Welcome to Compare Diabetes Drugs .com

The latest update included:

New Studies! All reported studies up to 30 August 2020 have now been included for the 6 original drug classes (DPP-4 inhibitors, GLP-1RAs, SGLT2 inhibitors, TZDs, SUs and the biguanide metformin), new experimental drugs (see below), along with numerous studies with basal insulins (e.g. insulin glargine, basaglar, toujeo, insulin degludec, insulin detemir, NPH insulin). Currently there are:

Total Studies: 307

Total Treatment Arms: 958

Total Patients: >170 000

New Cardiovascular Outcome Trials Section! New results, based on a comparison of 16 CV outcome trials, are now available here. Across these trials, a total of over 157 000 patients experienced more than 16 900 events. What can we learn from this vast amount of data, and how can we use that information to better design the next CV outcome trial?

New Clinical Trial Simulator! It is now possible to evaluate and optimise new trials using our Clinical Trial Simulator (available here). What two drug regimens would you like to see go head-to-head?

New Experimental Drugs! Using the latest clinical studies results, exciting new drugs currently in clinical development have been incorporated into the mathematical models. These include:

GLP-1 receptor agonists: Oral semaglutide (Novo), efpeglenatide/HM11260C (Sanofi/Hanmi) and ITCA 650 (Intarcia)

Dual GLP-1/GIP receptor agonists: tirzepatide/LY3298176 (Lilly) and NNC0090-2746 (Novo)

Dual GLP-1/GCG receptor agonist: MEDI0382 (MedImmune)

New Initial Combination Studies! The mathematic models have been successfully extended to both describe and predict initial combination therapies (71 new combination arms across 44 studies) using the individual monotherapy components (with some mechanistic insights incorporated). That is, the models can describe both the observed combinations, but can also be used to predict novel combinations. We hope to extend the website with examples of combination therapy in the near future.

A More Quantitative Path to Success! Clinical drug programs and studies can be smarter (higher probability of technical success) and faster (e.g. lower N). It's time to leave outdated and inefficient power calculations behind!

The goal of this work is to provide comparisons between the different drugs used to treat type 2 diabetes. Key measures are:

HbA1c: Glycated Hemoglobin, or HbA1c, provides the best, long term, gauge of blood glucose control. A diabetic patient may have a HbA1c of 75 mmol/mol (9%), whilst a normal HbA1c is below 42 mmol/mol (6%). Lower levels are better, and a common treatment goal is to reduce HbA1c to below 53 mmol/mol (7%).

Read More

All antihyperglycemic drugs reduce HbA1c, with reductions from 6 mmol/mol with the least efficacious drugs to 22 mmol/mol with the most efficacious (best) drugs.

Weight: In addition to blood glucose control, the different drugs may increase or decrease body weight. Since weight management is often important for patients, drugs that reduce weight are preferable.

Hypoglycemia: One of the main risks with antihyperglycemic drugs is hypoglycemia. Hypoglycemia, also known as low blood sugar, is when blood sugar decreases to below normal levels. This may result in a variety of symptoms including clumsiness, trouble talking, confusion, loss of consciousness, seizures, or death.

Results: The treatment effects illustrated below are based on 6 months treatment, for drug naive patients with a baseline HbA1c of 65 mmol/mol, and baseline weight of 90 kg. All treatment differences are versus placebo. For hypoglycemia, we report the relative risk (risk on drug / risk on placebo).

Read More

All treatment effects and treatment differences are dependent on multiple factors, including the duration of the study, the patient population and the baseline HbA1c. The mathematical models we developed appropriately adjust for these important factors (study differences), ensuring we truly compare 'apples with apples'. In addition, we can generate predictions for any combination of these factors, for any choice of comparator (i.e. instead of placebo, we could choose sitagliptin 100 mg, or semaglutide 1 mg).

In the graph below, you can select the endpoint of interest, and look at different classes of drugs. It is good to know that:

Drug Class: There are many ways antihyperglycemic drugs interact with the body to improve glycemic control. Drugs can be grouped based on this 'mechanism of action'.

