Article Text
PDF
Abstract
Objective To investigate the effectiveness of preoperative exercises interventions in patients undergoing oncological surgery, on postoperative complications, length of hospital stay and quality of life.
Design Intervention systematic review with meta-analysis.
Data sources MEDLINE, Embase and PEDro.
Eligibility criteria for selecting studies Trials investigating the effectiveness of preoperative exercise for any oncological patient undergoing surgery were included. The outcomes of interest were postoperative complications, length of hospital stay and quality of life. Relative risks (RRs), mean differences (MDs) and 95% CI were calculated using random-effects models.
Results Seventeen articles (reporting on 13 different trials) involving 806 individual participants and 6 tumour types were included. There was moderate-quality evidence that preoperative exercise significantly reduced postoperative complication rates (RR 0.52, 95% CI 0.36 to 0.74) and length of hospital stay (MD −2.86 days, 95% CI −5.40 to −0.33) in patients undergoing lung resection, compared with control. For patients with oesophageal cancer, preoperative exercise was not effective in reducing length of hospital stay (MD 2.00 days, 95% CI −2.35 to 6.35). Although only assessed in individual studies, preoperative exercise improved postoperative quality of life in patients with oral or prostate cancer. No effect was found in patients with colon and colorectal liver metastases.
Conclusions Preoperative exercise was effective in reducing postoperative complications and length of hospital stay in patients with lung cancer. Whether preoperative exercise reduces complications, length of hospital stay and improves quality of life in other groups of patients undergoing oncological surgery is uncertain as the quality of evidence is low.
- exercise
- surgery
- cancer
- review
- meta-analysis
Statistics from Altmetric.com
Introduction
In 2012, about 14.1 million new cases of cancer were diagnosed worldwide,1 with the most common types being lung (13%), breast (12%) and colorectal cancer (10%).2 Regardless of the type, cancer is a major global health problem.3 Surgery, in isolation or combined with chemotherapy and/or radiation, offers the only potential curative treatment for most malignancies.4 5 Complications from oncological surgery add to disease burden and distress for patients, often prolong hospital stay and may have negative consequences for subsequent quality of life. Therefore, effective strategies to minimise surgical complications hold potential to both improve patient outcomes and reduce healthcare costs.6–8
Recently, preoperative exercise has been suggested as a way to improve postoperative recovery.9 Previous systematic reviews have found that preoperative exercise may benefit patients with cancer through positive effects on function and physical outcomes, suggesting value in preoperative exercise, although from a limited evidence base.10–12 As this is a rapidly growing field, with the publication of additional trials,13–15 further analysis is warranted. Furthermore, explicit assessment of the potential sources of bias in the primary studies, meta-analysis of findings where possible and consideration of the strength of evidence will help to define the current evidence base.10–12 16
Knowledge of effective preoperative exercise programme on postoperative outcomes will allow clinicians to provide better guidance to patients undergoing oncological surgery and inform guideline developers and policymakers. In this systematic review and meta-analysis, we aimed to investigate the effectiveness of preoperative exercises interventions in patients undergoing oncological surgery on postoperative complications, length of hospital stay and quality of life.
Methods
The review protocol followed the methods recommended by the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P)17 and was registered on the international register for prospective reviews (http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42017055715) prior to conducting the review.
Data source and searches
A systematic electronic search was carried out in the MEDLINE (via Ovid), Embase (via Ovid) and PEDro (via http://www.pedro.org.au) databases from inception to 20 November 2016. The search was based on the Cochrane Handbook for Systematic Reviews of Interventions and used a combination of text words and Medical Subject Headings terms for randomised controlled trials (RCTs), preoperative exercise and cancer (online supplementary table 1). One review author (DS) conducted the first screening of potentially relevant records based on titles and abstracts, and two review authors (DS and PRB) independently performed the final selection of included trials based on full-text review. The search was complemented by screening the reference list of the included trials, relevant systematic reviews and clinical trials registries for unpublished studies. No language or publication restrictions were applied. Consensus between the two reviewers was used to resolve any disagreement.
