ORIGINAL ARTICLE

Using Chest Vibration Nursing Intervention to

Improve Expectoration of Airway Secretions and
Prevent Lung Collapse in Ventilated ICU Patients:
A Randomized Controlled Trial
Yu-Chih Chen1,2,3*, Li-Fen Wu1, Pei-Fan Mu4, Li-Hwa Lin1, Shin-Shang Chou1,3, Huei-Guan Shie5,6
Departments of 1Nursing and 5Respiratory Therapy, Taipei Veterans General Hospital; 2School of Nursing,
National Taipei College of Nursing; Schools of 3Nursing and 6Respiratory Therapy, Taipei Medical University;
and 4National Yang-Ming University School of Nursing, Taipei, Taiwan, R.O.C.

Background: Almost 80% of patients in the intensive care unit are intubated and on mechanical ventilation. Thus, their
airway clearance ability is compromised and their risk of lung collapse increased. A variety of interventions are used to
enhance airway clearance with the goal of preventing atelectasis and infection. The purpose of this study was to evalu-
ate the effect of a chest vibration nursing intervention on the expectoration of airway secretions and in preventing lung
collapse among ventilated critically ill patients.
Methods: This was a randomized, single-blind experimental study. A total of 95 patients were enrolled from 2 ICUs and
randomly assigned into either the experimental group (n = 50) or control group (n = 45). Patients in the control group
received routine positioning care, which consisted of a change in body position every 2 hours. Patients in the experimen-
tal group received routine positioning care plus the use of chest vibration nursing intervention for 72 hours. This inter-
vention consisted of placing a mechanical chest wall vibration pad on the patient’s back for 60 minutes when the patient
was in a supine position. The chest vibration intervention was performed 6 times a day. Outcome variables were dry spu-
tum weight (DSW) per 24 hours and lung collapse index (LCI); these were measured at 24, 48 and 72 hours.
Results: Patients who received the chest vibration nursing intervention had greater DSW and lower LCI after 24 hours.
Pre-test DSW and group could explain 48.2% of the variance in DSW at 24 hours. The LCI at 24, 48 and 72 hours were
all significantly improved in the intervention group compared to the control group. The previous LCI measured was the
most significant predictor of the next LCI measured. A significant difference was found between the control and experi-
mental groups in their 24-, 48- and 72-hour DSW and LCI after vibration, when monitored by the generalized estimating
equation in time sequence.
Conclusion: The results suggest that chest vibration may contribute to expectoration and thus improve lung collapse among
ventilated patients in an ICU. Chest vibration nursing intervention is a safe and effective alternative pulmonary clearance
method and can be used on patients who are on ventilators in ICUs. [J Chin Med Assoc 2009;72(6):316–322]

Key Words: airway clearance, chest vibration, critical care, dry sputum weight, lung collapse index

Introduction The literature shows that the incidence of lung col-

lapse in ventilated patients can reach 23–30% for
Almost 80% of patients in intensive care units are intu- those who have undergone upper abdominal opera-
bated and on mechanical ventilation, and thus have tion, 74% for those with acute spinal damage, 85% for
difficulty in keeping their airway clear. As a result, those with neuromuscular morbidities, and up to 90%
they face a very high risk of lung collapse complicated after cardiovascular operation.3–5 Lung collapse, if
by pneumonia because they cannot cough effectively.1,2 untreated, may progress to respiratory failure or acute

*Correspondence to: Dr Yu-Chih Chen, Department of Nursing, Taipei Veterans General Hospital,
201, Section 2, Shih-Pai Road, Taipei 112, Taiwan, R.O.C.
E-mail: [email protected] Received: March 13, 2009
● Accepted: April 23, 2009
●

