Abstract
BACKGROUND: Breathing through a tracheostoma results in insufficient warming and humidification of the inspired air. This loss of air conditioning, especially humidification, can be partially restored with the application of a heat and moisture exchanger (HME) over the tracheostoma. For medical professionals, it is not easy to judge differences in water exchange performance of various HMEs owing to the lack of universal outcome measures. This study has three aims: assessment of the water exchange performance of commercially available HMEs for laryngectomized patients, validation of these results with absolute humidity outcomes, and assessment of the role of hygroscopic salt present in some of the tested HMEs.
METHODS: Measurements of weight and absolute humidity at end inspiration and end expiration at different breathing volumes of a healthy volunteer were performed using a microbalance and humidity sensor. Twenty-three HMEs from 6 different manufacturers were tested. Associations were determined between core weight, weight change, breathing volume, and absolute humidity, using both linear and nonlinear mixed effects models.
RESULTS: Water exchange of the 23 HMEs at a breathing volume of 0.5 L varies between 0.5 and 3.6 mg. Both water exchange and wet core weight correlate strongly with the end-inspiratory absolute humidity values (r2 = 0.89/0.87). Hygroscopic salt increases core weight.
CONCLUSIONS: The 23 tested HMEs for laryngectomized patients show wide variation in water exchange performance. Water exchange correlates well with the end-inspiratory absolute humidity outcome, which validates the ex vivo weight change method. Wet core weight is a predictor of HME performance. Hygroscopic salt increases the weight of the core material. The results of this study can help medical professionals to obtain a more founded opinion about the performance of available HMEs for pulmonary rehabilitation in laryngectomized patients, and allow them to make an informed decision about which HME type to use.
- total laryngectomy
- heat and moisture exchanger
- HME
- comparison study
- water exchange
- humidity
- weight
- quality of life
Introduction
After total laryngectomy, breathing through the nose is inevitably replaced by breathing through the permanent tracheostoma, whereby inspired air is no longer optimally conditioned before reaching the trachea. The colder and dryer inspired air leads to pulmonary complaints such as increased mucus production and excessive coughing, and causes frequent airway infections.1 For pulmonary rehabilitation after total laryngectomy, heat and moisture exchangers (HMEs) are widely used to partially restore (this lack of) air conditioning. Patients who use these devices consistently have significantly fewer complaints of sputum production and coughing and report a better quality of life.2,3
HMEs consist of a functional core material (usually a foam, sometimes coated with hygroscopic salt) contained in a cassette. The functional core material traps and releases a small amount of water during exhalation and inhalation, respectively. With different types of HMEs currently available, the literature is growing rapidly on the effect of HMEs on clinical complaints,4–9 influence on respiratory function,2,10–12 in vivo humidity measurements,13–15 and comparison of in vitro measurements.16,17 However, no comparative data have been published as yet on the performance of the whole range of HMEs available for laryngectomized patients today. In vivo studies are not suitable for measuring a large variety of HMEs, as the measurements are too time-consuming and burdensome for patients. Furthermore, in vivo humidity measurements are technically difficult to perform.13,15 In vitro measurements, where either a mechanical lung model according to the ISO standards (International Standard Organization ISO 9360:2000 and 2001) or hygrometry is used, would allow the reliable assessment of a wide range of HMEs when performed in the same test rig; the heat and moisture capacity of these configurations, however, are probably not fully representative of human breathing.18–20
These issues were overcome with the recent development of an ex vivo method that enables measurement of water exchange performance of a variety of HMEs within a short timeframe, without the need to trouble patients, while still being universally feasible.21 In this method, the weight of an HME is measured twice: once at the end of inspiration and once at the end of expiration. The weight difference between end inspiration and end expiration (as function of the breathing volume) represents the water exchange performance during the breathing cycle.
The present study has three aims: ex vivo assessment of the water exchange performance of commercially available HMEs for laryngectomized patients, validation of these results with absolute humidity outcomes, and assessment of the role of hygroscopic salt present in some of the HMEs tested.
QUICK LOOK
Current knowledge
Breathing through a tracheostoma bypasses the normal heat and humidifying mechanisms of the upper airway, resulting in insufficient conditioning of the inspired air. A heat and moisture exchanger (HME) placed over the tracheostoma can increase tracheopulmonary humidity, improving patient comfort.
What this paper contributes to our knowledge
Twenty-three HMEs designed for laryngectomized patients demonstrated a wide variation in water exchange performance. The efficiency of water exchange correlated well with the end-inspiratory absolute humidity. Wet core weight proved to be a good predictor of HME performance.
