Comparative effectiveness of sodium-glucose co-transporter-2 inhibitors and glucagon-like peptide-1 receptor agonists in type 2 diabetes mellitus: randomized and real-word evidence

Objectives of the research
This PhD project will aim to examine the clinical effectiveness of GLP-1RAs and SGLT-2is combining RCTs and “real-word” evidence. Specific objectives are:
1. To conduct a network meta-analysis to assess the comparative effectiveness of GLP-1RAs and SGLT2is using information from RCTs
2. Conduct a longitudinal study using routinely collected primary care data to assess the effectiveness of SGLT2is and GLP-1RAs in type 2 diabetes in the real world
3. Synthesise the evidence from RCTs and real world studies together
4. Assess the comparative cost-effectiveness of SGLT2is and GLP-1RAs

Brief outline plan of research over the 3-7 years:
The project will involve the use of published aggregate-level data and available large existing datasets such as the Clinical Practice Research Datalink,

Assuming an intake of a full-time PhD student the following table describes a tentative research plan for the course of the PhD.

Time Task Objective
Months 1-10 Network Meta-analysis Evidence of the effectiveness of drugs from the trial
Months 10-20 Longitudinal studies Real World Evidence of the effectiveness of the drugs
Months 20-24 Evidence synthesis Combining RCT and observational evidence
Months 24- 32 Health economic analysis (one of the selected two data sources) Health care utilisation and health care costs, benefits of healthy lifestyle
Months 33 onwards Thesis writing

Methodology to be used:
1. Systematic literature review
2. Synthesis of available RCTs evidence with a network meta-analytical approach
3. Data management and manipulation of large primary care datasets
4. Combination of randomised and observational evidence
5. Health economic analysis

Background to Project
A range of glucose-lowering medications are currently available for the treatment of hyperglycaemia in patients with type 2 diabetes. In the last few years, two new classes of glucose-lowering drugs have been introduced and widely used [1]: glucagon-like peptide-1 receptor agonists (GLP-1RAs) and sodium-glucose co-transporter-2 inhibitors (SGLT-2i). The efficacy and safety of different formulations of GLP-1RAs and SGLT-2i have been extensively compared to other glucose-lowering therapies in several randomised clinical trials (RCTs) [2-3]. Very limited, on the other hand, is the evidence from direct comparisons between GLP-1RAs and SGLT-2i.

Current ADA/EASD guidelines suggest either GLP-1RAs or SGLT-2i as add-on to metformin (alone or in combination with other treatments) if glucose control is not achieved; however, no specific suggestions are given about which agent should be preferred [4]. In this view, it would be relevant to know which medication performs better and is associated with a lower risk of side effects to help decision makers in an informed choice.

When direct ‘head-to-head’ data are unavailable or limited, network meta-analysis (NMA) is considered the methodology of choice to estimate the comparative effectiveness of multiple treatments [5]. Recently, softwares to perform NMA have been made available in widely-used statistical softwares (i.e., Stata and R), widening their applications. Data from published RCTs can therefore be synthesised using NMA to evaluate differences across treatments and rank them according to a specific outcome.

For a better and wider understanding of treatment efficacy, however, data from RCTs should be considered along with “real-word” evidence (RWE). RWE supplements the knowledge obtained from RCTs, whose limitations are the generalisability to larger, more inclusive populations of patients [6]. Large primary care databases (i.e., Clinical Practice Research Datalink) have been used to evaluate the effects of treatments in “real-word” patients although the evidence about glucose-lowering effects of GLP-1RAs or SGLT-2i is very limited. Furthermore, recent methodological advancements make possible to “combine” evidence from RCT and RWE accounting for their differences [7].

A more precise estimate of GLP-1RAs or SGLT-2i effects obtained from RCTs and RWE data can also allow performing a health economics analysis to compare these new classes of treatments with other available glucose-lowering drugs.