Long-term Effects of Different Humidification Systems on Endotracheal Tube Patency: Evaluation by the Acoustic Reflection Method

All patients who stayed in the hospital over a 10-month period in the Medical Intensive Care Unit of Henri Mondor Hospital (Créteil, France) and who required continuous MV for more than 48 h were included. Patients likely to die within the next 48 h, with hemorrhagic disorders, and those in whom the trachea was intubated for more than 24 h before arrival were not included.

The experimental protocol was approved by the ethics committee of the French Society of Intensive Care Medicine. The requirement for informed consent was waived because this observational study was performed as part of our routine clinical practice.

Two consecutive periods of 5 months, with two distinct types of humidification, were compared. During the first period of the study, which lasted from August 26, 1998, to January 27, 1999, all the patients were ventilated with an HH (MR730 Respiratory Humidifier; Fisher & Paykel Healthcare, Auckland, New Zealand) with a dual heated-wire circuit. The setting permitted gas to be delivered at a temperature of 40°C at the Y piece and at 37°C in the humidification chamber. During the second period, which lasted from March 8 to August 2, 1999, all the patients were ventilated with a hygroscopic–hydrophobic HME (DAR SpA, Mirandola, Italy). HMEs were replaced every 48 h as recommended. 20 

During the two periods, the inner volume of the ETT and the resistance to inspiratory flow were measured three times a week (Monday, Wednesday, and Friday). The measurements were performed until the first extubation or the death of the patient. Tracheal suctioning was performed before each series of measurements.

The ETT area profile, A{x}, was inferred by the two-microphone acoustic reflection method as previously described. 16–19,21,22Briefly, the device used consisted of two microphones (piezoresistive pressure transducers 8510-B; Endevco France, Le Pré Saint-Gervais, France) and a horn driver mounted on a wave tube (overall length, 22 cm; ID, 0.35 cm) connected to the ETT connector. The other end of the wave tube was open to the atmosphere, allowing the subject to breathe spontaneously. A pressure wave is generated by the horn driver. The resulting pressure in the wave tube is recorded via  the microphones in a microcomputer to infer A{x} with a spatial increment, ΔL, of approximately 0.4 cm. This setup is derived from a two-microphone device furnished by Benson Hood Laboratories (Pembroke, MA) with a homemade wave tube (ID, 8 mm). The area was inferred using a deconvolution algorithm in the frequency domain. Such a setup was shown to be able to measure the area of a 50-cm-long ETT. 19The measurement procedure consists of connecting the wave tube, recording the pressure, and unconnecting the wave tube and takes less than 45 s. The ETT inner volume, Vol, is estimated by integrating the inner area along the longitudinal axis:

where L is the ETT length. The resistance of the ETT is estimated by using the Blasius formula along the ETT, as previously published. 21 

where L is the ETT length, μ is the viscosity, ρ is the density, and V̇ is the inspiratory flow.

Results were expressed as mean ± SD or as the number of patients affected. Because the duration of ventilation and the number of measurements varied among patients, we considered for each patient the mid duration of MV and the total duration of MV period studied (defined by the period from the first day of MV to the last), for comparison of the reduction in ETT volume and the increase in ETT resistance. ETT volume and resistance were expressed as a percentage of the in vitro  value obtained with a clean tube. We used the value obtained with a clean tube because in several patients the first value did not coincide with the first day of MV. To ensure that the individual differences in duration of the study period did not affect the results, we also calculated the mean daily reduction in ETT volume as the ratio of the total reduction to the number of days of the study period. Comparisons between groups (HH vs.  HME) were performed using the Mann–Whitney U test for continuous variables and the chi-square test with Yates correction when appropriate for categorical variables. A P  value below 0.05 was considered the limit of significance.

Forty-eight patients were initially enrolled during the first period (HH), and 34 were enrolled during the second period (HME). Fourteen patients in the HH group and 8 in the HME group were ventilated for less than 48 h (7 HH group and 4 HME group patients died, and 7 and 4, respectively, were extubated). Therefore, 34 patients in the HH group and 26 patients in the HME group were considered for the evaluation.

The two groups did not differ in terms of demographics, the presence of chronic obstructive pulmonary disease, markers of the severity of illness, mean fractional concentration of oxygen, and minute ventilation averaged throughout the study (table 1). Also, there was no significant difference between the HH and HME groups for half the duration (5.5 ± 3.3 vs.  4.8 ± 3.3 days; P = 0.40) or the total duration (10.5 ± 5.8 vs.  9.6 ± 6.3 days; P = 0.42) of the study period. All patients were orotracheally intubated with HI-LO tubes (Mallinckrodt Medical, Athlone, Saint-Louis, Ireland). ETT diameters ranged from 7.0 to 9.0 mm. Between the two groups, the average ETT diameter did not significantly differ (8.0 ± 0.3 vs.  7.9 ± 0.3; P = 0.55), nor did the value obtained with a clean tube in vitro  for ETT volume (14.0 ± 1.8 vs.  13.9 ± 1.8 ml; P = 0.79) or ETT resistance (2.6 ± 0.6 vs.  2.8 ± 0.6 cm H²O · l⁻¹· s⁻¹; P = 0.42).