Methods
The study was approved by the institutional review board of the Netherlands Cancer Institute-Antoni van Leeuwenhoek.
HME Devices
Twenty-three different commercially available HMEs intended for pulmonary rehabilitation of laryngectomized patients were ordered via local distributors. In vitro HME water loss and pressure drop values (ISO standards)18,19 were available for HMEs produced by two manufacturers. The other manufacturers, who were all approached for sharing the ISO standard values for their various HMEs, were unwilling or unable to provide these data. The available manufacturer details of the HMEs tested are shown in Table 1.
Most HMEs consisted of a core material and a relatively simple cassette. Some HMEs had a more elaborate cassette design potentially increasing HME performance with additional water storage: the four hands-free HMEs; the Provox Micron (Atos Medical, West Allis, Wisconsin), which has an additional electrostatic filter that, although hydrophobic, influences HME performance by preheating the inhaled air22; the Cyranose HME (Ceredas, Antony, France), which contains a metallic grid; and the Blom-Singer humidifier holder (InHealth Technologies, Carpinteria, California), which has a considerably larger cassette.
Water Exchange and Humidity Measurements
Water exchange (weight changes between inhalation and exhalation and vice versa), end-inspiratory absolute humidity (AHinsp), and breathing volume were measured using ex vivo weighing as described previously.21 In the test configuration (Fig. 1), the HME was mounted on a T-tube containing a fast heated capacitive hygrometer, absolute humidity sensor with a response time of 0.1–0.2 s,23 and a spirometer (MLT300 Flowhead, ADInstruments, Oxfordshire, United Kingdom). A healthy volunteer (first author, CvdB) breathed through the spirometer. The maximum flow of the spirometer was 300 L/min, so the volunteer was instructed to breathe at normal speed; in case of doubt, the flow signal could be inspected. For the weight measurements, a microbalance (MC210P, Sartorius, Göttingen, Germany; accuracy within 0.1 mg) was used. Before the start of weight measurements, each HME was prepared by the volunteer breathing through the HME until equilibrium of water saturation was reached. The length of this conditioning period varied between the different HMEs and was determined for each HME separately. HME weight measurements were performed 25 times, alternating at the end of inspiration and at the end of expiration, using three different breathing patterns (tidal, shallow, and deep breathing). Between each weight measurement, at least five breathing cycles (at tidal volume) were performed to recondition the HME. Of each HME type, 3 different HMEs were measured on separate days. The 23 HME types were measured in a randomized order. The first weight change was discarded as a result of the differing conditioning periods as well as the weight change between different breathing patterns, resulting in 21 weight changes for analysis per HME. Weight data points deviating > 75% from the previous and following weights in the data sequence were considered outliers and therefore excluded.
Core Weight and Water Uptake in Humid Conditions
In addition to the ex vivo weighing method, weighing of the core material of the HMEs themselves in different air humidity conditions was carried out to assess the amount of water accumulated in the devices, identical to the method described in 1988 by Ploysongsang et al.24 The weights of the HME core material were measured after conditioning in completely dry air and at different higher air humidity conditions up until approximately 55% relative humidity (RH) at 21°C. Higher levels of humidity were not used, to avoid nonlinear behavior in the hygroscopic HMEs.25 HMEs were kept in room conditions (between 30% and 40% RH), and long storage or storage at high humidity levels was avoided to prevent plastics of cassette and core material from slowly (time scale, days) absorbing water and getting heavier.
The HMEs were placed in each condition for approximately 4 h before being weighed. For each HME, the weight increase gradient (weight increase with increasing RH: mg/% RH) was calculated from the HME weight increase as a function of RH. Weights of the core material were obtained by dismantling the HME and subtracting cassette weight. Dry core weight is the core weight at 0% RH. Wet core weight is the core weight of HME under operating conditions (see conditioning above). The mean outcome of 3 HMEs per type was used for analysis.
