Division of Critical Care Medicine, Clínicas Hospital Brazil

Naue et al: Increased pressure support during vibrations in intensive care Increasing pressure support does not enhance secretion clearance if applied during manual chest wall vibration in intubated patients: a randomised trial Wagner da Silva Naue, Ana Carolina Texeira da Silva, Adriana Meira Güntzel, Robledo Leal Condessa, Roselaine Pinheiro de Oliveira and Silvia Regina Rios Vieira Division of Critical Care Medicine, Clínicas Hospital Brazil Questions: What is the effect of increasing pressure support during the application of manual chest wall compression with vibrations for secretion clearance in intubated patients in intensive care? Design: A randomised trial with concealed allocation, assessor blinding and intention-to-treat analysis. Participants: 66 patients receiving mechanical ventilation for greater than 48 hours. Intervention: All participants were positioned supine in bed with the backrest elevated 30 degrees. The experimental group received manual chest wall compression with vibrations during which their pressure support ventilation was increased by 10 cmh 2 O over its existing level. The control group received manual chest wall compression with vibrations but no adjustment of the ventilator settings. Both groups then received airway suction. Outcome measures: The primary outcome was the weight of the aspirate. Secondary outcomes were pulmonary and haemodynamic measures and oxygenation. Results: Although both treatments increased the weight of the aspirate compared to baseline, the addition of increased pressure support during manual chest wall compression with vibrations did not significantly increase the clearance of secretions, mean betweengroup difference in weight of the aspirate 0.4g, 95% CI 0.5 to 1.4. Although several other measures also improved in one or both groups with treatment, there were no significant differences between the groups for any of the secondary outcomes. Conclusion: Although increasing pressure support has previously been shown to increase secretion clearance in intubated patients, the current study did not show any benefits when it was added to chest wall compression with vibrations. Trial registration: NCT01155648. Journal of Physiotherapy : Physiotherapy, Chest wall compression, Mechanical ventilation, Hyperinflation, Randomised controlled trial Introduction Most patients in intensive care receive invasive ventilatory support, which typically relieves their work of breathing and improves their gas exchange. However, intubation for mechanical ventilation also has deleterious effects on mucus transport by ciliary mechanisms and by cough (Gosselink et al 2008, McCarren et al 2006). This can lead to the stasis of secretions in the airways, which can cause bronchial obstruction (Amato et al 2007). If bronchial obstruction in an airway is not reversed, the more distal airways will remain unventilated and become atelectatic. This may worsen hypoxia. Furthermore, the accumulation of bronchial secretions favours the multiplication of microorganisms in unventilated areas and subsequent development of pneumonia (Bhowmik et al 2009, Ntoumenopoulos et al 2002). Some physiotherapy techniques are intended to reverse these deleterious sequelae of intubation and bronchial obstruction by combating the accumulation of mucus. One such technique is manual chest wall compression with vibrations. This technique is achieved by a sustained isometric contraction of the physiotherapist s upper limbs, with an oscillating compressive force on the patient s thorax during expiration. It aims to facilitate the transport of mucus from peripheral to central airways, thereby facilitating clearance by aspiration with a suction catheter (Frownfelter 2004, McCarren et al 2006). Techniques that increase inspiratory tidal volume and therefore expiratory flow rates, such as hyperinflation via adjustment of the settings on a mechanical ventilator, may also help to mobilise secretions. One rationale for this is that such an intervention may increase ventilation to non-ventilated airways and thereby facilitate the cough mechanism, aiding the transport of mucus from peripheral to central airways (Lemes et al 2009, Savian et al 2006). Hyperinflation can be achieved using the mechanical ventilator by increasing pressure support. For example, Lemes and colleagues (2009) achieved significant increases in tidal volume by increasing pressure support to provide a peak airway pressure of 40 cmh 2 O. In randomised trials, this technique of ventilator hyperinflation increased the static compliance (Berney and Denehy 2002) and the amount of secretions obtained (Lemes 2007). This study is designed to compare the effectiveness of chest wall compression and vibration with and without a concurrent 10 cmh 2 O increase in inspiratory pressure support above the existing level via adjustment of the ventilator settings. Journal of Physiotherapy 2011 Vol. 57 Australian Physiotherapy Association 2011 21

