min and continued until no further increase in pressure could be obtained. At least five repeated determinations were made until three technically satisfactory and reproducible measurements were obtained (variations in PI max 10%). The highest recorded pressures maintained for 1 s were used for analysis (15). After 20 min of rest, patients were asked to breathe normally through a low-resistance breathing valve (0.1 kpa l 1 s, dead space 75 ml). Inspiratory flow was measured with a pneumotachograph (model 276, Jaeger) connected to a flow transducer (model DWD, Jaeger), and volume was obtained from mathematical integration of the flow signal. Pes and gastric pressures were measured with latex balloons that were 10 cm long and 3.5 cm in circumference attached to 100-cm-long fine polythene catheters (model 720199, Jaeger). One balloon was positioned in the midesophagus, and the other in the stomach. They were filled with 0.5 and 2 ml of air, respectively. The esophageal balloon was positioned in such a way that the difference in transpulmonary pressures was 0.2 kpa during inspiratory maneuvers against an occluded airway (4). A gastric balloon was advanced 65 cm from the nares. The esophageal and gastric balloons were passed through the nose after topical anesthesia with 10% lidocaine spray in the nose and nasopharynx. The balloon catheters were connected to differential pressure transducers (model DWD, Jaeger, 20 kpa) and amplified by carrier amplifiers (model PT, Jaeger). These were calibrated before each study with a water manometer. Pressure and volume signals were sampled at 100 Hz and analyzed via an analog-to-digital converter (Screenmate box, Jaeger) connected to a personal computer running LabView software (National Instruments, Austin, TX) and sampling at 100 Hz. Transdiaphragmatic pressure (Pdi) was obtained by digital subtraction of Pes from gastric pressure by using pressure at resting end expiration as the reference point. The transducers were shown to be linear over the range of pressures measured in this study, and the frequency response of the balloon and catheter-transducer system, determined by a pop test, was 14 Hz. Breathing pattern and mouth occlusion pressure were measured as previously described (14). Once the subjects felt comfortable with the mouthpiece and catheters and appeared relaxed, tidal breathing was recorded for a few minutes to ensure a steady state. P 0.1 was measured by the Whitelaw method (35). The inspiratory line was occluded without the subject s knowledge, approximately every 15 s for 0.5sby means of a pneumatic inflatable balloon (series 9327; Hans- Rudolph, Kansas City, MO). The mean of at least five measurements was determined. Before and after occlusions, ventilatory and pressure parameters were recorded. Their final parameters corresponded to the means of the values obtained both before and after P 0.1 measurements. From the breathing pattern record, we obtained tidal volume (VT), TI, and total cycle duration (TT). Expiratory time, duty cycle (TI/TT), mean inspiratory flow (ratio of VT to TI), and minute ventilation were calculated. The values from 10 cycles were averaged for the analysis of results (14). The mean breath Pes and Pdi were obtained by averaging the pressure values measured every 200 ms during the period of inspiratory flow. All Pes and Pdi measurements during quiet breathing were averaged from 1 min. Maximum Pdi (Pdi max) was measured at FRC during a maximal inspiratory effort against an occlusion, combined with a maximal expulsive effort in the abdomen (22). Each maximal effort lasted from 3 to 5 s, with a resting period of at least 1 min between efforts. The results were obtained with five efforts in all subjects. Great attention was paid to ensure that inspiratory efforts began from the end-expiratory VT. This was achieved by close observation of the flow and pressure tracings, as well as of the patient s breathing movements. The