Table of Contents

1. Zonisamide
2. Zonisamide: From Epilepsy to Tourette Syndrome
   1. One-Sentence Summary
   2. Quick Overview
   3. Why is This Prediction Reasonable?
   4. Clinical Trial Evidence
   5. Literature Evidence
   6. UK Market Information
   7. Safety Considerations
   8. Conclusion and Next Steps
   9. Disclaimer

Zonisamide: From Epilepsy to Tourette Syndrome

One-Sentence Summary

Zonisamide is an antiseizure medication marketed in the UK, and its original use is in epilepsy. The TxGNN model predicts it may be effective for Tourette syndrome with a very high score, but no clinical trials and no directly relevant publications support this prediction. The only related signal found is a review reporting that antiseizure medications can induce tic disorders.

Quick Overview

Why is This Prediction Reasonable?

Currently, detailed mechanism of action data is not available. Based on known pharmacology, zonisamide blocks voltage-gated sodium channels and T-type calcium channels, and it is thought to have weak dopaminergic and serotonergic effects. Its efficacy in epilepsy is established, and these actions could plausibly modulate the neural circuits involved in tics.

This link is speculative. The high TxGNN score is a computational prediction from a knowledge graph and is not clinical evidence. The mechanistic reasoning has not been tested in Tourette syndrome. One review found during the evidence search (PMID 36005856) points the other way, as it describes antiseizure medication-induced obsessive-compulsive and tic disorders.

Clinical Trial Evidence

Currently no related clinical trials registered.

Literature Evidence

Currently no related literature available for Tourette syndrome.

The review “Antiseizure medication-induced obsessive-compulsive disorder and tic disorder: a pragmatic review” (36005856, Expert Review of Neurotherapeutics, 2022) was retrieved under a different indication. It does not support efficacy, and it raises a possible safety signal for tic symptoms with antiseizure medications.

UK Market Information

This identifier follows the BNF classification code format rather than a standard PL number, so verify it against the MHRA register. BNF chapter 4.8.1 covers antiepileptic drugs. The licence indication text and NICE-appraised indications were not available in the data compiled.

Safety Considerations

Knowledge-graph contraindication signals. These come from drug-disease contraindication links in the TxGNN knowledge graph (PrimeKG), not from SmPC label text, and they need confirmation against the SmPC. The conditions flagged include:

• Kidney disease and nephrolithiasis, including acute urate nephropathy
• Liver disease and gallbladder disease
• Metabolic acidosis and acidosis disorders
• Neutropenia, agranulocytosis, and related congenital neutropenia syndromes
• Psychiatric conditions, including anxiety disorder, psychotic disorder, schizoaffective disorder, schizophreniform disorder, dysthymic disorder, neurotic disorder, and mental disorder broadly
• Diarrhoeal disease

Drug interactions. The interaction query returned 100 records, and the first 20 are summarised here.

• Major:
  • Metformin
  • Anticholinergic agents: atropine, hyoscyamine, scopolamine, methscopolamine, glycopyrronium, clidinium, dicyclomine, trospium, mepenzolate, propantheline
• Moderate: aprepitant, morphine, dexamethasone, cimetidine, clarithromycin, dronabinol, nabilone, metoclopramide, miconazole

Indication-specific signals (from other predicted indications in the same dataset):

• A case report of zonisamide-induced mania (PMID 2109869).
• The tic disorder review above.

Please refer to the SmPC and BNF for full safety information. Report suspected adverse reactions via the Yellow Card Scheme.

Conclusion and Next Steps

Decision: Hold

Rationale: The prediction rests on a model score alone (L5), with no registered trials or supporting publications, and the only related literature suggests antiseizure medications may provoke tics. The mechanistic argument is plausible but unproven.

To proceed, the following is needed:

• Mechanism of action data and a clearer mechanistic rationale for tic suppression
• A systematic literature search specific to zonisamide in Tourette syndrome and tic disorders, including any case series
• Review of the SmPC warnings and contraindications, which were not available in the data compiled
• Confirmation of the UK licence number, dosage form, and approved indication
• Preclinical or small exploratory clinical data before any formal trial is considered

Other predictions to consider. Several other predicted indications in the dataset have more supporting evidence, though none has registered trials supporting efficacy.