Supplementary file 1
Study selection
Eligible studies were RCTs and quasi-randomised controlled trials (qRCTs) evaluating the effectiveness of preoperative exercise interventions in patients scheduled to undergo any oncological surgery. Eligible comparisons included no intervention, placebo or minimal intervention. Trials comparing two active interventions were excluded. There were no restrictions on age, sex or source of patients. Trials were included when at least one of the following postoperative outcome measures was reported: (i) complications, (ii) length of hospital stay and (iii) quality of life.
Data extraction and quality assessment
Two review authors (DS and PRB) independently extracted data from the included trials using a preconstructed data extraction form. The following data were extracted: study design, sample size, age, gender, type of cancer, description of the intervention, number of sessions, duration of the intervention, adherence to intervention, type of control and outcomes. Where we found more than one publication for a study, we extracted data from all available publications. Discrepancies in data extracted by the two review authors were resolved by consensus. When medians and IQRs were reported, we calculated means and SDs, as previously reported.18
We extracted complication rates for each individual trial (ie, as the number of patients presenting with at least one complication). Table 1 reports the definition of complication as reported for each trial.
Definition and number of postoperative complication for each included trial
Risk of bias in included studies was assessed by two review authors independently (DS and PRB) using the Cochrane Risk of Bias assessment tool.19 Each domain of the risk of bias tool (ie, random sequence generation; allocation concealment; blinding of participant and personnel; blinding of outcome assessment; incomplete outcome data; selective reporting; and other bias) was classified as low, high or unclear risk of bias. Disagreements were resolved by discussion. ‘Blinding of outcome assessment’ was judged individually for each outcome measure. Outcomes that were self-reported by the participant (eg, quality of life) were rated as high risk of bias if participants were not blinded to the intervention.19
The quality of evidence was evaluated using the GRADE approach.20 The quality of evidence was downgraded by one level according to the following criteria: (i) limitation of study design (>25% of the included studies with high risk of bias (at least one of the domains judged as high risk)),21 22 inconsistency of results (statistically significant heterogeneity (I²>50%) or ≤75% of trials with findings in the same direction)23 and (iii) imprecision (wide CIs or the total number of participants was <400 participants).24 The indirectness criterion was not considered in this review because we included a specific population with relevant outcomes and direct comparisons. Where only single trials were available, evidence from studies with <400 participants was downgraded for inconsistency and imprecision (ie, sparse data) and rated as ‘low-quality’ evidence. The quality of the evidence could be further downgraded to ‘very low-quality’ evidence if limitations of study design were found. The quality of evidence was defined as high, moderate, low or very low.
Data synthesis and analysis
Whenever raw data were available, continuous outcomes were analysed by calculating the mean difference (MD) and 95% CIs. Relative risk (RR) and 95% CI were calculated for dichotomous outcomes. Data were pooled using random effects meta-analysis where possible (comparability of population and outcomes between trials). A qualitative synthesis and construction of descriptive summary tables were performed for studies that could not be pooled. The pooled estimates (MD or RR) defining the effect between preoperative exercise and control was considered ineffective when the 95% CI crossed the no effect line (no significant difference between the groups). For RR, values<1 favoured the preoperative exercise group, and for MD, negative values favoured preoperative exercise.
A post hoc sensitivity analysis was performed to provide a more homogenous pooled effect for trials that reported on pulmonary complication rates. In brief, five trials reported postoperative complication rates for patients undergoing lung cancer surgery. Of these, four trials reported on the number of pulmonary complications13 14 25–27 while one trial reported on the number of cardiopulmonary complications28 (table 1). Therefore, the single trial reporting on postoperative cardiopulmonary complications was removed from the sensitivity analysis.
Pooled estimates were obtained with Comprehensive Meta-Analysis software V.3 (Biostat, Englewood, New Jersey, USA).