316 J Chin Med Assoc • June 2009 • Vol 72 • No 6

© 2009 Elsevier. All rights reserved.
 Chest vibration intervention for ventilated ICU patients

respiratory distress syndrome, which would prolong Patients in the control group received routine posi-
ventilator use and increase mortality to 33% or 71%.6,7 tioning care, which included a change of position
A systematic review of nonpharmacologic protus- every 2 hours by the ICU nurses. The position turning
sive therapies found that a combination of more than sequence was left lateral, supine, right lateral and su-
1 chest physiotherapy procedure may help to reinflate pine. Patients in the experimental group received rou-
the collapsed lobe of a lung.8 Many studies have sug- tine positioning care plus mechanical chest vibration
gested that postural drainage combined with chest over 72 hours. The chest vibration nursing intervention
percussion, as well as lung hyperinflation plus suction, included placing a mechanical chest wall vibration
are the best ways to quickly solve lung lobe atelecta- pad on the patient’s back for 60 minutes when the pa-
sis.9,10 However, a head-down leg-elevated position is tient was turned into the supine position. The chest
harmful to the vital signs of unstable patients, and is vibration intervention was performed 6 times a day,
therefore not recommended for the critically ill.11 every 4 hours over the 72 hours. The vibration pad was
The manual performance of chest wall percussion or the placed from shoulder to sacrum. The mechanical chest
use of a hand-driven chest vibrator is labor-intensive wall vibration used was a Niagara vibrator type H.U.75,
and highly operator-dependent, with its efficacy be- frequency 70 Hz, 1,000–1,200 cycles/min. The vibra-
ing quite variable. Auto percussion or auto vibration tion wave was generated from the pad (40 × 60 cm) in
(1,000–1,200 cycles/min), which can be quantified spiral, vertical and horizontal directions. The patients
and timed, is more objective and should provide more lay on the pad with a blanket covering them. The vibra-
reliable data.12,13 Some studies have shown that high- tor was turned on and took just over half a second to
frequency chest compression leads to more mucus reach the maximum frequency and range of vibration.
clearance and better lung function compared with During vibration therapy, hemodynamic status and
conventional chest physiotherapy.8,14 vital signs were closely monitored, and if heart beat fluc-
The purpose of this study was to test the effective- tuated > 20 bpm, blood pressure fluctuated > 20 mmHg,
ness of a mechanical chest vibration pad linked with respiration rate fluctuated > 10 bpm, or oxygen satu-
repositioning every 2 hours when used on mechanically- ration dropped to lower than 95%, vibration would
ventilated critically ill patients with the aim of improv- be stopped immediately and the study process for that
ing pulmonary secretion clearance and preventing lung patient ended. However, none of the participants
collapse. experienced any of the above episodes in this study.
The mechanical ventilator settings for both groups
were adjusted by certified respiratory therapists in
Methods compliance with the prescription. Tidal volume was
set based on the patient’s weight (10 mL/kg). The
This was a randomized, single-blind experimental study. system was in the pressure control mode, with a pres-
Between April and July 2007, patients were enrolled sure level of 20–25 cmH2O, and a plus positive end
from 2 ICUs, 1 medical and surgical ICU and 1 neu- expiration pressure of 10–15 cmH2O as the plateau
rologic ICU, in a medical center in Taipei. The inclu- pressure for inspiration was set < 35 cmH2O. Every
sion criteria were age between 20 and 85 years, ventilator was equipped with heated-wire humidifier,
expected use of a ventilator for > 3 days, APACHE-II with the temperature set at 37°C and the moisture at
(Acute Physiology and Chronic Health Evaluation, 100%. The respiratory therapist checked the ventila-
version II) score of 15–40, ability to communicate in tor’s temperature and moisture every 8 hours. ICU
Mandarin or Taiwanese, and willingness to participate nurses assessed the patient’s breath sounds and suc-
in the study. Exclusion criteria included skin damage tioned out any secretion as needed.
to an area of the back, any tendency towards acute Demographic and clinical data were collected on en-
bleeding, presence of a chest drainage tube, fractured rolment and included sex, age, diagnosis on admission,
ribs or percutaneous emphysema, spinal surgery, un- medical history, number of days intubated before en-
stable intracranial pressure and patients who signed rolment, and APACHE II score. Before the experimen-
the “do not resuscitate” instruction. Simple random- tal interventions started, the outcome variables, 24-hour
ization was performed using a table of random num- dry sputum weight (DSW)15,16 and lung collapse index
bers, and eligible patients were randomly assigned (LCI),17 were evaluated. They were also measured at 24,
into either a control or an experimental study group. 48 and 72 hours after the study intervention began.
The institutional review board approved the study Twenty-four-hour DSW was calculated from
protocol and informed consent was obtained from all the 24-hour sputum collection,15,16 which had been
participants. dried using a heater set at 80°C for another 24 hours.