maximum value of Pdi max generated was used for further analysis. The TT di was calculated as previously reported by Bellemare and Grassino (6) (TT di Pdi/Pdi max TI/TT). According to Gaultier et al. (16) and assuming that PI increases linearly with time, the PI developed by the respiratory muscles during inspiration was calculated as PI 5 P 0.1 TI. The TT mu was determined by using the equation: TT mu PI/PI max TI/TT. Statistical analysis. The differences between the means of variables in study groups were analyzed by using one-way ANOVA. Post hoc analysis was performed by using the Bonferroni test for multiple comparisons. Correlations between PI and Pes, PI and Pdi, PI/PI max and Pdi/Pdi max, and TT di and TT mu were determined by Spearman correlation analysis. The agreement between TT di and TT mu was assessed by the method of differences against the means, according to Bland and Altman (9). These analyses were performed by using the software Statistical Package for the Social Sciences for Windows release 8.0 software (SPSS, Chicago, IL). In all cases, P

1544 NONINVASIVE RESPIRATORY MUSCLE EFFORT Table 1. General characteristics of the study groups Control Group NMD Group CWD Group UIP Group values of 0.05 were considered to be significant (1). Data are expressed as means SD. RESULTS n 11 17 12 14 Men, % 60 59 58 64 Age, yr 42 18 49 17 49 14 46 6 Weight, kg 74 6 65 10 a 75 6 e 76 7 e Height, cm 169 10 162 8 170 10 170 7 Current smoker, % 30 6 33 36 Pa O2, Torr 83.1 4.1 72.4 11.7 71.0 7.9 73.5 8.3 Pa CO2, Torr 35.9 2.4 41.4 6.6 b 39.7 4.5 34.6 2.7 e FVC liters 5.66 0.96 2.57 1.45 c 3.29 0.76 c 3.38 0.63 c %predicted 112 11 66 33 c 69 7 c 74 7 b FEV 1 liters 4.40 0.77 1.97 1.12 c 2.58 0.70 c 2.88 0.56 b,d %predicted 111 10 66 32 c 77 9 b 82 7 a FEV 1/FVC, % 77.9 4.3 78.4 11.6 78.0 6.4 85.2 6.8 FRC liters 2.64 0.44 1.76 0.23 c 1.66 0.28 c 1.68 0.22 c %predicted 99 8 72 7 c 64 6 c,d 64 7 c,d TLC liters 5.95 1.25 3.27 1.18 c 4.01 0.91 c 4.19 0.75 b %predicted 98 7 59 17 c 66 8 c 68 6 c Values are mean SD; n, no. of subjects. NMD, neuromuscular disease; CWD, chest wall disease; UIP, usual insterstitial pneumonia; Pa O2, arterial oxygen partial pressure; Pa CO2, arterial carbon dioxide partial pressure; FVC, forced expiratory vital capacity; FEV 1, forced expiratory volume in 1 s; FRC, functional residual capacity; TLC, total lung capacity. a P 0.05, b P 0.01, and c P 0.001 compared with control group; d P 0.05 and e P 0.01 compared with NMD group. Anthropometric characteristics and pulmonary function data of the patients and normal subjects are presented in Table 1. The NMD group had a significantly lower weight than the other three groups. No difference was found among the four groups in age, height, and smoking habit. Dynamic and static lung volumes were lower in the three groups of patients than in the control subjects. No difference was found between the patient groups and control subjects for arterial PO 2. However, the NMD group showed significantly higher arterial PCO 2 than control subjects and the UIP group. Breathing pattern, occlusion pressure, and inspiratory muscle measurements are shown in Table 2. PI max was lower in the NMD group than in the control, CWD, and UIP groups. Similarly, patients with NMD or CWD had a lower Pdi max than control subjects and UIP patients. No difference for Pdi was noted among the four groups of the study, although the Pdi/Pdi max was Table 2. Values of respiratory muscle function, breathing pattern, and occlusion pressure of the patients and normal subjects Control Group NMD Group CWD Group UIP Group PI max, kpa 10.60 3.42 5.08 1.73 c 8.55 1.32 e 9.96 2.16 f Pdi, kpa 1.31 0.38 1.41 0.58 1.12 0.36 1.55 0.39 Pes, kpa 0.94 0.28 1.13 0.56 0.81 0.29 1.14 0.36 Pdi max, kpa 10.66 2.66 5.18 2.40 c 6.69 0.76 c 10.32 1.80 f,g VT, liters 0.875 0.318 0.563 0.163 b 0.680 0.103 0.615 0.123 b TI, s 1.524 0.580 1.473 0.497 1.401 0.261 1.585 0.323 TT, s 3.288 1.204 