Results
The electronic search resulted in a total of 810 individual records. After screening citations by title and abstract, two review authors independently assessed 135 full-text articles. We included 17 articles reporting data from 13 different controlled trials (11 RCTs and 2 qRCTs) reporting data on 806 participants (figure 1).13–15 25–38
Flow chart showing process of selection of studies. RCT, randomised controlled trials.
Characteristics of included studies
The included trials investigated the efficacy of preoperative exercise on six different oncological surgeries: colon,29 liver resection for colorectal metastatic disease,15 30 oesophageal,31–33 lung,13 14 25–28 34 35 oral36 and prostate cancer resections.37 38 The sample size of the included studies ranged from 15 to 151 participants, with an average age of 63.3 (SD 10.2) years. The duration of the preoperative exercise ranged from 1 to 4 weeks, with the majority presenting a 2-week intervention period. The characteristics of the included trials are provided in table 2.
Characteristics of included studies
Risk of bias assessment
Around half of the trials had at least one risk of bias domain judged as high risk (table 3). These trials were at high or unclear risk of selection bias (allocation concealment), performance bias (blinding of participants and personnel) and detection bias (blinding of outcome assessment).
Risk of bias summary of the included trials
Study results
Colon cancer
One trial investigated the effect of a preoperative short-term intensive exercise programme versus exercise advice on patients undergoing colon cancer surgery (n=41).29 There was no effect of exercise on postoperative complications (RR 1.07, 95% CI 0.52 to 2.22), length of hospital stay (MD −5.40 days, 95% CI −16.72 to 5.92) or quality of life (MD −4.00, 95% CI −15.34 to 7.34). The quality of evidence was rated as very low (table 4).
Summary of findings and quality of evidence assessment (GRADE)
Colorectal liver metastases
Two reports presented data from one clinical trial investigating the effectiveness of a preoperative supervised exercise programme for patients undergoing liver resection for colorectal metastatic disease (n=34).15 30Compared with standard care, there was no evidence of effectiveness of the exercise programme to improve postoperative complication rates (RR 0.90, 95% CI 0.42 to 1.92) or length of hospital stay (MD −0.66 days, 95% CI −17.78 to 0.46). The quality of evidence was very low (table 4).
Oesophageal cancer
Three reports presented data from two trials investigating the effect of a preoperative respiratory muscle training programme on postoperative complications (n=83)32 33 and length of hospital stay (N=99)31–33 in patients undergoing oesophageal cancer surgery. For postoperative complications, data from one qRCT provided evidence of no effect of the preoperative respiratory muscle training over usual care (RR 1.18, 95% CI 0.71 to 1.97). Pooling of two trials did not reveal any difference in length of hospital stay (MD 2.00 days, 95% CI −2.35 to 6.35) (figure 2). The quality of evidence was low (table 4).
Mean difference for postoperative length of hospital stay (days) in controlled trials on efficacy of preoperative exercise for patients undergoing oesophageal and lung cancer surgery. Studies ordered chronologically. Negative values favour preoperative exercise. MD, mean difference.
Lung cancer
Eight reports presented data from five RCTs13 14 25 26 28 34 35 and one qRCT27 on postoperative complications (ie, pulmonary and cardiopulmonary complications; see table 1 for further information) and length of hospital stay in 432 patients undergoing lung cancer surgery. Five trials (n=332) were included in the meta-analysis to estimate the efficacy of preoperative pulmonary rehabilitation on postoperative outcomes. Pooling estimates provided moderate quality of evidence of a significant effect favouring preoperative pulmonary rehabilitation over control on postoperative complications (RR 0.52, 95% CI 0.36 to 0.74) and length of hospital stay (MD −2.86 days, 95% CI −5.40 to −0.33) (figures 2 and 3). The pooled effect of the post hoc analysis, after removing the study that reported postoperative cardiorespiratory complications, remained similar (RR 0.49, 95% CI 0.33 to 0.73; n=288). An additional trial (n=100) reported that preoperative pulmonary rehabilitation was effective in reducing postoperative complications and length of hospital stay compared with control, but no effect sizes were reported (table 4).35
Relative risk for number of post-operative complications in controlled trials on efficacy of preoperative exercise for patients undergoing lung cancer surgery. Studies ordered chronologically. Values <1 favour preoperative exercise. RR, relative risk.
