Insulin Detemir

證據等級: L5

預測適應症: 10 個

Insulin Detemir: From Diabetes Mellitus (Insulin-Dependent) to Type 1 Diabetes Mellitus

One-Sentence Summary

Insulin detemir is a long-acting basal insulin analog used for insulin replacement therapy in diabetes mellitus. The TxGNN model’s top-ranked prediction is Type 1 Diabetes Mellitus, supported by 50 clinical trials and 19 publications — but this is not a novel repurposing signal: the evidence itself shows type 1 diabetes is insulin detemir’s already-established, on-label use, not a new indication. This candidate should be treated as a data-quality/model-artifact case rather than a genuine repurposing opportunity, and the remaining nine TxGNN candidates (ranks 2–10) are mechanistically weak or unsupported.

Quick Overview

Item	Content
Original Indication	Not formally recorded in this dataset (data gap) — insulin detemir is a long-acting basal insulin analog for diabetes mellitus (type 1 and type 2)
Predicted New Indication	Type 1 Diabetes Mellitus (flagged as the drug’s original indication, not a true new use — see caveat below)
TxGNN Prediction Score	99.77%
Evidence Level	L1
US Market Status	Not Marketed
Number of NDAs	0
Recommended Decision	Proceed with Guardrails

⚠ Important Caveat on This Prediction

The evidence pack’s own rationale for the top-ranked prediction explicitly states this is not a genuine repurposing candidate:

“此為 insulin detemir（長效基礎胰島素類似物）之原始核准適應症，非真正意義上的老藥新用” — This is insulin detemir’s original approved indication, not true drug repurposing.

Type 1 diabetes is the condition insulin detemir was developed and approved to treat. The TxGNN model surfaced it as a top prediction because the drug-disease relationship is strongly embedded in the knowledge graph — not because it represents a new therapeutic opportunity. The large clinical trial and literature base below confirms an established use, not an emerging hypothesis.

Of the other nine TxGNN candidates in this evidence pack, none has clinical trial or literature support: eight are rated L5 (model prediction only, recommendation “Hold”), two of those (drug-induced localized lipodystrophy, pressure-induced localized lipoatrophy) are flagged in the rationale as likely reversed-direction artifacts (insulin injection is a known cause of these conditions, not a treatment), and two rare-disease candidates (thiamine-responsive dysfunction syndrome, pancreatic agenesis) are L4 mechanistic extensions with no dedicated studies.

Why is This Prediction Reasonable?

Formal mechanism-of-action data (original_moa) is marked as a data gap in this dataset. However, the evidence pack’s rationale field describes the mechanism: insulin detemir binds the insulin receptor, and its myristic acid (C14 fatty acid) side chain allows reversible albumin binding, which slows subcutaneous absorption and produces a stable, prolonged basal insulin effect. This provides steady basal glucose control that directly substitutes for the absolute insulin deficiency characteristic of type 1 diabetes.

Because type 1 diabetes is the pathology insulin detemir was designed to treat, the “predicted new indication” and the drug’s original indication are the same condition. The reasoning that would normally justify a repurposing hypothesis (shared pathway, adjacent disease biology) does not apply here — there is no repurposing logic to evaluate, only confirmation of on-label mechanism-of-action fit.

Clinical Trial Evidence

Trial Number	Phase	Status	Enrollment	Key Findings
NCT01697657	Phase 3	Completed	131	Randomized, multinational crossover trial comparing hypoglycemia frequency: insulin detemir + aspart vs. NPH + aspart in basal-bolus T1D regimens
NCT00184665	Phase 3	Completed	501	2-year efficacy/safety comparison of detemir vs. NPH insulin in T1D (HbA1c, hypoglycemia, antibodies)
NCT03220425	Phase 3	Completed	752	6-month multicenter comparison of detemir (2400 nmol/mL formulation) vs. NPH in basal-bolus T1D regimen
NCT00474045	Phase 3	Completed	470	Randomized trial of detemir vs. NPH (both with aspart) in pregnant women with T1D
NCT00322257	Phase 3	Terminated	596	Inhaled mealtime insulin vs. subcutaneous aspart, both combined with detemir, in T1D (pulmonary safety focus)
NCT00604344	Phase 3	Completed	401	48-week Japanese trial comparing detemir and NPH human insulin in basal-bolus regimen
NCT00312156	Phase 3	Completed	347	Efficacy/safety comparison of detemir vs. NPH in children and adolescents with T1D
NCT00542399	Phase 4	Completed	50	Once- vs. twice-daily detemir dosing in children/adolescents with T1D
NCT00537303	Phase 4	Completed	296	Step-wise addition of insulin aspart to once-daily detemir plus oral antidiabetics
NCT02922179	N/A (Observational)	Completed	103,951	Large real-world descriptive study of long- and intermediate-acting insulin users