Three patients experienced substantial ETT obstruction (two using the Fisher & Paykel humidifier and one using a DAR HME). Two patients, one in each group, needed emergency reintubation, and in one patient (No. 25) in the HH group, extubation and reintubation was decided because the acoustic method had shown a 10.3% ETT volume reduction and a 30.5% increase in Blasius resistance. Serial measurements obtained from this patient with the acoustic method during the course of MV are presented in figure 1. The incidence of obstruction expressed per day of MV was 2 in 441 for HH and 1 in 285 for HME. For those three patients, only the first episode of MV was used in the analysis.

At mid duration of the study period of MV, no statistical difference was observed in volume reduction between the two devices. At the end of the study period, the mean reduction in ETT volume was significantly greater in the HME group than in the HH group (−5.1 ± 2.5 vs. −3.3 ± 2.9%; P = 0.008;fig. 2). Expressed as a mean daily reduction in ETT volume, the difference was also significant (−0.4 ± 0.3 vs. −0.8 ± 0.8%/day; P = 0.008).

Similar results were observed in the progressive increase of the ETT Blasius resistance during MV. No significant difference was observed between the two groups at mid duration of MV. At the end of the study period, the mean increase in ETT Blasius resistance was significantly greater in the HME group than in the HH group (8.4 ± 12.2 vs.  19.4 ± 17.7%; P = 0.001;fig. 3). When expressed as the mean daily increase in ETT resistance, it was also significant (1.0 ± 1.4 vs.  3.2 ± 3.7%/day; P = 0.002).

The main finding of this study is that during long-term MV, i.e. , after more than 4 or 5 days, the type of humidification system used may significantly modify the ETT patency. Accordingly, we observed after 9 or 10 days of MV a significantly greater reduction in ETT volume and a greater increase in Blasius resistance with the HME than with the HH.

These results confirm that permanent deposits of bronchial secretions on the inner wall of the ETT can gradually obstruct the lumen of the ETT. This deposit was less pronounced with the HH than with the HME. The HH system is an active device during all phases of the respiratory cycle and can deliver heated humidity to the airway. By contrast, the HME device is a passive device that keeps humidity and heat in the airway. Mucociliary transport can be altered by changes in ciliary function or physical properties of the mucus, which depend in part of the gas conditioning in mechanically ventilated patients. Konrad et al.  23showed that mucociliary transport can be altered after 3 days of MV in intubated patients and suggested that several weeks of MV of patients with severe respiratory insufficiency would have more pronounced effect on ultrastructural changes. Recently, Nakagawa et al.  24reported in a randomized controlled trial that the type of humidification device influenced mucociliary transport by cough. In this study, the authors 24studied the transportability by cough in vitro  and found that it was diminished after 72 h of MV in the HME group. These data 23,24support the fact that the impact of the humidification device on the mucosal structure and function can exist and may need several days of MV to be detected. 5 

Endotracheal tube patency was evaluated because its status can be influenced by progressive deposit of secretions, the quality of which is influenced by gas conditioning. ETT patency thus indirectly reflects the quality of gas conditioning in the patient airways. Several studies presented in table 2have already shown the presence of partial obstruction of the ETT after few days of MV. 9,16,21,25–29 

Villafane et al. , 9reported that during prolonged MV (after 6 ± 3 days of MV), the patency of artificial airways can be affected by the type of airway humidification. In this study, 9the overall reduction in ETT mean diameter evaluated by the hydraulic method was significantly greater with a hydrophobic HME than with an HH (−6.5 vs. −1.5%; P < 0.01). Contrary to the current study, in the study of Villafane et al. , 9no significant difference was observed in the overall reduction in ETT mean diameter between a hygroscopic HME and the HH (−2.5 vs. −1.5%; not significant). We hypothesize that the reason for such a discrepancy between the two studies is the mean duration of MV. The current study looked at prolonged periods of MV (10 ± 6 days), greater than in the study of Villafane et al.  9(6 ± 3 days). Indeed, in the current study, we did not find any difference between the two devices at mid duration of the study period of MV (5 ± 3 days), this duration being quite similar to the total duration of the study of Villafane et al.  9Both studies suggest that there is no substantial difference between the two types of humidification devices used, as long as the patients are mechanically ventilated for a short duration (5 or 6 days).

