Background
The incidences of colorectal cancer (CRC) are 5–10 times higher in Europe, North America and Oceania than in countries in Africa, south Asia and Central America [
1], and the incidence in Norway is among the highest in the world [
2]. Established risk factors for CRC are age, family history of CRC, inherited syndromes (Familial adenomatous polyposis, Lynch syndrome) and inflammatory bowel disease. In addition, several modifiable lifestyle-related risk factors are associated with CRC. Those include smoking, body fatness, abdominal fatness, diabetes, physical inactivity and an unhealthy diet (high consumption of alcohol, red and processed meat, and low consumption of foods containing dietary fibre) [
3,
4]. World Cancer Research Fund (WCRF)/American Institute for Cancer Research (AICR) estimates that about 45% of all CRC cases could be prevented by improved lifestyle [
3].
About 40% of CRC patients [
5] have at least one concomitant disease (e.g. hypertension, cardiovascular disease (CVD), diabetes, chronic obstructive pulmonary disease or other malignancies) at the time of diagnosis and increased risk of developing additional comorbidities after CRC diagnosis [
6‐
10]. These comorbid conditions may preclude or reduce effect of treatment, and consequently reduce disease-specific and total survival [
8,
11,
12].
While it is well established that an unhealthy diet increases risk of CRC (e.g. see the latest update from World Cancer Research Fund, 2011 [
4]) there are few studies that have focused on the effect of diet on disease outcomes and survival [
13‐
15]. In paucity of data, health authorities in most countries recommend the same diet to CRC survivors (i.e. patients living with a CRC diagnosis, including those who have recovered) as to people without a cancer diagnosis [
3].
Inflammation and oxidative stress are central underlying disease mechanisms in cancer and several other chronic diseases. Recent research suggests that there are two major molecular pathways leading to CRC, both of which involve inflammation and oxidative stress as major driving forces. The majority of CRC cases may be due to molecular events that result in chromosomal instability, while about 20-30% of CRCs are due to gene hypermethylation (called CpG island methylator phenotype (CIMP)) [
16‐
18]. A large proportion of the CRC cases due to CIMP display microsatellite instability [
18,
19]. In total, about 70 mutations in different genes have been identified as relevant for these two pathways to CRC, and it is assumed that each individual CRC tumor accumulates an average of 9 CRC pathogenic mutations out of this total pool of 70 mutations [
16].
The heterogeneous pathogenesis of CRC comply with the hallmarks of cancer defined by Hanahan and Weinberg [
20] and the cancer genome landscape as defined by Vogelstein et al [
21]. Underlying these hallmarks of cancer, Hanahan and Weinberg proposed that genome instability and inflammation are two underlying driving forces [
20]. These two processes or mechanisms are closely intertwined, since inflammation is a major cause of oxidative stress, and oxidative stress is a major cause of genome instability. Although inflammation and oxidative stress ultimately may be related to all CRC cases, the degree of inflammation and oxidative stress may vary significantly with the molecular signature present in the individual CRC patient [
22].
In clinical trials and various models systems, we have identified a number of plant foods (e.g. berries, nuts, spices, coffee and specific fruits and vegetables) with the potential of dampening inflammation and oxidative stress [
23‐
29]. Furthermore, a number of studies have also suggested that adherence to a prudent diet (e.g. Mediterranean diet) reduce inflammation and oxidative stress [
30,
31]. We suggest that a prudent diet rich in specific plant-foods may be beneficial for CRC patients, especially those CRC cases with molecular signatures creating major inflammation and oxidative stress.
No intervention studies have investigated the role of diet in disease outcomes and survival in CRC-patients after diagnosis. Furthermore, no previous diet intervention study has focused on dampening inflammation and oxidative stress in this cancer population. This paper presents the background and design of a randomized controlled food-based diet intervention that examines the effects on disease outcomes and survival in CRC survivors. The diet intervention includes foods and drinks that have been suggested to dampen inflammation and oxidative stress. While specific anti-inflammatory and antioxidant-rich foods are emphasized in each food category, the complete intervention is fully in accordance with the prudent diet recommended by the Norwegian food-based dietary guidelines (NFBDG) [
32] (i.e. a diet similar to the Mediterranean diet).
Objectives
Outcomes are inconsistently defined in many clinical cancer trials [
33,
34]. For the primary outcomes, we have used the proposed guidelines for outcomes as described by Punt et al [
34]. The two primary outcomes are (to be assessed when all patients have completed 5, 10, and 15 years, respectively, of follow-up after baseline):
1.
Disease-free survival (DFS) (events are defined as detection of local recurrence or metastasis or any second cancer or death from any cause)
2.
Overall survival (OS) (event is defined as death from any cause)
Secondary outcomes are:
I.
Time to recurrence (events are defined as detection of local recurrence or metastasis)
II.
CVD -free survival (events of CVD (ICD-10; chapter I) or death from any cause)
III.
CRC-specific survival (death due to CRC)
IV.
Total cancer-specific survival (death due to CRC or any other cancer)
V.
Inflammatory disease-specific survival (death due to inflammatory disease)
VI.
