Type 2 diabetes: two drugs on day one

Clinical Rounds, the GPAtlas podcast. Endocrinology. 18 minutes. Published 31 August 2026. Free to listen.

Clinical source: our own Ocean article "Type 2 Diabetes & Insulin therapy (T2DM)", last reviewed 15 April 2026.

Transcript

Sarah: Hello, and welcome to GPAtlas Clinical Rounds. I'm Sarah, and this is the podcast where we take one thing you'll genuinely see this week, and work out together what to do about it.

Ben: And I'm Ben. I'm the one who's been through the guidance, so when Sarah tells you what she did, I'll tell you whether the evidence agrees with her. Today, the new diagnosis of type two diabetes, and the fact that what most of us were taught to do first is no longer what we're asked to do.

Sarah: So. Fifty-two year old man, came in for a medication review, and his HbA1c has come back at fifty-eight. New diagnosis. No heart disease, no kidney disease, nothing else on his list. He's a bit overweight, that's it. And what I would have done, without thinking, for the whole of my career, is start metformin, tell him about his diet, and book him back in three months to see whether it worked. That is now the wrong first move.

Ben: Three things. What NICE actually changed in February. Why the old sequence was doing him harm rather than nothing. And what to say to a man who came in for a review and is leaving with two new tablets.

Sarah: So what changed. NICE now asks for dual therapy from diagnosis. Metformin, and an SGLT2 inhibitor. Not metformin, wait, measure, add. Both, from the start. And the bit that surprised me most is that this isn't reserved for the high-risk patient. He hasn't got heart failure. He hasn't got kidney disease. He's an ordinary new diagnosis, and he still gets both.

Ben: And the reasoning is about what those three months were costing him. The old sequence treats the diagnosis as a trial: start one drug, see if it holds, escalate when it doesn't. But the cardiovascular and renal protection from an SGLT2 inhibitor doesn't depend on how high the sugar is. It's a separate benefit. So every month he spent on metformin alone was a month of protection he wasn't getting, and the reason we withheld it was tradition rather than evidence. Practically, you still introduce them stepwise so you can tell which one upset his stomach, but the destination is both, and it is not conditional on failure.

Sarah: So that's the first thing. New diagnosis means two drugs, not one. And you are not waiting for metformin to fail before you add the second.

Sarah: Now, the patient who does have heart disease already, because that's a different conversation again. If he'd walked in with established cardiovascular disease, the guidance goes further still, and the starting position is triple therapy. Metformin, an SGLT2 inhibitor, and subcutaneous semaglutide, up to one milligram a week. Not after two failures. First line.

Ben: And that is worth sitting with, because it reverses the instinct most of us have, which is that the sicker patient gets the cautious approach. Here the sicker patient gets more, sooner, because he has the most to lose from waiting. If you take one idea away today, make it that one: in type two diabetes, established heart disease is a reason to escalate immediately, not a reason to be careful.

Sarah: And a practical note that's ours rather than the hospital's. Check his kidney function before you start, and check it again after. And warn him about the genital thrush, because if you don't mention it, he will stop the tablet and not tell you why. That one conversation, thirty seconds at the start, saves the prescription.

Sarah: Recap. Established heart disease means escalate sooner, not slower. And mention the thrush before he finds out himself.

Sarah: So what do I actually say to him? Because this is the hard part, honestly. He came in for a review and he's leaving with a diagnosis and two tablets, and that feels like a lot. What I've found works is not leading with the sugar. I say, we've found this early, and there are two things that protect you. One works on the sugar. The other one protects your heart and your kidneys, and it does that whatever your sugar does. So they're doing different jobs, and that's why you're getting both.

Ben: Which happens to be true, rather than a form of words. That's exactly why the guidance changed. And it reframes the second tablet from an escalation, which sounds like bad news, into a protection, which sounds like the reason he came. Worth adding that this sits alongside everything else, not instead of it: the diet conversation, the blood pressure, the statin decision. The drugs got simpler and earlier. Nothing else went away.

Sarah: One tablet for the sugar, one for the heart and kidneys. Different jobs. That's the sentence.

