High frequency chest compression (HFCC) is used for treatment and prevention of the lung diseases characterized by impaired mucus clearance and/or cough, where patients are at risk for acquiring acute bronchitis or pneumonia. The HFCC treatment frequencies may be prescribed according to the manufacturers' generic guidelines or may be determined for each individual patient by a “tuning” method that measures, at the mouth, the air volume displacement and the associated airflows produced at each frequency. Tuning is performed while the patient is breathing normally during the HFCC system operation. After measurements for several breaths at one frequency have been collected, the program randomly selects and measures another frequency until the entire frequency range of the machine being tuned has been sampled. Frequencies range from 6 to 21 Hz for the sine waveform machines and from 6 to 25 Hz for the square waveform machines. Each group of flow signals is digitized and analyzed by the program. For each frequency, the HFCC flow velocities and volumes are computed and averaged. These average flows and volumes are rank ordered; the three frequencies with the highest flows and the three frequencies producing the largest volumes are selected for prescription. If the same frequency is selected as one of the three best frequencies for both flow and volume, the next ranked frequency is selected randomly for flow or volume. Significant differences exist between patients and HFCC machines. In a series of 100 cystic fibrosis (CF) patients with varying degrees of lung disease, we found that the best-ranked frequencies varied from patient to patient and did not correlate with patients' age, gender, height, weight, or spirometry parameters. With the sine waveform, the highest HFCC airflows were between 13 and 20 Hz 82% of the time and the largest HFCC volumes were between 6 and 10 Hz 83% of the time. With the square waveform, both the highest average HFCC flow rates and the largest volume average HFCC displacements were between 6 and 14 Hz. Nevertheless, in this sample of 100 consecutive tunings, every frequency from 6 and 20 Hz was a best frequency for at least one patient. These findings provide the basis for recommending a tuning protocol to be used for prescribing frequencies with the various HFCC machines, because they are different from one another. If a patient's tuning cannot be done, it may be useful to prescribe the best frequencies based on the waveform machine he or she uses.
That there might be a best frequency—or a number of best frequencies—for high frequency chest compression (HFCC) was suggested by Malcolm King and associates,1 who studied several frequencies in nine dogs and found 13 Hz to be optimum. In another study2 of seven dogs treated for 30 minutes of HFCC at 13 Hz, researchers concluded that HFCC was effective in enhancing peripheral and central mucus clearance. In a third study3 of dogs, investigators concluded that “high-frequency ventilation by rapid chest wall compression enhances tracheal mucus clearance when compared with spontaneous breathing, whereas high-frequency oscillation at the mouth does not.”
Hansen and Warwick confirmed the value of HFCC therapy in studies of patients with cystic fibrosis (CF) with an HFCC device that produced a mean volume of cleared mucus of 3.3 mL per session, compared with 1.8 mL for a conventional manual therapy.4 They developed a method of measuring, at the mouth, the induced airflow and the integrated volume displacement with HFCC compression frequencies from 5 to 25 Hz.5 During their “tuning” of the square wave pulses over this range of frequencies, they found that no frequency was always best and that each frequency was sometimes best. They also found the best frequencies for induced airflows were often different from the best frequencies for volume displacement.
Their arbitrary compromise was to select the three frequencies with the largest average volume displacements and the three with the highest average airflows and to use each frequency for five minutes of HFCC compression based on the time noted for King's dogs. The effectiveness of this decision was confirmed in a clinical study6 where 94% of the regression lines for vital capacity (FVC) and one-second forced expiratory volume (FEV1) became more positive during self-administered HFCC therapy, as compared with slopes when manual chest physical therapy was used. Because we discovered a different tuning pattern of best frequencies with the change from the square to sine waveform machines, we designed this comparison of 100 consecutive tunings with both waveforms used for treatment of our CF patients.
HFCC has been shown to enhance mucus clearance and to help maintain—even to improve—pulmonary function in patients with chronic lung diseases. HFCC therapy has been applied to the chest wall with square and sine waveforms delivered to an inflatable vest. Both waveform compressions produce transient airflow bias in the airways toward the mouth. Mucus clearance is enhanced by several mechanisms:
The airflow spikes produce a “sweeping” effect in the airways.7
The high frequency oscillations transmitted to the airways break up both DNA and mucus.8
The high frequency oscillations can act as a physical mucolytic.9
The oscillation of the airways dislodges adherent mucus from the bronchial walls.10
The water content, and thus the fluidity, of secretions is increased by increasing intracellular production of ATP, which stimulates water transport into the airway lumen.11
Because the frequency at which the chest is compressed cannot be used to predict the displaced volume and the airflow spikes that result from each compression, we developed a tuning technique to measure, at the mouth, the airflow velocities and the simultaneous volume displacements that can be calculated by integration.5 An enhanced version of this program made it possible to easily identify the frequencies that produce the highest airflows and largest volumes.