Cardiovascular (CVD)-specific survival (death due to CVD)
VII.
New morbidity of other diet-related chronic diseases (e.g. ischemic coronary heart disease, cerebrovascular disease, thromboembolic disease, type 2 diabetes, obesity, hypertension and chronic obstructive pulmonary disease)
VIII.
Dietary intake and nutritional status
IX.
Physical activity and function
X.
Nutrition biomarkers (e.g., carotenoids, fatty acids, 25-hydroxy vitamin D)
XII.
Anthropometric measures (e.g. weight, waist and hip circumference)
XIII.
Biomarkers for inflammation and oxidative stress (e.g. isoprostanes, cytokines)
XIV.
Transcription- and epigenetic profiles
XV.
Biomarkers for cardiovascular disease, metabolic syndrome, type 2-diabetes, thromboembolic disease and cancer (e.g. blood pressure, total/LDL-cholesterol, HbA1c, CRP, IL-6, IL-10, TNFα)
XVI.
Health related quality of life and fatigue
The secondary outcomes will be assessed after 5, 10, and 15 years and described in detail in subsequent reports. In addition, intervention effects on secondary outcomes VII-XVI will also be assessed at 6 months, 1 year and 3 years follow-up.
Discussion
The primary aims of the CRC-NORDIET are to study whether a healthy diet rich in anti-inflammatory and antioxidant-rich foods and based on the NFBDG can improve DFS and OS in CRC patients. To our knowledge, this is the first randomized controlled trial designed to investigate the effect of a dietary intervention on these outcomes in CRC patients, and to investigate the potential role of diet in dampening of inflammation and oxidative stress in these patients. The multiple strategies used to achieve compliance to the intervention during the first year, followed by the 14 years maintenance and follow-up period make the design of this intervention unique.
Data on role of diet on disease outcomes and survival in CRC survivors is limited. To date, several cohort studies, but no RCTs have investigated the effects of food-based dietary interventions on these outcomes. Data from a US cohort study with stage III colon cancer patients suggest that high intakes of red and processed meat, fat, refined grains and dessert, i.e. a Western dietary pattern, after diagnosis are associated with a significantly reduced disease-free and OS [
13]. Similar findings are reported in a Canadian cohort study with CRC patients staged I-III, where patients with the highest intake of processed meat the previous year before diagnosis had an 82% increased risk of recurrence or death compared with patients with the lowest intake [
15].
Prospective cohort studies have consistently reported that physical activity after colorectal cancer diagnosis reduces risk of mortality. In a meta-analysis of six prospective cohort studies, including 7522 CRC survivors, the authors observed that the most physical active survivors had a 42% lower risk of total mortality compared to those who were least active. The risk reduction of cancer-specific mortality was 39% [
94]. No RCT has so far confirmed that physical activity impacts mortality in CRC survivors.
Interventions designed to investigate the effect of diet separated from other lifestyle factors (smoking, physical activity, weight regulation) are needed in order to investigate whether there is a causal relationship between diet and survival as well as disease-related outcomes. Our intervention is intended to change the dietary habits towards a diet in agreement with the NFBDG. These dietary guidelines are developed to prevent chronic diseases, including cancers, in the general population. Several large cohort studies have shown that there is a consistent inverse association between adherence to cancer prevention guidelines and cancer-specific and all-cause mortality [
89,
95]. Among cancer survivors, reduction in total mortality between highest versus lowest score in adherence to diet recommendations has been documented in five different cohort studies, ranging from 24% to 36% (follow-up period from 3.7 to 13.6 years) [
88‐
93]. Association of adherence to American Cancer Society guidelines and reduction in death attributed to cancer has been shown to be 25 and 26% in men and women, respectively [
89]. Hastert et al documented an association of adherence to the WCRF/AICR guidelines and reduction in cancer-specific mortality of 61% in respondents with the highest compared to the lowest WCRF/AICR score (follow-up time of 7.7 years) [
95].
Furthermore, NFBDG include advice regarding red and processed meat, dietary fibre, dairy products and garlic, all of which are related to risk of CRC. Whether these dietary factors also may have effect on survival and disease outcomes, remain unclear. With improvement in cancer survival, these perspectives are increasingly important. The main objective of the present study is to test if diet will improve survival and cancer-related outcomes, mediated through reduced inflammation and oxidative stress. To strengthen this assumption, we have chosen to select specific foods within the NFBDG which have been identified as anti-inflammatory or antioxidant-rich in previous preclinical and clinical studies [
23‐
29,
37‐
39,
41‐
48,
53,
55,
96].
