Why Combination Therapy Fails in Practice

Guidelines increasingly suggest combining an SGLT2 inhibitor with a GLP-1 in type 2 diabetes. A pragmatic trial tried to test that in ordinary care and ran into something before it could measure any outcome: people did not fill the prescriptions. At four months, 84% of the single-drug group had filled against 53% of the combination group.^[1]

What happened

Participants were randomized to monotherapy or dual therapy and followed for a median of ten months. The gap in fill rates narrowed over time but did not close — 87% against 68% overall. Among those who did fill, 22% of the monotherapy group and 49% of the dual therapy group discontinued a study medication, mostly because of side effects.^[1]

Half of the people who started two drugs stopped at least one of them. Quality-of-life scores showed no change in either group, which is unsurprising when so much of the intended treatment never happened.

The trial did not find that combination therapy failed. It found that combination therapy largely did not occur.

Why this population makes it worse, not better

The participants were insured, and the study team had actively worked to facilitate medication uptake during the feasibility phase.^[1] This is not a story about people who could not afford treatment. It is what happened under conditions substantially better than average.

That reframes the finding. If half of an insured, supported, trial-enrolled population cannot sustain two drugs, the guideline recommendation describes something that mostly does not survive contact with a pharmacy counter.

It fits a pattern we keep finding

This is the third distinct measurement in our recent reading showing that the gap between prescribed and taken is where most of the lost effect lives.

• Only 58% reach maintenance dosing and 21% never leave the starting dose, in UK primary care.
• Discontinuation runs from 36.9% to 55.8% depending on drug and formulation, in the same dataset.
• Half of combination starters stop one drug, here.

Trials report what happens to people who take the drug. Guidelines assume that. Neither describes the ordinary experience, and the difference is not small — it is most of the effect. Our piece on why real-world results differ from trial results covers the same gap from the other end.

What to take from it

If a combination has been suggested to you, the useful questions are not about efficacy. They are: what will this cost each month once both are running, which one do we stop first if the side effects are intolerable, and what does the plan look like if only one of them is sustainable?

It is a small, early report — 173 participants in the feasibility phase of a larger trial — and it measured something that larger, cleaner trials systematically design out.