Research Patients screened for participation (n = 1304) Excluded (n = 1238) haemodynamically unstable or not initiating breaths (n = 1237) lack of consent (n = 1) Randomised (n = 66) (n = 34) (n = 32) Baseline Measured secretions obtained with aspiration, pulmonary mechanics, haemodynamics and oxygenation (n = 34) (n = 32) Lost to follow-up (n = 0) Experimental Group usual care manual chest vibrations with increased pressure support aspiration of airway Control Group usual care manual chest vibrations aspiration of airway Lost to follow-up (n = 0) 6 hours Measured secretions obtained with aspiration, pulmonary mechanics, haemodynamics and oxygenation (n = 34) (n = 32) Design and flow of participants through the trial. Characteristics of participants. Therefore, the research questions of this study were: 1. In patients receiving mechanical ventilation in intensive care, does the addition of an increase in pressure support during manual chest wall compression and vibration increase the amount of secretions obtained? 2. Does it improve peak inspiratory pressure, tidal volume and dynamic compliance? 3. Does it have adverse haemodynamic effects? Method Design A randomised trial with assessor blinding of the primary outcome, concealed allocation and intention-to-treat analysis was undertaken at the Clínicas Hospital in Porto Alegre, Brazil, between May 2008 and May 2010. Participants were recruited from the Intensive Care Unit. To achieve concealed allocation, each random allocation was concealed in an opaque envelope until a patient s eligibility to participate was confirmed. Outcomes were measured immediately after the intervention. Characteristic Randomised (n = 66) Exp (n = 34) Con (n = 32) Gender, n male (%) 15 (44) 13 (41) Age (yr), mean (SD) 64 (15) 65 (19) APACHE II, mean (SD) 26 (7) 23 (7) Time ventilated (d), mean (SD) 8 8 (6) Pathology, n (%) COPD 7 (21) 5 (16) Bronchopneumonia 9 (26) 16 (50) Heart failure 6 (18) 3 (9) Stroke 8 (24) 2 (6) Septicaemia 14 (41) 12 (38) Other 18 (53) 17 (53) APACHE II = Acute Physiology and Chronic Health Evaluation, COPD = Chronic obstructive pulmonary disease, Other = immunosuppressed, acquired immune deficiency syndrome, neoplasm Journal of Physiotherapy 2011 Vol. 57 Australian Physiotherapy Association 2011

Naue et al: Increased pressure support during vibrations in intensive care Mean (SD) for outcomes for each group, mean (SD) difference within groups, and mean (95% CI) difference between groups. Groups Difference within groups Difference between groups Pre Post Post minus Pre Post minus Pre Exp (n = 34) Aspirate weight (g) 2.6 (3.0) Peak (cmh 2 O) 20.9 (4.2) Tidal volume (ml) 465 (88) Cdyn (ml/cmh 2 O) 32 (9) Heart rate (bpm) 92 (21) Resp rate (br/min) (6) MAP (mmhg) 91 (18) SpO 2 (%) 96.9 (2.5) Con (n = 32) 1.3 (1.2) 21.8 (3.4) 5 (119) 36 (11) 91 (19) 20 87 (17) 97.6 (2.8) Exp (n = 34) 3.5 (3.8) 21.2 (4.5) 521 (120) 35 (10) 96 (20) 93 (20) 96.9 (3.1) Con (n = 32) 1.7 (1.6) 21.7 (3.5) 555 (145) 38 (13) 96 (18) (6) 91 (14) 97.2 (3.3) Exp Con Exp minus Con 0.9 (2.1) 0.3 (0.9) 56 (69) 3 4 (10) 0 (6) 2 (12) 0.0 (2.0) 0.5 (1.5) 0.2 (1.7) 33 (101) 2 (9) 6 (9) 2 3 (12) 0.4 (1.9) 0.4 ( 0.5 to 1.4) 0.5 ( 0.2 to 1.1) ( 20 to 65) 1 ( 3 to 4) 1 ( 6 to 3) 2 ( 4 to 1) 1 ( 7 to 5) 0.4 ( 0.6 to 1.4) Shaded row = primary outcome, bpm = beats per minute, br = breaths, Cdyn = Dynamic compliance of the respiratory system, Resp = respiratory, MAP = mean arterial pressure, SpO 2 % = percentage saturation of oxyhaemoglobin estimated by pulse oximetry Participants Patients who were intubated and had received mechanical ventilation for at least 48 hr in the Intensive Care Unit and who were initiating spontaneous breaths were eligible to participate. Exclusion criteria were: ventilator associated pneumonia, positive end-expiratory pressure greater than 10 cmh 2 O, haemodynamic instability (defined as mean arterial pressure less than 60 cmh 2 O), contraindications to an increase in the applied inspiratory pressure (eg, pneumothorax, undrained