• Absence epilepsy (L3) has the strongest evidence. It has a clear T-type calcium channel rationale and zonisamide-specific reports (PMIDs 15847848 and 24907183), and these are likely small or retrospective studies.
• Manic bipolar affective disorder (L4) has indirect literature, including a pilot study in bipolar depression and a report of zonisamide-induced mania.
• Fibromyalgia (L4) has one withdrawn trial with zero enrolment.

  Disclaimer

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

Table of Contents

1. Zolmitriptan
2. Zolmitriptan: From Migraine to Migraine with Brainstem Aura
   1. One-Sentence Summary
   2. Quick Overview
   3. Why is This Prediction Reasonable?
   4. Clinical Trial Evidence
   5. Literature Evidence
   6. UK Market Information
   7. Safety Considerations
   8. Conclusion and Next Steps
   9. Disclaimer

Zolmitriptan: From Migraine to Migraine with Brainstem Aura

One-Sentence Summary

Zolmitriptan is a 5-HT1B/1D receptor agonist (a triptan), originally used for the acute treatment of migraine. The TxGNN model predicts it may be effective for migraine with brainstem aura, but there are currently no registered clinical trials and only general migraine and triptan literature (20 publications, mostly reviews) supporting this direction, with no subtype-specific data.

Quick Overview

Why is This Prediction Reasonable?

Zolmitriptan is a selective serotonin 5-HT1B/1D receptor agonist with both central and peripheral activity. Preclinical work shows it can inhibit trigeminovascular activation, and it is established for the acute treatment of migraine with or without aura. Detailed mechanism-of-action data were not available in the data compiled for this report, so this description draws on the published pharmacology literature.

Migraine with brainstem aura is a migraine subtype that shares the core pathophysiology of other migraines, including trigeminovascular activation and CGRP release. On this basis the mechanism is plausible.

There is, however, an important caveat. Triptans have traditionally been used with caution, or avoided, in brainstem aura (formerly basilar-type migraine) because of theoretical concerns about vasoconstriction. The supplied literature consists of general migraine and triptan reviews and gives no subtype-specific efficacy or safety data. The very high TxGNN score mainly reflects proximity to the already approved migraine indication, not independent evidence for this subtype.

Two other predictions (atrophoderma vermiculata and ulerythema ophryogenesis) have no identifiable mechanistic link, no trials and no literature. They appear to be graph-based artefacts and are not pursued further here.

Clinical Trial Evidence

Currently no related clinical trials registered.

Literature Evidence

The publications below are general migraine or triptan papers. None studies zolmitriptan specifically in migraine with brainstem aura.

UK Market Information

The compiled data show one valid single-ingredient authorisation. The identifier is in a non-standard format and is not a typical MHRA PL number, so it should be verified against the MHRA register. The dosage form and approved indication fields were blank. Please check the current SmPC, the BNF and any NICE guidance for the licensed indications.

Safety Considerations

• Contraindications (from the knowledge graph, not label text): The TxGNN knowledge graph links zolmitriptan with contraindications in the following areas. These should be confirmed against the SmPC.
  • Cardiovascular: hypertension, coronary artery disease, myocardial infarction, myocardial ischaemia, Prinzmetal angina, Wolff-Parkinson-White syndrome, peripheral vascular disease, Raynaud disease, ischaemic bowel disease
  • Cerebrovascular: stroke, transient ischaemic attack, cerebral infarction, brain ischaemia, vertebrobasilar insufficiency, intracranial arteriosclerosis
  • Migraine subtype: familial or sporadic hemiplegic migraine
  • Other: kidney disease, liver disease, gallbladder disease, haemoglobinopathy

  The cerebrovascular and hemiplegic migraine entries are particularly relevant to a brainstem aura indication.

• Drug Interactions: 82 interactions were recorded. The main ones are:
  • Major: lorcaserin, dolasetron, palonosetron, ondansetron, granisetron, dexfenfluramine, fenfluramine, sibutramine (mostly serotonergic or 5-HT3 antagonist combinations)
  • Moderate: morphine (including liposomal), cimetidine, obeticholic acid
• Case report signal: A published case describes acute myocardial infarction after zolmitriptan use.

