The seven principles, in order of priority

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 insulin must continue throughout, outside of specialist trial settings such as beta-cell transplant protocols.
2. Start low, go slow. Semaglutide starts at 0.25 mg weekly and titrates up over weeks. Tirzepatide starts at 2.5 mg. Liraglutide starts at 0.6 mg daily. Titration intervals are typically 4 weeks, sometimes longer, never faster than the tolerability of GI side effects 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 is euglycaemic hyperketonaemia, not classic DKA (see Part 2, Pasqua 2025: 2 of 24 completers, 8 per cent, had this event pattern).
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 data, dose changes, side effects, ketone patterns. Frequency drops to monthly once stable.
7. Minimum effective dose. Push only as far as the clinical goals require. The tirzepatide cohort showed glycaemic benefit plateaus at 2.5 to 5 mg; further weight loss comes at higher doses but glycaemic gain is small. Use the smallest dose that gets you where you want to go.

Insulin reduction bands, the starting framework

These are individualised starting bands, informed by Holt 2021 ADA/EASD consensus, Saeed 2024 Clinical Diabetes companion, DTS Consensus 2025, and ADJUST-T1D (which achieved 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 is a GNL operational framework that adds stratification based on John Pemberton’s clinical synthesis; it is not a published protocol. It is a conversation starter, not a rule.

These bands are rough starting points. Individual insulin responses vary significantly. They are not clinical rules.

The algorithm: 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 (MDI, standard pump, AID)

The reduction band above tells you how much. This section tells you where the reduction lands. The model draws directly on the 2025 DTS Consensus (Recommendations 11, 15 and 27), the ADJUST-T1D insulin-adjustment guidance for AID users, and the Sept 2025 GNL-Shah reference matrix. The framework was discussed with Professor Shah in Episode 17 of the GNL Podcast, which was the inspiration for this focused programme of work; the specific operational model below has not yet been reviewed by him against the April 2026 evidence update, and is being sent for his feedback alongside the rest of this guide.

Track 1, MDI (multiple daily injections)

Apply the chosen reduction proportionally across basal, carb ratios, and correction factors.

• Balanced regimen (approximately 50 per cent basal, 50 per cent bolus). Reduce basal, carb ratios and correction factors proportionally. If the band is 20 per cent, 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.
• Basal-heavy regimen (approximately 75 per cent basal). The basal share is larger, so the cautious-first approach is basal down approximately 20 per cent; bolus components (carb ratios and corrections) relaxed by approximately 40 per cent. The total TDD reduction lands at the chosen band.
• Bolus-heavy regimen (approximately 70 to 75 per cent bolus). Bolus down approximately 20 per cent; basal down approximately 40 per cent. This is the mirror of the basal-heavy case.

Track 2, standard (non-AID) pump

Same principles as MDI, with the flexibility pumps give. Insulin-to-carb ratios (ICR), insulin sensitivity factors (ISF) and basal profiles are all programmable. Apply the reduction across all three, proportional to the basal:bolus split, as in Track 1. Temporary basal rates can be useful for 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 that GLP-1RA drives, and Recommendation 12 calls for algorithm-level work to adapt. Today, the adjustment is a partnership between the user, the algorithm, and the care team.

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

The 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. ADJUNCT ONE and TWO, in the pre-AID era at higher liraglutide doses, did show hyperglycaemia-with-ketosis signals. The mitigations are: never stop basal insulin; maintain home ketone monitoring capability; check ketones during illness, unexplained hyperglycaemia, nausea, or reduced food intake; avoid aggressive over-reduction; keep CGM alerts on. Any concerning ketone reading is a reason to contact your diabetes care team immediately.

2. Hypoglycaemia

Hypoglycaemia is driven by aggressive insulin reduction running behind the physiology, not by GLP-1RA itself. Use CGM with alerts on. On AID systems, start with higher glucose targets and lower them gradually as tolerance confirms. The 2 to 4 week review cadence is specifically designed to catch hypo patterns before they become severe.

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. Management: smaller and more frequent meals; avoid large fatty meals during titration; moderate fibre early on; prioritise hydration; if side effects persist, slow titration rather than force through. Discontinuation for persistent GI events in the RCT evidence ranges 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 (see Part 4). Practical protection framework, drawn from general sports-nutrition guidance applied to this context: protein around 1.5 g per kg of body weight 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 the Jensen 2024 JAMA Network Open RCT, which showed that exercise added to GLP-1RA-induced weight loss preserves bone mineral density better than GLP-1RA alone in obesity (directional support only for T1D; the primary population was non-T1D). Discuss resistance-training plans with your diabetes care team so pre-exercise glucose targets can be agreed.

The first 12 weeks, a rough timeline

• Week 0. Baseline CGM review, ketone-monitoring education, AID settings snapshot, reduction band agreed with care team, starting dose prescribed.
• Weeks 1 to 4. Lowest starting dose (for example semaglutide 0.25 mg weekly). Insulin reduction as agreed. CGM review at 1 and 2 weeks. Watch for GI tolerance. If tolerance is fine and CGM shows no excess hypo, consider dose escalation at week 4.
• Weeks 4 to 8. Dose escalated (for example semaglutide 0.5 mg weekly). Second insulin reduction step, if the band was 20 per cent or above. Second CGM review cycle. Weight trajectory starts to emerge.
• Weeks 8 to 12. Dose may escalate further (for example semaglutide 1 mg weekly). CGM review cycles land monthly. TDD reduction approaching the band; weight loss becoming clinically visible.
• Week 12 and beyond. Minimum-effective-dose conversation. If TIR, TDD and weight goals are met, stay at current dose. Further escalation only when a clinical reason exists.