Thirty-five RCTs (n=4,145) were included in the review .
Study quality was limited. Only 11 RCTs reported using a satisfactory method of randomisation, 5 described allocation concealment, 16 blinded observers, and follow-up was complete in 13. Many trials were small and underpowered, and none provided full details about the intervention (e.g. patient position)
Prophylactic respiratory physiotherapy versus no intervention (13 RCTs, n=1,411).
Pneumonia (6 RCTs): one RCT, with a pneumonia rate of 37% in the control group, found a statistically significant benefit for the intervention group who received deep breathing, directed cough and postural drainage (RD 23.6%, 95% CI: 7, 40; NNT 4, 95% CI: 3, 14). The other 5 RCTs, which had pneumonia rates of 2 to 5% in controls, found no significant difference between the groups.
Atelectasis (9 RCTs): 2 RCTs, with atelectasis rates of 39% and 77% in the control groups, found a statistically significant benefit for the intervention group who received deep breathing and directed cough, with or without postural drainage (RD 24%, 95% CI: 5, 43; NNT 4, 95% CI: 2, 18; and RD 18%, 95% CI: 5, 31; NNT 6, 95% CI: 3, 19). The other 7 RCTs, which had rates of 20 to 25% in controls, found no significant difference between the groups.
Unspecified pulmonary complications (8 RCTs): one four-arm RCT found a statistically significant benefit for the three intervention groups who received incentive spirometry, deep breathing and directed cough, or IPPB (RD 25.5 to 26.3%; NNT 4). The other RCTs found no significant difference between the groups. There was marked heterogeneity in event rates in both groups for this outcome, with incidence rates of complications ranging for 0 to 50%.
Oxygenation (5 RCTs) and vital capacity (3 RCTs): no significant differences were reported between the groups for these outcomes.
Prophylactic respiratory physiotherapy versus active control (22 RCTs, n=2,734).
No RCTs reported a significant difference between the groups in rates of pneumonia or atelectasis. Five RCTs reported significant differences in vital capacity (4 RCTs), unspecified pulmonary complications (one RCT) and/or oxygenation (one RCT); their findings favoured a range of different interventions.
Reported adverse effects included intolerable abdominal discomfort associated with bilevel positive airway pressure (15% and 29% of patients, respectively, in 2 RCTs), claustrophobia and nose ulcers with CPAP (9% and 4%, respectively, in one study), abdominal distension with IPPB (18% in one RCT) and incisional hernia during chest physiotherapy in a single patient in one RCT. Four trials noted that no adverse events had occurred and 26 did not mention adverse events.