J Chin Med Assoc • June 2009 • Vol 72 • No 6 317

 Y.C. Chen, et al

Table 1. Demographic and baseline characteristics of the 95 study participants*

Control group Experimental group

χ2 p
(n = 45) (n = 50)

Male 30 (66.7) 39 (78.0) 1.53 0.22

Diagnosis
Sepsis 16 (35.6) 20 (40.0) 0.19 0.65
Respiratory failure 16 (35.6) 19 (38.0) 0.06 0.80
Surgery 18 (40.0) 11 (22.0) 3.61 0.06
Past history
CVA 9 (20.0) 17 (34.0) 0.23 0.13
COPD 7 (15.6) 8 (16.0) 0.00 0.95

t test p

Age (yr) 66.8 ± 19.8 73 ± 15.6 −1.84 0.07

PaO2/FiO2 ratio 265 ± 205.4 337.9 ± 164.9 −7.22 0.09
APACHE II 25.4 ± 6.6 23.1 ± 7.2 1.47 0.14
Days in hospital 16.6 ± 23.2 11.5 ± 10.4 1.39 0.17
DSW (mg/24 hr) 5.42 ± 3.98 5.74 ± 6.23 −0.31 0.76
LCI 2.09 ± 0.79 2.26 ± 0.63 −1.15 0.25
*Data presented as n (%) or mean ± standard deviation. CVA = cerebrovascular accident; COPD = chronic obstructive pulmonary disease; APACHE II = Acute
Physiology and Chronic Health Evaluation version II; DSW = dry sputum weight; LCI = lung collapse index.

Clinically, the nurse suctioned the sputum into a suction group; 87 were in the general surgical-medical ICU,
bottle, and every morning, at 7 a.m., the investigator and 8 were in the neurologic ICU. Table 1 presents the
weighed the bottle, stirred it evenly, and drew out demographic and baseline characteristics of the partic-
10 mL, which was sent to the laboratory for dry weight ipants, which had no significant differences between
analysis. The weight of the dried sputum was then mul- the 2 groups.
tiplied to give the 24-hour amount. LCI was evaluated In the experimental group, the DSW over the first
by 1 respiratory physician and 1 nurse practitioner inde- 24 hours after intervention was significantly increased
pendently using a 4-point scale (0 = normal expansion, compared to that of the control group (Table 2). The
1 = single lobe collapsed, 2 = 2 lobes collapsed, 3 = mul- mean DSW of the experimental group was also greater
tiple lobes collapsed)17 and based on changes in routine than that of the control group at 48 and 72 hours,
chest X-ray and the patients’ clinical presentation every but the difference did not reach statistical significance.
morning. If there was disagreement between the 2 LCI An intragroup comparison across the experimental
readings, a face-to-face discussion was held by the physi- group showed that the best result for sputum excre-
cian and the nurse practitioner to achieve consensus. tion was 24 hours after the chest vibration interven-
SPSS version 17.0 (SPSS Inc., Chicago, IL, USA) tion, which was significantly more than the pre-test
was used for data analysis. Univariate and multivariate average of 5.74 ± 6.22 mg. The LCI in the experimen-
analyses were used to compare the variables between tal group was significantly improved compared to that
the 2 groups. Multivariate longitudinal regression analy- of the control group at 48 and 72 hours (Table 2).
ses were performed using the generalized estimating The intragroup comparison also showed that the LCI
equation (GEE). Continuous variables were compared in the experimental group was significantly improved
using Student’s test for normally distributed variables. after intervention while that of the control group
The χ2 test was used to compare categorical variables. showed no significant differences at the 3 time points.
Statistical significance was considered at a p value In addition, analyses of the repeated relationships
≤ 0.05 for all comparisons. were performed by GEE. The DSW of the experi-
mental group, monitored by GEE in the time
sequence, yielded a β value of significance of 5.419
Results (p = 0.000). At 24 hours after chest vibration in the
control and experimental groups, the predictive value
In total, 95 patients participated in the study, includ- for pre-test DSW was 2.985 ± 0.854, which was statis-
ing 45 in the control and 50 in the experimental tically significant (p = 0.000). At 48 hours, the value