Additional Equipment
HMEs were placed in an airtight box during weight measurement to prevent water evaporation. A freeze drying chamber (FDC206, SpeedVac system, Savant Instruments, Farmingdale, New York) was used to create a vacuum for 0% RH conditioning. For the wet conditions, a Plexiglas climate room containing a electromotor-driven propeller for air mixture (26 × 42 × 16 cm2 as described previously by Zuur et al)15 was used. Room conditions were recorded during the measurements, using a commercial calibrated humidity sensor (Testo, Almere, Netherlands) with an accuracy of ± 0.6°C and ± 2.5% RH. Calibration of the absolute humidity sensor was performed as described previously,15 using the Testo sensor as reference humidity sensor. Calibration of the spirometer was performed according to the recommendations of the manufacturer. Spirometer data were recorded and analyzed with Powerlab software (ADInstruments), and humidity values were registered with data acquisition software (Acquis 2.8, Anästhesie-Technik, Göttingen, Germany) and exported to a spreadsheet (Excel, Microsoft, Redmond, Washington). The body temperature of the healthy volunteer was checked at every ex vivo measurement with an electronic aural thermometer (Genius2, Kendall, Tyco Healthcare Group, Mansfield, Ohio).
Data Normalization and Statistical Analysis
Weight changes and AHinsp data were normalized to a reference ambient humidity of 5 mg/L as described previously.21 A summary of the normalization formulas is given in Appendix 1 (see the supplementary materials at http://www.rcjournal.com).
Assessment of the association between water exchange and average breathing volume was determined using a linear mixed effects model for each HME type (three HMEs of one type together). The association between AHinsp and inspiratory breathing volume was determined using an exponential decay nonlinear least squares regression as described previously.21
For the core weight experiments, weighted r2 values were calculated using weighted Pearson correlations with inverse variances as weights. Aikake information criterion was used to compare the associations between inspirational absolute humidity and both wet and dry core weights in two weighted linear regressions.
Results
The water exchange (weight change between inhalation and exhalation) as a function of the breathing volume for all HMEs is shown in Figure 2. The graphs represent the exchange of water that was condensed onto and evaporated from the HME during respiration per HME type. Of the total weight change data points (1,449), 21 (1.4%) points were excluded as outliers according to the criterion given in the methods section. The parameters of the model fits shown in Figure 2 are supplied in Appendix 2 (see the supplementary materials at http://www.rcjournal.com). Ambient absolute humidity during the measurements ranged between 7–12 mg/L, and averages and SD per HME type are given in Appendix 2 in Table 5 (see the supplementary materials).
A breathing volume of 0.5 L was chosen for comparison between the HMEs (vertical dashed line, Fig. 2), because this is the average tidal breathing volume previously reported for laryngectomized patients.12 For each of the tested HMEs, the water exchange at the breathing volume of 0.5 L can be found in Table 2. The water exchange capacity ranged from 0.5 to 3.6 mg. Most HME types had a standard error of about 0.1, but the standard error of the hands-free HMEs tended to be slightly larger (up to 0.16).
Figure 3A presents an overview of all HME performances from Table 2 for the water exchange including the standard errors (vertical bars) ranked from high to low performance.
In Table 2, the values of AHinsp at a breathing volume of 0.5 L are reported. Details of the measurements of AHinsp and the parameters of the exponential decay model21 are given in Appendix 2, Table 4 (see the supplementary materials at http://www.rcjournal.com). Figure 3B shows the HME performances ranked according to AHinsp.
The correlation between water exchange and AHinsp values for each HME tested is shown in Figure 4. The calculated inverse variance weighted r2 is 0.89. See Table 5 and Fig. 9 in the supplementary material for an in-depth analysis of the correlation shown with a Bland-Altman plot.
The last 4 columns of Table 2 list the results for the weight measurements of the HMEs in different humidities: the weight increase gradient (weight increase per % RH), dry core weight, wet core weight, and water uptake (the difference between wet and dry core weight). The correlations between AHinsp and wet core weight are shown in Figure 5 for all HMEs, as well as for a selection of HMEs with relatively simple cassettes that are likely to have no relevant HME effect of their own (see selected HMEs in methods section). The overall weighted variance r2 was 0.79, and 0.87 for the simple cassette HMEs. The correlations between AHinsp and dry core weights for all HMEs and the simple cassette HMEs were 0.69 and 0.85, respectively. The Aikake information criterion of the weighted linear regressions of AHinsp as predicted by wet core weight was 29.9, whereas the Aikake information criterion when using dry core weight was 38.0, suggesting that wet weight is a better predictor of AHinsp. Water exchange also correlates with core weight (see Table 2), but the r2 values are lower because of the larger standard error of the water exchange values.
The relation between HME water uptake and water exchange values is illustrated in Figure 6 for the simple cassette HMEs. The data points for the HMEs with a hygroscopic (salt-containing) core material and the types specified by the manufacturer as not containing such material are marked differently (see Table 1 for details). It is clear that, for hygroscopic HMEs, the water exchange improves when more water is absorbed onto the HME.