The purpose of this study was twofold: first, to determine if any patient characteristics could predict the best frequencies and second, to see if there were any differences in the flows and volumes generated between the square and the sine waveforms. The hypothesis we tested was that the magnitude of the flows and volumes generated during HFCC are dependent on the frequency used, and patient characteristics such as age, sex, height, weight, and pulmonary function are independent of the device used.
Patients with CF who routinely use HFCC therapy for airway clearance were selected for this study. They were instructed to bring their own custom-fit vest for the study. The patients' weight, height, and age were recorded. Spirometry was performed with a Sensormedics Vmax22 pulmonary unit (Sensormedics Corp, Yorba Linda, CA).
Patients were tuned to sine and square waveform devices* in random order using the tuning session protocol in place at our institution.5 While performing HFCC, the patient wears a nose clip and breathes through a mouthpiece attached to a pneumotachograph (Hans Rudolph Inc, Kansas City, MO). The frequencies of the HFCC pulse waveform were switched in random order after four breaths were accepted by the computer program and until all frequencies over the effective frequency range of each waveform were tested: 6 to 20 Hz for sine waveform and 6 to 25 Hz for square waveform.
The flow signals were digitized and analyzed by a program that extracts the HFCC-induced flow signals that are superimposed over the tidal breathing cycle. The average HFCC airflow induced at each frequency was computed and then, by integration of the flow signal, the average volume of air displaced at each frequency was calculated. The three frequencies that produced the highest average flow and the three that produced the largest average volume were selected as the best frequencies for that pulse waveform.
The data were analyzed for each HFCC frequency by linear regression, with either airflow or volume as the dependent variable and FEV1, FVC, height, weight, gender, and age as the independent variables. Further analysis was done by multiple linear regressions with either airflow or volume as the dependent variable and HFCC frequency, FEV1, FVC, height, weight, gender, and age as the independent variables. Stepwise regression with a forward selection procedure was used to identify the models that contained those variables that better explained the variance in airflow and volume. The SAS statistical system (SAS, Inc, Cary, NC) was used for these analyses.
The composite characteristics of the 100 patients who participated in the study (56 women, 44 men) are detailed in Table 1.
Pattern of Frequencies Selected for Flow and Volume by Flow Waveform
The percentage of times each HFCC frequency was selected as a highest airflow and a largest volume for the sine waveform is shown in Figure 1. With the sine waveform, the highest airflows and the largest volumes are discordant. The largest volumes were found most often in the low frequencies and highest flows in the highest frequencies. For the sine waveform, the rank order of six frequencies most often selected by volume was 8, 6, 9, 7, 10, and 12. These frequencies were selected 25, 22, 14, 13, 6, and 5% of the time, respectively (85% total). For the sine waveform, the rank order of six frequencies most often selected by airflow was 19, 18, 20, 17, 15, and 13. These frequencies were selected 20, 17, 13, 9, 8, and 7% of the time, respectively (74% total). Thus, the best frequencies for flow and volume were completely discordant.
The percentage of times each HFCC frequency was selected as a highest airflow and a largest volume for the square waveform is shown in Figure 2. With the square waveform, the highest airflows and the largest volumes were concordant, and with both, the most often selected were in the low frequencies of 6 to 14 Hz. For the square waveform, the rank order of six frequencies most often selected by volume was 6, 7, 8, 9, 10, and 11. These frequencies were selected 30, 27, 19, 9, 5, and 3% of the time, respectively (93% total). The six frequencies most often selected by airflow were 7, 6, 10, 8, 11, and 14. These were selected 28, 18, 8, 8, 7, and 6% of the time, respectively (75% total).
This discordance and concordance of best flow and best volume frequencies yield differences in the number of times a best frequency for one parameter is a best frequency for the other parameter (Figure 3).
Linear regression of flow and volume data vs age, sex, and height showed no significant correlations, with regression lines nearly parallel to the abscissa and with R2 values ranging from .002 to .02.
