Clearance of airway secretions, or airway hygiene, is a normal physiological process needed for the preservation of airway patency and the prevention of respiratory tract infection. Impaired clearance of airway secretions can result in atelectasis and pneumonia, and may contribute to respiratory failure prompting admission to an intensive care unit (ICU).

Physical methods to augment the clearance of secretions are often used in the ICU. This review focuses on mechanical methods and pharmacological agents commonly used to maintain airway hygiene in the ICU, and their effect on clinical outcomes of critically ill patients. The impact of oral hygiene, tracheal suctioning, bronchoscopy, mucus-controlling agents, and kinetic therapy on the incidence of nosocomial respiratory infections, length of stay in the hospital and the ICU, and mortality will be discussed. Where possible, we have distilled available data into recommendations for airway hygiene in ICU patients.

Literature for this article was identified by searching the PubMed database (1966 to present). English-language articles of relevance were selected and reviewed. Additional resources were obtained through the bibliographies of reviewed articles. The following search terms were used: airway hygiene, oral hygiene, mucociliary clearance, tracheal suctioning, bronchoscopy, mucolytics, chest physiotherapy, kinetic therapy, continuous lateral rotational therapy, selective digestive decontamination, nosocomial pneumonia, hospital-acquired pneumonia, ventilator-associated pneumonia, and pneumonia.

A summary of recommendations and associated strength of supporting evidence for strategies used in the reduction of nosocomial pneumonia is shown in Table 1. The following grading system described by Kollef 1 was used to assess strength of evidence: A, supported by at least two randomized, controlled investigations; B, supported by at least one randomized, controlled investigation; C, supported by nonrandomized, concurrent-cohort investigations, historical-cohort investigations, or case series; U, undetermined or not yet studied in clinical investigations.

A general tenet of hospital-acquired pneumonia and ventilator-associated pneumonia (VAP) is that infections of the lower respiratory tract are preceded by colonization or infection of the upper airway. Thus, most hospital-acquired pneumonias stem from micro- or macro-aspiration of infected secretions from the upper airway. Methods that reduce oropharyngeal colonization have been postulated to reduce infections of the lower respiratory tract in the critically ill. Surprisingly, nasopharyngeal and oropharyngeal hygiene often receives little attention from intensivists.

Many risk factors have been linked to airway colonization, including severity of illness, length of hospitalization, prior or concomitant antibiotic use, malnutrition, endotracheal intubation, azotemia, underlying pulmonary disease, inadequate oral hygiene, substandard infection control practises during bathing, tracheal suctioning, enteral feeding, and endotracheal tube manipulation. Furthermore, critical illness is associated with alterations in oral mucosal-cell-surface fibronectin and carbohydrate expression patterns, disruption of the mucosa by endotracheal tubes and suction catheters, and increased secretion of mucus, all of which provide binding sites for pathogenic bacteria 2345. Gram-negative bacteria and Staphylococcus aureus replace the normal upper-airway flora of patients hospitalized for more than 5 days 6. Not surprisingly, many patients (50%) have pathogenic bacteria in their oropharynx on admission to the ICU 7. The periodontal areas, oropharynx, sinuses, stomach, and trachea are colonized by the end of the first week in most mechanically ventilated patients 789. Tracheal and oropharyngeal colonization is associated with VAP, whereas gastric colonization in the setting of acid suppression therapy may not be 7810.

Oral antiseptic rinses such as chlorhexidine gluconate reduce the rate of nosocomial pneumonia in critically ill patients 111213. The twice-daily use of chlorhexidine gluconate 0.12% oral rinse reduced respiratory tract infections by 69% and antibiotic use by 43% in cardiac surgical patients postoperatively without affecting antibiotic resistance patterns 11. The impact was greatest in patients intubated for more than 24 hours and who had the highest degree of bacterial colonization 13. The cost of nursing care did not significantly increase with the use of chlorhexidine oral rinse, and the liquid form was easier and quicker to apply than an antibiotic paste 1113. Despite the reduction in bacterial colonization and/or nosocomial pneumonia, the use of oral antiseptic rinses has not been shown to affect the duration of mechanical ventilation or survival, although it may modestly reduce the length of stay in the hospital 1214.

