Patients with type 2 diabetes (T2D) are at increased risk of developing several cancer types. Among patients with cancer, there is a common perception that patients with T2D have a poorer survival than those without T2D. However, the studies reporting these findings have several major weaknesses. Additionally, no study to-date has analysed survival in individuals with and without T2D, and with and without cancer, to address the question 'do individuals with T2D and cancer have reduced survival over and above what would be expected by T2D and cancer acting separately?' This is referred to as a diabetes-cancer interaction.
To address this interaction, it is necessary to have a large dataset of information on patients with T2D linked with cancer diagnosis, and also linked with subsequent cause of death. We plan to bring together three linked datasets covering the period: 02/01/1998- 31/12/2015.
We will investigate the relationship between T2D and cancer, and its impact on survival. This will involve complex statistics.
The findings will inform the impact of diabetes on survival in patients with cancer, and facilitate updating clinical guidelines.
Studies report that patients with cancer and type 2 diabetes (T2D) have poorer survival than patients with cancer without T2D. These studies had major weaknesses (small sample sizes, biases, confounding), and importantly, no study tested the hypothesis that T2D adversely impacts on survival, over and above the independent negative effect of T2D on survival. We propose a novel study to evaluate the effect of T2D on survival in a matched non-cancer cohort; testing the interaction between cancer and T2D on survival.
Methods: We will delineate CPRD data extracts for individuals with incident cancer (02/01/1998- 31/12/2015) linked with NCRAS - cancer cohort - and matched (1: 5) by year of birth (±2 years), sex and general practice to derive a non-cancer cohort. Derived cohorts will be linked with ONS to determine cause of death. Within each cohort, we will stratify by pre-existing T2D.
Primary outcomes: Any cause death (cancer and non-cancer cohorts); cancer deaths in cancer cohorts
Secondary outcome: Cardiovascular deaths
Models: We will develop time-to-death models from cancer diagnosis or matched index date for the non-cancer cohort. A priori, we will stratify by gender; smoking status; IMD deprivation quintiles; and Charlson comorbidity index; and model competing risks of death.
Death from any cause
Death from cancer
Cardiovascular death
Darren Ashcroft - Chief Investigator - University of Manchester
Nasra Alam - Corresponding Applicant - University of Manchester
Alison Wright - Collaborator - University of Manchester
Andrew G Renehan - Collaborator - The Christie NHS Foundation Trust
Martin Rutter - Collaborator - University of Manchester
Matthew Sperrin - Collaborator - University of Manchester
HES Admitted Patient Care;NCRAS Cancer Registration Data;ONS Death Registration Data;Patient Level Index of Multiple Deprivation