INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mechanical ventilation is usually delivered via an endotracheal tube linking the patient to the ventilator. One of the most common techniques to wean a patient from mechanical ventilation is to assess his/her ability to breathe spontaneously through the endotracheal tube after disconnection from the ventilator, with administration of humidified oxygen (1). Because of the shape of the humidifier, this test is often referred to as a T-piece trial and can last up to two hours in order to test the patient's tolerance to spontaneous breathing (1, 2). However, the endotracheal tube has a finite respiratory resistance (3), which could theoretically represent an additional load and impair the patient's ventilation.

The resistance of the endotracheal tube induces a specific work of breathing which has been assessed in bench studies (8), in nonintubated normal volunteers breathing through tubes (11, 12) or during mechanical ventilation in intubated patients (6). For this reason, it has been claimed that breathing through an endotracheal tube constitutes a considerable ventilatory challenge. In intubated patients, several studies have measured the work of breathing dissipated in the whole circuitry of the ventilator, but the specific contribution of the endotracheal tube has been ignored (13, 14). In addition, in many studies a catheter was introduced at the carina, which, by itself, may markedly change the impedance-flow characteristics of the circuit. No study has specifically addressed the question of the part of the work related to the endotracheal tube during a T-piece trial, however.

The endotracheal tube being an additional resistance, the patient work of breathing could decrease after extubation. Natural upper airway, however, may offer a similar resistance to breathing after extubation. Studies that evaluated the overall work of breathing before and after extubation lead to conflicting results. Brochard and coworkers found a decrease in the power of breathing after extubation but no change in the work expressed in J/L (15). Others found an increase of the work of breathing expressed in J/L after extubation (16, 17). These authors found no tracheal or glottic abnormalities and hypothesized that the upper airway above the glottis may play a role in this increased patient's work of breathing.

The aim of the present study was to compare the work of breathing before extubation when patients were spontaneously breathing through their endotracheal tube, and after tracheal extubation. The influence of a 2-h T-piece trial was also assessed. We only considered successfully extubated patients for subsequent analysis. The resistance and the work of breathing dissipated in the endotracheal tube was measured by using the acoustic reflection technique (18, 19) and a similar technique was used for the upper airway after extubation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

The experimental protocol was approved by the institutional ethics committee and informed consent was obtained from each patient. The study was initially carried out during 19 attempts of spontaneous breathing in 16 patients ventilated for more than 24 h, orotracheally intubated, and thought ready for a T-piece trial before extubation. Before the beginning of the study, all patients were receiving pressure support ventilation. They were fully conscious, and sitting in their bed at 60° from horizontal. Their position was kept constant throughout the study with the neck kept in the trunk axis. During the spontaneous breathing trial, the endotracheal tube was connected to a T-piece providing humidification and oxygen. After tracheal extubation, the patients received nasal oxygen.

Five attempts failed (one failure of the T-piece, four re-intubations) and the relevant characteristics of the 14 patients who were not reintubated in the following 48 h and retained for the analysis are given in Table 1. For two patients (numbers 12 and 13), the second and successful extubation only was considered.

Measurements

Flow was recorded with a Fleisch no. 2 pneumotachograph connected to a differential pressure transducer (MP45, ± 2 cm H₂O; Validyne, Northridge, CA) placed between the end of the endotracheal tube and the T-piece, or between a mouth piece and the T-piece after extubation. Integration of the flow signal provided tidal volume. For the purpose of bacteriological safety and airway humidification, a filter (Hygrobac; DAR, Mirandola, Italy) was positioned between the patient and the pneumotachograph. Airway pressure was measured between the filter and the pneumotachograph with a differential pressure transducer (MP45, ± 70 cm H₂O; Validyne, Northridge, CA). Because pressure was measured after this filter, the equation for inspiratory pressure drop (PF) through the filter as a function of flow had been determined on a bench for subsequent correction of measured airway pressure, resistance and work of breathing of the upper airway. This equation was: PF(cm H₂O) = 1.786 V(L/s)^1.096. Esophageal and gastric pressures were measured using a double balloon catheter (Marquat, Boissy Saint Léger, France). The esophageal and gastric balloons were filled with 1 ml of air. Appropriate placement of the esophageal balloon was assessed with an occlusion (20). Each balloon was connected to a differential pressure transducer (SDX 001 ± 70 cm H₂O; Sensym, Santa Clara, CA). Pressure and flow signals were digitized at 128 Hz and sampled using an analogic/numeric system (MP100; Biopac Systems, Santa Barbara, CA).