318 J Chin Med Assoc • June 2009 • Vol 72 • No 6

 Chest vibration intervention for ventilated ICU patients

Table 2. Changes in dry sputum weight (DSW) and lung collapse index (LCI)*

Control group Experimental group

t test p
(n = 45) (n = 50)

DSW (mg/24 hr)

24 hr 5.39 ± 4.39 8.70 ± 6.48 −2.94 0.004
48 hr 4.09 ± 2.96 4.65 ± 3.44 −0.86 0.392
72 hr 3.56 ± 3.10 4.04 ± 3.43 −0.72 0.471
LCI
24 hr 1.78 ± 0.82 1.52 ± 0.65 1.69 0.096
48 hr 1.56 ± 0.89 1.18 ± 0.72 2.24 0.028
72 hr 1.60 ± 0.91 0.96 ± 0.73 3.75 0.000
*Data presented as mean ± standard deviation.

Table 3. Changes in dry sputum weight with chest vibration, mon- Mean
EMDSWUL EMDSW
itored by generalized estimating equation (n = 95)
15 EMDSWLL CMDSW
Standard CMDSWUL EMDSW
β p CMDSWLL CMDSW
Dry sputum weight (mg)
error

Group 10
Control vs. experimental 5.419 0.586 0.000

Time sequence
24 hr post-test vs. pre-test −1.864 0.594 0.957 5
48 hr post-test vs. pre-test −1.331 0.594 0.025
72 hr post-test vs. pre-test −0.032 0.647 0.004

Groups in time sequence (experimental vs. control) 0

24 hr post-test vs. pre-test 2.985 0.845 0.000
48 hr post-test vs. pre-test 0.241 0.999 0.809 Pre-test 24 hr 48 hr 72 hr
72 hr post-test vs. pre-test 0.159 1.008 0.875
Measurement time
Figure 1. Changes at pre-test, and 24, 48 and 72 hours for dry
sputum weight (DSW) across the 2 groups. EMDSWUL = experimen-
was 0.241 ± 0.999, which was not statistically signifi- tal group mean DSW upper limit; EMDSW = experimental group
cant (p = 0.809). At 72 hours, the value was 0.159 ± mean DSW; EMDSWLL = experimental group mean DSW lower limit;
1.008, which was also not statistically significant CMDSWUL = control group mean DSW upper limit; CMDSW = con-
(p = 0.875) (Table 3, Figure 1). trol group mean DSW; CMDSWLL = control group mean DSW lower
The results of chest vibration, as evaluated by LCI limit.
and monitored by GEE in time sequence, showed
an average β value of 2.089 (p = 0.000). At 24 hours
Table 4. Changes in lung collapse index with chest vibration,
after chest vibration in the control and experimental
monitored by generalized estimating equation (n = 95)
groups, the predictive value for pre-test X-ray status
was −0.429 ± 0.161 (p = 0.008). At 48 hours, the pre- β
Standard
p
dictive value was −0.547 ± 0.188 (p = 0.004), while at error
72 hours, the value was −0.811 ± 0.211 (p = 0.000) Group
(Table 4, Figure 2). Control vs. experimental 2.089 0.117 0.000
Stepwise regressions were performed by stratifying
Time sequence
patients based on their characteristics, and grouping
24 hr post-test vs. pre-test −0.311 0.129 0.016
categories in order to find predictors for DSW and
48 hr post-test vs. pre-test −0.533 0.143 0.000
LCI at 24, 48 and 72 hours. Firstly, the results showed 72 hr post-test vs. pre-test −0.489 0.153 0.001
that the previous DSW measured was the most signif-
icant predictor of the next DSW measured and ex- Groups in time sequence (experimental vs. control)
24 hr post-test vs. pre-test −0.429 0.161 0.008
plained 39.6%, 19.1% and 48.4% of the variance in
48 hr post-test vs. pre-test −0.547 0.188 0.004
DSW at 24, 48 and 72 hours, respectively (Table 5).
72 hr post-test vs. pre-test −0.811 0.211 0.000
Secondly, grouping was a significant predictor of DSW

