INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Tracheotomy is commonly performed on mechanically ventilated patients with prolonged difficulties in weaning from the ventilator. Many reasons are often mentioned for the decision to perform tracheotomy on these patients, very few of which are supported by convincing evidence (1). Conversely, tracheotomy carries a risk of immediate and long-term complications (4, 5).

Failure of weaning from the ventilator often results from an imbalance between respiratory muscle capacity and the loads imposed on the respiratory system (6). Information about the effect of tracheotomy on the work of breathing (WOB) may explain a possible influence of tracheotomy on the weaning process, and may therefore help in making individual decisions to perform this invasive procedure. Indeed, both the resistance of the tracheotomy cannula and its internal dead space (volume) may be smaller than those of an endotracheal tube (ETT), and tracheotomy may therefore ease the respiratory workload. On the other hand, the relative influence of the artificial airway in difficult-to-wean patients may be relatively small, and the total effort of breathing may be essentially dictated by the respiratory mechanics of the patient.

The purpose of the prospective study described here was to evaluate the effects of tracheotomy on breathing pattern and patients' efforts to breathe by examining these two parameters before and shortly after the procedure (7). In order to correlate the results observed in patients with the resistive properties of the prosthesis, a bench study was also performed, using the patients' ETTs after their removal.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Eight consecutive adult patients hospitalized in the medical intensive care unit of our hospital, requiring prolonged mechanical ventilation and who were candidates for tracheotomy according to current recommendations, were included in the study (3, 16). The patient group was quite representative of medical patients converted to tracheotomy, with indications for the procedure that were neurological (coma), respiratory (chronic obstructive pulmonary disease [COPD]), or neuromuscular (diaphragmatic dysfunction). All patients were studied while spontaneously breathing in the pressure-support (PS) ventilation mode. The study was performed between November 1993 and June 1994. All patients had given their informed consent to participate in the study, which was approved by the ethics committee for research of our institution.

Clinical Measurements

Measurements were made on the day before and 6 h after the end of the tracheotomy procedure. Airflow was recorded with a Fleisch No. 2 pneumotachograph (Fleisch, Lausanne, Switzerland) connected to a differential pressure transducer (MP 45, ± 2 cm H₂O; Validyne Corp., Northridge, CA). The flow signal was integrated to yield tidal volume (VT). The pneumotachograph was inserted between the endotracheal tube and the Y-piece of the ventilator. Airway pressure (Paw) was measured with a differential pressure transducer (SDX 001, ± 70 cm H₂O; Sensym, Santa Clara, CA).

Esophageal pressure (Pes) was recorded with a balloon catheter connected to a differential pressure transducer (Validyne MP 45, ± 50 cm H₂O). Gastric pressure (Pga) was similarly recorded in four patients. The position and the validity of the Pes and Pga measurements were checked by analyzing the shape of both curves, using the occlusion technique, and, as previously described, after the drinking of water and by gently pushing on the patient's abdomen (17, 18).

All signals were sampled and digitized at 200 Hz, and data were entered into a microcomputer. VT, respiratory rate (RR), and minute ventilation (E) were obtained from the flow signal. The WOB per breath was computed from Pes/VT loops, and was expressed as WOB per liter of ventilation (WOB/L, in J/L), or as power of breathing (WOB/ min, J/min). The inspiratory WOB per breath was computed according to the following principles: Work per breath was calculated from a Campbell diagram, by computing the area under the inspiratory Pes/ VT curve on the one hand and under the static Pes/VT curve of the chest wall on the other hand (19). The Pes values at zero-flow points were considered as the beginning and end of expiration. The theoretical value for chest wall compliance, which is 4% of the predicted value of the vital capacity (VC) per centimeter of water, was used to trace the static Pes/VT curve for the chest wall (19). Some degree of error in estimating chest wall compliance probably occurs in some patients when estimation is based on this theoretical value. This error, however, is the same when different periods of breathing are compared, and therefore does not invalidate comparisons. The relaxation curve of the chest wall was superimposed on the complete diagram, assuming that the end-expiratory elastic recoil pressure of the chest wall was equivalent to the Pes level at the beginning of inspiratory effort. The beginning of the sharp negative deflection of the Pes curve was taken as the onset of effort. Any difference between this initial Pes level and the zero-flow point indicated the presence of intrinsic positive end- expiratory pressure (this difference was referred to as PEEPi) (8, 20- 22). Separate calculations of the resistive WOB and elastic WOB were also made, and were expressed per liter of ventilation (rWOB/L and eWOB/L, respectively). These latter calculations were made with transpulmonary pressure (PTP) measurements, calculated as Paw minus Pes, and with measurements of volume.