Assessment of Patient's Effort

Inspiratory work of breathing done by the patient was computed from esophageal pressure and tidal volume loops as previously described (15, 21). In brief, the inspiratory work per breath was calculated from a Campbell diagram by computing the area enclosed between the inspiratory esophageal pressure-tidal volume curve on the one hand, and the static esophageal pressure-volume curve of the chest wall on the other, using theoretical value for the chest wall compliance. Resistive and elastic parts of the work of breathing were computed from transpulmonary pressure measurements. The beginning of inspiration was separated from the elastic recoil pressure by an amount equal to intrinsic positive end expiratory pressure (PEEPi), computed on the esophageal pressure recording as the negative deflection in esophageal pressure from the onset of inspiratory effort to the point of zero flow. This method can be influenced by expiratory activity but in none of the patients could substantial expiratory muscle activity be detected, as defined by an expiratory rise in gastric pressure (> 0.5 cm H₂O) (22). The pressure-time index of the respiratory muscles was calculated as the integral of the esophageal pressure recording versus time during the inspiratory effort, from the onset of pressure deflection, referenced to the chest wall relaxation curve, to the end of inspiratory flow. Each value was the mean of four recording trials over 1 min.

Endotracheal Tube Assessment

The longitudinal area profile of endotracheal tubes was measured with the two transducer acoustic reflection method as described earlier (18, 19). Two microphones (piezoresistive pressure transducers 8510-B; Endevco France, le Pré St. Gervais, France) were inserted into the wall of a plexiglass tube of 0.8 cm inner diameter. The first microphone was 5 cm from the proximal tube opening, the second 12 cm. A small horn-driver (25 W; Bouyer, Montauban, France) located 6.5 cm from the second microphone provided a broad band pulse of short duration generated by a computer. The proximal tube opening was connected to the endotracheal tube and the distal opening was left open to the atmosphere or connected to the T-piece if the patient needed oxygen. The pressure signal generated by the horn-driver was recorded by the two microphones, amplified, band-pass filtered (10 Hz to 10 kHz dual-decade filter FV625A, 48 dB/Oct. Butterworth, NF; Equipements Scientifiques, Garches, France) and digitized at 40 Hz per channel (DMA TM 100; Labmaster, Dipsi industrie, Asnières, France). The cross-sectional area of the endotracheal tube was calculated as the mean of ten acquisitions and plotted against the axial distance to provide the longitudinal profile of the area. This allowed us to deduce the actual diameter of the endotracheal tube. The inspiratory pressure drop in the endotracheal tube was computed from this diameter and from the flow with the Blasius resistance formula (23). The inspiratory work of breathing dissipated in the endotracheal tube was computed from the pressure drop-tidal volume loop (24). Each value of endotracheal tube resistance and work of breathing was the mean of four recording epochs of one minute.

Upper Airway Assessment

The distance between the nostril and the glottis was measured before the T-piece trial by fiberoptic bronchoscopy. This measurement was used to place a piezoelectric pressure sensor (Gaeltec, Dunvegan, Isle of Skye, UK) just above the glottis before extubation. After extubation, supra-glottic pressure was recorded with this in situ sensor. The position of the transducer was not directly checked at this time; the shape and the amplitude of the signal both during spontaneous breathing and swallowing were analyzed to ensure a correct positioning. The signal was digitized and sampled as described above. Pressure drop in the supraglottic airway was calculated by the difference between supraglottic pressure and airway pressure at the mouth piece corrected for the presence of the filter. Resistance of the supra-glottic airway was calculated and the inspiratory supraglottic work of breathing was computed from the pressure drop to tidal volume loop. Each value of upper airway work of breathing was the mean of four one-minute long recordings.

After extubation, the longitudinal profile of the airway area was obtained with the same acoustic reflection device (18, 19) as for endotracheal tube measurement, except that the inner diameter of the plexiglass tube was 1.9 cm. The patients breathed through the tube via a mouth piece. A schema of the different set-ups used for measurements is given in Figure 1.

View larger version (19K): [in this window] [in a new window]   Figure 1.   Schema indicating the two set-ups used for measurements. (Upper panel ) set-up used before and after tracheal extubation to record flow and pressure signals. (Lower panel ) set-up used before and after tracheal extubation to measure the longitudinal area profile of the endotracheal tube or the upper airway.