haemothorax, subcutaneous emphysema), osteoporosis, peak airway pressure greater than 40 cmh 2 O, neurosurgery, and a relative who was unwilling to consent to the patient s participation. Intervention All participants received usual medical and nursing care while in the Intensive Care Unit. This included position changes second hourly, aspiration of the airway as needed, chest wall vibrations with compression twice a day. Clinical data including gender, age, baseline Acute Physiology and Chronic Health Evaluation II (APACHE II) scores, comorbidities, start and end dates of mechanical ventilation, presence or absence of ventilator-associated pneumonia, type of ventilator and mode of ventilation were recorded at baseline. After randomisation, all participants were positioned supine in bed with the bedhead elevated 30 deg. In this position, their airway was aspirated once with a 12-gauge suction catheter with a vacuum pressure of 40 cmh 2 O. Two hours later, haemodynamic and pulmonary measures were recorded. The participants artificial airway was then aspirated 3 times with an open suction system, for 12 sec, at intervals of 30 sec, with the same catheter and vacuum pressure. The aspirate was collected in a vial and stored for weighing. Haemodynamic and ventilator measures were recorded 1 min later. These were the baseline measures. Approximately six hours later, all participants were again positioned in supine with the bedhead elevated 30 deg and had their airway aspirated once, as described above. Two hours later, haemodynamic and pulmonary measures were recorded. Experimental group participants then received manual chest wall compression with vibrations for 5 min to each hemithorax by a physiotherapist. During the application of these manual techniques, the ventilator settings were altered so that inspiratory pressure support increased by 10 cmh 2 O above the existing level. Control group participants received the same regimen of compression with vibration of the chest wall, but without any change in their ventilator settings. In all participants, the airway was then aspirated 3 times with an open suction system, for 12 sec, at intervals of 30 sec, again with the same catheter and vacuum pressure. The aspirate was collected in a vial and stored for weighing. The haemodynamic and pulmonary measures were recorded 1 min later. Outcome measures The secretions obtained with each aspiration were collected and stored in a collection flask and weighed on an electronic scale by an investigator blinded to whether the sample was from the experimental or control group. The pulmonary measures recorded were: peak inspiratory pressure, endexpiratory pressure, and tidal volume, each measured via the mechanical ventilator. Dynamic compliance was calculated as the tidal volume divided by the difference between the peak inspiratory pressure and the endexpiratory pressure. The haemodynamic measures recorded were: heart rate, respiratory rate, mean arterial pressure, and oxyhaemoglobin saturation measured by peripheral pulse oximetry. Journal of Physiotherapy 2011 Vol. 57 Australian Physiotherapy Association 2011 23

Research Data analysis The minimal important difference in secretions aspirated with a single treatment has not yet been established. We therefore nominated 0.7 g as the between-group difference we sought to identify. Assuming a SD of 1 g, 68 participants (34 per group) would provide 80% power, at the 2-sided 5% significance level, to detect a 0.7 g difference between the experimental and control groups as statistically significant. Continuous data were summarised as means and standard deviations and categorical data were summarised as frequencies and percentages. Normal distribution of the data was confirmed with the Kolmogorov-Smirnov test. Between-group differences in change from baseline were analysed using unpaired t-tests. Mean differences (95% CI) between groups are presented. Within-group changes were analysed using a paired samples t test. Chi-squared or Fischer s exact test were used for categorical variables. Data were analysed by intention to treat. Results through the trial Recruitment and data collection were carried out between May 2008 and May 2010. During the study period, 1304 patients were screened for eligibility. Sixty-six met the eligibility criteria and were randomised: 34 in the experimental group and 32 in the control group. The flow of participants through the trial and the reasons for the exclusion of some participants are illustrated