Lifestyle interventions are feasible in patients with colorectal cancer with potential short-term health benefits: a systematic review

Participants: adults ≥18 years of age with non-metastatic CRC.

Interventions: modification of ≥1 lifestyle factor (weight, diet, physical activity [PA], smoking and alcohol).

Primary outcomes: to assess feasibility of performing lifestyle interventions in CRC patients (e.g. recruitment, adherence and follow-up/retention rates)

Secondary outcomes: summarise any documented short (e.g. quality-of-life measurements) and long-term health outcomes (e.g. cancer recurrence, cancer-specific survival)

Study design: randomised controlled trials (RCTs)

In June 2015, a clinical librarian searched the following databases to identify relevant publications: Dynamed, Cochrane Database of Systematic Reviews, OVID MEDLINE, OVID EMBASE, PEDro. Keywords used are the following: colon cancer, rectal cancer, colorectal cancer, obesity, diet, excess weight, smoking, alcohol, physical activity, exercise, randomised controlled trials, systematic reviews, lifestyle factors, lifestyle intervention, recurrence, disease-free survival, overall survival, short-term outcomes, long-term outcomes. Exclusion keywords were: risk factors, screening uptake, screening awareness. Clinicaltrials.​gov website was searched for relevant registered trials. There were no restrictions on date or language of publication. Expert opinions, letters and comments were excluded.

The resultant citations (abstracts) were independently examined with duplicates removed (SJM and AB). After screening, the full text was obtained and those papers independently assessed for eligibility, with attention paid to reference lists of reviews. Each reviewer then independently allocated relevant papers to the four lifestyle factor sections: alcohol, diet/excess weight, PA and smoking. A paper could be allocated to more than one section if more than one lifestyle factor was assessed. For further information, the authors were contacted by e-mail and any reviewing disagreements were resolved by discussion or by conferring with one further reviewer (ASA or NM).

The Jadad scale [19] was applied to each paper to assess the methodological quality of the included trials. The quality scale involved three items: randomization, double blinding and withdrawals or dropout. The score ranged from 0 to 5: 0–2 for randomization, 0–2 for blinding and 0–1 for withdrawals or dropout, with a score of 2 or less considered methodologically poor [20]. Data were collected independently by the two reviewers and crosschecked.

With the success of enhanced recovery after surgery programmes that focus on early mobility after surgery [37], researchers assessed pre-operative PA interventions (‘pre-habilitation’). The two published studies support the feasibility of performing a pre-operative physical activity intervention in patients with CRC; however, both studies contained mixed benign and malignant populations [23, 29]. In addition, colonic cancer accounted for the majority of CRC patients recruited. The exercise intervention was stationary cycling in both studies with mean duration of intervention completed 7.4 and 3.8 weeks respectively. Low dropout rates were reported (10.5 and 26%) with one study reporting an 84% follow-up rate, suggesting the interventions are tolerated well by patents. In relation to physical fitness improvements after the cycling intervention, one study reported no differences in 6-min walk test (6MWT) between groups at the two time points tested: pre-surgery (at the end of pre-habilitation) or post-surgery [mean 9.6 weeks, SD 3.4] [23]. In comparison, the other study found significantly improved cardiorespiratory function with an increased peak power (26% increase versus 0% in controls), reduced heart rate and oxygen uptake in their intervention group only [29]. It is possible that the study of Carli et al. was contaminated by their control group, increasing their 6MWT (pre-habilitation −10.6 m versus controls +8.7 m), a phenomenon described by Courneya et al. in 2003, where controls can be indirectly incentivised to exercise more [25]. Alternatively, the authors argued that the controls were not an optimal comparative group as their walking programme encouraged a minimum of 30 min/day, which means many will have achieved more than the weekly recommended UK guidelines for PA [38]. As a result, this group went on to re-analyse their data, combining the two groups to determine the overall affect of pre-habilitation on 6WMT, finding increases in 33%, no change in 38% and decreases in 29% [24]. These results translated in to significantly improved re-operative rates (29% in those that had deteriorated compared to the increased or no change groups combined, 18 and 2% respectively) and earlier recovery to baseline function 2–4 months after undergoing surgery (32 versus 77 versus 59%).