Discussion
In this study, the ex vivo water exchange performance of 23 presently obtainable HMEs was assessed, the results of which were then validated with absolute humidity measurements at the end of inspiration (AHinsp).
Most strikingly, the performance of the various devices proved to be highly heterogeneous. The HME water exchange capacity (of the 23 HMEs tested) ranged between 0.5 and 3.6 mg at a tidal volume of 0.5 L. Comparison studies of other HME types (for temporarily tracheostomized patients and in mechanical ventilation settings) also showed a considerable variability in the performance of HME values.20,26,27
The high correlation (Fig. 5, r2 = 0.89) between water exchange and AHinsp confirms our previous results based on only 6 different HMEs from one manufacturer, which shows that water exchange is a valid measure for HME performance and means a further and broader validation of the ex vivo weighing method.21 The present study might have revealed even more interesting results if the manufacturers' ISO specifications for all tested HMEs had been available for comparison with the ex vivo results. Unfortunately, such a comparison was not possible, because the manufacturers approached were unable or unwilling to provide these data, and because the data that were available were measured under different ISO conditions (see Table 1). We have therefore used the AHinsp as the accepted standard for validation, as this quantity has been validated using in vivo studies against ISO observations, but it is possible that the water exchange actually is a better quantity because it measures the average performance over the full inspiration and expiration, whereas the AHinsp is only measured at end inspiration (see also Table 5 and Fig. 9 in Appendix 2, available in supplementary materials at http://www.rcjournal.com). In particular, for the best performing HMEs, the ranking according to the water exchange outcome differs from the ranking according to AHinsp. This might be due to technical difficulties with either of the methods, but it might also be a true result because the water exchange measures the averaged performance over a breathing cycle and AHinsp is an instantaneous observation.
For 2 Atos Medical HMEs, we found slightly different water exchange values from those reported previously, with notably slightly higher values for the Provox Normal and XtraFlow, resulting in a better correlation with the AHnsp (see Table 2).21,28 The present results are indeed probably more reliable as a result of a learning curve effect associated with the application of any new technique and, in particular, because we were more careful in this study not to exceed the flow limitation of our spirometer.
The current study has been performed with a mouth-breathing healthy volunteer instead of stoma breathing of a laryngectomized subject. In both situations, the expired air is almost completely saturated with water, but the saturation is more complete in the volunteer (99% RH vs 92% RH).29,30 The use of a laryngectomized patient might have slightly influenced the absolute results, but the relative ordering of the HMEs would not be different, as can also been seen in Figure 7, where the results of this study are compared with actual in vivo observations.
This ex vivo method enables an HME comparison study without performing measurements clinically in laryngectomized patients, which with 23 different HMEs would be a near impossible undertaking. As mentioned in our previous papers, an advantage of the ex vivo HME-weighing method is that it can be carried out by a single volunteer. The reason for this is that the spirometer included in the test configuration registers all breathing variations of the volunteer, so that the inhaled volume—the primary confounder when comparing HMEs—can be taken into account properly and will not be unintentionally influenced by the volunteer. Moreover, in a previous study with 6 different volunteers, the intervolunteer variation was negligible.21
This study also provides some additional insight into how HMEs function. Although clinicians might consider it to be a rather simple device (just a piece of foam), the theory of HME performance is actually quite a complicated combination of thermodynamics and flow mechanics. An HME must be able to store (and release) a considerable amount of heat, required to condense and evaporate the water in the expired/inspired air. If the HME is unable to do this, the temperature increase/decrease inside the HME will slow the condensation/evaporation process.31,32 The most important parameters of the HME core material are therefore probably heat capacity (to store evaporative heat), its structure (to ensure sufficient contact with the air flowing through the device), and heat conductivity. Heat capacity is determined by the heat capacity index of the chosen material and its quantity. Figure 5 (see also Table 2) shows that the amount of core material predicts HME performance very well for HMEs that do not have an additional HME effect from the cassette (point above the fit line in Fig. 5). Points below the line show core materials that do not participate in the HME effect (dead weight). However, most points are on the fit line, which suggests that all core material participates in the absorption of evaporative heat. The observed ex vivo variations therefore are likely a result of other parameters, such as differences in the heat capacity index of the core material and/or to HME effects of the cassette. This correlation points toward total heat capacity, determined by the amount of core material, as the most important HME design parameter. HME optimization requires that as much (thin) core material as possible is stored in the limited available space (the HME should stay cosmetically acceptable) without an unacceptable increase in breathing resistance.