Ben: Before we go further, one thing about making the diagnosis, because it's easy to be too quick. In an asymptomatic person, one raised HbA1c is not enough. You repeat it to confirm. And there is a list of situations where the HbA1c will lie to you, falsely low or falsely high: pregnancy, recent blood loss or transfusion, haemoglobinopathy, haemolysis, advanced kidney disease, and rapid-onset diabetes. In any of those you use a fasting or random glucose instead. Diagnosing on a single number in the wrong patient is how people end up labelled for life on a test that was never valid for them.

Sarah: And there's one presentation that should stop you cold rather than start a diabetes pathway. New diabetes with unexplained weight loss in someone over sixty goes down the urgent pancreatic cancer pathway. Not into a diabetes review in six weeks. That's a hard one to hold in mind, because a new diagnosis with weight loss looks like textbook diabetes, and the weight loss feels like it explains itself. It doesn't.

Ben: One HbA1c is not a diagnosis. And new diabetes plus weight loss over sixty is a cancer pathway.

Ben: Now the numbers you're steering by. There are three and they're worth separating. Forty- eight millimoles per mole if he's on lifestyle alone or metformin. Fifty-three if he's on something that can cause hypos. And you consider insulin if the HbA1c stays at fifty-eight or above. Forty-eight, fifty-three, fifty-eight.

Sarah: And the counterweight to all three, which matters more as your list ages. Avoid over- treating in frailty. In an older, frail adult the harm from hypoglycaemia, the falls, the fractures, the cardiac events, the admissions, usually outweighs any gain from tight control. So you de-escalate the drugs that cause hypos. I've taken more people off gliclazide in the last two years than I've started, and that is not me being lax, that is the guidance.

Ben: Kidney function then drives most of the rest of the regime, drug by drug. Metformin: review the dose below an eGFR of forty-five, stop below thirty. And watch for gastrointestinal upset, where switching the preparation or retitrating often rescues it, and for B12 deficiency on long-term use. The SGLT2 inhibitors go much further down than people assume. Dapagliflozin is licensed in chronic kidney disease down to an eGFR of fifteen, and our article notes it as the least costly option NICE supports.

Ben: And the point people miss when the sugars come good: you keep the SGLT2 inhibitor for the kidney and heart protection even when the HbA1c target is met. It isn't there as a glucose drug you can retire once the number looks nice.

Sarah: Forty-eight, fifty-three, fifty-eight. And you keep the SGLT2 at target.

Sarah: Sick day rules, and I want to give these properly because they're the ones that get written on a leaflet and never spoken aloud. Never stop insulin when unwell. Never. That's the one people get backwards. What you do temporarily suspend if he's dehydrated is the metformin, the SGLT2 inhibitor, the ACE inhibitor or ARB, anti-inflammatories, and the GLP-1. Monitor glucose and ketones more often, every one to two hours, and keep the fluids going.

Ben: And the reason the SGLT2 inhibitor is on that pause list is worth spelling out, because it is the single most dangerous thing on today's prescription. An SGLT2 inhibitor can precipitate ketoacidosis with only modestly raised, or even near-normal, glucose. So a normal reading does not exclude it. In any unwell patient on one of these, you check ketones whatever the glucose is. Euglycaemic diabetic ketoacidosis kills people because the glucose looked fine and everybody relaxed.

Sarah: Insulin stays. The rest pause. And check ketones whatever the glucose says.

Ben: Two prescribing rules that are absolute. Never run a GLP-1, or tirzepatide, alongside a DPP-4 inhibitor. There's no added benefit, so you stop the gliptin when you start the GLP-1. And do not newly initiate exenatide, Byetta or Bydureon, which is discontinued in the UK. Anyone still on it moves to semaglutide, dulaglutide or tirzepatide. Same for insulin detemir, Levemir: don't initiate it, because UK supply ends in December twenty twenty-six, so switch existing patients to glargine or degludec in good time rather than in a hurry.

Ben: And there's a stop rule on the GLP-1 that I think is under-applied. You only continue it beyond six months if there's a beneficial metabolic response: a fall in HbA1c of at least eleven millimoles per mole, and weight loss of at least three per cent of initial body weight. Both. Not either. If both aren't met, you stop and move to the next option. The common error is continuing on the weight loss alone.

Sarah: And if it does come to insulin, it's more doable in primary care than people fear. First- line basal is human NPH insulin, once or twice daily. You continue the metformin, and you review the sulfonylurea, reducing or stopping it to limit hypos. Then he self-titrates to a fasting glucose of five to seven. One caution: when you add a GLP-1 to a sulfonylurea or to insulin, you reduce the older drug to limit hypos, and a GLP-1 with insulin needs specialist support.