3.149 1.021 3.133 0.523 3.109 0.263 P 0.1, kpa 0.152 0.073 0.256 0.166 a 0.304 0.115 a 0.320 0.102 a TI/TT 0.464 0.053 0.476 0.118 0.449 0.056 0.510 0.097 VT/TI, l/s 0.626 0.302 0.411 0.149 a 0.498 0.094 0.411 0.130 a PI, kpa 1.03 0.36 1.70 0.84 2.09 0.79 b 2.41 0.57 c,d Pdi/Pdi max 0.128 0.043 0.317 0.159 c 0.165 0.041 e 0.156 0.055 f PI/PI max 0.103 0.042 0.358 0.183 c 0.247 0.087 a 0.254 0.076 a TT di 0.058 0.015 0.145 0.061 c 0.074 0.021 f 0.077 0.018 f TT mu 0.046 0.016 0.163 0.071 c 0.114 0.050 a 0.127 0.039 b Values are mean SD. PI max, maximal static inspiratory pressure; Pdi, mean transdiaphragmatic pressure; Pes, mean esophageal pressure; Pdi max, maximal transdiaphragmatic pressure; VT, tidal volume; TI, inspiratory time; TT, total time; P 0.1, mouth occlusion pressure at 100 ms beginning inspiration; TI/TT, duty cycle; VT/TI, mean inspiratory flow; PI, mean inspiratory pressure; TT di, tension-time index of the diaphragm; TT mu, tension-time index of inspiratory muscles. a P 0.05, b P 0.01, and c P 0.001 compared with control group; d P 0.05, e P 0.01, and f P 0.001 compared with NMD group; g P 0.001 compared with CWD group.

NONINVASIVE RESPIRATORY MUSCLE EFFORT 1545 Table 3. Correlation analysis between different pressures in control subjects and patients with restrictive disorders Control Group NMD Group CWD Group UIP Group PI-Pes 0.816 0.681 0.021 0.229 PI-Pdi 0.806* 0.439* 0.004 0.474 PI max-pdi max 0.815 0.738 0.720* 0.833 PI/PI max-pdi/pdi max 0.952 0.895 0.238 0.218 TT mu-tt di 0.915 0.892 0.077 0.081 Values are Spearman s correlation coefficient. * Correlation is significant at the 0.05 level; correlation is significant at the 0.01 level. greater in NMD patients than in control subjects and CWD or UIP patients. NMD and UIP patients were characterized by a more shallow breathing pattern and a lower VT than control subjects. There was no significant difference in respiratory timing or mean inspiratory flow between groups. PI estimated from P 0.1 was higher in the UIP and CWD groups than in NMD and control groups. However, the fraction of PI max developed by the respiratory muscles for breathing at rest (PI/PI max ) was significantly higher in the three groups of patients compared with control subjects. Mean TT di was greater in the NMD group than in control subjects or patients with CWD or UIP. However, TT mu was higher in the three groups of patients than in control subjects. In the control subjects, positive correlations were found between PI and Pes (r 0.816, P 0.04), between PI and Pdi (r 0.806, P 0.005), between PI/PI max and Pdi/Pdi max (r 0.952, P 0.001), and between TT di and TT mu (r 0.915, P 0.001). In the NMD group, PI correlated with Pes (r 0.681, P 0.004) and Pdi (r 0.439, P 0.048), PI/PI max with Pdi/Pdi max (r 0.895, P 0.001), and TT di with TT mu (r 0.892, P 0.000). No significant correlation between these parameters was found in either the CWD or UIP groups (Table 3). Figure 1 shows the plot of the difference between PI and Pes and their mean, with the limits of agreement, in the four study groups. In control subjects, the mean difference between PI and Pes was 0.089 0.156 kpa, and the limits of agreement were 0.224 and 0.401 kpa. In the NMD group, the mean PI and Pes difference was 0.444 0.420 kpa, and the limits of agreement were 0.396 and 1.284 kpa. In the CWD group, the mean PI and Pes difference was 1.285 0.807 kpa, and the limits of agreement were 0.329 and 1.449 kpa. Finally, in the UIP group, the mean PI and Pes difference was 1.268 0.629 kpa, and the limits of agreement were 0.009 and 2.527 kpa. Fig. 1. Difference between mean inspiratory pressure (PI) and mean esophageal pressure (Pes) during spontaneous breathing against the mean of these 2 variables in control subjects (A) and in neuromuscular disease (NMD; B), chest wall disease (C), and usual interstitial pneumonia (D) patients. Solid line, mean difference; dashed lines, 2 SDs around the mean.