Selective decontamination of the oropharynx, and gastric and subglottic area, or selective digestive tract decontamination (SDD) with several different regimens have been proposed as a method for decreasing VAP. These SDD regimens have included one or more of the following antibiotics: polymixin, tobramycin/gentamycin, vancomycin, and amphotericin B suspension. More than 50 randomized controlled trials and 12 meta-analyses evaluating the SDD prophylaxis have been published 151617181920212223242526272829. These trials were conducted in unselected medical and surgical patients who required 2 days or more of mechanical ventilation and in selected patients including liver transplant, cardiac surgical, and burn patients. Most randomized control trials found a beneficial effect in reduction of bacterial colonization and/or VAP; however, effects on length of stay in the ICU and mortality are inconsistent. Although some analyses suggest a reduction in mortality in selected patients 2427, others have demonstrated only a modest effect with the use of combined (topical and systemic) SDD therapy 181923. Survival advantage seems to be limited to surgical ICU patients treated with a combination of parenteral and topical prophylactic antibiotics 182430. A modified SDD regimen that includes oral vancomycin reduces the rate at which isolates of methicillin-resistant Staphylococcus aureus are found 3132. The impact of this regimen on the incidence of nosocomial pneumonia is less clear 3132. Furthermore, a transient increase in isolation of vancomycin-resistant Enterococcus has been observed with this regimen 32. Failure of SDD strategies to reduce morbidity and mortality consistently in critically ill patients may be related to suboptimal study design, the emergence of antibiotic-resistant bacteria, and an alteration in hosts' normal bacteriologic flora 33. Additional toxicity and increased healthcare costs related to SDD 172734 have further dampened the enthusiasm for this preventive strategy. On the basis of available evidence, SDD with topical antibiotics should not be routinely used in the ICU. If combined parenteral and topical SDD is used in surgical/trauma patients, clinicians should be vigilant in monitoring for the emergence of new multidrug-resistant organisms.

A recent randomized controlled trial of oral topical iseganan, an antimicrobial peptide with a broad in vitro activity against aerobic and anaerobic Gram-positive and Gram-negative bacteria and yeasts, was stopped early because of a higher, although not statistically significant, rate of VAP and death among mechanically ventilated ICU patients in the iseganan arm 35. The widespread use of biocides such as chlorhexidine in hospital, domiciliary, industrial, and other settings raises concern for the development of antibiotic-resistant infection. Fortunately, no clinically significant alteration of antibiotic resistance patterns has been noted in epidemiologic studies 3637. Therefore, chlorhexidine gluconate oral rinse, an inexpensive and easily applied agent, should be a routine aspect of care of the critically ill patient.

Airway hygiene is impaired in critically ill patients as a result of depressed cough reflex and ineffective mucociliary clearance from sedation, high inspired oxygen concentrations, elevated endotracheal tube cuff pressure, and tracheal mucosal inflammation and damage 383940. Accordingly, care of intubated patients includes tracheal suctioning to facilitate the removal of airway secretions. Routine suctioning via endotracheal tubes is often performed on the basis of the assumption that it maintains airway patency and prevents pulmonary infection. However, tracheal suctioning induces mucosal injury, exposing the basement membrane and thereby facilitating bacterial adhesion in vitro 41 and in low-birthweight babies 42. Suctioning has also been associated with many deleterious effects including decreased arterial oxygen tension (12 to 20 mmHg) 434445, which may be more pronounced in cardiac surgical patients 43 and in patients with hypoxemic respiratory failure requiring high levels of positive end-expiratory pressure 46, in patients with cardiac arrhythmias (7 to 81%) 4547, and in cardiac arrest in patients with acute spinal cord injury 48. These detrimental effects of tracheal suctioning, however, may be mimimized through the use of manual hyperinflation, preoxygenation, sedation, and optimal suctioning technique 495051. Normal saline is frequently instilled into the trachea before endotracheal suctioning under the assumption that it may help dislodge secretions and facilitate airway clearance. However, Ackerman and colleagues have shown that intratracheal instillation of 5 ml of normal saline adversely effects oxygen hemoglobin saturation and has no effect on the clearance of secretions 5253. In addition, a 5 ml saline instillation dislodges fivefold more viable bacterial colonies from the endotracheal tube than does the insertion of a tracheal suction catheter alone 54 and is associated with prolonged deoxygenation 5556. Limiting the frequency and duration of tracheal suctioning and limiting the use of saline instillation may prevent its adverse effects without affecting the duration of mechanical ventilation, length of stay in the ICU, mortality in the ICU, and incidence of pulmonary infection 455758. Thus, tracheal suctioning of intubated patients should be performed on an as-needed basis that is defined by the quantity of secretions obtained, not at prescribed, set intervals.