in Figure 1. Baseline characteristics of the participants were similar between the allocated groups (Table 1). Interventions to the experimental group were provided by the Intensive Care Unit physiotherapist, who had seven years of clinical experience, including four years in intensive care. The Intensive Care Unit of the Clínicas Hospital in Porto Alegre, Brazil, was the only centre to recruit and test patients in the trial. The Intensive Care Unit has 25 adult medical-surgical beds and a throughput of 1117 patients per year. All randomised participants completed the trial, including both interventions as randomly allocated and all outcome measures. No participant in either group had adverse haemodynamic or ventilatory changes during the intervention to such an extent that they necessitated cessation of the intervention. Effect of intervention Group data for all outcomes for the experimental and control interventions are presented in Table 2, while individual data are presented in Table 3 (see eaddenda for Table 3). The weight of the aspirate was significantly greater after physiotherapy in the experimental group, compared to baseline. However, the control group also showed a small increase and overall the difference in effect between the experimental and control groups was not statistically significant, mean difference 0.4g (95% CI 0.5 to 1.4). After the interventions, peak airway pressure did not significantly differ between the experimental and control groups. Tidal volume was significantly greater after physiotherapy in the experimental group, compared to baseline. However, the control group also showed a small increase and overall the difference in effect between the experimental and control groups was not statistically significant, mean difference ml (95% CI 20 to 65). Similarly, dynamic compliance improved significantly after physiotherapy in the experimental group, but the change was not significantly greater than in the control group, mean difference 1 cmh 2 O (95% CI 3 to 4). Heart rate increased significantly in both groups from baseline, but the between-group difference in this change was not statistically significant. The changes in respiratory rate were clinically unimportant, with no statistically significant difference between the groups in the change during the intervention, mean difference 2 breaths per minute (95% CI 4 to 1). The changes in mean arterial pressure and oxyhaemoglobin saturation were also not statistically significantly different between the experimental and control groups. Discussion Several authors have described the use of hyperinflation to prevent lung collapse, re-expand atelectatic areas, increase oxygenation, improve lung compliance and facilitate the movement of secretions from the small to the larger central airways (Denehy 1999, Savian et al 2006, Singer et al 1994). These effects appear to occur due to an increase in the tidal volume generated by the hyperinflation that further expands the normal alveoli through the interdependence mechanism, which also re-expands collapsed alveoli (Stiller 2000). Lemes and colleagues (2009) provided data to support this using a randomised crossover trial. A ventilator-induced increase in pressure support improved the volume of secretions aspirated and the static compliance of the respiratory system. Although the difference in the intervention arms in both the Lemes study and the current study was the use of ventilator-induced hyperinflation, the other interventions applied to both groups differed. In the Lemes study, positioning was the only other intervention. In the current study, both groups received positioning and chest wall compression with vibrations. Thus, while Lemes and colleagues (2009) found that ventilator hyperinflation improved the amount of secretions aspirated above the effect of positioning alone, the current study found it did not significantly improve the amount of secretions above the effect of positioning and chest wall vibrations with compression. This result may have been influenced by the difference in the average baseline sputum production of the two groups, which was relatively large. The current study used chest wall vibrations with compression in both groups and therefore can only examine its effect as uncontrolled data. Notwithstanding this, both groups increased the amount of secretions aspirated after the interventions, with the within-group change being statistically significant in the experimental group. Unoki and colleagues (2005) also examined the effect of manual chest wall compression in a randomised crossover trial. Chest wall compression had a modest and statistically nonsignificant effect on the volume of secretions aspirated. Even with uncontrolled data, it is valuable to see the effect of chest wall compression with vibration isolated from the effects of other techniques. Most other studies of chest wall compression have included it with techniques such as postural drainage and percussion. Ntoumenopolous and colleagues (2002) and Vieira and colleagues (2009) have 24 Journal of Physiotherapy 2011 Vol. 57 Australian Physiotherapy Association 2011