Diehl et al.  29evaluated the changes in the work of breathing induced by tracheotomy in eight ventilator-dependent patients and compared the characteristics of the removed ETT with those of a new clean ETT. The authors observed significant extra work of breathing induced by the removed ETT, which had been left in place for 31 ± 19 days of MV. There was an increase in resistive work of breathing of 28% over the work measured with a clean new ETT. These studies suggest that the mean diameter or volume of ETT trended downward from day to day, and a marked difference was observed after a long-term period of MV. Progressive reduction in the ETT diameter accompanies permanent mucus deposition. Until the hydraulic 9and acoustic methods, 16,19no accurate method was available to monitor its accumulation (table 2). Peak airway pressure, which has frequently been used as an index of gas flow resistance in the ventilated patient, does not allow one to distinguish the contribution of the ETT resistance to the overall resistance and is highly dependent on flow pattern and patient spontaneous activity. 22,25,26 

The current investigation is the first study to directly assess the effects of HME and HH in vivo  on the patency of the ETT in patients undergoing prolonged MV. The mean values for volume reduction for the period of the study were moderate (< 10%), though in some patients a reduction greater than 15% was observed (n = 1 in the HH group, n = 6 in the HME group; P = 0.02). Similarly, an increase in Blasius resistance by more than 30% was observed in several patients (n = 3 in the HH group, n = 7 in the HME group; P = 0.06). ETT patency reduction can increase the mechanical load of respiration in ventilated patients and may markedly reduce the efficiency of pressure-targeted modes because part of the ventilatory assistance is used to overcome the resistive load due to the ETT. As shown in a previous study by Heyer et al. , 21a correct estimation of the ETT-related work of breathing may be especially important when patients are weaned from MV. Indeed, the level of pressure support and the tolerance of a T piece trial are among the main predictors used to determine whether tracheal extubation is feasible. Also, the use of an HME increases dead space and adds an elastic load to respiratory work. 10,30–32HMEs are increasingly employed because they are easy to use, require less manipulation, and may be less expensive than HHs. 14,33Uncertainties have persisted regarding the efficacy and the potential risks of HMEs. Indeed, tracheal obstruction with the first generation of HMEs exhibiting only hydrophobic properties 6–8occurred in some studies because of inadequate airway humidification. The low incidence of obstruction expressed per day of MV observed in our study (3 in 726) confirms that this risk is no longer a major concern with the new HMEs. HMEs provide bacterial filtration 34and minimize circuit contamination. They may need to be changed only every 5 or 7 days, 14,33without any adverse effects for the patients, which may reduce costs. Our study, however, raises concern about their efficacy over prolonged periods of MV, i.e. , more than 4 or 5 days. The heated-wire circuits with the HH eliminate condensation in the circuit and the need for water traps and their maintenance, which has been considered as a major shortcoming of this type of system, but increase their cost. For the HH, the 40°C setting ensures a 37°C gas temperature when the gas reaches the patient because cooling occurs across the 8- to 10-cm flexible catheter mount and the 6 cm of protruding ETT. This follows the recommendation of the manufacturer and is supported by the literature as mimicking normal upper airway function. 5This study shows that obstruction still exists with HHs, however, and it may be worthwhile to explore the causes of this process.

The current study also indicates that the acoustic reflection method can be performed in routine clinical practice to monitor the ETT patency. In one patient, the acoustic measurement prompted the decision for reintubation, to avoid a risk of emergency obstruction. A similar technique has recently been proposed for checking the ETT placement. 35 

One limitation of this study is its nonrandomized nature. To assess the respective efficacy of the two systems, performing a randomized study design was first considered. However, ventilating patients with the two types of humidification systems in the department at the same time was a source of confusion for the personnel and made the necessary training for HH manipulation more difficult. Therefore, we decided against performing a randomized controlled trial and preferred to study two successive periods. No changes in the policy of MV or in the prevention, diagnosis, or management of nosocomial pneumonia occurred during this period. The only concern could be that two different periods of the year may be associated with different climatic conditions. Although this has not been demonstrated, one could imagine that this may influence the two types of humidification systems. The two periods, however, included a similar duration of cold periods (winter season) and warm periods (summer season). Several recent reports have suggested that such a study design could give results similar to those of randomized controlled trials. 36 

Our results confirm the findings of Villafane et al. , 9who reported that during prolonged MV, the patency of artificial airways can be affected by the type of airway humidification. More importantly, the results of this study suggest that during long-term MV (i.e. , around 9 or 10 days), HH should be the technique of choice for adequate humidification. For a short period of MV, HME may be used safely.

The authors thank the physiotherapists of their unit, Nicole Jackson, R.R.T., and Sylvie Lely, R.R.T., Hopital Henri Mondor, Créteil, France, for their invaluable help in this study. They also thank the nurses and the rest of the personnel of the intensive care unit for their active participation. This article is dedicated to the memory of Alain Harf (M.D., Ph.D., Professor, Service de Réanimation Médicale, Service de Physiologie, Hôpital Henri Mondor, Assistance Publique-Hôpitaux de Paris, Université Paris 12, and Institut National de la Santé et de la Recherche Médicale Unit °492, Créteil, France), who died before seeing this work published.