This draws attention to the role of hygroscopic salts that are used to improve the performance of some HMEs. The best performing HMEs in this test are hygroscopic HMEs. Two of the HMEs that are supposed to be hygroscopic according to manufacturers' specifications (Servox and Prim-Air Phon I high flow, Table 1) had only an average performance and a weight increase gradient close to zero (Table 2), while showing no water uptake (Fig. 6); these HMEs behave as if they do not contain any hygroscopic salts in/on the core material. The fact that the performance of hygroscopic and nonhygroscopic HMEs with the same wet core weight is comparable also refutes the common notion that hygroscopic salt plays a part in or is required for the quick storage and release of water during breathing. Figure 6 shows the true explanation for the function of hygroscopic salt. It increases the weight of the HME by attracting a layer of water, and, because water has a high heat capacity, the performance of the HME improves. The amount of water can be large; for the Provox XtraMoist HME, for instance, water constitutes 50% of the total wet core weight.
It is important to note that, although a large water uptake enhances the HME performance, it can also have undesirable side effects. Breathing resistance may increase if the pores in the foam become too small, and, in extreme cases (such as entering very cold outside air from a warm room), excessive water might condense in the HME and may start dripping into the trachea or on the clothes/skin.
To understand the meaning of these ex vivo results for clinical practice, the issue is to what extent these HMEs can bridge the physiological humidity gap between nose and stoma breathing.13,30,33–35 The physiological tracheal climate during nose breathing is known for healthy volunteers and for head and neck cancer patients with a temporary tracheotomy.29,30 As laryngectomized patients are head and neck cancer patients, the subglottic humidity value during nose breathing in this patient group (29.3 mg/L at 1 cm behind the temporary tracheostoma)30 can be considered the target humidity value in the upper trachea of laryngectomized patients. In Figure 8 (available in the supplementary materials at http://www.rcjournal.com), the HMEs tested are placed in order of their AHinsp values next to the target AHinsp value of nose breathing (all values are standardized to a reference ambient humidity of 5 mg/L). Also included in the figure are the considerably higher in vivo values of three HMEs published previously.12,22 These are due to the fact that the trachea itself has a considerable HME effect too, which influences the in vivo measurements where the humidity is measured in the stoma/trachea approximately 1 cm behind the HME. This effect is absent in the ex vivo setup, where the absolute humidity sensor was placed downstream of the HME outside the body. Extrapolating the trend of the in vivo values to those of the ex vivo measurements, as shown in Figure 7, one can see that the best performing HMEs come closer to the target value for optimal physiological climate conditions in the trachea. However, there is clearly still some room for improvement of the water exchange capacity even with the best performing HMEs.
The better understanding of HME performance achieved in the present study should help professionals in choosing the right HME for their patients. It might furthermore trigger the development of new HMEs for laryngectomized patients with performance that restores the physiological situation in the trachea even more, leading to an even greater reduction of clinical complaints and improvement of laryngectomized patients' quality of life.
Conclusions
The 23 HMEs for laryngectomized patients tested show wide variation in water exchange performance. Water exchange correlates well with the end-inspiratory absolute humidity outcome, which validates the ex vivo weight change method. In addition, (wet) core weight is a good predictor of HME performance for HMEs with a simple cassette. Hygroscopic salt increases the weight of the core material and therefore the performance of the HME. The results of this study can help medical professionals to obtain a more founded opinion about the performance of available HMEs for pulmonary rehabilitation in laryngectomized patients, and allow them to make an informed decision on which HME type to use.
Acknowledgments
We acknowledge Klaus Züchner PhD for providing the heated capacitive hygrometer.23
Footnotes
- Correspondence: Frans JM Hilgers MD PhD, Department of Head and Neck Oncology and Surgery, Netherlands Cancer Institute-Antoni van Leeuwenhoek, Plesmanlaan 121 1066 CX Amsterdam, Netherlands. E-mail: f.hilgers{at}nki.nl.
Supplementary material related to this paper is available at http://www.rcjournal.com.
This research was funded in part by an unrestricted research grant of the manufacturer of some of the tested HMEs, Atos Medical. The study sponsor had no involvement in the study design; in the collection, analysis, and interpretation of data; in the writing of the manuscript; or in the decision to submit the manuscript for publication. The authors have disclosed no other conflicts of interest.
- Copyright © 2014 by Daedalus Enterprises