Multiple Regression Statistical Correlates by Flow Waveform
Flow and Volume Statistical Correlates for the Sine Waveform
For airflow measured at the mouth, by multiple regression analysis a model including age, height, FEV1, and HFCC frequency as explanatory variables was statistically significant, with P = .0007 (P < .05 for all variables), but with an r2 of only .18. The highest airflows were generated in subjects with FEV1 greater than 120% of predicted.12
For volume measured at the mouth, by multiple regression analysis a model including height, FVC, and HFCC frequency as explanatory variables was statistically significant, with P = .001 (P < .05 for all variables), but with a r2 of only .16. HFCC frequencies in the low range resulted in the largest volume displacements in subjects with moderately compromised FVC between 50 to 80% of predicted.5
Flow and Volume Statistical Correlates for the Square Waveform
For airflow measured at the mouth, by multiple regression analysis a model including age, height, and HFCC frequency as explanatory variables was statistically significant, with P = .0007 (P < .05 for all variables), but with an r2 of only .18. The highest airflows were generated in subjects with FEV1 greater than 120% of predicted.5
For volume measured at the mouth, by multiple regression analysis a model including height, FVC, and HFCC frequency as explanatory variables was statistically significant, with P = .001 (P < .05 for all variables), but with an r2 of only .16. HFCC frequencies in the lower range resulted in the largest volumes in subjects with moderately compromised FVC from 50 to 80% of predicted.5
For both airflow and volume displaced at the mouth, regardless of waveform used for HFCC, the generated regression models that best fit the data left unexplained a large amount of the variability in the data. Our study failed to identify other subject characteristics that could account for the unexplained variability. It is possible that some of this variability could be explained by functional and structural changes in the lung that were not quantitatively assessed. This will require further study.
HFCC therapy has become the standard for airway clearance for patients with CF and other chronic lung diseases. According to the website www.thevest.com, as of July 2006, more than 46,000 prescriptions had been written for home use; more than 2,500 are currently in use in hospitals. The number of prescriptions in use in hospitals is unknown. Although many studies have demonstrated the effectiveness of airway clearance with both waveforms, there has been no study of optimum frequencies for the use of either.
This study shows that the choice of frequencies based on flows or volumes measured at the mouth will be very different for the two waveforms. Although it is clear from clinical observations that all frequencies are effective for airway clearance, this study has shown that the frequencies are not equally effective for airflows or volumes measured at the mouth and that the patterns of highest airflows and largest volumes are different.
It is not known whether it is better to prescribe for highest airflow frequencies or for largest volumes. The choice might be of little importance for patients using the square waveform, because the usual best frequencies for that waveform are the same for both airflows and volumes. However, if either airflows or volumes are superior for airway clearance, a clear choice would have to be made for the sine waveform that has the best frequencies for volume displacements in the low hertz range and the best frequencies for airflows in the high hertz range. These two ranges for the sine waveform do not overlap. There is a significant gap between the higher frequencies, which are best for flow, and the lower frequencies, which are best for volume.
These suggestions go beyond conclusions drawn from limited animal studies by King and other Canadian researchers in that all frequencies were tested on 100 patients with cystic fibrosis, and the results were analyzed for the contribution of age, sex, height, weight, and pulmonary function. Also, two waveforms were studied and compared on the same subjects. This comprehensive study sets a new standard to be met or to be improved upon by future research.
A resolution of the question of whether flow or volume frequency prescription is better for airway clearance cannot be made with the square waveform alone, but the clear difference observed with the sine waveform could permit a study to answer this question.
Currently, little effort is made to prescribe vest frequencies based on tuning of the patient and the vest. However, if one were not convinced, as we are on the basis of our clinical observations, that tuning the vest and the patient increases the effectiveness of HFCC airway clearance, then one could arrange a double-blinded or crossover study of the effectiveness of prescribing the best frequencies and the worst frequencies. The clinical implications of these findings, as well as their implications for the interpretation of past clinical studies and in planning for future clinical studies, needs to be investigated.
These differences may be due to the devices that produce the different waveforms. The square waveform is produced by pulses of air generated when a rotating valve alternately permits a pulse of air to enter the vest the patient is wearing and then vents the pressure to the ambient air. The sine waveform vest is inflated continuously by a biased airflow to a selected and fixed pressure, and the sine waveform pulse pressures are generated by forward and backward movement of a diaphragm.