J Chin Med Assoc • June 2009 • Vol 72 • No 6 319

 Y.C. Chen, et al

Mean Table 6. Stepwise regression analysis for prediction of lung

EMXraySUL EMXrayS
CMXrayS collapse index at the 3 time points after chest vibration nursing
EMXraySLL
intervention (n = 95)
3.00 CMXraySUL EMXrayS
CMXraySLL CMXrayS R2 β F p
2.50
At 24 hr 8.124 < 0.000
X-ray (score)

2.00 Pre-test LCI 0.115 0.354

Grouping 0.178 0.252
1.50 classification
1.00 At 48 hr 45.256 < 0.000
LCI at 24 hr 0.587 0.723
0.50
DSW at 24 hr 0.623 −0.205
0.00 CV history 0.647 −0.155

Pre-test 24 hr 48 hr 72 hr At 72 hr 40.766 < 0.000

Measurement time LCI at 48 hr 0.450 0.604
Grouping 0.521 0.274
Figure 2. Changes at pre-test, and 24, 48 and 72 hours for X-ray classification
lung collapse index (LCI) across the 2 groups. EMXraySUL = experi-
mental group mean X-ray score upper limit; EMXrayS = experimen- LCI = lung collapse index; DSW = dry sputum weight; CV = cardiovascular.

tal group mean X-ray score; EMXraySLL = experimental group mean

X-ray score lower limit; CMXraySUL = control group mean X-ray score and cardiovascular history were significant predictors
upper limit; CMXrayS = control group mean X-ray score; CMXraySLL = explaining the variance in LCI at 48 hours (Table 6).
control group mean X-ray score lower limit.

Discussion
Table 5. Stepwise regression analysis for prediction of dry
sputum weight at the 3 time points after chest vibration The performance of effective and safe pulmonary
nursing intervention (n = 95) nursing care always poses a challenge for nurses tak-
R2 β F p
ing care of ventilated ICU patients in light of their
critically ill condition. Since manual percussion is no
At 24 hr 34.864 < 0.001 longer used to help excrete sputum, the present study
Pre-test DSW 0.396 0.640 was conducted to provide empirical support for the
Group classification 0.482 −2.93
effectiveness of vibration in preventing lung collapse.
At 48 hr 9.993 < 0.001 The results showed that for ventilated ICU patients,
DSW at 24 hr 0.191 0.426 routine positioning care combined with 60 minutes
History of surgery 0.245 0.232 of chest-wall deep vibration performed every 4 hours
History of COPD 0.288 −0.208
by auto vibrator at 1,000–1,200 cycles/min when pa-
At 72 hr 42.291 < 0.001 tients were in supine position, plus suction if necessary,
DSW at 48 hr 0.484 0.728 was able to achieve a significant difference in 24-hour
Pre-test status 0.530 −0.216 DSW compared with a control group that received
of LCI
routine care only. The LCI in the experimental group
DSW = dry sputum weight; COPD = chronic obstructive pulmonary disease; also improved significantly at 48 and 72 hours com-
LCI = lung collapse index.
pared to that in the control group. Our results are con-
sistent with those of previous studies.1,2,5,18,19
at 24 hours. Thirdly, the patient’s surgical history and A significant difference was found between the con-
chronic obstructive pulmonary disease (COPD) were trol and experimental groups at 24, 48 and 72 hours
significant predictors of DSW at 48 hours. Finally, it with regard to DSW after vibration, as monitored
was found that pre-test LCI was a significant predic- by GEE in time sequence. The predictive value for
tor of DSW at 72 hours. pre-test DSW was statistically significant at 24 hours
The previous LCI measured was the most significant after chest vibration started, but not at 48 and 72
predictor of the next LCI measured and explained hours. Thus, there was a significantly greater expecto-
11.5%, 59% and 45% of the variance in LCI at 24, 48 ration effect at 24 hours after auto chest vibration in
and 72 hours, respectively. Group was a significant the ventilated patients. At 48 and 72 hours, there was
predictor for LCI at 24 and 72 hours. DSW at 24 hours a continuing but only limited increase in excretion

