Starting GLP-1RA in Type 1 Diabetes Safely

Before you read further

Everything in this Part is a framework for a conversation with your diabetes care team, not a prescription. The reduction bands, the regimen-specific guidance, the AID-system notes are starting-point territory. Individual responses vary; insulin doses, pump settings and AID targets must be agreed with your specialist team before they go live. GLP-1RA and GLP-1/GIP use in T1D remains off-label.

Seven principles, in order of priority

The order matters. The first four are non-negotiable safety pillars; the last three are how the dose ends up in the right place.

1. Basal insulin is never stopped. T1D physiology means the beta cells are not there to be stimulated. GLP-1RA does not rescue absent beta cells. Basal continues throughout, outside specialist trial settings.
2. Start low, go slow. Semaglutide starts at 0.25 mg weekly and titrates over weeks; tirzepatide starts at 2.5 mg; liraglutide starts at 0.6 mg daily. Titration intervals are typically 4 weeks, never faster than gastrointestinal tolerance allows.
3. Reduce insulin proactively, not reactively. The reduction comes before the hypos, not after. A mild avoidable hypo at week 2 is a signal to reduce further; a severe hypo at week 2 means the starting reduction was too small.
4. Monitor ketones. Home ketone monitoring is standard of care during titration. Check during illness, unexplained hyperglycaemia, nausea, or reduced food intake. The risk pattern is euglycaemic hyperketonaemia, not classic DKA (Pasqua 2025 saw this pattern in 2 of 24 completers).
5. Use CGM with alerts on. Low and high alerts stay on throughout titration. This is not optional.
6. Review every 1 to 2 weeks during titration. Clinician-led review of CGM, dose, side effects and ketones. Frequency drops to monthly once stable.
7. Minimum effective dose. Push only as far as the clinical goals require. Glycaemic benefit on tirzepatide plateaus at 2.5 to 5 mg; further weight loss comes at higher doses but glycaemic gain is small.

Insulin reduction bands, the starting framework

These bands are individualised starting points, drawn from the ADA/EASD 2021 framework, the Saeed 2024 Clinical Diabetes companion, the DTS Consensus 2025, and ADJUST-T1D (which delivered a 22 units per day TDD reduction at 26 weeks, roughly 25 per cent of typical TDD in a BMI-above-30 cohort). The four-row mapping below adds operational stratification from John Pemberton’s clinical synthesis; it is a conversation starter, not a published protocol or a rule.

The algorithm is straightforward: start with the smallest reduction compatible with baseline risk, then review CGM every 1 to 2 weeks and iterate with your care team.

The three-track onboarding model

The reduction band tells you how much. This section tells you where the reduction lands. The model draws on the 2025 DTS Consensus (Recommendations 11, 15 and 27) and the ADJUST-T1D insulin-adjustment guidance for AID users.

Track 1, MDI

Apply the chosen reduction proportionally across basal, carb ratios and correction factors. On a balanced regimen (roughly 50 per cent basal, 50 per cent bolus) at a 20 per cent band, that is basal down 20 per cent, carb ratios relaxed to deliver 20 per cent less bolus for the same carb count, corrections similarly relaxed. On a basal-heavy regimen (around 75 per cent basal), the cautious-first approach is basal down 20 per cent and bolus components relaxed by around 40 per cent; on a bolus-heavy regimen (70 to 75 per cent bolus), the mirror applies. The total TDD reduction lands at the chosen band either way.

Track 2, standard pump (without AID)

Same principles as MDI, with the flexibility a pump gives. Insulin-to-carb ratios, insulin sensitivity factors and basal profiles are all programmable. Apply the reduction across all three, proportional to the basal-to-bolus split. Temporary basal rates can be useful in the early days of titration, particularly if nausea is reducing food intake. Review weekly for the first 4 weeks.

Track 3, AID system

AID algorithms respond to reduced insulin demand differently, and none of them can automatically absorb a 20 to 30 per cent shift in physiological requirement without manual settings adjustment alongside. The DTS Consensus (Recommendation 27) explicitly states that AID systems should accommodate the TDD reduction GLP-1RA drives. Today, the adjustment is a partnership between the user, the algorithm and the care team. Tap a system for its specific approach.

The iLet is the most opinionated system and needs the carb-entry workaround during titration; the Tandem X2 is the most flexible because multiple personal profiles are first-class; Omnipod 5 and CamAPS FX sit between. None of this substitutes for clinician-led CGM review every 1 to 2 weeks during titration.

Four safety pillars during titration

1. Ketosis and DKA

The risk pattern is euglycaemic hyperketonaemia, not classic hyperglycaemic DKA. Pasqua 2025 reported 2 of 24 completers (8 per cent) with this pattern; ADJUST-T1D had one high-ketone event attributed to infusion-set failure; the tirzepatide cohort had no DKA. The mitigations are: never stop basal insulin; keep home ketone monitoring; check during illness, unexplained hyperglycaemia, nausea or reduced food intake; avoid over-rapid reduction; keep CGM alerts on. Any concerning ketone reading is a reason to contact your diabetes care team straight away.

2. Hypoglycaemia

Hypoglycaemia is driven by insulin reduction running behind the physiology, not by GLP-1RA itself. CGM with alerts on, higher AID targets at the start lowered as tolerance confirms, and the 1 to 2 week review cadence are the safety net.

3. Gastrointestinal tolerance

Nausea, vomiting and diarrhoea are the dominant adverse events, dose-related, typically worst in the first 1 to 2 weeks after each dose escalation. Smaller and more frequent meals; avoid large fatty meals during titration; moderate fibre early on; prioritise hydration; and if side effects persist, slow titration rather than push through. Delayed gastric emptying also stretches the meal-bolus timing window; pre-bolus intervals that worked before GLP-1RA often need to shorten, split or move to post-meal during titration. Real-world discontinuation for persistent gastrointestinal events runs from 5.5 per cent (ADJUST-T1D) to 22 per cent (ADJUNCT ONE, pre-AID era).

4. Lean mass and bone

Weight loss from any cause carries a lean-mass risk if nutrition and exercise are not adjusted. T1D also carries a baseline fracture risk that is 2 to 5 times the general population, and bone is a live safety question in this class (Part 4). The practical protection framework is drawn from general sports-nutrition guidance applied to this context: protein around 1.5 g per kg per day, split across 3 to 4 meals; resistance training 2 to 4 sessions per week; vitamin D, iron and B12 monitoring. The strongest interventional support for protection comes from Jensen 2024 (JAMA Network Open), where exercise added to GLP-1RA-induced weight loss preserved bone mineral density better than GLP-1RA alone in obesity. Discuss resistance-training plans with your diabetes care team so pre-exercise glucose targets can be agreed.

The first 12 weeks

The shape is consistent across regimens. Week 0 is the baseline visit: CGM review, ketone-monitoring education, AID settings snapshot, reduction band agreed, starting dose prescribed. Weeks 1 to 4 stay on the lowest dose (semaglutide 0.25 mg weekly is the canonical example), with the agreed insulin reduction in place and CGM reviews at 1 and 2 weeks; if tolerance is fine and CGM shows no excess hypo, dose escalation is on the table at week 4. Weeks 4 to 8 step the dose up (semaglutide to 0.5 mg) and bring a second insulin reduction step if the band was 20 per cent or above; weight trajectory starts to emerge. Weeks 8 to 12 may escalate further (semaglutide to 1 mg) and CGM reviews land monthly. By week 12 the conversation moves to minimum effective dose: if TIR, TDD and weight goals are met, stay there. Further escalation needs a clinical reason.