Naue et al: Increased pressure support during vibrations in intensive care shown that a combination of physiotherapy techniques can reduce the risk of ventilator associated pneumonia in mechanically ventilated patients in intensive care. However, Patman and colleagues (2008) found that physiotherapy did not prevent, or hasten recovery from, ventilator-associated pneumonia in patients with acquired brain injury. While this is valuable information that can be applied clinically, authors such as Hess (2007) have commented that the effects of the individual techniques in these complex physiotherapy interventions are indistinguishable, and therefore the current study and others that allow the effect of individual techniques to be separated from the overall physiotherapy regimen can help advance our understanding of which techniques are effective. The increase in peak inspiratory tidal volume caused by hyperinflation may improve expiratory flow rates and therefore assist in shifting secretions from smaller airways to the larger central airways, thereby reducing the resistance in the airways and leading to an increase in tidal volume (Choi and Jones 2005, Santos 2010). Although there was a significant within-group improvement in tidal volume in the group that received ventilator-induced hyperinflation, this was not significantly greater than the improvement in the control group in the current study. Berney and Denehy (2002) demonstrated a significant increase in lung compliance after hyperinflation in a randomised crossover trial. Savian and colleagues (2006) later published similar results, attributing the increase in pulmonary compliance to improved distribution of ventilation and the subsequent recruitment of collapsed lung units. Although the within-group improvement in lung compliance in the experimental group was statistically significant, this was not significantly greater than the improvement in the control group in the current study. One limitation of this study was the sample size. Although formal power calculations were performed a priori and a desirable sample size was recruited, some outcomes still have confidence intervals that include the possibility of clinically worthwhile effects particularly in the beneficial direction. Therefore, ventilator-induced hyperinflation should be investigated further. Another limitation is that only one outcome albeit the primary outcome was assessed by a blinded investigator. Also, there were baseline differences in some groups that were large enough to have possibly influenced the final outcomes to a clinically meaningful degree. In summary, although the addition of ventilator-induced hyperinflation appears to have an effect on the amount of sputum aspirated and the compliance of the respiratory system over the effect of positioning alone (Lemes et al 2009), the current study did not show similar benefits when increased pressure support was added to positioning and chest wall compression with vibration. eaddenda: Available at JoP.physiotherapy.asn.au Table 3. Ethics: The Clínicas Hospital Ethics Committee(s) approved this study (number 07504). All participants gave informed consent before data collection began. Support: This study was supported by the Fundo de Incentivo a Pesquisa e Eventos (FIPE) Research and Event Inventive Fund. Acknowledgements: The authors are grateful to the patients, nurses, and officers of the Division of Critical Care Medicine of Clínicas Hospital for their assistance in the conduct of this work. Competing interests: None declared. Correspondence: Wagner da Silva Naue, Department: Division of Critical Care Medicine, Clínicas Hospital / Rio Grande do Sul Federal University, Ramiro Barcelos, 2350. Porto Alegre / RS, Brazil. Email: wnaue@hcpa.ufrgs.br References Amato MB, Carvalho CR, Isola A, Vieira S, Rotman V, Moock M, et al (2007) [Mechanical ventilation in Acute Lung