(View plain language abstract)
Since outline, the main changes in our application have been to address panel’s queries, strengthen rationale for our design, and more realistically estimate the project outputs.
The World Cancer Research Fund (WCRF), and work from the present researchers, have shown excess weight (commonly expressed as elevated body mass index, BMI) is a risk factor for several cancers. By extension, elevated BMI at or after cancer diagnosis might be associated with a poor prognosis, and indeed, many studies show such associations. This is a key rationale for weight management strategies in cancer survivorship programmes.
However, interpretation of analyses in the cancer post-diagnosis setting is susceptible to several major sources of confounding and biases. Such was the extent of these concerns that the WCRF 2014 report on Breast Cancer Survivorship concluded that evidence was suggestive but limited.
We argue that the optimum setting to test such associations is as secondary analyses of trial data, where confounding and biases can be minimised. Even better is to preform individual patient data (IPD) meta-analyse across these trials to increasing estimate precision. The cancers of interest here are colorectal (CRC) and endometrial (EC). We found two reported IPD meta-analyses of BMI and prognosis in colon cancer in adjuvant chemotherapy trials (see confounding issues under objectives), and found no IPD meta-analysis of the associations of BMI and outcome in rectal cancer or EC.
Hypothesis and objectives
We will test the hypothesis that elevated BMI, measured at or after diagnosis in patients with CRC and EC, are associated with reduced overall survival (OS) or disease-free survival (DFS) compared with normal weight.
To reduce the confounding of chemotherapy dose capping/reduce chemotherapy intensity/pharmacokinetic variability in obese patients, the ‘clean’ primary cohorts within the trials will consist of surgery only chemotherapy-naïve non-metastatic patients.
Settings and methods
Primary outcome measures are five-year OS; secondary endpoints are five-year DFS. We established a ‘federation’ of principal investigators from published trials. The confirmed collaborative of chemotherapy-naïve participants with measured BMI numbers 7,675 patients with CRC and 1,847 women with EC – these datasets represent twice as large as anything hereto gathered.
Based on these samples, for CRC, we have 92% power for OS and 98% power for DFS; and for EC, 93% power for OS and 97% power for DFS. The power calculation showed that these numbers are underpowered for interactions by gender, sub-site or histological sub-type, but concurrently demonstrated that to achieve sufficient power would require unattainable and unavailable sample sizes (x4 fold). Trial leads will attend a set-up workshop to establish mechanisms to share data.
The federation approach means that for some trial data can stay at source and the analyses comes to the data. We will test relationships between peri- and post-diagnosis BMI and OS, and DFS, using BMI as categorical and continuous variable, exploring linear and non-linear relationships. We will explore for the obesity paradox. A priori stratified analyses and sensitivity analyses will be performed.
Given the novel approach of a large-scale within-trials IPD meta-analysis, embedded in a low risk of confounding analytical platform, this project will deliver a significant new contribution to the two WCRF themes on relationships between food, nutrition and physical activity (here, obesity) with cancer prognosis and outcome in cancer survivors.
Specifically, through two high-quality publications, the findings will offer unbiased information to weight management strategies in survivors of colorectal and endometrial cancers.