Adjunctive Therapy in T1D, Open Questions and What’s Next

The paediatric gap

ISPAD 2024 is unambiguous: no Phase 3 RCTs exist for GLP-1RA, tirzepatide or SGLT2i in paediatric T1D. Paediatric GLP-1RA approvals are for obesity or T2D only (liraglutide, exenatide, dulaglutide from age 10 and above). None are approved for paediatric T1D. The chapter language “could be considered adjunctive therapy in specific populations of children and adolescents with T1D and obesity” sits inside a gap-first overarching statement that “larger trials are needed”. The DTS Consensus 2025 also flags this as a major research gap.

Known concerns in this age group cluster around four issues: gastrointestinal side effects in up to 30 per cent of paediatric patients on GLP-1RA for obesity or T2D indications; an unknown effect on peak bone mass accrual in adolescence (the window during which most bone is laid down); an unknown effect on growth and puberty; and adherence patterns that differ from adult populations.

There is no short-form paediatric answer that is safer than “speak to your paediatric diabetes team”. Any paediatric off-label use should be led by a specialist paediatric diabetes team, with explicit discussion of the evidence gap, ketone education, CGM with alerts, and growth and puberty monitoring. Evidence grade for paediatric T1D adjunct therapy is D by absence of Phase 3 data.

Tirzepatide in T1D, the Phase 3 gap

Tirzepatide (dual GLP-1/GIP agonist) in T1D is currently anchored on a single-centre retrospective observational cohort of 26 adults (Akturk and Shah 2025; Part 2) plus the Klein and Shah 2025 genetic-obesity substudy. Neither is an RCT. The signal is encouraging: larger weight loss than monotherapy GLP-1RA, meaningful TIR gains, no DKA, one peroneal nerve palsy linked to rapid weight loss. Evidence grade C.

Two Phase 3 trials are in flight. SURPASS-T1D-1 (NCT06914895) tests tirzepatide in adults with T1D, weekly, multi-centre, placebo-controlled. SURPASS-T1D-2 (NCT06962280) is a parallel design. Until these read out, head-to-head claims that tirzepatide is incrementally better than semaglutide in T1D rest on within-subject observational data, and the 2025 DTS Consensus is explicit on this point.

Long-term bone and lean mass

Mechanistically, Sherk, Schauer and Shah 2020 showed that GLP-1, GIP and GLP-2 each acutely suppress bone resorption (measured by CTX) in healthy adults and in people with T1D. That is mechanistic evidence, not fracture-outcome evidence. In T2D meta-analyses, GLP-1RA and DPP-4i are not associated with increased fracture risk. T1D itself carries a fracture risk that is 2 to 5 times the general population. Long-term fracture end-points with GLP-1RA or tirzepatide in T1D have not been measured.

The strongest interventional protection signal comes from Jensen 2024 (JAMA Network Open), where exercise added to GLP-1RA-induced weight loss preserved BMD better than GLP-1RA alone. Population was obesity, not T1D, so the signal is directional support rather than direct evidence. The practical protection framework (1.5 g per kg per day protein, resistance training 2 to 4 times per week, vitamin D, iron and B12 monitoring) is drawn from general sports-nutrition guidance applied to this context and remains Grade D for T1D-specific outcomes until proper T1D trials run.

The questions that need answering are fracture end-points in adult T1D over 2 to 5 years on GLP-1RA or tirzepatide; bone-density trajectories in adolescents during the peak-bone-mass window; and head-to-head body-composition preservation across semaglutide, tirzepatide and liraglutide with structured resistance training protocols.

AID algorithm adaptation

Current AID algorithms were designed for a non-adjunct context. When TDD falls by 20 per cent or more over 12 weeks because of GLP-1RA, the algorithm’s insulin-history buffer, its learning behaviour, and its ceiling of maximum basal suppression are all being asked to absorb a shift they were not explicitly designed for. In practice, manual settings adjustment (Part 3, Track 3) sits alongside the algorithm response. The DTS Consensus 2025 (Recommendations 12 and 13) calls directly for algorithm-level work.

Three open lines of work matter: automated detection of GLP-1RA-driven TDD shifts, so the algorithm can re-tune without manual profile-switching; per-system comparison of adaptation timing across Tandem Control-IQ, Omnipod 5, CamAPS FX, Medtronic 780G and iLet in GLP-1RA users; and faster insulin analogue compatibility, since DTS Recommendation 16 flags the kinetic alignment between ultra-rapid insulin and delayed gastric emptying as a design opportunity.

Sick day on adjunctive therapy, the red-flag flow

The sick-day question lands often once GLP-1RA is in the regimen. The pattern that matters is not classic DKA; it is euglycaemic hyperketonaemia, where blood glucose can sit near-normal while ketones climb. Pasqua 2025 saw this in 2 of 24 completers. The framework below is population-level. The personalised plan, including correction-insulin steps and the threshold at which the team wants the next call, stays with the diabetes team. Insulin is never stopped.

Equity of access

ADJUST-T1D was 88 per cent non-Hispanic White and 88 per cent privately insured. Real-world uptake of GLP-1RA in T1D, off-label, is shaped heavily by supply, price, insurance and geography. Registry-based monitoring of uptake, outcomes and discontinuation across race, ethnicity and insurance status is an urgent adjacent question. The DTS Consensus raises this; most national diabetes societies have not yet addressed it in formal policy.

Other open questions

Beyond the headline gaps, four more sit on the active research list. A weekly GLP-1RA plus AID Phase 3 programme is needed before regulatory approval is realistic; ADJUST-T1D was N=72 and industry has not yet committed. C-peptide stratification is clinically actionable if the ADJUNCT post-hoc signal (absent C-peptide predicts intolerance) is replicated prospectively. The Klein and Shah 2025 genetic-obesity exploratory signal needs validation in a larger multicentre cohort. And combination with SGLT2i (liraglutide plus dapagliflozin showed additive glycaemic benefit but two DKA events in a small trial) is the next frontier if DKA risk can be managed.

Live trials list

Public ClinicalTrials.gov registrations as of April 2026, the small set most relevant to a person with T1D tracking the field. Tap each for the registry link.

A live trial-tracking page is in planning at GNL. In the meantime, ClinicalTrials.gov search terms “type 1 diabetes” + “GLP-1” or “tirzepatide” will surface current entries.