Protocol

Recordings were performed at the beginning of the T-piece trial for subsequent computation of patient inspiratory work of breathing. The longitudinal area profile of the endotracheal tube was obtained. After 2 h of spontaneous breathing, the same recordings were performed. If the attending physician judged the patient's status to be satisfactory, tracheal extubation was performed. Recordings were performed again, immediately after extubation. Flow and airway pressure were recorded via a mouth piece, the patient wearing a nose-clip. The supraglottic pressure was recorded with the in situ pressure sensor. The esophageal and gastric catheter and the supraglottic pressure transducer were then removed and the longitudinal area profile of the airway was recorded with a mouth piece but without a nose-clip.

Statistical Analysis

Data are expressed as mean value ± standard deviation. Two-by-two comparisons were done using a two-tailed paired t-test. A p value below 0.05 was considered as the limit of significance.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A complete set of measurements could be obtained in 11 patients, while, in three, it was not possible to measure the supraglottic pressure and to record the cross sectional area of the airway. An example of flow and pressures tracings (Patient 5) before and after extubation is given in Figures 2 and 3.

View larger version (13K): [in this window] [in a new window]   Figure 2.   Flow and pressures before extubation. Example of flow and pressures tracings recorded in a representative patient (no. 5) before extubation. From top to bottom: flow (Flow), airway pressure (Paw), esophageal pressure (Pes), and gastric pressure (Pga). The grid lines separate expiration (E) from inspiration (I).

View larger version (16K): [in this window] [in a new window]   Figure 3.   Flow and pressures after extubation. Example of flow and pressures tracings recorded in Patient 5 after extubation. From top to bottom: flow (Flow), airway pressure (Paw), supraglottic pressure (Pglott), esophageal pressure (Pes), and gastric pressure (Pga). The grid lines separate expiration (E) from inspiration (I).

Work of Breathing Before and After Extubation

Breathing pattern and indexes of respiratory effort for the patients are given in Table 2. Mean values for the pressure time index are shown in Figure 4, separating patients into two groups according to the size of their tube. The total work of breathing of the patients was statistically identical between the end of the T-piece trial and the immediate post extubation time and neither transpulmonary elastic (p = 0.09) nor resistive (p = 0.14) work of breathing significantly changed after extubation (Figure 5). Regarding other parameters, only tidal volume significantly increased after extubation (p = 0.04).

View larger version (14K): [in this window] [in a new window]   Figure 4.   Mean (± SEM) values of the pressure-time index for the eight patients having an endotracheal tube of small size (< 8.0 mm of internal diameter), and for the six patients having a larger endotracheal tube size (8.0 mm or above), at the beginning of the T-piece trial, its end, and after extubation. There were no significant differences among the three periods in the two groups.

View larger version (16K): [in this window] [in a new window]   Figure 5.   Evolution of patient's work of breathing. Changes in mean work of breathing expressed in J/L, between the beginning of the T-piece trial, its end, and immediately after extubation. Triangles represent total work of breathing, squares represent transpulmonary elastic work of breathing, and circles represent transpulmonary resistive work of breathing. Standard deviations are vertically shown.

Work of Breathing During the T-piece Trial

We questioned whether the duration of the 2-h trial had an influence on our measurements. Again, indexes of respiratory effort and breathing pattern did not change from the beginning to the end of the T-piece trial (Table 2). The same can be said for the transpulmonary elastic (p = 0.11) and resistive (p = 0.22) work of breathing.

Study of the Endotracheal Tube

The work of breathing dissipated against the endotracheal tube before extubation amounted to 0.17 ± 0.10 J/L and 2.28 ± 1.90 J/min and represented 11.0 ± 3.9% of the total work of breathing when expressed in J/L and 10.5 ± 3.4% when expressed in J/min.