Closed (in-line), repeated-use endotracheal suction devices have become commonplace in ICUs. These devices eliminate the need for disconnection from mechanical ventilation and do not require single-hand sterile technique of open suction methods. Available data on the effect of these catheters on tracheal colonization are conflicting 5960. A single study has reported that in-line suction systems are associated with a reduction in the incidence of VAP 61, but most studies and a recent meta-analysis find no beneficial effect on nosocomial pneumonia 59626364. Their cost-effectiveness, however, is questionable, with some studies demonstrating cost savings 6566 and others suggesting higher costs associated with closed systems 62636768. Despite the failure of closed endotracheal suctioning systems to reduce VAP or mortality, these devices may be preferable because of their efficiency and smaller number of suction-induced complications 65 and cost-effectiveness among patients requiring more than 4 days of mechanical ventilation 66. Routine changes of these devices are not required 69.

A more recent development is the use of endotracheal tubes with a dorsal suction channel that opens immediately before the inflatable cuff in the subglottic area. Suction can be applied via this port continuously or intermittently for purposes of removing secretions from the subglottic space. A recent meta-analysis reported that these endotracheal tubes reduce early-onset VAP in patients expected to require more than 72 hours of mechanical ventilation. Early-onset VAP was defined as pneumonia occurring 5 to 7 days after intubation. Duration of mechanical ventilation, the length of ICU stay, and mortality were not different when intention-to-treat data were summarized 70. However, subglottic suctioning does not alter oropharyngeal or tracheal bacterial loads 71, nor does it significantly reduce the duration of mechanical ventilation, length of ICU stay, or mortality 7072737475. Furthermore, continuous aspiration of subglottic secretions can cause severe tracheal mucosal damage 76. Endotracheal tubes with a dorsal suction channel are significantly more expensive than standard endotracheal tubes, increase nursing workload, and therefore have the potential to increase ICU costs 77. A case cost analysis model suggests a potential cost-saving associated with the use of subglottic suctioning primarily if these tubes decrease the length of the ICU stay 77, which has not been observed 72737475. The use of endotracheal tubes that allow subglottic suctioning outside populations with a high incidence of early-onset VAP cannot be recommended at present.

Distilled water, normal saline, hypertonic and hypotonic saline have long been used to 'loosen' thick airway secretions and promote airway hygiene. However, formed mucus does not readily incorporate topically applied water 78. Increased sputum production and clearance attributed to bland aerosols, especially hypertonic saline, may be due to the irritant nature of the aerosol, which may cause bronchoconstriction rather than mucolysis 798081.

N-Acetylcysteine (NAC) is a mucolytic agent that breaks disulfide bonds of mucus, reducing its viscosity and elasticity, and has anti-inflammatory properties in experimental models 7982. As a result of its mucolytic properties, many practioners advocate its use in the ICU for assistance in the clearance of airway secretions. Although there are limited data on the use of NAC in ICU patients, it is important for clinicians to recognize the potential deleterious effects of this practice. In vitro, NAC may antagonize aminoglycoside and β-lactam antibiotics 8384. Additionally, NAC at concentrations less than 10% inhibits the growth of Pseudomonas strains in vitro 84, potentially causing false-negative sputum cultures. Delivery via aerosol thins airway secretions but does not change pulmonary function or sputum volume in patients with stable chronic bronchitis 8586. Furthermore, aerosolized NAC can cause bronchoconstriction and inhibit ciliary function 8788. Concomitant administration of a bronchodilator partly ameliorates NAC-induced bronchospasm 89. Gas exchange may worsen acutely after NAC administration, possibly because liquefied secretions gravitate into smaller airways 90. Tracheal instillation of NAC through an endotracheal tube or bronchoscope may induce the rapid accumulation of liquefied secretions that must be suctioned immediately to prevent asphyxia 91. Direct tracheal instillation of NAC is more effective than aerosol inhalation in the treatment of atelectasis caused by mucoid impaction 9293; however, it remains unclear whether undirected instillation of NAC results in delivery to areas of mucus accumulation. Thus, despite its widespread use, few data are available to support NAC as a mucolytic agent. Dornase alfa (recombinant human DNase) is used as a mucolytic in cystic fibrosis and other bronchiectatic conditions 9495. Case reports of its use in status asthmaticus 96, acute respiratory distress syndrome 97, and mechanically ventilated pediatric patients 9498 have been published but no prospective efficacy studies are available to support its use in adult ICU patients for airway hygiene.