Because the static chest wall line is displaced from the zero-flow point, the presence of PEEPi increases the area enclosed in the Campbell diagram, and increases the calculated WOB. When no expiratory abdominal activity is detected, PEEPi implies that the inspiratory muscles must generate sufficient force to counteract the opposing positive recoil pressure generated by dynamic hyperinflation before inspiratory flow begins, and it therefore acts as an inspiratory threshold load. In the four patients with recordings of Pga, we checked the absence of an increase in Pga during expiration as previously described (23). In four patients, however, such recordings were not obtained. For these patients we calculated WOB by using the two possible extreme assumptions about expiratory muscle activity, to see whether this influenced the results in terms of net effect on WOB. We first calculated inspiratory WOB by assuming that all PEEPi was due to dynamic hyperinflation (i.e., with no expiratory muscle activity), and then also calculated inspiratory WOB without taking into account the PEEPi, assuming that the entire PEEPi value (and its subsequent change) could be due to expiratory abdominal muscle activity and was not part of the inspiratory work. We refer to this second calculation, with lower values of WOB, as WOB_low.

WOB/min (in J/min) was obtained by multiplying the mean work per breath by RR. Power divided by E yielded WOB/L (in J/L). Ten to 30 breaths were used to compute average values, depending on breath-to-breath variability.

Because measurements of WOB do not take into account inspiratory efforts that are unable to trigger the ventilator, analysis of desynchronization also included measurements of the pressure-time product per breath (PTP/b) and per minute (PTP/min) for the inspiratory muscles. For assisted breaths, PTP/b was calculated from the beginning of the inspiratory deflection to the end of inspiratory flow relative to the Pes tracing, assuming that elastance of the chest wall was linear within the range of VT. The static chest wall line was used as the reference for the area calculation. For ineffective efforts, we considered a horizontal line as the reference for the calculation, since virtually no variation in volume was detected. PTP/min was calculated as PTP/b multiplied by RR. In this way, a distinction was made between the ventilator respiratory rate (RR) and the patient respiratory rate (RRp). The percentage of ineffective inspiratory efforts was therefore calculated as (1  RR/RRp) × 100 (24).

Occlusion pressure (P0.1) was also determined, using the slope of airway pressure-versus-time curves at the beginning of the inspiratory effort, during triggering of the ventilator demand valve, as previously described (25).

Protocol

Three series of measurements were repeated on the day before and 6 h after tracheotomy, at three different PS levels. PS was initially set at the same level as used before tracheotomy (baseline pressure support: PS-B), and was also reduced (PS5) and increased (PS+5) by 5 cm H₂O in a randomized order; recordings were made at the end of each 15-min period under each PS condition. Because some patients exhibited respiratory distress while breathing spontaneously without respiratory assistance, this ventilatory modality was not studied in our patients. One may consider, however, that the low level of PS to which the patients were subjected was close to that of spontaneous unassisted breathing (8.5 ± 3.9 cm H₂O). External PEEP and FI_O2 were kept constant during all measurements. FI_O2 was chosen at a level sufficient to obtain more than 90% arterial oxyhemoglobin saturation. The trigger pressure threshold for ventilation was fixed at 1 cm H₂O and kept constant before and after tracheotomy. Different ventilators were used in the study, but measurements for a given patient were made with the same ventilator.

Tracheotomy was performed surgically and with the patient under general anesthesia produced with Propofol (Zeneca Pharma, Cergy, France) and Fentanyl (Janssen, Boulogne-Billancourt, France). The initial dosages were respectively 2 mg/kg and 2 × 10³ mg/kg, with intravenous administration. The same drugs were readministered at half the initial dosage every 15 min during the procedure. The inner cannula diameter was kept similar to that of the ETT, and was 8 mm in seven of the patients and 7 mm in one. Resolution of anesthesia was complete in all patients at the time of the second set of measurements (usually 6 h after the end of the tracheotomy procedure).

In Vitro Measurements

Removed ETTs were kept after tracheotomy for the purpose of bench testing. Their flow-impeding characteristics were compared in vitro with those of unused ETTs and to tracheotomy tubes of the same inner diameter. The same transducers and acquisition system were used as in the patient study. Six sets of measurements were made during phasic ventilation of a lung model connected to a ventilator via the different cannulas. In this model, the resistance to airflow was entirely due to the cannula or to the tube. For a given used ETT, VT and RR were chosen to resemble the values observed during the clinical measurements at the three different PS levels, before and after tracheotomy, and a sinusoidal flow was chosen. The same breathing pattern was used for comparing the resistive work induced by the used tube with that induced by a new, clean tube and by a new tracheotomy cannula. A Veolar ventilator (Hamilton Medical, Bonaduz AG, Switzerland) was used for these measurements. The total work exerted by the lung model, including the cannula, was calculated by plotting the volume signal versus the airway pressure signal, and was then partitioned into its resistive and its elastic components. The resistive work of breathing (rWOB) was averaged for each PS assistance level and expressed in J/L.