Please refer to the SmPC and BNF for full safety information. Report suspected adverse reactions via the Yellow Card Scheme.

Conclusion and Next Steps

Decision: Hold

Rationale: The prediction is mechanistically plausible but is driven mainly by similarity to the approved migraine indication. There are no clinical trials and no subtype-specific data, and the traditional caution about triptans in brainstem aura, together with vascular contraindications recorded in the knowledge graph, raises a real safety question.

To proceed, the following is needed:

• Subtype-specific efficacy and safety evidence for triptans in migraine with brainstem aura (or a well-designed study)
• Review of the current SmPC and BNF warnings and contraindications for zolmitriptan, which were not available in the data compiled for this report
• Verification of the UK marketing authorisation record (number format, dosage form and approved indication)
• A cardiovascular and cerebrovascular risk assessment for the target population
• Detailed mechanism-of-action data from DrugBank to support the mechanistic-link analysis

  Disclaimer

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

Table of Contents

1. Ziprasidone
2. Ziprasidone: From Bipolar Disorder and Schizophrenia to Major Affective Disorder
   1. One-Sentence Summary
   2. Quick Overview
   3. Why is This Prediction Reasonable?
   4. Clinical Trial Evidence
   5. Literature Evidence
   6. UK Market Information
   7. Safety Considerations
   8. Conclusion and Next Steps
   9. Disclaimer

Ziprasidone: From Bipolar Disorder and Schizophrenia to Major Affective Disorder

One-Sentence Summary

Ziprasidone is an atypical antipsychotic, originally used for schizophrenia and manic episodes in bipolar disorder. The TxGNN model predicts it may be effective for major affective disorder (chiefly depression), with 7 clinical trials and 20 publications currently supporting this direction. Several of the trials are small or exploratory, and only a few are randomised, placebo-controlled studies.

Quick Overview

The top-ranked prediction was trichotillomania (99.83%), but it has no supporting trials or literature, so it is rated L5 and Hold. Major affective disorder is the best-supported prediction, and this report focuses on it.

Why is This Prediction Reasonable?

Currently, detailed mechanism of action data is not available in the data compiled for this report. Based on known pharmacology, ziprasidone antagonises dopamine D2 and serotonin 5-HT2A receptors, acts as a 5-HT1A agonist, and inhibits serotonin and noradrenaline reuptake. Its efficacy in bipolar mania and schizophrenia is established.

Mood disorders overlap clinically and biologically with bipolar disorder. Depressive and mixed states occur within bipolar illness, and several atypical antipsychotics are already used to augment antidepressants in major depression. The combined receptor profile above gives a plausible basis for antidepressant effect, particularly as an add-on in patients who respond inadequately to antidepressants.

Bipolar mania is already an approved indication, so the genuinely new contribution is mainly in unipolar and psychotic depression. The evidence is exploratory rather than confirmatory.

Clinical Trial Evidence

No trials are listed in other registries (e.g. ISRCTN or EU CTR) in the data compiled for this report.

Literature Evidence

UK Market Information

This is a BNF-style product code (BNF chapter 4.2.1, antipsychotic drugs) rather than a standard PL number. Please check the MHRA register and the SmPC for the licensed indications and authorisation details. No NICE appraisal data were available.

Safety Considerations

• Contraindications suggested by the knowledge graph (not label text):
  • Cardiac: long QT syndrome (familial and acquired), torsades de pointes, myocardial infarction, sudden cardiac arrest, heart conduction disease.
  • Metabolic: diabetes mellitus, diabetic ketoacidosis, hyperglycaemia, hyperlipidaemia, obesity, potassium deficiency.
  • Haematological: neutropenia, agranulocytosis.
  • Neurological: epilepsy, basal ganglia disease.
  • Organ function: liver failure, kidney disease.
  • Other: orthostatic hypotension, female breast carcinoma.
• Drug Interactions: 100 interactions are recorded. Examples classed as Major are hydrocortisone, amphotericin B (including the lipid complex), bupropion and morphine. Moderate examples include metformin, pioglitazone, epinephrine, atropine and famotidine.
• QTc prolongation: The reviews in the data compiled for this report note this as a key concern, particularly in children and adolescents.