320 J Chin Med Assoc • June 2009 • Vol 72 • No 6

 Chest vibration intervention for ventilated ICU patients

compared to the control group; however, this did not The chest vibration nursing intervention in this
reach statistical significance. study was designed to be simple and easily carried out
A significant difference was also found between by the nurses who would perform these procedures.
the control and experimental groups for LCI at 24, This is different from the study of Templeton and
48 and 72 hours after chest vibration, when moni- Palazzo, who applied very complicated chest physio-
tored by GEE in time sequence. The statuses of lung therapy to their critically-ill ventilated patients, which
collapse at the 3 time points were all able to predict included inflating the lung manually, vibration, suction
pre-test LCI. We believe that chest vibration made a in a sitting position, inspiration and muscle move-
significant improvement to the rate of lung collapse at ment, postural drainage and ventilated suction.7 Such
24 hours for the ventilated patients with sputum an approach places an extremely heavy burden on the
retained in their airway because of the increase in spu- patients and nurses. In contrast, our intervention cre-
tum secretion. Later, at 48 and 72 hours, since there ated no extra burden on either the patients or nurses.
was no continuing increase in sputum secretion, the Our study showed that chest vibration nursing inter-
improvement at 24 hours remained significantly dif- vention is able to reduce lung collapse among critically-
ferent relative to the pre-test state of lung collapse. ill and mechanically-ventilated patients and does this
In the regression analysis, pre-test DSW and group- quickly within 24 hours; furthermore, patients’ con-
ing classification were the 2 significant predictors of dition continues to improve with intervention up to
24-hour DSW. This result supports the idea that chest 72 hours.
physiotherapy can have an immediate effect on the The present study shows that there is an obvious
first day. In addition, the predictive factors of DSW at effect of the intervention when it is used on ventilated
48 hours included DSW at 24 hours, the postoperative adult patients. Although intragroup matching was per-
status of the patient and COPD history. In other words, formed between the experimental and control groups
the DSW at 48 hours seemed to be affected by the of patients with cerebrovascular accident and COPD
amount of sputum excreted the day before, whether history, regression analysis found that these patients
the patient had been operated on, and whether the required a specific type of chest physiotherapy that fits
patient had a history of COPD. The 2 most impor- their unique needs; this is because it is necessary to
tant predictors of DSW at 72 hours were the DSW at consider their overall poorer ability to excrete sputum
48 hours and pre-test LCI (when patients were enrolled and inflate their lungs.
into the study). In summary, whether monitored at 24, One limitation of this study is that the study par-
48 or 72 hours, the DSW of that day was always an im- ticipants were from 2 units in 1 hospital, which limits
portant predictor of the following day’s DSW. In addi- the generalizability of the study to other types of units
tion, the DSW at 48 and 72 hours were affected by a and healthcare sectors in Taiwan. Nevertheless, this
previous history of pulmonary disease and pre-test lung study shows that chest vibration nursing intervention
collapse condition. Thus, it is clear that chest physio- is a safe and effective alternative method of pulmonary
therapy needs to be performed when patients have pul- clearance and can be used on patients who are on ven-
monary morbidities or have had serious lung collapse. tilators in an ICU. The present study established a stan-
When monitoring LCI at 24, 48 and 72 hours, we dardized chest vibration nursing intervention based on
found that the predictive factors for lung collapse at a literature review and clinical observation. The study
24 hours included the pre-test status and grouping clas- results confirm the feasibility of this approach in an
sification. This finding confirms the idea that there ICU setting. The addition of this intervention to con-
is an immediate effect of chest vibration on the first ventional positioning care appears to be better in pre-
day of treatment. The predictive factors of LCI at venting lung collapse than conventional positioning
48 hours included the LCI and DSW at 24 hours, and care alone. Thus, positioning care with the use of
history of cerebrovascular accident. The predictive auto vibration performed every 4 hours is an effective
factors of LCI at 72 hours included lung collapse status intervention.
at 48 hours and grouping classification. These results
indicate that, whatever the time point, the LCI of that
day was always a predictor of the following day’s LCI. References
In addition, LCI at either 24 or 72 hours was also
related to chest vibration, which again confirms the 1. Ciesla ND. Chest physical therapy for patients in the intensive
care unit. Phys Ther 1996;76:609–25.
effectiveness of chest vibration in these circumstances,
2. Ntoumenopoulos G, Presneill JJ, McElholum M, Cade JF.
especially for patients with serious chest morbidities Chest physiotherapy for the prevention of ventilator-associated
or lung collapse. pneumonia. Intensive Care Med 2002;28:850–6.

