Oesophageal adenocarcinoma (OAC) is known to be an aggressive cancer with very poor prognosis. Barrett’s oesophagus (BO), which refers to a change in the lining of the oesophagus as a complication of acid reflux, is the only known precursor to OAC. There is interest in the role of chemoprevention in this context - medications which may prevent malignant progression of BO. These candidate chemopreventive medications (CCMs), include acid-suppressive drugs called proton-pump-inhibitors (PPIs), aspirin and statins.
The most recent European medical guidelines have recommended the use of PPIs as chemoprevention in the management of BO. However, the evidence underpinning these recommendations is limited. Although previous research studies have suggested that use of these medications is associated with a reduced risk of OAC, most studies are at risk of bias leading to incorrect results.
We plan to carry out a future research study using English healthcare databases (including the clinical practice research datalink, CPRD) to see if CCMs are effective in patients with BO. We will use methods which reduce bias and are more likely to generate results which reflect the truth. To inform this future research we will conduct a feasibility study. This aims to calculate the number of patients with new diagnosis of BO, the number of prevalent users (prescriptions issued before diagnosis of BO) of each of these medications, the number of new users (incident users) after diagnosis of BO, the number of high versus low dose PPI users after diagnosis of BO, and overall progression rate to cancer.
Barrett’s oesophagus (BO) is the precursor to oesophageal adenocarcinoma (OAC). There is clinical interest in chemoprevention (with proton pump inhibitors, aspirin and statins) to alter its natural history. Previous observational research has suggested significant reductions in the incidence of OAC in patients with BO among users of CCMs. However, most studies are at serious or critical risk of bias. Departure from the tenets of a randomised controlled trial when conducting observational research is likely a greater threat to validity than unmeasured confounding. The explicit alignment of the protocol for an observational study to that of the hypothetical “target” trial can improve causal inference. Before proceeding to a target trial emulation study, we propose a feasibility study. The objectives are to:
1. Determine the number of eligible patients contributing to CPRD Aurum from 1990 diagnosed with BO
2. Determine the number of additional patients from 1990 diagnosed with BO in HES APC
3. Determine the overall annual incidence rate for progression to oesophageal cancer.
4. Determine the overall annual oesophageal cancer-free mortality rate.
5. Quantify the use of individual CCMs in patients with BO (separately for prevalent and new users).
6. Quantify the numbers of high and low dose new PPI users in patients with BO.
7. Undertake power calculations for each CCM comparison.
The intended public health benefit of this feasibility study is to develop an understanding of how to optimally use routine data sources to determine and describe patients with BO, their medication use, and their overall annual rate of malignant progression. This data will inform a well-designed study to evaluate the effect of medications which may prevent malignant progression of BO.
Barrett’s oesophagus, oesophageal cancer, all-cause mortality.
Mie Thu Ko - Chief Investigator - University of East Anglia
Mie Thu Ko - Corresponding Applicant - University of East Anglia
Allan Clark - Collaborator - University of East Anglia
Kathryn Richardson - Collaborator - University of East Anglia
Leo Alexandre - Collaborator - University of East Anglia
HES Admitted Patient Care