Miglitol

證據等級: L5

預測適應症: 10 個

Miglitol: From Type 2 Diabetes Mellitus to Type 1 Diabetes Mellitus (Insulin Adjunct)

Note on candidate selection: This evidence pack lists 10 TxGNN-predicted indications for miglitol. The top 9 (by raw TxGNN score, e.g. focal stiff limb syndrome, classic stiff person syndrome) have zero clinical trial or literature support and are flagged by the model’s own rationale as likely knowledge-graph artifacts (indirect paths through diabetes-comorbidity nodes, not direct pharmacology). Only Type 1 Diabetes Mellitus (rank 10) has real supporting evidence and a non-Hold recommendation. This report focuses on that candidate as the only actionable one in the pack.

One-Sentence Summary

Miglitol is an alpha-glucosidase inhibitor originally used to control postprandial hyperglycemia in type 2 diabetes. The TxGNN model — and a body of older clinical literature — suggests it may also be useful as an adjunct to insulin therapy in Type 1 Diabetes Mellitus, with 7 clinical trials (1 directly relevant, Phase 3) and 16 publications currently supporting this direction.

Quick Overview

Item	Content
Original Indication	Type 2 Diabetes Mellitus (postprandial glycemic control) — general pharmacological knowledge; no Taiwan/US regulatory license data available
Predicted New Indication	Type 1 Diabetes Mellitus (adjunct to insulin)
TxGNN Prediction Score	99.60%
Evidence Level	L2
US Market Status	未上市 (Not Marketed)
Number of NDAs	0
Recommended Decision	Proceed with Guardrails

Why is This Prediction Reasonable?

Miglitol is an alpha-glucosidase inhibitor: it delays intestinal breakdown and absorption of complex carbohydrates, blunting the postprandial blood glucose spike. This mechanism does not depend on residual insulin secretion, which is why it is pharmacologically plausible in Type 1 Diabetes Mellitus (T1DM) as well as Type 2.

In T1DM, exogenous insulin dosing often fails to fully control the sharp postprandial glucose rise, even with intensive insulin therapy. Adding an alpha-glucosidase inhibitor like miglitol slows carbohydrate absorption, flattening this peak and — in several of the studies below — reducing the insulin dose needed around meals and the associated risk of post-meal hypoglycemia from over-correction.

This is not a novel mechanistic hypothesis: the same drug class (miglitol, and its research-era compounds BAY-m-1099/BAY-o-1248) has been studied as an insulin adjunct in insulin-dependent diabetes since the late 1980s, well before TxGNN’s prediction. The model’s high score here reflects a real, previously documented off-label pattern rather than a purely novel signal — unlike the other 9 predictions in this pack, which the model’s own rationale identifies as indirect knowledge-graph paths with no supporting pharmacology.

Clinical Trial Evidence

Trial Number	Phase	Status	Enrollment	Key Findings
NCT00213109	Phase 3	Completed	N/A	Open-label trial evaluating efficacy and safety of miglitol in insulin-treated Type 1 Diabetes patients. Directly relevant, but open-label (non-placebo-controlled), which limits evidence strength.

Six additional trials were returned by the search (NCT02475499, NCT02476760, NCT02456428, NCT06449235, NCT03492580, NCT01697592) but were assessed as not relevant — they study incretin-based drugs, canagliflozin, or omarigliptin in Type 2 diabetes, not miglitol in T1DM, and are excluded from this table.

Literature Evidence

PMID	Year	Type	Journal	Key Findings
21869539	2011	Cohort	Endocrine Journal	Miglitol 25→50mg TID added to intensive insulin therapy in 11 T1DM patients; assessed effect on insulin dose, weight, hypoglycemia, and incretin response.
24843410	2010	Cohort/Combination therapy	J Diabetes Investigation	Combination of miglitol + insulin in T1DM; addresses uncontrolled postprandial glucose rise despite intensive insulin therapy.
2180090	1990	Clinical study (placebo-controlled)	S Afr Med J	50mg miglitol vs placebo in 11 insulin-dependent diabetics; significantly lowered post-meal glucose increments (p<0.001).
2663321	1989	Clinical study (single-blind, placebo)	Diabetes Research	Miglitol (BAY-m-1099) vs placebo in 13 insulin-dependent diabetics; significantly reduced postprandial glucose AUC (p<0.01).
3311550	1987	Clinical study	Clin Pharmacol Ther	Miglitol (BAY-m-1099) reduced meal-time insulin requirements in IDDM patients.
2060451	1991	Crossover study	Diabetes Care	Alpha-glucosidase inhibition (miglitol) evaluated as adjunct to insulin, including timing of insulin administration relative to meals.
3130257	1988	Clinical study	Eur J Clin Invest	Two alpha-glucosidase inhibitors (including miglitol precursor BAYm1099) evaluated for glycemic control and insulin requirements in IDDM.
3277827	1988	Clinical study	Diabetes Res Clin Pract	Two alpha-glucosidase inhibitors (including BAY m 1099/miglitol) improved postprandial metabolic control in insulin-dependent diabetics.
3520133	1986	Clinical study	Klinische Wochenschrift	BAYo1248 and BAYm1099 (miglitol) significantly improved postprandial glucose tolerance and reduced insulin requirements vs placebo.
3286168	1988	Clinical study	Diabetes Res Clin Pract	Evaluated timing of preprandial insulin combined with alpha-glucosidase inhibition (BAY-m-1099/miglitol) in IDDM.

Six further publications (PMID 8261749, 11460577, 12073790, 33268615, 20307399, 3653827) were returned but were either general reviews not specific to miglitol, studies of unrelated drugs (e.g. SGLT2 inhibitors), or lacked usable abstract content — excluded from the table above.

US Market Information

Miglitol currently has no marketing license on record for this jurisdiction (market status: 未上市 / Not Marketed, 0 NDAs). No product/dosage form data is available to tabulate.

Safety Considerations

Please refer to the package insert for safety information.

Conclusion and Next Steps

Decision: Proceed with Guardrails

Rationale: The mechanism (delayed carbohydrate absorption, insulin-independent) is directly applicable to T1DM, and this use is backed by one completed Phase 3 open-label trial plus ~10 supportive clinical studies spanning 1986–2011 — sufficient for L2 evidence. However, most of the supportive literature is decades old, uses miglitol’s pre-approval code name (BAY-m-1099), and no contemporary placebo-controlled RCT or formal safety/MOA documentation is available.

To proceed, the following is needed:

TFDA/FDA package insert data — key warnings, contraindications, and DDI profile (currently a blocking data gap)
Formal, current mechanism-of-action documentation from DrugBank or equivalent
A contemporary placebo-controlled RCT of miglitol as insulin adjunct in T1DM, given the existing evidence base predates modern insulin therapy standards
Clarification of regulatory pathway, since miglitol currently has no license status in this jurisdiction
Disclaimer

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