Twenty-eight RCTs (number of participants unclear, range one to 457) were included: 15 parallel group trials and 13 crossover trials. Eleven low level studies were included for gaps in the evidence for which there were no RCTs. Jadad scores ranged from 1 to 5 (average 2) and PEDro scores ranged from 4 to 9 (average 6). Criteria most commonly not fulfilled by the included studies were appropriate randomisation, adequate concealment of treatment allocation and blinding.
Hyperinflation (three RCTs): One RCT (n=17 participants) reported no effects of hyperinflation on oxygenation, heart rate, blood pressure, mixed venous oxygen saturation and dynamic compliance following oxygen saturation. One RCT (n=15) found that hyperinflation improved lung compliance and decreased airway resistance for up to 30 minutes following suction. A second RCT (n=20) compared manual with ventilator hyperinflation and found no significant difference in sputum weight, but both methods improved static compliance following suction. One RCT (n=16) found that suctioning through an adaptor resulted in significantly higher pressure of arterial oxygen values for up to one hour post-section compared with off-ventilator suction.
Open suction versus close suction (12 RCTs): Two randomised cross over studies (n=19) reported no significant difference between open and closed suction on arterial oxygen saturation, but a beneficial effect in favour of closed suction on end-expiratory lung volume (WMD 12.69dL, 95% CI 9.30 to 16.09). Five RCTs showed no difference between open and closed suctioning on ventilator associated pneumonia; three of these trials also reported no difference in terms of incidence of ventilator associated pneumonia per 1,000 patient days. There was no difference between open suction and closed suction for oxygenation (two RCTs), length of hospital stay (two RCTs) and duration of mechanical ventilation (two RCTs).
Frequency of changing closed suction systems (two RCTs): There was no difference between 24 hour and 48 hour changes of closed suction systems for incidence of ventilator associated pneumonia, mortality, length of stay in intensive care unit, duration of mechanical ventilation and colonisation of catheter tips.
Subglottic suctioning (two RCTs, one observational study, one case report): There was evidence to support the use of subglottic suction for the prevention of pneumonia in patients expected to require greater than 72 hours of mechanical ventilation.
Minimally invasive suctioning (two RCTs): The two studies had methodological limiations that made it impossible to draw conclusions.
Saline instillation (three RCTs): Use of saline with suctioning may cause a decrease in oxygen saturation although this change may not be clinically significant.
Medications administered during suctioning (three RCTs, one observational study): One RCT found that remifentanil (15ng/mL) was associated with less coughing with suctioning and greater decreases in mean arterial pressure and heart rate compared to lower doses (5ng/mL or 10ng/mL). This was confirmed in an observational study. One RCT found that a homeopathic remedy reduced secretions, facilitated earlier intubation and decreased intensive care unit stay in patients with chronic obstructive pulmonary disease who had failed extubation due to profuse tenacious stringy secretions.
Outcomes assessed only in single studies or observational studies were clinical indication for suctioning patients (one observational study), hyperoxygenation (one observational study), infection control issues (two case series), airway pressures (one RCT, one bench study), size of suction systems (one RCT, one bench study) and continuous versus intermittent suctioning (one observational study).