Statistical Analysis

Data are expressed as mean ± SD. Because each level of PS before tracheotomy corresponded to the same level after tracheotomy, paired comparisons were made. Values were compared through Wilcoxon's nonparametric test for small samples. A value of p  0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Characteristics of the Patients

During a 10-mo period, eight patients were included and completed the study. Detailed clinical data and outcome are shown in Table 1. The mean duration of tracheal intubation before tracheotomy was 31 ± 19 d. During the same period, two additional patients were included but could not complete the study because of local hemorrhage following the tracheotomy procedure.

Clinical Results

The main findings for WOB are presented in Figure 1 and Table 2. A significant reduction in WOB/L was observed at PS-B and PS5, and a near-significant reduction was observed at PS+5 (p = 0.05).

View larger version (13K): [in this window] [in a new window]   Figure 1.   Individual values for WOB, expressed in J/L, at baseline pressure support (PS-B) before and after tracheotomy (T). The difference was significant at p  0.05.

Comparisons between WOB_low values obtained before and after tracheotomy are shown in Table 2. A significant reduction was similarly observed for PS-B and PS5.

Resistive and elastic work decreased significantly at PS-B and PS5 (Table 2).

A significant reduction in P0.1 was observed for all levels of PS (Figure 2, Table 2). The measurements of PEEPi also indicated a significant reduction for all PS levels (Figure 3, Table 2). No variation in expiratory abdominal activity, as judged by analysis of the shape of the Pga curve, was observed in the four patients on whom this measurement was made. There was only a trend toward a reduction in RR and E, whatever the PS level (Table 2).

View larger version (14K): [in this window] [in a new window]   Figure 2.   Individual values for P0.1 at baseline pressure support (PS-B) before and after tracheotomy (T). The difference was significant at p  0.05.

View larger version (16K): [in this window] [in a new window]   Figure 3.   Individual values for PEEPi at baseline pressure support (PS-B) before and after tracheotomy (T). The difference was significant at p  0.05.

With regard to the analysis of synchronization, one patient (Patient 1) had ineffective inspiratory efforts at more than 2/min at all three PS levels before tracheotomy (Table 3). The percentage of ineffective inspiratory efforts increased when the PS level was increased. The PTP/b related to ineffective inspiratory efforts was similar whatever the PS level before tracheotomy, whereas the PTP/min for ineffective inspiratory efforts, expressed as a percentage of the PTP/min for effective efforts, was much higher for the PS+5 level. After tracheotomy, a reduction was observed in all of these parameters. Some mild levels of desynchronization (ineffective inspiratory efforts at less than 3/min) were observed in two other patients for PS+5 before tracheotomy, but disappeared after tracheotomy.

In Vitro Results

The mean values of rWOB measured with the different tubes and cannulas are shown in Table 4. The lower values were observed for the tracheotomy tubes and the higher for the removed ETTs. The mean percentage reduction in rWOB with tracheotomy tubes was 21% versus new ETTs at the PS-B level. The mean reduction from tracheotomy tubes to removed ETTs amounted to 43% at the same PS level. There was significant extra work induced by the removed ETTs, with an increase in rWOB of 28% over the work with the new ETTs.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Tracheotomy is widely performed in intensive care units. Little is known, however, about its effects on weaning from and the duration of mechanical ventilation. This study was the first to evaluate, in the clinical environment, the effects of tracheotomy on parameters that may have a major influence on the issue of discontinuation of mechanical ventilation. It mainly demonstrated a significant reduction in WOB and in P0.1. Part of this reduction resulted from substantial extra work imposed by used ETTs that had been left in place for prolonged periods.

It is important to determine the mechanisms by which WOB can be reduced after tracheotomy. WOB can be represented as the sum of resistive work (applied to the lungs and mainly to artificial and natural airways) and elastic work (applied to the lungs and the chest wall). A diminution in the resistive component of WOB was the first factor explaining the overall reduction in WOB with tracheotomy. Such a reduction can be due to the following factors:

1. A difference in length exists between tracheotomy cannulas and ETTs. This is confirmed by our in vitro measurements comparing the resistance of new endotracheal and tracheotomy tubes of the same inner diameter. The effect of curvature, which may slightly disadvantage the tracheotomy tube in terms of resistance (26), is largely overcome by the length effect.
2. A progressive and insidious reduction in the inner diameter of the ETT was evidenced in vitro by the differences observed in removed versus new ETTs. A reduction in inner diameter may rapidly affect the resistive load: for example, in the case of a laminar stream, Poiseuille's law indicates that the pressure decrease along the tube is equal to K · / r⁴ (where K is a constant, the flow, and r the radius of the tube). In the case of a turbulent stream, the decrease in pressure is K' · ²/D⁵ (where K' is a constant and D is the effective diameter). It should be pointed out that obstruction of the ETT was never clinically suspected in any patient before tracheotomy. Such data are in accordance with the findings in a previous study by Wright and coworkers, in which an augmentation in ETT resistance was found with in vivo as compared with in vitro measurements (27), and are also in accordance with the findings in a recent study by Villafane and coworkers in our group (28). In the latter study, this phenomenon was explained by a permanent deposit of secretions on the inner wall of the ETT.
3. It is also possible that tracheotomy facilitated and improved the efficacy of endotracheal suctioning (2). This could explain a change in resistance related to the main bronchi and to the lowest portion of the trachea.

We also observed a reduction in the eWOB, which could have resulted from a change in PEEPi, contributing to the total reduction in WOB. The amount of PEEPi depends on RR and on E in patients with obstructive disease, but these latter parameters, although slightly decreased, were not significantly modified after tracheotomy. It is therefore more likely that the diminution in PEEPi observed after tracheotomy was related to a diminution in expiratory resistance of the airway, as discussed earlier. Our patients could also have displayed a change in expiratory abdominal activity after tracheotomy. Indeed, this pattern of ventilation has been shown as a possible cause of PEEPi (23). Unfortunately, no systematic recording of Pga was done, and we therefore cannot formally address this point for all patients. In the four patients for whom recordings of Pga were available, however, this pattern of breathing was not observed. Moreover, we calculated WOB_low for the other patients and then substituted the calculated values for the measured values for the overall comparison of WOB before and after tracheotomy. This was an extreme assumption in which we considered that all measured PEEPi values would have been positive owing to active expiration. A significant reduction was again observed for PS-B and PS5, indicating a benefit in terms of inspiratory effort.

Additionally, a significant effect of the anesthesic drugs used for the tracheotomy procedure might explain a decrease in WOB, although this is unlikely in view of the short half-life of elimination of fentanyl and propofol and because the patients were perfectly awake at the time of the study. More importantly, we observed a diminution in WOB expressed in J/L, with no significant change in E, which can hardly be explained by a diminution of central inspiratory drive alone. Also, our results were consistent with the flow-impeding characteristics of the cannula as assessed in vitro.

We also observed a significant reduction in P0.1 after tracheotomy. P0.1 was first presented as an index of central respiratory drive (29, 30). It has also been proposed as a predictor of successful weaning from ventilation (9, 12, 14). Two factors may explain the diminution in P0.1. The first is the modification of the mechanical characteristics of the respiratory system that occurs with tracheotomy, reducing hyperinflation and resistances. The second factor is the diminution of the inspiratory central drive in response to this modification and, although it is less likely for the reasons discussed earlier, to the anesthesic agents used for tracheotomy.

In three patients, and especially in one, we observed an improved synchrony with the ventilator after tracheotomy. This could have been related to a reduction in the resistive load and in PEEPi. In accord with previous findings by others, this patient showed a higher rate of desynchronization when higher levels of PS were employed (24, 31, 32).

The effect of tracheotomy on the duration of mechanical ventilation is unknown. To date, only one study has prospectively evaluated the effect of tracheotomy in a small group of trauma patients, suggesting a benefit in favor of early tracheotomy as compared with prolonged translaryngeal intubation (2). Several other reasons, however, are often mentioned for the decision to perform tracheotomy on medical patients, and these reasons have to be balanced against the risk of the procedure (4). Improved patient comfort has been reported in one study, resulting from a facilitation of communication, mobility, oral alimentation, and suctioning of secretions (2). No other benefit, such as a reduction in the risk of laryngeal or tracheal injury or a reduction in the risk of nosocomial pneumonia, could be demonstrated. Tracheotomy is thus frequently performed during weaning from mechanical ventilation without strong arguments, substantiated by clinical studies, to support it. Accordingly, in a review of the literature, Heffner promoted an anticipatory approach in which tracheotomy would be performed after at least 7 d of intubation, when extubation appeared to be distant, and when it was likely to produce an important benefit (3). Our results also indirectly suggest that detecting and/or preventing insidious ETT obstruction may offer an alternative to tracheotomy in some patients.

In summary, the present study contributes physiologic information about the respiratory benefit of tracheotomy, and also emphasizes the importance of the insidious obstruction of ETTs. This may justify the development of methods to detect and quantify this complication (33, 34). It can also justify the evaluation of devices designed to remove obstructions from ETTs (35), and could argue against the current practice of keeping the same ETT in place for a prolonged period. In tracheotomized patients, changing the tracheotomy cannula is easily performed in a systematic fashion, which can be considered another advantage of this technique.