Please refer to the SmPC and BNF for authoritative safety information. Report suspected adverse reactions via the Yellow Card Scheme.

Conclusion and Next Steps

Decision: Proceed with Guardrails

Rationale: Several Phase 2 randomised, placebo-controlled trials and multiple systematic reviews cover ziprasidone in depression. However, there is no completed Phase 3 randomised evidence, and the one trial labelled Phase 3 is an eight-patient open-label study. This supports L2 evidence and further investigation, not routine use.

To proceed, the following is needed:

• Published results and effect sizes from the completed Phase 2 trials, especially NCT00555997, NCT00490542 and NCT00633399
• An adequately powered Phase 3 randomised trial in unipolar or psychotic depression
• Baseline and follow-up ECG (QTc) monitoring, electrolyte checks, and exclusion of patients with cardiac risk factors
• Review of the recorded drug interactions, especially the Major-rated ones
• Confirmation of the licensed UK indications and product details from the MHRA register and SmPC
• Mechanism of action data, which are currently missing

  Disclaimer

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

Table of Contents

1. Zidovudine
2. Zidovudine: From HIV Infection and AIDS to Feline Acquired Immunodeficiency Syndrome
   1. One-Sentence Summary
   2. Quick Overview
   3. Why is This Prediction Reasonable?
   4. Clinical Trial Evidence
   5. Literature Evidence
   6. UK Market Information
   7. Safety Considerations
   8. Conclusion and Next Steps
   9. Disclaimer

Zidovudine: From HIV Infection and AIDS to Feline Acquired Immunodeficiency Syndrome

One-Sentence Summary

Zidovudine is a nucleoside reverse transcriptase inhibitor originally used to treat HIV infection and AIDS in people. The TxGNN model predicts it may be effective for feline acquired immunodeficiency syndrome, a veterinary condition caused by feline retroviruses. There are no registered clinical trials and 19 publications for this prediction, all animal or veterinary studies. There is no human evidence for this indication.

Quick Overview

Why is This Prediction Reasonable?

Detailed mechanism of action data are not available in the data compiled for this report. Zidovudine is, however, a nucleoside reverse transcriptase inhibitor (NRTI). It is a prodrug that is phosphorylated inside cells and blocks the viral enzyme that copies retroviral RNA into DNA. Its efficacy in HIV infection and AIDS is established.

Feline immunodeficiency virus (FIV) is a lentivirus closely related to HIV, and its reverse transcriptase is similar. FIV infection in cats is widely used as an animal model of HIV. Zidovudine has therefore been tested in infected cats for decades, and the prediction fits the drug’s known pharmacology.

The prediction is not a new human use. It describes veterinary use in cats, and in the literature zidovudine in cats is mostly a research tool. Zidovudine is also known to cause blood toxicity in cats, which limits its practical value.

Clinical Trial Evidence

Currently no related clinical trials registered.

Literature Evidence

UK Market Information

This is the only authorisation in the data compiled for this report. It is a single-ingredient product with valid status. The code suggests the BNF antiretroviral (HIV infection) section, which fits zidovudine’s original use. The data do not record a licensed indication, manufacturer or NICE appraisal, so confirm these against the SmPC.

Safety Considerations

• Contraindication associations (knowledge-graph data, not label text): liver failure, liver disease, kidney disease, pancreatitis, neutropenia, agranulocytosis, anaemia (including macrocytic anaemia), myopathy and other muscle diseases, myositis, dermatomyositis, lactic acidosis, fatty liver disease, hypertriglyceridaemia and gallbladder disease. Haematological and muscle toxicity are the main concerns. Check the SmPC for the formal contraindications.
• Drug interactions: The interaction database lists 100 interacting drugs, and the data compiled for this report show 20. Most are moderate: amphotericin B products, corticosteroids (betamethasone, budesonide, dexamethasone, hydrocortisone, prednisolone, prednisone, triamcinolone), statins (rosuvastatin, simvastatin), naltrexone and orlistat. Cimetidine and clarithromycin are minor. The level for pantoprazole, omeprazole, morphine and metformin is unknown.
• Animal toxicity: Zidovudine caused blood toxicity in FeLV-infected cats in the toxicity study above (PMID 2163339).