Ben: Six months. Eleven millimoles and three per cent. Both, or you stop.

Sarah: The rest of the review, briefly, because the glucose is genuinely not the main event. Blood pressure: an ACE inhibitor or ARB first line, especially with albuminuria, target under a hundred and forty over ninety, or under one thirty over eighty if the ACR is seventy or above. And never an ACE inhibitor combined with an ARB. Lipids: atorvastatin twenty milligrams for primary prevention, eighty if there's established cardiovascular disease. Then HbA1c every three to six months until stable and six-monthly after. And the annual review, which is HbA1c, blood pressure, lipids, ACR, eGFR, feet, eyes, BMI, and injection sites if relevant.

Sarah: Two conversations worth having on purpose. Remission is real and he should hear the number: aim for five to ten per cent loss, and substantial loss, around ten to fifteen kilograms, can drive remission. There's an NHS programme to refer into. And driving, if he drives for a living: glucose above five to drive, check within two hours before and every two hours on a long journey, don't drive below four, and after a hypo wait at least forty- five minutes after the glucose is back to five or above. Not fifteen. Forty-five.

Ben: And two things to ask about that won't be on the screen. Ask whether he's buying semaglutide privately, because it may not appear on the record at all. And there's a rare eye complication, an optic neuropathy, at about one in ten thousand, that he should know to report as sudden visual loss. If he fasts for Ramadan, offer a pre-fasting risk assessment and education, and be clear that he breaks the fast immediately if he has a hypo.

Sarah: Ask what he is buying online. It will not be on your screen.

Sarah: Two practical things before we finish, and they're both the kind of detail that decides whether treatment works rather than whether it's correct. If he ends up on insulin, look at where he's injecting. Lipohypertrophy is easy to miss and easy to cause, and it makes absorption completely unpredictable, so a dose that was right on Monday does something else on Thursday. Advise rotating sites about a centimetre apart and alternating body areas. It takes thirty seconds to examine and it explains a lot of erratic control that otherwise gets blamed on the patient.

Sarah: And check his feet before you choose the drug, not after. Canagliflozin is the one to avoid in active foot disease or peripheral arterial disease. So if he's sitting there with a neuropathic ulcer, that choice is already made for you. It's a small thing, but it's the difference between a prescription that helps and one you have to unpick later.

Sarah: And honestly, the thing I'd most want a registrar to take from all of this is that almost none of what we've talked about is about the sugar. It's the blood pressure, the statin, the kidneys, the feet, the eyes, and whether he can still drive to work. The glucose is the label on the problem. It isn't the problem.

Ben: Look at the injection sites. And look at the feet before you pick the drug.

Ben: Right, exam corner. Sarah, and everybody driving, a few seconds. Fifty-eight year old woman, newly diagnosed type two diabetes, HbA1c sixty-one. She had a myocardial infarction four years ago. Normal renal function, no heart failure. According to current NICE guidance, what is the first-line drug treatment? A, metformin alone, reviewed at three months. B, metformin plus an SGLT2 inhibitor. C, metformin plus an SGLT2 inhibitor plus a GLP-1 receptor agonist. Or D, lifestyle measures alone for three months first.

Ben: It's C. She has established atherosclerotic cardiovascular disease, and that puts her in the triple therapy group from the start. B would be right for somebody without that history, and it's the trap, because it's the correct new answer for the wrong patient. A is the old sequence the guidance moved away from. And D delays every one of her protections for three months in the person on this list who can least afford it.

Sarah: So. Three things for Monday. One. A new diagnosis of type two diabetes means two drugs, metformin and an SGLT2 inhibitor, and you are not waiting for the first to fail. Two. Established cardiovascular disease means three, from the start, because the sickest patient has the most to lose from waiting. And three. Tell him about the thrush before he stops the tablet and doesn't tell you why.

Sarah: One thing to reflect on, if you're logging this. Think about the people you diagnosed in the last year who are still on metformin alone. Not because it failed, but because that was the sequence. Are any of them due a conversation? The full transcript and the references are on the episode page. Ben and I are synthetic voices. The medicine isn't. See you next week.

More from Clinical Rounds

Every episode of Clinical Rounds, free to listen with a full transcript.