J Chin Med Assoc • June 2009 • Vol 72 • No 6 321

 Y.C. Chen, et al

3. Bowton DL. Nosocomial pneumonia in the ICU—year 2000 13. Shin S, Ku Y, Park MS, Moon SY, Suh JS. Blood flow resist-
and beyond. Chest 1999;115:28S–33S. ance with vibration and its effect on blood cell migration. Clin
4. Krause MF, Hoehn T. Chest physiotherapy in mechanically ven- Hemorheol Microcirc 2004;30:353–8.
tilated children: a review. Crit Care Med 2000;28:1648–51. 14. Brierley S, Adams C, Suelter J, Gooch T, Becker B. Safety
5. Manza RM, Carvalho CRF, de Romanholo BMS, Vieira JE. and tolerance of high frequency chest wall oscillation (HFCWO)
Chest physiotherapy during immediate postoperative period in hospitalized critical care patients. Respir Care 2003;48:
among patients undergoing upper abdominal surgery: random- 1112.
ized clinical trials. Sao Paulo Med J 2008;126:269–73. 15. Kempainen RR, Williams CB, Hazelwood A, Rubin BK, Milla
6. Barker M, Adams S. An evaluation of a single chest physiotherapy CE. Comparison of high-frequency chest wall oscillation with
treatment on mechanically ventilated patients with acute lung differing waveforms for airway clearance in cystic fibrosis. Chest
injury. Physiother Res Int 2002;7:157–69. 2007;132:1227–32.
7. Templeton M, Palazzo MG. Chest physiotherapy prolongs dura- 16. Placidi G, Cornacchia M, Polese G, Zanolla L, Assael BM,
tion of ventilation in the critically ill ventilated for more than Braggion C. Chest physiotherapy with positive airway pressure:
48 hours. Intensive Care Med 2007;33:1938–45. a pilot study of short-term effects on sputum clearance in pa-
8. McCool FD, Rosen MJ. Nonpharmacologic airway clearance tients with cystic fibrosis and severe airway obstruction. Respir
therapies. Chest 2006;129:250S–9S. Care 2006;51:1145–53.
9. Fink JB. Positioning versus postural drainage. Respir Care 2002; 17. Deakins K, Chatburn RL. A comparison of intrapulmonary
47:769–77. percussive ventilation and conventional chest physiotherapy for
10. Thomas J, Cook DJ, Brooks D. Chest physical therapy the treatment of atelectasis in the pediatric patient. Respir Care
management of patients with cystic fibrosis. A meta-analysis. 2002;47:1162–7.
Am J Respir Crit Care Med 1995;151:846–50. 18. Bradley JM, Moran FM, Elborn JS. Evidence for physical ther-
11. Tyler ML. Complications of positioning and chest physiotherapy. apies (airway clearance and physical training) in cystic fibrosis:
Respir Care 1982;26:458–66. an overview of five Cochrane systematic reviews. Respir Med
12. Krasovitski B, Kimmel E. Shear stress induced by a gas bubble 2006;100:191–201.
pulsating in an ultrasonic field near a wall. IEEE Trans Ultrason 19. Dosman CF, Jones RL. High-frequency chest compression: a
Ferroelectr Freq Control 2004;51:973–9. summary of the literature. Can Respir J 2005;12:37–41.

322 J Chin Med Assoc • June 2009 • Vol 72 • No 6