Please refer to the SmPC and BNF for full safety information. Report suspected adverse reactions in people via the Yellow Card Scheme.

Conclusion and Next Steps

Decision: Hold

Rationale: The prediction is mechanistically plausible, but the evidence is animal and veterinary only, with no registered trials. Cats are not a human indication, and zidovudine’s blood toxicity in cats is a concern. The prediction does not support a UK human repurposing programme.

To proceed, the following is needed:

• Mechanism of action data to confirm the mechanistic link
• A clear decision on the intended setting. This is a veterinary or research-model use rather than a human indication, so any development would fall under veterinary regulation
• Controlled veterinary efficacy and safety data, including dose and haematological toxicity monitoring
• Approved indication text and SmPC safety information for the UK authorisation, because the licence record is incomplete

Other predicted indications for zidovudine include AIDS-related complex and congenital HIV. These fall within its existing HIV use and are not true repurposing.

Disclaimer

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

Table of Contents

1. Zanamivir
2. Zanamivir: From Influenza to Pyelonephritis
   1. One-Sentence Summary
   2. Quick Overview
   3. Why is This Prediction Reasonable?
   4. Clinical Trial Evidence
   5. Literature Evidence
   6. UK Market Information
   7. Safety Considerations
   8. Conclusion and Next Steps
   9. Disclaimer

Zanamivir: From Influenza to Pyelonephritis

One-Sentence Summary

Zanamivir is a neuraminidase inhibitor, originally used to treat influenza. The TxGNN model predicts it may be effective for pyelonephritis, but this is a model prediction only, with no clinical trials and no publications supporting it.

Quick Overview

Why is This Prediction Reasonable?

Currently, detailed mechanism of action data is not available in the data compiled for this report. Zanamivir is known to be a viral neuraminidase inhibitor, and its efficacy in influenza is established.

Pyelonephritis is typically a bacterial infection of the kidney. No plausible mechanistic link to viral neuraminidase inhibition has been identified, and the high TxGNN score appears to be a graph-based prediction without supporting clinical or literature evidence. On current information, the prediction is not mechanistically credible.

Other top-ranked predictions (for example disorders of tyrosine and phenylalanine metabolism, dengue susceptibility, aspergillosis susceptibility, Legionnaires’ disease susceptibility, HIV susceptibility and schistosomiasis susceptibility) show the same pattern. None has a supporting clinical trial, and the few publications retrieved concern influenza antiviral resistance and are unrelated to those conditions.

Clinical Trial Evidence

Currently no related clinical trials registered.

Literature Evidence

Currently no related literature available for pyelonephritis.

UK Market Information

The BNF classification for this product is within the antiviral chapter (influenza, section 5.3.4).

Safety Considerations

• Contraindication-related disease associations (knowledge graph, not label text): epilepsy, chronic obstructive pulmonary disease, severe early-onset COPD, bronchial disease and delirium. These come from graph associations and should be verified against the SmPC. Inhaled zanamivir carries a recognised risk of bronchospasm in people with airways disease.
• Drug Interactions (8 recorded):
  • Insulin human (inhalation, rapid acting): Moderate
  • Clarithromycin, loratadine, montelukast, ketotifen, fexofenadine, ibuprofen and diclofenac: severity level not classified

Please refer to the SmPC and BNF for full safety information. Report suspected adverse reactions via the Yellow Card Scheme.

Conclusion and Next Steps

Decision: Hold

Rationale: The prediction rests on model output alone (L5), with no registered trials or relevant publications. Pyelonephritis is a bacterial infection, and there is no plausible mechanistic link to a viral neuraminidase inhibitor.

To proceed, the following is needed:

• Mechanism of action data and a credible mechanistic rationale for the new indication
• Preclinical or clinical evidence for activity in the predicted condition
• Approved indication text and dosage form details from the UK licence record
• SmPC warnings and contraindications for a formal safety review
• Assessment of route compatibility (inhaled powder versus the route likely needed for a kidney infection)

  Disclaimer

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

Table of Contents

1. Zafirlukast
2. Zafirlukast: From Asthma to Bronchitis
   1. One-Sentence Summary
   2. Quick Overview
   3. Why is This Prediction Reasonable?
   4. Clinical Trial Evidence
   5. Literature Evidence
   6. UK Market Information
   7. Safety Considerations
   8. Conclusion and Next Steps
   9. Disclaimer

Zafirlukast: From Asthma to Bronchitis

One-Sentence Summary

Zafirlukast is a cysteinyl leukotriene receptor antagonist, originally used to treat asthma. The TxGNN model predicts it may be effective for bronchitis (score 99.93%), but no clinical trials or publications were found for this indication, so the prediction rests on the model alone. A second prediction, obstructive lung disease, is supported by 18 publications (mostly reviews) but no clinical trials.

