However, only two included studies reported costs associated with preoperative intervention23 and 24 and only one reported

a reduction in costs in the intervention group.23 Future research should also aim to include measures of cost effectiveness to allow clinicians, policy-makers and researchers to justify resource use in this population. The majority of studies included in this review had good methodological quality and only a moderate risk of bias. The largest risk of bias came from the lack of blinding, which is difficult to achieve in the setting of non-pharmacological clinical research.44 TNF-alpha inhibitor It is critical that study designs attempt to provide methods of blinding, including: sham education or rehabilitation; blinding participants to study hypotheses; and centralising assessment of outcome assessors to

minimise the risk of bias associated with non-blinding.44 The lack of concealed allocation also introduced bias into the included studies. There also may be clinical differences in people who undergo coronary artery bypass graft surgery alone versus combined GSK126 coronary artery bypass graft and valvular surgery, though these populations were analysed together. The inhomogeneity of the interventions was a limitation of this review. Also the long-term physical function outcomes of people undergoing cardiac surgery could not be attributed to their preoperative or hospital management in studies that included a follow-up period of weeks or months. During this time, it is possible that a proportion of people attended cardiac rehabilitation following cardiac surgery, which improves physical outcomes and mortality.45 Subjective measures such as pain, quality of life and anxiety were not included in this review. Finally, it was not possible to include all relevant articles in the meta-analyses, as studies did not use homogenous variables.

In conclusion, preoperative interventions reduce the risk of postoperative pulmonary complications, reduce hospital length of stay in older populations and may shorten time to extubation in people undergoing cardiac surgery. Preoperative intervention did not significantly affect ICU length of stay. The clinical significance of these improvements was small, except in the case of inspiratory PDK4 muscle training where hospital length of stay was reduced by a pooled mean difference of 2.1 days. No clear conclusions could be drawn regarding the effect of preoperative intervention on physical function or the cost-effectiveness of preoperative intervention. Further research would help in establishing the clinical significance and implications of these findings. What is already known on this topic: People undergoing cardiac surgery recover in hospital for several days postoperatively. At this time, they risk developing pulmonary complications, which typically prolong length of stay in hospital.