1546 NONINVASIVE RESPIRATORY MUSCLE EFFORT In control subjects and the NMD group, mean difference between TT mu and TT di and the limits of agreement were 0.011 0.0058 ( 0.023 and 0.001) and 0.0178 0.0488 ( 0.047 and 0.082), respectively (Fig. 2). TT mu values corresponding to the fatigue threshold were extrapolated from the values of TT di obtained by Bellemare and Grassino (6, 7). Diaphragmatic fatigue threshold value was extrapolated by regression of TT mu against TT di. As shown in Fig. 3, TT mu isopleths of 0.17, 0.20, and 0.24 corresponded to TT di isopleths of 0.12, 0.15, and 0.20, respectively. DISCUSSION The major findings of this study were the following: 1) in healthy subjects and in patients with NMD, there was a significant relationship between PI, as estimated from mouth occlusion pressure, and Pes; 2) PI max was related to Pdi max in normal subjects and patients with restrictive disorders; 3) consequently, we found a high correlation between TT di and TT mu in neuromuscular patients, whereas these indexes were not related in patients with CWD or UIP. Fig. 2. Difference between tension-time index of inspiratory muscles (TT mu) and tension-time index of the diaphragm (TT di) against the mean of these 2 variables in control subjects (A) and patients with NMD (B). Solid line, mean difference; dashed lines, 2 SDs around the mean. Fig. 3. Relationship between duty cycle duration [inspiratory time/ total cycle duration (TI/TT)] and transdiaphragmatic pressure-tomaximum transdiaphragmatic pressure ratio (Pdi/Pdi max) (A) and between TI/TT and inspiratory pressure-to-maximum inspiratory pressure ratio (PI/PI max) (B) in healthy subjects (control) and patients (NMD group). As the translation of neural drive into inspiratory driving pressure is modulated by the lung volume, patients with restrictive disorders, who breathe at low lung volumes, improve the effectiveness of their inspiratory muscles as pressure generators (24). However, this is not the sole explanation for higher P 0.1 in these patients with respect to healthy subjects. A P 0.1 increase, inversely related to the deterioration of muscle strength, has been described in patients with several NMD (3, 5, 14). When muscles fail to generate tension, the muscle spindles or Golgi tendon organs are excited (14). In these cases, the respiratory system detects the muscle weakness and increases the nervous drive. Kyphoscoliotic patients maintain near-normal ventilation through increased central drive in response to lower lung and chest wall compliance (10). In nonhypoxemic and nonhypercapnic patients with diffuse lung fibrosis, an increase in P 0.1 has been described. Although patients with chronic interstitial lung disease have a relatively low neuromuscular coupling attributed to a reduced inspiratory muscle strength

NONINVASIVE RESPIRATORY MUSCLE EFFORT 1547 (17), in these patients, P 0.1 only correlated with lung elastance, indicating that, as lung elastance increases with the progression of interstitial lung disease, there is a compensatory increase in P 0.1 that maintains the inspiratory flow (17, 24). Thus, whereas both vagal mechanisms and mechanoreceptors in the chest wall sensitive to reduced rib cage expansion should contribute to increased respiratory drive in CWD or UIP patients (11), the deterioration of muscle strength is the main cause of the changes in the central ventilatory drive in patients with NMD (3, 14). Our study shows a close relationship between PI and Pes, as well as between PI and Pdi in healthy subjects. The relationship between these parameters is also significant in NMD patients, although the discrete correlation coefficient observed in these patients could be attributed to several factors. PI determination assumes that PI increases linearly over time during the occlusion. Because this is not always the case in humans, this method may lead to an overestimation of PI. In anesthetized humans, the ratio of P 0.1 to the pressure at 1 s of inspiration is constant, even when respiration is stimulated with enough CO 2 to triple the peak pressure (34); therefore, the shape of the wave is assumed to remain