Although the square waveform is no longer marketed, there are enough machines in use to study the clinical effectiveness of prescription of frequencies by tuning by comparisons of sputum produced by use of both waveforms with and without tuning. An alternative triangle waveform machine is being made now, using a newly developed, different rotating valve. One published study of that machine suggests that the triangle waveform pulses of air is equal to or better than the sine waveform at clearing mucus from the airways.13 That study used the frequencies for the square waveform, because the square and triangle machines use valves to direct pressure pulses into the vest for compression and to divert the air that had entered the vest toward the room atmosphere.
The readers concerned with the clinical application of HFCC for treatment of their patients should carry away the understanding that almost all of the prior studies of HFCC use have assumed that all HFCC machines are equal, because specific details of how the systems were used have not been described in sufficient detail for the study to be reproduced. In this paper, however, the patent for the tuning procedure is identified, along with how each patient's vest was selected (the vest used was the one he or she had been using effectively for home HFCC therapy). Although the program for conducting the study is not offered to readers, its operations are detailed enough that any clinic that wants to do tuning can have a working program written in a day.
Because our prior paper, in which we compare the second and third generation waveforms, unfortunately fails to meet these standards, we have footnoted the details that will make it possible to repeat that study.†
The changes in pulsar airflows and volumes measured at the mouth differ greatly by the square and sine wave machines. These differences are not explained by the patient characteristics investigated.
The sine and square waveforms differ in the pattern of HFCC frequencies that result in the highest airflows and largest volumes.
These differences suggest that the HFCC frequencies may best be prescribed by tuning each vest machine together with each vest a patient wears on a regular basis.
The high concordance of volumes and airflows with the square waveform suggests that for the square waveform, either can be used for prescription.
The absence of concordance of volumes and airflows with the triangle waveform suggests that research will be needed for these machines to discover whether one or the other or some combination of both will be most effective for airway clearance.
Future studies comparing HFCC machines with other methods of airway clearance should be preceded by such studies so as not to handicap the HFCC machines in the studies.
These hypotheses are testable; the selection of a HFCC machine should be a joint decision of patients, parents, physicians, and respiratory therapists.
*The machines and vests used in this study were Model 102 for the square waveform and Model 103 for the sine waveform, both made and distributed by Advanced Respiratory Inc of St Paul, MN. All the vests were the Minnesota version of the company's standard vest. These vests were fitted to cover the whole of the rib cage from the top of the sternum to the end of the rib cage and were adjusted to be comfortable. The pressure used was the pressure the patient used at home, because that pressure was comfortable for all of the frequencies. No studies were done on either machine to determine if there might be an optimum pressure for any frequency.
†For both machines, we used the vest/jacket that each patient used at home with Model 103, ignoring the possibility that the optimum vest/jacket might be different with the different waveform. We used the Model 103 home treatment protocol (i.e. the frequencies were used in ascending frequency, each for five minutes per frequency followed by a pause to perform three coughs). Because the Model 103 has continuous chest compression even when not delivering HFCC, one of the tubes was disconnected so that the user would be able to take deep breaths and could perform optimum coughs. Because the triangle waveform machine vents to room air when the HFCC stops, the tubes could remained attached.
The pressures prescribed for the study were selected by each patient. With the sine waveform Model 103, the frequency and pressure interaction is such that a patient's maximum was determined by starting with the lowest “best” frequency and pressure 3, and after a few breaths increasing the pressure to one frequency higher, and so forth until patient notices that with the new pressure, it is harder to breathe. The prior pressure became the pressure to be used at frequency. For the next two frequencies, the pressure was dropped by one point and for the highest frequencies was dropped by two points.
For the triangle waveform, the patients used frequency 6 Hz to find the highest comfortable pressure and used that pressure for all six frequencies, because the triangle waveform pulses vented to the room and did not increase the vest pressure. The triangle waveform used the square waveform frequencies. In addition, when the triangle waveform compressions stopped, the venting of the vest/jacket to the room air reduced the residual vest/jacket pressure, allowing the tubes to remain connected when the three coughs were made after each frequency.
From the University of Minnesota Cystic Fibrosis Center, Minneapolis, MN.
Research support for this study provided by the Cystic Fibrosis Center Grant from the United States Cystic Fibrosis Foundation and the Annalisa Marzotto Chair for Cystic Fibrosis Patient Care.