Four aspects that may be of importance for achieving lifestyle changes and facilitate compliance to the intervention: 1) timing of intervention, 2) choice of motivational approach to achieve lifestyle changes, 3) duration of intervention, and 4) use of incentives and methods to achieve sustainable changes. Previous studies have shown that cancer patients in general are particularly motivated to change dietary habits at the time of diagnosis, often reported as the teachable moment [
97‐
99]. Interventions designed to include this teachable moment have shown to be successful [
97‐
99]. In our trial, we introduce the dietary intervention within a few months from surgery, thereby expecting we reach the patients within the time frame of this teachable moment. Furthermore, principles from MI [
61] are implemented in each of the dietary counselling sessions by trained registered clinical dietitians. Previously published trials that have succeeded in changing lifestyle behaviours in cancer patients are based on theoretical frameworks and theories, including social cognitive behaviour therapy and use of MI. It is emphasized that the patient defines her/his own goals to increase the chances that he or she will succeed in changing dietary habits. We suggest that it is important to focus on a few realistic goals at the time, instead of aiming at changing the whole diet immediately. In addition, follow-up by the same registered clinical dietitian during the entire intensive intervention period may be of importance for the commitment to the intervention goals.
In order to increase the chances of sustainable lifestyle changes and compliance to the intervention, our intervention consists of a one year intensive period and a subsequent maintenance period which lasts for until 14 years. Taking into account that the teachable moment may vary among the patients and it may take time to establish new sustainable dietary habits, the inclusion of a maintenance period will probably be beneficial with regard to an increased long-term adherence to the intervention. Previous published trials that have failed in compliance from the patients may have had too short time frame of diet intervention.
Different strategies are reported to be effective in promoting lifestyle changes in cancer survivors [
14,
97,
98,
100,
101]. Interventions focusing on individual counselling [
14,
98,
100], and also interventions with a mixed strategy of individual in-person counselling, telephone counselling and mailed materials [
97,
101] have been shown to be effective in health behaviour change among CRC survivors. Thus, during the first year, the CRC-NORDIET study offers individualized counselling, free foods, a discount card on healthy foods, access to a login-restricted web page, printed materials, cooking courses and inspiration day, which may all be effective incentives to follow the NFBDG.
In placebo-controlled RCTs, an intervention is tested by comparing one group of individuals who receive the intervention with a control group who receives a placebo. This type of placebo-controlled RCT is most often not possible when studying food-, or exercise-based interventions, since placebo-foods or placebo-exercise do not exist. In addition, no food based intervention can be analysed thoroughly without considerations regarding energy intake and energy expenditure. We have therefore selected to give the intervention group and the control group the same advice on physical activity. We include careful monitoring of physical activity to control for any confounding effects of physical activity. Of ethical reasons, we also include standard dietary advice (i.e. standard clinical care) in the control group, as well invitations to group meetings and feed-back reports on health status. Thus, while the control group in the present study is not identical to a placebo group, this particular study design was used in order to isolate the effect of diet intervention on CRC patients, and to reduce drop-outs from the control group, which is a common concern in long-term intervention trials.
Sample size estimation is not straightforward in RCTs with complex diet intervention and long term hard outcomes. This is especially the case when no similar trials have been previously published. By using the best available information from scientific literature and Norway Cancer Registry on survival rates in the control group and expected reduction in mortality rates in the intervention group, we have performed power calculations on the two primary outcomes after 5, 10 and 15 years after baseline. We conclude that 250 patients in each group would give us a reasonable chance (at 80% power) to detect any significant effects (see Methods and Design section for details) after 5, 10 and 15 years.
In a similar study, testing the effects of two different 6 months adjuvant cytostatic protocols (i.e. the MOSAIC study [
102]), 2246 patients who had undergone curative resection for stage II and III colon cancer, were recruited. After a median follow-up for 38 months, fewer cancer-related events was observed in the alternative treatment group compared to the standard treatment group (HR 0.77,
p = 0.002). The main reason for the lower number of patients required in our CRC-NORDIET study compared to the MOSAIC study is due to an older population with more expected events (50–80 years versus 19–75 years) and a longer follow-up time (10 and 15 years versus 3 years).
We have also performed a number of power estimations on secondary outcomes (data not shown). In general, the CRC-NORDIET study is expected to have enough statistical power to detect significant effects in a majority of these intermediate outcomes (to be published in relevant reports). These power calculations on primary and secondary outcomes are also supported from the RCTs with physical activity intervention in CRC and breast cancer patients; Friedenreich and Courneya detected significant effects on intermediate outcomes (e.g. inflammation biomarkers) as well as disease outcomes in RCTs with 200–250 patients per group [
103‐
105].
Conclusion and perspectives
The CRC-NORDIET study investigates whether a diet aimed at dampening inflammation and oxidative stress and in full accordance with the NFBDG will improve survival and disease outcomes in CRC patients. This RCT is unique in several aspects related to the interventions as well as outcomes. Since previous research on the role of diet for CRC survivors is limited, the study is important in order to improve health outcomes and survival in this population.
Acknowledgements
The authors would like to thank the hospital personnel from Oslo University Hospital and Akershus University Hospital, Norway for assistance during recruitment of patients. In addition, we would like to thank all collaborative partners, technicians and students that have, and will contribute to this study. A specific list of persons will be acknowledged in the relevant reports that will be published from the study.
We would also like to thank Norgesgruppen, Mills, Stabburet, Haugen-Gruppen Norway, Norsk Kaffeinformasjon, Balholm and TINE for generous support of foods. These companies were not involved in the study design and they will not be involved in the collection, analysis, and interpretation of data, or in the publications that will result from this study.