Quick Overview

Why is This Prediction Reasonable?

Detailed mechanism of action data is not available in the data compiled for this report. Based on established pharmacology, zafirlukast is a selective, competitive antagonist of the CysLT1 receptor. It blocks the effects of the cysteinyl leukotrienes LTC4, LTD4 and LTE4, which drive bronchoconstriction, airway oedema, mucus secretion and eosinophil recruitment.

Asthma and bronchitis are both inflammatory airway conditions in which mucus hypersecretion and airway narrowing play a part. A mechanistic link is therefore plausible. However, this is a computational prediction only. The very high score may partly reflect the overlap between asthma and other airway diseases rather than genuine new-indication evidence. No primary data were available to verify the link.

For context, the second-ranked prediction, obstructive lung disease (score 99.17%, evidence level L4), is supported by reviews such as the Busse et al. 1999 review of leukotriene pathway inhibitors in asthma and COPD. This evidence is indirect, and the prediction is largely explained by the existing asthma indication.

Clinical Trial Evidence

Currently no related clinical trials registered for bronchitis.

Literature Evidence

Currently no related literature available for bronchitis.

For reference, the literature retrieved for the second prediction (obstructive lung disease) includes the following. All are reviews or pharmacology overviews, and none are primary trials.

UK Market Information

The identifier above follows a BNF-style code format rather than a standard PL number, and the BNF section 3.3.2 (leukotriene receptor antagonists) is the relevant classification. Please confirm the authorisation details with the MHRA.

Safety Considerations

• Contraindications (from knowledge graph associations, not label text): liver disease, cirrhosis of liver, gallbladder disease, insomnia, anxiety, anxiety disorder, neurotic disorder and dysthymic disorder. These should be checked against the SmPC.
• Drug Interactions: 100 interactions were identified. Those listed as moderate include aprepitant, acetylsalicylic acid, metronidazole, budesonide (including nasal), saxagliptin, eliglustat, naltrexone, nateglinide, nitisinone, repaglinide, rosuvastatin, simvastatin and tinidazole. Sibutramine is listed as minor. Pantoprazole, mesalazine, morphine, metformin and omeprazole have an unknown severity level.

Please refer to the SmPC and BNF for full safety information. Report suspected adverse reactions via the Yellow Card Scheme.

Conclusion and Next Steps

Decision: Hold

Rationale: The bronchitis prediction has a very high model score but no supporting trials or literature (evidence level L5). The UK record also lacks indication and safety label data, so the benefit-risk balance cannot be assessed.

To proceed, the following is needed:

• Clinical trial or observational evidence for zafirlukast in bronchitis, with the type of bronchitis (acute, chronic or eosinophilic) specified
• Confirmed mechanism of action data and a clearer link between asthma and bronchitis
• UK SmPC warnings, contraindications and approved indication text, particularly for hepatic safety
• A review of the interactions with statins, repaglinide, nateglinide and corticosteroids
• Further evaluation of obstructive lung disease (such as COPD) as a separate research question

  Disclaimer

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

Table of Contents

1. Yohimbine
2. Yohimbine: From an Unspecified Original Indication to Migraine Disorder
   1. One-Sentence Summary
   2. Quick Overview
   3. Why is This Prediction Reasonable?
   4. Clinical Trial Evidence
   5. Literature Evidence
   6. UK Market Information
   7. Safety Considerations
   8. Conclusion and Next Steps
   9. Disclaimer

Yohimbine: From an Unspecified Original Indication to Migraine Disorder

One-Sentence Summary

Yohimbine is a selective alpha-2 adrenoceptor antagonist marketed in the UK as a 5 mg tablet, although the data compiled for this report do not record an approved indication. The TxGNN model predicts it may be effective for migraine disorder, but there are no registered clinical trials and only 7 publications, almost all preclinical or indirect. The evidence is weak, and the pharmacology points in an unfavourable direction, because migraine-related alpha-2 research mainly involves agonists such as clonidine.