Oral cancer
One trial reported the effects of mouth opening exercise training with follow-up telephone calls compared with control on postoperative quality of life (n=60).36 Preoperative individual exercise training with follow-up telephone calls enhanced postoperative quality of life in patients with oral cancer (no estimates reported) (table 4).
Prostate cancer
Two RCTs reported on the effectiveness of preoperative pelvic floor muscle exercise over control on self-reported quality of life following radical prostatectomy. One trial (n=118) assessed quality of life using the International Continence Society Male Short form and reported a significant mean difference at 1 month (MD −3.70; P=0.002) and 3 months (MD −4.10; P=0.002) postoperative, favouring the exercise group.37 The other trial (n=16) assessed quality of life using the mental and physical component of the Prostate Cancer Index at the postoperative follow-up. There was no significant difference between the groups for either the mental or physical component scores.38 The quality of evidence was very low (table 4).
Discussion
We found moderate-quality evidence that preoperative pulmonary rehabilitation was effective in reducing postoperative complications and length of hospital stay in patients undergoing lung cancer surgery. For other oncological surgeries, including colon, liver resection for colorectal metastatic disease, oesophageal, oral and prostate cancer, the effectiveness of preoperative exercise on postoperative complications, length of hospital stay and quality of life was uncertain due to the limited number of trials included and the low quality of evidence.
Comparison with other studies
While the findings from our review are somewhat in line with previous reviews, our study provides higher-quality evidence and a more realistic estimate of the effects of preoperative exercises in people undergoing cancer-related surgical procedures. Although several reviews have investigated the effectiveness of a preoperative exercise programme on postoperative outcomes across a range of conditions,10 12 16 39–41 a meta-analysis was only performed by three reviews.12 16 41 These reviews reported a reduction on postoperative complication that ranged from 45% to 67% and reduced length of stay of 4–5 days in the preoperative exercise group compared with control. The pooled effects reported from these reviews were mostly higher than the ones reported on our review; however, this could be due to the high risk of bias introduced by the inclusion of non-randomised studies (eg, case series, cohort studies), the inclusion of two active interventions (eg, strength and endurance training vs breathing exercise for lung expansion) and the inclusion of populations undergoing non-cancer-related surgery (eg, cardiac and abdominal surgery). Therefore, it is possible that the effects had been overestimated in previous studies.
Meaning of the study
Postoperative complication is a major concern for patients undergoing oncological surgery. Our review found that preoperative exercise compared with a control intervention on patients undergoing lung cancer surgery reduces postoperative complication rates by 48% and length of hospital stay by almost 3 days, and should therefore be considered as a preoperative standard care. Most of the preoperative exercises investigated in the included trials (83%) incorporated both aerobic (eg, walking) and respiratory muscle training exercises (eg, breathing exercises), which were performed for 1–2 weeks preoperatively. The frequency of the exercises varied from three times a week to three sessions daily. Trials that performed a larger number of sessions per week reported a larger effect size, suggesting a dose–response relationship.
Two trials investigated the effectiveness of preoperative exercise for patients undergoing oesophageal cancer surgery. The interventions described in these two trials were based on respiratory muscle training only, reporting evidence of no effect. Future studies should consider the inclusion of aerobic and resistance training components to their exercise intervention as they may positively impact on the postoperative outcomes.
Our review provides preliminary evidence that a preoperative exercise programme that incorporates aerobic, resistance and respiratory muscle training improves postoperative outcomes in patients undergoing cancer surgery by reducing postoperative complication rates and length of hospital stay. Such findings may also impact on healthcare costs and on patients’ quality of life, and consequently have important implications for patients, healthcare professionals and policymakers. However, futures studies are required to confirm our results, investigate cost-effectiveness and test the feasibility of implementation.