Study of the Upper Airway

The mean resistance of the supraglottic airway was 0.79 ± 0.40 cm H₂O · s/L at 0.25 L/s and 0.92 ± 0.56 cm H₂O · s/L at 0.50 L/s (nonsignificantly different between the two flows). These values were significantly smaller than endotracheal tube resistance (1.43 ± 0.31 cm H₂O · s/L at 0.25 L/s; p = 0.008 and 2.41 ± 0.52 cm H₂O · s/L at 0.50 L/s; p < 0.001). This was consistent with the acoustic longitudinal area profile of the endotracheal tubes (Figure 6) and of the supraglottic airway (Figure 7): in all but two patients (nos. 10 and 13) the cross sectional area of the upper airway, including the glottis, was larger than that of the endotracheal tube. In all but one patient the smallest part of the acoustic profile was the glottis. In one patient (no. 10), the narrowest part of the upper airway was located in the pharynx. The mean value of the glottic area was 1.55 = 0.59 cm² whereas the mean value of the cross sectional area of the smallest part of the upper airway was 1.49 ± 0.63 cm². The latter was significantly larger than the mean value of the cross sectional area of the largest area of the endotracheal tube (0.97 ± 0.14 cm²; p = 0.02).

View larger version (14K): [in this window] [in a new window]   Figure 6.   Acoustic area profile of an endotracheal tube. Example of the acoustic area profile of an endotracheal tube (Patient 13). The internal diameter given by the manufacturer was 7.5 mm. Each point is the mean of 10 recordings. Standard deviations are vertically shown. The first 5 cm of the tracing are due to the acoustic measurement device. The tracheal end of the tube is located at 35 cm.

View larger version (14K): [in this window] [in a new window]   Figure 7.   Acoustic area profile of upper airway. Example of an acoustic area profile of the upper airway (Patient 13). Each point is the mean of 10 recordings. Standard deviations are vertically shown. The first 5 cm of the tracing are due to the acoustic measurement device. The glottis is the narrowest part of the upper airway. It is located at 22 cm. Note that the scale is ten times smaller than in Figure 4 for the endotracheal tube in the same patient.

The work of breathing dissipated in the supraglottic airway after extubation was 0.12 ± 0.10 J/L and 1.69 ± 1.97 J/min. These values were not significantly lower (p = 0.14 and 0.35) than those of the work of breathing lost against the endotracheal tube. The work of breathing of the supraglottic airway was 7.0 ± 4.3% of the total work of breathing in J/L or in J/min.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The main result of this study was that the work of breathing was statistically identical before and after extubation and that, immediately after extubation, the resistance of the supraglottic airway was in a normal range and significantly lower than the resistance of the endotracheal tube. During the T-piece trial, the work induced by the endotracheal tube was around 10% of the total work performed by the patient. Furthermore, the work of breathing did not significantly change between the beginning and the end of a 2-h T-piece trial.

Patient Work of Breathing

One of the goals of our study was to look for a possible increase in the work of breathing after extubation, as reported by other authors (16, 17). This group of investigators suggested a role for the supraglottic airway in explaining the increase of work of breathing noticed after extubation (17). The mean resistance value of the supraglottic airway we found was similar to or even lower than that of normal subjects in other studies. At a flow of 0.30 L/s, Séries and coworkers reported a pharyngeal resistance of 1.00 cm H₂O · s/L (25) and Maltais and coworkers a pharyngeal resistance of 1.30 cm H₂O · s/L (26). We also recorded the longitudinal area profile of the upper airway after extubation with the acoustic reflection technique (18) and found no dramatic narrowing of the upper airway. In all but one patient (no. 10), the narrowest part of the airway was the glottis. In two other patients who experienced post extubation respiratory distress, we found that the contribution of the upper airway to total work of breathing was in a very low range (6% and 1%), with an enlargement of the upper airway. Therefore, the supraglottic airway was not impaired after extubation and did not play a role in a possible increase of patient's work of breathing.

In our study, work of breathing was higher than the values reported in other studies for successfully extubated patients (15, 16, 21). In studies assessing work of breathing as a weaning parameter, the limit values of work of breathing for successful weaning ranged from 1.00 J/L (27) to 1.40 J/L (28). Ishaaya and coworkers, who showed an increase in work of breathing after extubation, reported a mean value of work of breathing of 1.49 J/L in their patients (17). However, two of their patients needed endotracheal reintubation. The larger values of work of breathing measured here were partly explained by the equipment used for measurement which included the presence of a bacteriological filter. It added a resistance and a dead space which might increase the work of breathing (29). The same equipment was used in all situations, however, and we believe it did not influence the comparison made.