constant in normal subjects. However, apart from the shape of the occlusion pressure wave exhibiting substantial variability (32), there are several situations in which a change in the shape of the wave has been demonstrated or suspected. In exercise, healthy subjects and COPD patients seem to change the shape to a nonlinear profile (19, 23). Calculations of pressure waveforms in conscious normal subjects also indicate that there are probably changes in shape when an inspiratory resistive load is applied at rest (13, 34). Moreover, P 0.1 can also be partially produced by the relaxation of the expiratory muscles (34). The difference between inspiratory muscle contraction and expiratory muscle relaxation cannot be inferred from mouth occlusion pressure changes, and this, therefore, represents a potential additional factor influencing the Pdi vs. PI relationship. The main finding of our study is that PI correlates with Pes in NMD, but not in CWD and UIP. The airway pressure tracings on the occlusion pressure efforts did not show less linearity for the CWD and UIP patients than for the normal and NMD patients, although their short duration does not allow us to rule out this phenomenon. CWD and UIP patients show neither more plateauing nor even a decreasing swing on their Pes curves compared with the normal subjects. Thus, as the Pes curves look perfectly normal in CWD and UIP, inaccuracy of PI could be attributed to an excessive elastance, relief of expiratory muscle activity, or some other cause of overestimation of inspiratory efforts by P 0.1. In fact, differences in the PI-Pes data might be due to the lower FRC in both CWD and UIP groups, together with the stiffer chest wall and lungs, respectively. Both circumstances could contribute passively to increases in P 0.1 without any increased muscle activation. Figure 1, C and D, clearly shows that the degree of overestimation of respiratory effort by PI in CWD and UIP increases as the level of effort increases. That observation could support the concept that excessive lung or chest wall elastance makes PI unreliable as an index of effort, because sicker patients would have more elastance at any particular lung volume and because patients making greater efforts would have higher elastance because their end-inspiratory volume would be higher, regardless of how sick they were. Because PI max is related to Pdi max in all study groups, a relationship between PI/PI max and Pdi/Pdi max and, consequently, between TT mu and TT di is observed in control subjects and NMD patients. In NMD patients, we obtained a high value for the ratio of the mean PI per breath over the PI max (0.36 0.18). A similar PI-to-PI max ratio has been described in severely ill patients during recovery from acute respiratory failure (0.42 0.11), which is considered a good model to approximate the condition before the institution of mechanical assistance (37). After discontinuation of mechanical ventilation, patients breathe against a high-inspiratory load and develop a pattern of rapid, shallow breathing to maintain minute ventilation (33). In some of these patients, inspiratory muscles perform work that may lead to fatigue, failure of weaning trial, and reinstitution of mechanical ventilation (37). In accordance with this, our results suggest that many NMD patients also breathe against a high-inspiratory load, and their inspiratory muscles perform a task that likely could lead, for at least some of them, to fatigue. The limits of agreement between TT mu and TT di were wider in the NMD group than in the normal group. One explanation of this may be the different contribution of the inspiratory rib cage muscles to VT in healthy subjects and neuromuscular patients. In fact, TT mu is an index that, as estimated from mouth pressure, assesses the overall contribution of the inspiratory and expiratory muscles, whereas TT di represents only the diaphragm. Patients with NMD generally display abnormal thoracoabdominal breathing patterns characterized by the asynchrony of the rib cage and abdominal displacements (26). This is important because, in some neuromuscular patients, most of the