Read More

In addition to than basal insulin regimens, there are 6 main drugs classes. These include the biguanide metformin, DPP-4 inhibitors, GLP-1 agonists, SGLT2 inhibitors, sulfonylureas (SU) and thiazolidinediones (TZDs). Dual GLP-1/GIP and GLP-1/GCG receptor agonists are also in development. The graphs below can be used to show comparisons of the drugs within each drug class if you wish.

Drug name: The graphs below show the generic drug name, not the brand name. For example, metformin is the generic drug name, whilst the brand name in the USA is Glucophage. We use the generic name, since the brand name may differ in different parts of the world.

Utility: Utility is the name given to a combined outcome, incorporating HbA1c, weight, risk of hypoglycemia, and the dosing regimen (daily tablets, daily injection or weekly injection). The highest utility is achieved with the largest reductions in HbA1c and weight, no increased risk of hypoglycemia, and a daily pill.

Read More

Each drug can be compared on each endpoint separately (e.g. reduction in HbA1c), but it can be very useful to combine results across endpoints to get an overall score. Called a Clinical Utility Index (CUI) or Multi-Criteria Decision Analysis (MCDA), a composite score is determined using weights for each endpoint (more important => higher weighting), and a scale for each endpoint (from worst to best outcomes). Our simple CUI/MCDA example shown below uses the following weights: HbA1c = 60%, Weight = 15%, Hypoglycemia = 15% and Regimen = 10%. Additional endpoints and different weighting/scales can be used, so these results should be seen as illustrative of what can be done.

Current Selection : HbA1c and All Drugs (best outcomes are towards the right)

Change in HbA1c (mmol/mol)

Our Interactive Graph! On the following graph, the best outcomes are towards the top of each scale. Click on the axis label (e.g. "Drug", "HbA1c" etc.) to sort the data by that scale. Try hovering your mouse over the drugs or lines to see each drug separately (works best on desktop computer)

The original analysis has now been published in the highly regarding Journal of Clinical Pharmacology and Therapeutics. The title (with link to PDF), citation and abstract are shown below.

A Model-Based Meta-Analysis of 24 Antihyperglycemic Drugs for Type 2 Diabetes: Comparison of Treatment Effects at Therapeutic Doses.
Maloney A, Rosenstock J, Fonseca V
Clin Pharmacol Ther. 2018 Nov 20. doi: 10.1002/cpt.1307

Model-based meta-analysis was used to compare glycemic control, weight changes, and hypoglycemia risk across 24 antihyperglycemic drugs used to treat type 2 diabetes. Electronic searches identified 229 randomized controlled studies comprising 121,914 patients. To ensure fair and unbiased treatment comparisons, the analyses adjusted for important differences between studies, including duration of treatment, baseline glycated hemoglobin, and drug dosages. At the approved doses, glycemic control was typically greatest with glucagon-like peptide 1 receptor agonists (GLP-1RAs), and least with dipeptidyl peptidase-4 (DPP-4) inhibitors. Weight loss was highly variable across GLP-1RAs but was similar across sodium-glucose cotransporter 2 (SGLT2) inhibitors. Large weight increases were observed with sulfonylureas and thiazolidinediones. Hypoglycemia risk was highest with sulfonylureas, although gliclazide was notably lower. Hypoglycemia risk for DPP-4 inhibitors, SGLT2 inhibitors, and thiazolidinediones was generally very low but increased slightly for both GLP-1RAs and metformin. In summary, important differences between and within drug classes were identified.

The current analysis includes data from 307 studies. The circular network diagram below shows the number of unique studies for each drug at every node (the outer circles), and the width of the connecting lines is proportional to the number of studies comparing each pair of drugs. Our work 'harvests' the enormous amount of knowledge contained in this ‘dense’ network of treatment comparisons. Try hovering your mouse over the drug name to focus on the network for that drug.

We hope you found this information useful. Please check out our other pages to learn more. Back to Top

Subscribe to our Newsletter

If you would like to be informed when we add new results or features to this website, please subscribe (around 1-3 emails per year). We will not pass on your details to any third parties. Any feedback is also much appreciated!