Injury Jornal Brasiliero de Pneumologia 33 Suppl 2S: S119 127. Berney S, Denehy L (2002) A comparison of the effects of manual and ventilator hyperinflation on static lung compliance and sputum production in intubated and ventilated intensive care patients. Physiotherapy Research International 7: 100 108. Bhowmik A, Chahal K, Austin G, Chakravorty I (2009) Improving mucociliary clearance in chronic obstructive pulmonary disease. Respiratory Medicine 103: 496 502. Choi JS, Jones AY (2005) Effects of manual hyperinflation and suctioning in respiratory mechanics in mechanically ventilated patients with ventilator-associated pneumonia. Australian Journal of Physiotherapy 51: 25 30. Denehy L (1999) The use of manual hyperinflation in airway clearance. European Respiratory Journal 14: 958 965. Frownfelter DD, Dean E (2004) Fisioterapia Cardiopulmonar Principios e Práticas (3 edn). Reivinter. Gosselink R, Bott J, Johnson M, Dean E, Nava S, Norrenberg M, et al (2008) Physiotherapy for adult patients with critical illness: recommendations of the European Respiratory Society and European Society of Intensive Care Medicine Task Force on Physiotherapy for Critically Ill Patients. Intensive Care Medicine 34: 1188 1199. Hess DR (2007) Airway clearance: physiology, pharmacology, techniques, and practice. Respiratory Care 52: 1392 1396. Lemes DA, Zin WA, Guimarães FS (2009) Hyperinflation using pressure support ventilation improves secretion clearance and respiratory mechanics in ventilated patients with pulmonary infection: a randomised crossover trial. Australian Journal of Physiotherapy 55: 249 254. Lemes DA, Guimarães FS (2007) O uso da hipeinsuflação como recurso fisioterapêutico em unidade de terapia intensiva (The use of hyperinflation as a physical therapy resource in intensive care unit). Revista Brasileira de Terapia Intensiva 19: 1 5. McCarren B, Alison JA, Herbert RD (2006) Manual vibration increases expiratory flow rate via increased intrapleural pressure in healthy adults: an experimental study. Australian Journal of Physiotherapy 52: 267 271. Ntoumenopoulos G, Presneill JJ, McElholum M, Cade JF (2002) Chest physiotherapy for the prevention of ventilatorassociated pneumonia. Intensive Care Medicine 28: 850 856. Patman S, Jenkins S, Stiller K (2008) Physiotherapy does not prevent, or hasten recovery from, ventilator-associated pneumonia in patients with acquired brain injury. Intensive Care Medicine 35: 258 265. Pisi G, Chetta A (2009) Airway clearance therapy in cystic fibrosis patients. Acta Biomedica 80: 102 106. Journal of Physiotherapy 2011 Vol. 57 Australian Physiotherapy Association 2011 25

Research Santos LJ (2010) Efeitos da manobra de hiperinsuflação manual associada à pressão positiva expiratória final em pacientes submetidos à cirurgia de revascularização miocárdica Revista Brasileira de Terapia Intensiva : 98. Savian C, Paratz J, Davies A (2006) Comparison of the effectiveness of manual and ventilator hyperinflation at different levels of positive end-expiratory pressure in artificially ventilated and intubated intensive care patients. Heart and Lung 35: 334 341. Singer M, Vermaat J, Hall G, Latter G, Patel M (1994) Hemodynamic effects of manual hyperinflation in critically ill mechanically ventilated patients. Chest 106: 1182 1187. Stiller K (2000) Physiotherapy in intensive care: towards an evidence-based practice. Chest 118: 1801 1813. Unoki T, Kawasaki Y, Mizutani T, Fujino Y, Yanagisawa Y, Ishimatsu S, et al (2005) Effects of expiratory rib-cage compression on oxygenation, ventilation, and airwaysecretion removal in patients receiving mechanical ventilation. Respiratory Care 50: 1430 1437. Vieira DF (2009) Implantação de Protocolo de Prevenção de Pneumonia Associada à Ventilação Mecânica: Impacto do cuidado não farmacológico. Cochrane Database of Systematic Reviews: 149. Submitting randomised trials to Journal of Physiotherapy Authors are reminded that Journal of Physiotherapy accepts only registered trials. All clinical trials submitted for publication must have been registered in a publicly-accessible trials register. We will accept any register that satisfies the International Committee of Medical Journal Editors requirements (such as The Australian Clinical Trial Registry at www.actr.org.au or the US trial register at ClinicalTrials.gov). Authors must provide the name and website address of the register and the trial registration number on submission. 26 Journal of Physiotherapy 2011 Vol. 57 Australian Physiotherapy Association 2011