This is the lifestyle area where most research has been performed. Courneya et al. (2003) published the first RCT assessing PA intervention in CRC survivors (74% colon cancer) that had completed their treatment at least 3 months previously [25]. For approximately 16 weeks, patients were randomised to a home-based exercise programme (cycling, swimming or walking) and weekly telephone calls whilst the controls received no exercise prescription but did receive weekly phone calls. Recruitment was only 35% with a high retention rate of 92%, suggesting the intervention was well tolerated. At follow-up testing, there were no differences in QoL measurements or PA between groups (FACT-C mean difference −1.3, 95% CI −7.8 to 5.1, p = 0.679). In addition to the study not being powered, the authors proposed that contamination by the control group could explain the results (documented at 51.6%) and explored this by performing an ancillary analysis comparing those patients that had increased physical fitness to those that had decreased. This time significantly improved QoL scores in the increased physical fitness group were found (6.5, 95%CI 0.4–12.6, p = 0.038).

The next four trials support the findings of Courneya et al. with a home-based telephone-guided PA intervention [26, 30‐33]. The patient populations vary with some containing mixed cancer populations, and it is not always stated what percentage of the CRC patients recruited had rectal cancer. The exact interventions vary and can contain aerobic and strength components, making direct comparisons between studies difficult. However, low dropout rates (9–19%) with good adherence and retention rates suggest these interventions are feasible and well tolerated by CRC patients.

The majority of these studies have documented improvements in their physical parameters in CRC patients that had undergone adjuvant PA intervention. In the RENEW trial, older and overweight cancer survivors had significantly increased levels of activity, with the mean function scores and lower extremity function declining less rapidly in comparison to the control group that had delayed intervention after 12 months [31, 32]. At 2-year follow-up, the rate of decline in physical function had significantly slowed (seen in both the immediate and delayed intervention groups), but increased in the year after the intervention finished in the immediate group, displaying a protective effect of PA. Other studies have reported improved 6MWT (increased by 186.9 ft versus 81.9 in controls, p = 0.006), VO2 max (mean difference 0.16 L/min, 95% CI 0.1–0.2; p < 0.001) and leg press strength (29.7 kg, 23.4 to 34.9, P < 0.0001) [30, 33]. In addition, one study reported improved fatigue levels in the intervention group (EORTC QLQ C20; −6.6 points, 95% CI −12.3 to −0.9; p = 0.02). [30, 33]. At 1 year, the intervention group had.

The most recent publication on one of the largest powered trials on colon cancer participants found increased time in moderate PA at 12 months (30 min a day more; p = 0.003) [26]. In addition, the intervention group was more likely to achieve the Australian Physical Activity Recommendations (16.4 versus 9.2%; p = 0.047). Within the telephone counselling, advice was given about ‘limiting sedentary habits such as watching television’, and further work from this group found reduced sedentary time in both groups, but this difference was not significant [36]. However, a subgroup analysis found only the intervention group decreased sedentary time at 12 months in the >60 years of age, male and non-obese. The remaining three papers did not report any beneficial outcomes with adjuvant PA interventions in CRC patients which may reflect the different PA interventions. One group started in hospital training (walking on ward, stair climbing and strength training) that continued after discharge (×5/week) [27, 28]. There were no differences between groups at 30 and 90 days after surgery in sit-to-stand test and 6MWT. The authors suggested that post-operative exercise is not beneficial in the CRC population, but this study’s intervention was primarily based on strength exercises with the majority of the aerobic activity performed in the hospital to allow quick recovery of mobility post-operatively, rather than as part of a targeted progressive intervention.