pressure is generated by the rib cage inspiratory musculature, with a significant contribution by the respiratory muscles of expiration. In fact, the mean gastric pressure/pdi found in the NMD group was lower than in the control group (0.20 0.15 vs. 0.28 0.08), suggesting that neuromuscular patients have an increased recruitment of the inspiratory rib cage muscles other than the diaphragm. The finding that, in NMD patients, the fatigue threshold of the inspiratory muscles as assessed by TT mu exceeds the values of TT di is probably due to an increased recruitment of rib cage muscles. Several authors found that tension-time index of rib cage muscles of 0.30 corresponds to the fatigue threshold of rib cage muscles (38). Indeed, as recruitment of rib cage muscles increased, the fatigue threshold of the inspiratory

1548 NONINVASIVE RESPIRATORY MUSCLE EFFORT muscles defined by TT mu approached the values of the fatigue threshold of the rib cage muscles only. In COPD patients, a higher fatigue threshold of the respiratory muscles than that obtained in our study has been reported (28). It could be due to a still increased recruitment of rib cage muscles. Moreover, in COPD patients, hyperinflation might favor rib cage muscles over the diaphragm, and the pressure-generating ability of rib cage muscles would, therefore, be enhanced. In conclusion, this study presents a noninvasive and easily determined clinical PI, based on the measurements of mouth occlusion pressure and TI, in three types of restrictive respiratory diseases. Whereas, in patients with CWD or UIP, the agreement was too poor to allow for satisfactory estimation, PI seems reliable for estimating the respiratory muscle effort during spontaneous breathing in normal subjects and in patients with neuromuscular disorders. Such an index is not reliable in patients with kyphoscoliosis of lung interstitial disease. The authors thank Asunción Alvarez, Pilar Librán, and Carmen Suarez for outstanding technical support. DISCLOSURES This study was supported by Grant 07/031/96 from Comunidad Autónoma de Madrid, Spain. REFERENCES 1. Altman DG, Gore SM, Gardner MJ, and Pocock SJ. Statistical guidelines for contributors to medical journals. Br Med J 286: 1489 1493, 1983. 2. [Anon]. Clinical indications for noninvasive positive pressure ventilation in chronic respiratory failure due to restrictive lung disease, COPD, and nocturnal hypoventilation. A Consensus Conference Report. Chest 116: 521 534, 1999. 3. Baydur A. Respiratory muscle strength and control of ventilation in patients with neuromuscular disease. Chest 99: 330 338, 1991. 4. Baydur A, Pangiotis K, Behrakis K, Zin WA, Jaeger M, and Milic-Emili J. A simple method of assessing the validity of the esophageal balloon technique. Am Rev Respir Dis 126: 788 791, 1982. 5. Begin R, Bureau MA, Lupien L, Bernier JP, and Lemieux B. Pathogenesis of respiratory insufficiency in myotonic dystrophy. Am Rev Respir Dis 125: 312 318, 1982. 6. Bellemare F and Grassino A. Effect of pressure and timing of contraction on human diaphragm fatigue. J Appl Physiol 53: 1190 1195, 1982. 7. Bellemare F and Grassino A. Evaluation of human diaphragm fatigue. J Appl Physiol 53: 1196 1206, 1982. 8. Bellemare F and Grassino A. Force reserve of the diaphragm in patients with chronic obstructive pulmonary disease. J Appl Physiol 55: 8 15, 1983. 9. Bland JM and Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1: 307 310, 1986. 10. Conti G, Rocco M, Antonelli M, Bufi M, Tarquini S, Lappa A, and Gasparetto A. Respiratory system mechanics in the early phase of acute respiratory failure due to severe kyphoscoliosis. Intensive Care Med 23: 539 544, 1997. 11. DiMarco AF, Kelsen SG, Cherniack NS, and Goethe B. Occlusion pressure and breathing pattern in patients with interstitial lung disease. Am Rev Respir Dis 127: 425 430, 1983. 12. European Community for Coal and Steel. Standarization of the lung function tests. Bull Eur Physiopath Respir 19, Suppl l5: 1 95, 1983. 13. Gallagher CG, Sanii R, and Younes M. Response of normal subjects to inspiratory resistive unloading. J Appl Physiol 66: 1113 1119, 1989. 