Unanswered questions and future research
The results of our systematic review and meta-analysis clearly warrant the development of new high-quality RCTs aimed at investigating the effect of a preoperative exercise programme for most major oncological procedures. Since preoperative pulmonary rehabilitation was effective in reducing postoperative complication and length of hospital stay in patients with lung cancer, it is possible that similar exercise and educational programmes could benefit patients undergoing other oncological surgeries, although the literature is currently lacking. Most of the included trials were not appropriately powered to detect a between-group difference on the postoperative outcomes reported, considering that the sample size ranged from 15 to 151 participants in the included studies. Therefore, future studies should also include an appropriate sample size, performing a formal sample size calculation.
There is evidence that preoperative quality of life in patients with cancer is poor and decreases even further after surgery.42 In addition, despite the evidence suggesting that exercise improves physical and mental health in patients with cancer,43 there are only a limited number of trials investigating the effect of preoperative exercise on patients’ quality of life. Therefore, the effect of preoperative exercise on quality of life at short-term and long-term postoperation should be explored in future trials. Lastly, if the efficacy of preoperative exercise is confirmed by future high-quality trials, the cost-effectiveness of preoperative exercise interventions should also be investigated.
Strengths and weaknesses of the study
Different from previous reviews conducted in the field, our review followed rigorous methodological criteria to provide the most robust evidence on the effectiveness of preoperative exercises in people undergoing cancer-related surgery. Some of the strengths of our review include that it was prospectively registered on PROSPERO and followed the methods recommended by the Cochrane Handbook for Systematic Reviews of Intervention and written in accordance with the PRISMA statement, including the use of GRADE to consider the quality of the evidence. We used a broad search strategy, which was able to identify all trials investigating the effectiveness of preoperative exercise on diverse oncological procedures.
Our review has some limitations. First, the limited number of trials, small sample size, poor reporting and the low quality of evidence prevented pooling of some included surgical oncological trials and outcomes. Second, the type of exercise, duration and reporting on adherence to the exercise programme of the trials included in our systematic review and meta-analysis varied throughout. Third, the outcome measures assessed, including complication rates and quality of life, were not standardised across trials, with some not reporting the estimates. Finally, we did not have an adequate number of studies to assess a funnel plot, so it is possible that publication bias exists. The review authors extracted complication rates for each individual trial as the number of patients presenting with at least one complication, despite the fact that some patients had more than one complication. Furthermore, the definition of postoperative complication varied between the included trials (table 1). This needs to be considered when interpreting the results. Finally, there is very limited evidence on short-term and long-term effect on postoperative quality of life as most trials did not assess this outcome.
Summary and conclusions
There is moderate-quality evidence that preoperative exercise reduced postoperative complication rates by 48% and length of hospital stay by 3 days in people who undergo a surgical procedure due to lung cancer. For other oncological surgeries, including colon, liver resection for colorectal metastatic disease, oesophageal, oral and prostate cancer, there is very limited evidence that preoperative exercise may be beneficial for reducing surgical complications and length of stay, and improving quality of life.
What is already known?
Preoperative exercise in people with cancer has been suggested to improve postoperative recovery, but there is currently insufficient synthesis of high-quality evidence to inform a change in clinical practice.
What are the new findings?
Preoperative pulmonary rehabilitation was effective in reducing postoperative cardiopulmonary complications and length of hospital stay in patients undergoing lung cancer surgery.
For other oncological surgeries, the effectiveness of preoperative exercise on postoperative complications, length of hospital stay and quality of life was uncertain due to the limited number of trials included and the low quality of evidence.
References
Footnotes
Contributors DS, PRB, MH, MS and JY conceived and designed the study, analysed and interpreted the data, and contributed by drafting the article. MH, MS and JY provided statistical expertise. DS and PRB collected and assembled the data. DS is guarantor. All authors participated in the revision and final approval of the manuscript.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.