Literature Evidence

PMID	Year	Type	Journal	Key Findings
36623517	2023	RCT	Lancet Diabetes Endocrinol	EXPECT non-inferiority trial: insulin degludec vs. detemir (both + aspart) in pregnant women with T1D
29477399	2018	Systematic Review / Network Meta-analysis	Value Health	Comparative efficacy/safety of basal insulin regimens in adults with T1D
33662147	2021	Systematic Review (Cochrane)	Cochrane Database Syst Rev	Ultra-long-acting insulin analogues for people with T1D
21878861	2011	Systematic Review / Meta-analysis	Pol Arch Med Wewn	Detemir vs. NPH insulin in T1D: glycemic control outcomes
36763996	2022	Review / Meta-analysis	Clin Ther	Efficacy and tolerability of degludec vs. other long-acting basal insulins (incl. detemir) in T1D/T2D
15516157	2004	Review	Drugs	Insulin detemir: review of use in T1D and T2D management
15691219	2005	Review	BioDrugs	Spotlight on insulin detemir in T1D and T2D
17326333	2006	Review	Vasc Health Risk Manag	Insulin detemir in the treatment of T1D and T2D
20539842	2010	Review	Vasc Health Risk Manag	Update on treatment of T1D and T2D, focused on insulin detemir
18454569	2008	Review	Paediatr Drugs	Insulin analog preparations (incl. detemir) in children/adolescents with T1D

US Market Information

No marketing authorization records are present in this dataset. taiwan_regulatory.total_licenses = 0 and licenses is empty, consistent with the recorded market status of Not Marketed in this jurisdiction. No product/NDA table can be constructed from available data.

Safety Considerations

Please refer to the package insert for safety information. (Key warnings, contraindications, and drug-interaction data are all marked as data gaps in this evidence pack — notably DG001, a Blocking severity gap on TFDA label warnings/contraindications, which prevents any S1 safety pre-assessment.)

Conclusion and Next Steps

Decision: Proceed with Guardrails

Rationale: Type 1 diabetes mellitus is supported by an extensive Phase 3 RCT and systematic-review base (L1), but this reflects insulin detemir’s existing approved use, not a validated new indication — so “proceed” here means proceeding with the understanding that no genuine repurposing opportunity exists in the top-ranked candidate, and the guardrail is against mistaking model rank for novelty. The nine other TxGNN candidates in this pack (ranks 2–10) do not clear the evidence bar: eight are L5/Hold with no trial or literature support, two are flagged as probable reversed-direction (adverse-effect) artifacts, and two rare-disease candidates are speculative mechanistic extensions only.

To proceed, the following is needed:

Resolve DG001 (Blocking): obtain TFDA label warnings/contraindications before any safety pre-assessment can occur
Resolve DG002 (High): obtain formal DrugBank MOA data to properly ground mechanism-based candidate scoring
Populate drug.original_indications so future TxGNN runs can auto-detect and exclude “already-approved indication” false positives like this one
If genuine repurposing signal is the goal, deprioritize rank 1 and instead investigate whether any lower-ranked candidates (e.g., pancreatic agenesis, thiamine-responsive dysfunction syndrome) warrant targeted literature/trial searches beyond the mechanistic extrapolation already provided
Disclaimer

This content is for research purposes only and does not constitute medical advice. Clinical validation is required before any clinical application.