14. García-Río F, Pino JM, Gómez L, Alvarez-Sala R, Villasante C, and Villamor J. Regulation of breathing and perception of dyspnea in healthy pregnant women. Chest 110: 446 453, 1996. 15. García-Río F, Prados C, Díez Tejedor E, Díaz Lobato S, Alvarez-Sala R, Villamor J, and Pino JM. Breathing pattern and central ventilatory drive in mild and moderate generalised myasthenia gravis. Thorax 49: 703 706, 1994. 16. Gaultier C, Boule M, and Girard F. Inspiratory force reserve of the respiratory muscles in children with chronic obstructive pulmonary disease. Am Rev Respir Dis 131: 811 815, 1985. 17. Gorini M, Spinelli A, Ginanni R, Duranti R, Gigliotti F, Arcangeli P, and Scano G. Neural respiratory drive and neuromuscular coupling during CO 2 rebreathing in patients with chronic interstitial lung disease. Chest 96: 824 830, 1989. 18. Gross D, Grassino A, Ross WRD, and Macklem PT. Electromyogram pattern of diaphragmatic fatigue. J Appl Physiol 46: 1 7, 1979. 19. Hayot M, Ramonatxo M, Matecki S, Milic-Emili J, and Prefault C. Noninvasive assessment of inspiratory muscle function during exercise. Am J Respir Crit Care Med 162: 2201 2207, 2000. 20. Herrera A and Garcia-Rio F. Exploración funcional de los músculos respiratorios. Arch Bronconeumol 36: 146 158, 2000. 21. Katzenstein A. Idiopathic interstitial pneumonia: classification and diagnosis. In: Current Topics in Lung Pathology. IAP Monograph Series, edited by Churg AM and Katzenstein A. Baltimore, MD: Williams & Williams, 1993, p. 1 31. 22. Laporta D and Grassino A. Assessment of transdiaphragmatic pressure in humans. J Appl Physiol 58: 1469 1476, 1985. 23. Loiseau A, Loiseau P, Dubreuil C, and Pujet JC. Shape change of the occlusion-pressure wave during exercise. Eur Respir J 3: 1179 1185, 1990. 24. Milic-Emili J. Recent advances in clinical assessment of control of breathing. Lung 160: 1 17, 1982. 25. NHLBI Workshop. Respiratory muscle fatigue. Report of the Respiratory Muscle Fatigue Workshop Group. Am Rev Respir Dis 142: 474 480, 1990. 26. Perez A, Mulot R, Vardon G, Barois A, and Gallego J. Thoracoabdominal pattern of breathing in neuromuscular disorders. Chest 110: 454 461, 1996. 27. Quanjer PH, Tammeling GJ, Cotes JE, Pedersen OF, Peslin R, and Yernault JC. Lung volumes and forced ventilatory flows. Eur Respir J 6, Suppl 16: 5 40, 1993. 28. Ramonatxo M, Bouladr P, and Préfaut C. Validation of a noninvasive tension-time index of inspiratory muscles. J Appl Physiol 78: 646 653, 1995. 29. Ramonatxo M, Mercier J, Cohendy R, and Préfaut C. Effect of resistive loads on pattern of respiratory muscle recruitment during exercise. J Appl Physiol 71: 1941 1948, 1991. 30. Roussos C. Respiratory muscle fatigue in the hypercapnic patients. Bull Eur Physiopath Respir 15: 117 123, 1979. 31. Roussos CS and Macklem PT. The respiratory muscles. N Engl J Med 307: 786 797, 1982. 32. Siafakas NM, Chang HK, Bonora M, Gautier H, Milic- Emili J, and Duron B. Time course of phrenic activity and respiratory pressures during airway occlusion in cats. J Appl Physiol 51: 99 108, 1981. 33. Tobin MJ, Perez W, Guenther SM, Semmes BJ, Mador MJ, Allen SJ, Lodato RF, and Dantzker DR. The pattern of breathing during successful and unsuccessful trials of weaning from mechanical ventilation. Am Rev Respir Dis 134: 1111 1118, 1986. 34. Whitelaw WA and Derenne JP. Airway occlusion pressure. J Appl Physiol 74: 1475 1483, 1993.

NONINVASIVE RESPIRATORY MUSCLE EFFORT 1549 35. Whitelaw W, Derenne JP, and Milic-Emili J. Occlusion pressure as a measure of respiratory center output in conscious men. Respir Physiol 23: 181 189, 1975. 36. Wilson SH, Cooke NT, Moxham J, and Spiro SG. Sternomastoid muscle function and fatigue in normal subjects and in patients with chronic obstructive pulmonary disease. Am Rev Respir Dis 129: 460 464, 1984. 37. Zakynthinos SG, Vassilakopoulos T, and Roussos C. The load of inspiratory muscles in patients with needing mechanical ventilation. Am J Respir Crit Care Med 152: 1248 1255, 1995. 38. Zocchi L, Fitting W, Majani U, Fracchia C, Rampulla C, and Grassino A. Effect of pressure and timing of contraction on human rib cage muscle fatigue. Am Rev Respir Dis 147: 857 864, 1993.