β₂-Adrenergic agonists increase ciliary beat frequency in experimental models, raising the possibility that they may be useful for airway hygiene 99. Salbutamol increases large-airway mucociliary clearance, both in stable patients with chronic obstructive pulmonary disease and in healthy subjects 100, although this may not be true for smaller airways 101. There are no data from ICU patients. Thus, despite the simplicity and attractiveness of this approach, there are no data to support the use of inhaled β₂-adrenergic agonists as adjuncts for airway hygiene in ICU patients.

Immobility impairs cough and mucociliary clearance in patients receiving mechanical ventilation, thereby promoting retention of secretions 102103. Kinetic therapy with beds that intermittently or continuously rotate patients along their longitudinal axis by 40° or more have gained acceptance in the care of the critically ill patients 104.

A modest body of data regarding the effect of kinetic beds for continuous lateral rotation therapy (CLRT) to facilitate airway hygiene is available. Some of these studies found that CLRT reduced the incidence of pneumonia and decreased the length of stay in the ICU in heterogeneous groups of critically ill patients receiving mechanical ventilation 105106107108109. Conversely, other studies detected no meaningful effect on measures of care in the ICU 110111. Differences in diagnostic criteria for nosocomial pneumonia and variable use of broad-spectrum antibiotics in treatment and control groups may be responsible for the diametrically opposed results of these studies. Importantly, three of the five studies demonstrating benefit from CLRT were supported by kinetic bed manufacturers 105106107. Data from these studies support the use of CLRT in critically ill patients. The increased cost of CLRT is offset by reduced nosocomial pneumonia and antibiotic use and a decreased length of stay in the hospital and the ICU 106108109112.

Chest physiotherapy, including gravity-assisted drainage, chest wall percussion, chest wall vibrations, and manual lung hyperinflation, aids in the re-expansion of atelectatic lung 113114115 and increases peak expiratory flow rates 116. Chest physiotherapy treatment is as effective as early bronchoscopy in patients with acute atelectasis 114. A single study examined the role of chest physiotherapy in patients at high risk for post-operative respiratory failure. Hall and colleagues randomized 301 spontaneously breathing patients at high risk for respiratory complications to receive incentive spirometry (IS) or IS plus chest physiotherapy after abdominal surgery 117. The incidence of respiratory complications was not significantly different between two study groups. The inclusion of physiotherapy, however, required 30 minutes of staff time per patient. A recent meta-analysis examined the role of various chest physiotherapy techniques including IS, deep breathing exercise, chest physical therapy (cough, postural drainage, percussion/vibration, suction, and ambulation), intermittent positive-pressure breathing, and continuous positive airway pressure (CPAP) 118. In patients who underwent abdominal surgery, any type of lung expansion technique was better than no intervention in the prevention of pulmonary complications. Although no particular intervention seemed superior, IS may require the least staff time, while CPAP may be particularly beneficial in patients who have limited ability to participate with other physiotherapy techniques 118. In contrast, others reported poor adherence to an independent use of IS among patients recovering from surgery and superiority of encouragement by hospital personnel to the use of IS 119120. Current evidence does not support the use of IS in the prevention of complications after cardiac or abdominal surgery 121122. When nosocomial pneumonia is used as a specific endpoint, only one study showed the beneficial effect of physiotherapy 123. Routine chest physiotherapy therefore cannot be recommended for prophylaxis against respiratory complications in spontaneously breathing patients.

There are mixed data about chest physiotherapy in mechanically ventilated patients. A single study of patients receiving mechanical ventilation for 48 hours or more showed a reduced incidence of VAP 124, whereas a second trial in trauma patients was unable to detect an effect on nosocomial pneumonia rates 125. Furthermore, chest physiotherapy may cause cardiac arrhythmias, bronchospasm, and transient hypoxemia, and may prolong the duration of mechanical ventilation 126127128. These complications, as well as increased staff use and cost, combined with a paucity of data regarding beneficial effects on the prevention of nosocomial pneumonia, should limit the use of chest physiotherapy to the ICU patients with acute atelectasis, exacerbations of bronchiectasis, and conditions that are characterized by excessive sputum production and impaired cough.