Quick Overview

Why is This Prediction Reasonable?

Detailed mechanism-of-action data are not available in the compiled data. From pharmacology records, yohimbine is a selective alpha-2 adrenoceptor antagonist with moderate affinity for alpha-1 adrenoceptors. It also binds several serotonin receptor subtypes (5-HT1A, 1B, 1D, 1E, 1F, 2B, 5A and 7). Many of these serotonin receptors are relevant to existing migraine treatments.

Preclinical work links noradrenergic and alpha-2 signalling to cortical spreading depression (the biological basis of migraine aura), trigeminovascular nociception and meningeal artery tone. This makes the pathway plausibly relevant.

However, the direction of effect is a concern. Alpha-2 agonists such as clonidine are the agents studied in migraine, and an antagonist such as yohimbine could act in the opposite direction. The literature is largely animal or in vitro work, and no clinical study has tested yohimbine in migraine. The link is therefore indirect.

Other predictions in the model’s top ten (migraine with brainstem aura, headache disorder, migraine susceptibility) largely reuse the same migraine literature and add no independent support. Several others (ADHD, transient ischaemic attack, glaucoma) have mechanistic evidence suggesting yohimbine could worsen rather than improve the condition.

Clinical Trial Evidence

Currently no related clinical trials registered.

Literature Evidence

UK Market Information

Only one single-ingredient product is listed, and it is recorded as valid. The compiled data do not give the manufacturer, dosage form or approved indication.

Safety Considerations

• Conditions flagged in the knowledge graph: The TxGNN knowledge graph (drug–disease contraindication links, not label text) lists anxiety disorder, neurotic disorder, dysthymic disorder, kidney disease, pyoureter, liver disease, gallbladder disease, hypotensive disorder, heart disease and cardiovascular disease.
• Drug interactions (20 records):
  • Major: iobenguane (I-131).
  • Moderate: lorcaserin, troglitazone, doxepin (oral and topical) and procarbazine.
  • Minor: morphine (including liposomal).
  • Receptor-binding records (serotonin, alpha-2 adrenoceptors) are also on file and are not interactions in the clinical sense.
• Context from the pharmacology records: One source states that yohimbine is no longer recommended as a treatment for erectile dysfunction.

Warning and contraindication text from the product information was not available. Please refer to the SmPC and BNF for full safety information. Report suspected adverse reactions via the Yellow Card Scheme.

Conclusion and Next Steps

Decision: Hold

Rationale: There are no clinical trials, and the literature is almost entirely animal, in vitro or indirect. Alpha-2 agonists, not antagonists, are the agents associated with migraine research, so the direction of effect is questionable. The very high prediction score appears to reflect model proximity rather than independent evidence. The knowledge graph also flags cardiovascular, hepatic, renal and anxiety-related conditions, which are relevant to a migraine population.

To proceed, the following is needed:

• Product information (SmPC) warnings, contraindications and the approved indication for the UK-listed tablet
• Detailed mechanism-of-action data, specifically whether alpha-2 antagonism could plausibly help rather than aggravate migraine
• Preclinical studies testing yohimbine itself in migraine models (for example spreading depression and trigeminovascular models)
• A cardiovascular and psychiatric safety assessment for the target population before any clinical study is considered

  Disclaimer

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

Table of Contents

1. Xylitol
2. Xylitol: From an Unspecified Original Indication to Meningococcal Infection
   1. One-Sentence Summary
   2. Quick Overview
   3. Why is This Prediction Reasonable?
   4. UK Market Information
   5. Safety Considerations
   6. Conclusion and Next Steps
   7. Disclaimer

Xylitol: From an Unspecified Original Indication to Meningococcal Infection

One-Sentence Summary

Xylitol is a sugar alcohol sold in the UK as part of a mucin/xylitol/sorbitol lozenge, but no approved indication is recorded in the data compiled for this report. The TxGNN model predicts it may be effective for meningococcal infection, but there are currently no clinical trials and no publications supporting this direction.