Although there were large interindividual differences, on average the work of breathing was identical before and after extubation (Figure 4). The 2-h T-piece trial preceding extubation did not play a role in this result because work did not change during this time. In a previous study (30) we reported that work of breathing did not differ before and after extubation when it was expressed in J/L. However, work of breathing decreased after extubation when it was expressed in J/min because of a decrease in minute ventilation. The difference with this study could be explained by the increase in tidal volume observed in the present work. This could be due to a shorter delay between extubation and measurements; we performed our recordings as soon as the patients were comfortable compared to a 30-min delay in the preceding study (30). Krieger and coworkers (31) have shown an increase in tidal volume and minute ventilation immediately after extubation, while, 30 min later, these parameters were similar to pre-extubation values. Another explanation could be the larger dead space of our measurement device (29). Lastly, the patients had a pressure transducer above their glottis which could have induced behavioral changes. Contrary to these results, Ishaaya and coworkers (17) reported an increase in the work of breathing after extubation. In their study, two among eight patients were reintubated, however, whereas all of the 14 patients of our study were successfully extubated. Indeed, we had the opportunity to assess two other patients who necessitated reintubation of their trachea, one immediately and one within 24 h following extubation. In these two patients, the work of breathing dramatically increased after extubation despite no abnormalities in their upper airway: from 1.79 to 3.62 J/L and from 1.86 to 2.43 J/L. If we had included these two patients in our analysis, the mean work of breathing would have been increased after extubation, albeit not significantly. In these two patients, however, the post-extubation changes in work of breathing were not directly caused by a change in upper airway resistance. We also carefully controlled for the position of the patients, a factor which may influence resistance of the upper airway.

Since the work of breathing at the end of the T-piece trial was identical to the work of breathing after extubation, the T-piece trial mimics well the work of breathing the patients will perform after extubation. Therefore, a 2-h T-piece trial is probably a clinically relevant assessment of the capability of ventilated patients to be extubated. The use of pressure support ventilation at the end of the weaning period is probably needed mainly to compensate for the resistance and the dead space of the circuitry of the ventilator and not of the endotracheal tube (13, 16, 29, 30). Because the T-piece does not alter patient's effort per se, it can be inferred from our data that this period might last a sufficient time to allow a proper assessment of the patient's tolerance to spontaneous breathing.

Work of Breathing Dissipated in the Endotracheal Tube and in the Upper Airway

Our study provides the first measurement of the actual work of breathing dissipated in the endotracheal tube during a T-piece trial. Measurements were made with a noninvasive technique and we did not use a catheter positioned at the carinal end of the endotracheal tube to record the pressure as described in other studies (6, 9, 14), since it markedly modifies the pressure drop-flow relationship in the tube (23). For instance a catheter with a 2 mm external diameter can increase the pressure drop by 58% in a 7.5 mm ETT, by 52% in an 8.0 mm ETT, by 47% in an 8.5 mm ETT, and by 42% in a 9 mm ETT, and thus greatly modify the related work.

The proportion of work of breathing dissipated in the endotracheal tube during the T-piece trial was about 10 to 11%. It is possible that the large values of total work found in this study may tend to underestimate this percentage. This was the same for the upper airway after extubation, however.

After extubation, recordings were made during mouth breathing and the nose was shunted. As mentioned above, the mean resistance value of the supraglottic airway was similar to that of normal subjects in other studies (25, 26) but not significantly different from the endotracheal tube in terms of related work of breathing. This was in part explained by two patients (nos. 8 and 9) whose supraglottic resistance was larger than their endotracheal tube resistance (0.90 versus 1.37 cm H₂O · s/L and 1.00 versus 1.40 cm H₂O · s/L at a flow of 0.25 L/s) and by the increase in tidal volume after extubation. Since the endotracheal tube bypasses the glottis, the glottis and the trachea have to be taken into account to compare the work induced by the "artificial airway" and the "natural airway." Some authors have reported a mean resistance of 0.50 cm H₂O · s/L (32). If this glottic resistance was added to the supraglottic resistance, the total resistance would be very similar to the endotracheal tube resistance. Moreover, nasal resistance has been reported to range from 3.5 to 4.40 cm H₂O · s/L at a flow of 0.30 L/s (25, 26). It is thus likely that the work of breathing dissipated in the upper airway does not substantially differ from those dissipated in the endotracheal tube, and may furthermore be even higher if the patient breathes through the nose.

Conclusion

Our study showed that in successfully extubated patients removal of the endotracheal tube does not alter the patient's work of breathing. A T-piece trial reflects well the work of breathing the patient will perform after extubation when successful and probably constitutes a reliable assessment of the capability of an intubated patient to be liberated from the ventilator.