Limited data are available about flutter valves, cornet type devices, or intrapulmonary percussive ventilation in ICU patients. Cough assist devices, which generate a strong expiratory flow through the application of positive pressure followed by the application of negative pressure, have been shown to aid in the removal of secretions 129130 and may avoid intubation in patients with neuromuscular disease 131. However, the long-term benefit of these devices in other ICU populations is unclear. A small study of pneumatic high-frequency chest compression in long-term mechanically ventilated patients found it to be just as efficacious as percussion and postural drainage 132, and another study of patients immediately after cardiopulmonary bypass surgery reported reduced sternotomy pain with cough 133.

Atelectasis occurs when alveolar closing volume rises above functional residual capacity and is rarely due to proximal airway obstruction by mucus. Critically ill patients have elevated closing volumes as a result of increased lung water, advanced age, and low functional residual capacity due to respiratory muscle weakness, sedation, and supine positioning 134. Fiberoptic bronchoscopy is frequently requested and often performed for 'pulmonary toilet' or treatment of atelectasis in critically ill patients; however, its utility in improving gas exchange and preventing nosocomial pneumonia is limited. Only one randomized study compared the efficacy of fiberoptic bronchoscopy with chest physiotherapy for the treatment of acute atelectasis 114. Thirtyone consecutive patients with acute lobar atelectasis were randomly assigned to receive immediate fiberoptic bronchoscopy with airway lavage followed by chest physiotherapy every 4 hours for 48 hours or to receive chest physiotherapy alone. Patients who were assigned to chest physiotherapy alone underwent 'delayed fiberoptic bronchoscopy' if their atelectasis persisted at 24 hours. Radiographic resolution of lung volume loss with bronchoscopy was similar to that after the first chest physiotherapy treatment in the control group (38% versus 37%). Although half of the patients who were randomized to chest physiotherapy alone underwent 'delayed' bronchoscopy, this study provided the impetus for teaching pulmonologists-in-training that there is rarely a role for bronchoscopy for atelectasis. In addition, bronchoscopy in patients with acute or incipient respiratory failure is not without adverse effects. Deterioration of gas exchange, barotrauma from elevated airway pressures, increased myocardial oxygen demand, and intracranial pressure have been reported 135136137138. Nevertheless, subgroup analysis 114, case series 137139140141142143, and clinical experience support a role for bronchoscopy for some patients with atelectasis: specifically, critically ill patients with acute wholelung, lobar, or segmental atelectasis without air bronchograms 114 who are unable to maintain airway hygiene independently and remain symptomatic after 24 hours of aggressive chest physiotherapy (every 4 hours). Clinical experience indicates that atelectasis recurs frequently after bronchoscopy because the cause of compromised airway hygiene continues. Thus, failure to resolve the primary problem should not be an indication for repeated invasive intervention in the airways. There is no role for empiric 'cleaning of the airways' with a bronchoscope.

The rationale for airway hygiene is the prevention of pneumonia and respiratory failure. Current practice reflects clinical experience and expert opinion and not the results of controlled clinical trials. Several evidence-based recommendations can now be made about airway hygiene in critically ill patients. First, oropharyngeal decontamination with oral biocide rinses reduces the incidence of nosocomial pneumonia and should be part of the routine care of the mechanically ventilated patient. Careful monitoring of antibiotic resistance patterns after the initiation of routine biocide use is prudent at this time. Second, tracheal suctioning should be performed only on an 'as needed' basis. Closed (in-line) suction catheter devices should be used and routine changes of these devices are not necessary. Third, aerosolized mucus-controlling agents, such as NAC, and routine instillation of saline have no demonstrable effect on the clearance of airway secretions and should not be routinely used in critically ill patients. Fourth, kinetic therapy decreases the rate of nosocomial pneumonia, and may reduce the length of stay in the ICU and the hospital. The high cost of kinetic beds may be offset by a shortening of stay and a decreased use of systemic antibiotics. Fifth, chest physiotherapy should be limited to patients with acute atelectasis and/or excessive sputum production who are unable to conduct airway hygiene independently. Sixth, the therapeutic role of bronchoscopy in acute atelectasis is limited and transient, and should be reserved primarily for patients with acute atelectasis involving more than a single lung segment in the absence of air bronchograms who remain symptomatic after 24 hours of chest physiotherapy.

CLRT = continuous lateral rotation therapy; CPAP = continuous positive airway pressure; ICU = intensive care unit; IS = incentive spirometry; NAC = N-acetylcysteine; SDD = selective digestive decontamination; VAP = ventilator-associated pneumonia.

The authors declare that they have no competing interests.

All authors participated in literature review and manuscript preparation.