The third paper that did not report improvements aimed to assess two different methods of promoting diet and PA: tailored print communication (TPC) only, telephone motivational interviewing (TMI) only, combination and controls [22]. Using self-reporting, the authors found that none of the four groups had an increase in PA at 1 year follow-up. This study was not powered, and the health behaviour interventions had substantially less patient contact compared to the other RCTs (e.g. TMI consisted of quarterly calls versus biweekly calls by CanChange) [26].

Only one group has specifically assessed dietary interventions in rectal cancer, finding both short- and long-term patient benefits [34, 35]. Each patient underwent neo-adjuvant chemoradiotherapy for 1.5 months, surgery 3–5 weeks later, followed by adjuvant chemotherapy. During their radiotherapy, the authors randomised the patients into three groups: group 1 had 6 weeks of individualised nutrition counselling and education sessions using regular foods, group 2 had usual diet with high-protein supplements added in (40 g/protein/ day) or group 3 maintenance of usual diet. On completion of radiotherapy, groups 1 and 2 found significant increases in energy and protein intake, but at 3 months, only group 1 documented significant reductions in radiotherapy toxicity and improved nutritional intake/status and QoL. At 5-year follow-up [35], they found that again, only group 1 had a sustained improvement in dietary interventions with 91% maintaining adequate nutritional status. In addition to higher QoL scores, late radiotherapy toxicity was also significantly lower (9 versus 59% group 2 versus 65% group 3; p = 0.001). However, the most intriguing finding was that both disease-free survival and disease-specific survival were found to be significantly longer in group 1 after adjustment for age and disease stage (median survival: group 1 with 7.3 years versus group 2 with 6.5 years versus group 3 with 4.9 years; p < 0.01).

The next three RCTs performed dietary interventions in the adjuvant setting as part of a multiple behavioural intervention (e.g. physical activity, counselling), which is a pragmatic approach to lifestyle change [22, 26, 31, 32]. The time after diagnosis that the study started recruiting varied from the earliest (12 months) to the latest (8 years) with different patient populations in each study (one mixed cancer populations, one colon cancer patients only, one colon and rectal but numbers in each not stated). All interventions were based upon dietary counselling with home-based telephone guidance and reported low dropout rates (0–12%) with good follow-up rates (77.6 to 89%) supporting that these interventions were feasible and well tolerated.

Two of these three studies reported dietary improvement at follow-up [31, 32]. The RENEW Trial reported improved BMI and dietary quality in their older, overweight cancer patients (n = 91) [BMI −0.56 kg/m² (95%CI −0.75 to −0.36; p < 0.001); Diet Quality Score 5.2 (95%CI 3.4 to 7.0; p < 0.001)]. These changes were maintained at the 2-year follow-up. In addition, their control group underwent similar improved dietary changes when they underwent their delayed intervention 12 months after the intervention group.

CanChange is the largest study in colon cancer patients finding good adherence to their telephone-guided counselling and [26] a significantly reduced BMI at 6 and 12-month follow-up as well as a reduction in fat intake (by 8.5% at 6 months and 7% at 12 months; p = 0.001 and p = 0.006) and increased vegetable intake (0.4 servings a day at 6 months; p = 0.001). Although this improvement intake seems small, it has the potential to be clinically relevant, as an increase of one portion of fruit and vegetables daily has been shown to reduce cancer.

NC Strides differs from the other two trials as it compared two types of behavioural intervention approaches to modifying lifestyle habits in patients with CRC [22]. TPC consisted of sending out four personalised newsletters. In comparison, TMI participants (telephone motivational interviewing intervention) received four telephone calls with specially trained motivational counsellors. After 1 year, there was a significant increase in fruit and vegetable consumption in the entire study population, but when the CRC patients were separately analysed (healthy controls excluded), there were no significant consumption improvements. One explanation could be selection bias as some participants had been recruited from another trial that had provided dietary advice at its completion.