Quick Overview

Why is This Prediction Reasonable?

Currently, detailed mechanism of action data is not available, and no original indication is recorded. Based on the information compiled, no direct mechanistic link between xylitol and meningococcal infection has been established. The very high TxGNN score reflects a knowledge-graph computational prediction only.

One speculative, unverified rationale is that xylitol has been reported to reduce the adherence of some respiratory bacteria (for example Streptococcus pneumoniae and Haemophilus influenzae) to nasopharyngeal epithelium and to reduce biofilm formation. Whether this extends to Neisseria meningitidis colonisation is not supported by the data compiled for this report.

Invasive meningococcal disease is life-threatening. Xylitol would not be expected to replace antibiotic treatment or vaccination.

UK Market Information

This is a single-ingredient authorisation record that is currently valid. The identifier shown is as supplied in the compiled data and is not in the standard MHRA “PL” format, so it should be verified against the MHRA register.

Safety Considerations

• Knowledge-graph contraindication signal: The TxGNN knowledge graph (PrimeKG) links xylitol to a contraindication edge for dental fluorosis. This comes from the knowledge graph and not from product label text, so it should be interpreted cautiously.

No drug interaction records were found in the data compiled for this report.

Please refer to the SmPC and BNF for full warnings, contraindications and interaction information. Report suspected adverse reactions via the Yellow Card Scheme.

Conclusion and Next Steps

Decision: Hold

Rationale: The prediction rests only on a computational model. There are no registered trials or publications, no mechanism of action data, no recorded original indication, and no label-based safety data. Meningococcal infection is a serious condition where unproven treatments carry significant risk.

To proceed, the following is needed:

• Mechanism of action data (for example from DrugBank)
• The approved indication, dosage form and SmPC for the authorised lozenge product
• Label-based warnings and contraindications from the SmPC and BNF
• Preclinical or observational evidence on xylitol and Neisseria meningitidis adherence or nasopharyngeal carriage
• Review of the dental fluorosis contraindication signal
• Clarification of the intended route and formulation, as route compatibility has not been assessed

  Disclaimer

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

Table of Contents

1. Voriconazole
2. Voriconazole: From Invasive Fungal Infections to Multidrug-Resistant Tuberculosis
   1. One-Sentence Summary
   2. Quick Overview
   3. Why is This Prediction Reasonable?
   4. Clinical Trial Evidence
   5. Literature Evidence
   6. UK Market Information
   7. Safety Considerations
   8. Conclusion and Next Steps
   9. Disclaimer

Voriconazole: From Invasive Fungal Infections to Multidrug-Resistant Tuberculosis

One-Sentence Summary

Voriconazole is an azole antifungal, originally used for fungal infections such as aspergillosis, candidaemia and fusariosis. The TxGNN model predicts it may be effective for multidrug-resistant tuberculosis, but there are no registered clinical trials and only 3 publications, none of which show that voriconazole treats tuberculosis itself.

Quick Overview

Why is This Prediction Reasonable?

Currently, detailed mechanism of action data is not available from the data compiled for this report. Based on general pharmacology, voriconazole is a triazole antifungal that inhibits fungal CYP51 (lanosterol 14-alpha-demethylase), and its efficacy in invasive fungal infections is established.

No direct antimycobacterial mechanism is established for voriconazole. Azoles have shown some in vitro activity against Mycobacterium tuberculosis cytochrome P450 enzymes (CYP121 and CYP125), but this is preclinical and is not supported by the data compiled here. The high TxGNN score most likely reflects proximity to other anti-infective drugs in the knowledge graph rather than validated efficacy.

The retrieved literature mainly describes fungal disease occurring alongside, or mimicking, tuberculosis. In those reports voriconazole treats the fungal infection, not the TB. This prediction should therefore be regarded as hypothesis-generating only.

Clinical Trial Evidence

Currently no related clinical trials registered.

Literature Evidence

Other predicted indications (tuberculoma, inactive tuberculosis) were supported only by case reports or reviews of fungal infections (cerebral phaeohyphomycosis, chronic pulmonary aspergillosis, paracoccidioidomycosis) that mimic or follow TB. These do not support a TB indication. Cysticercosis, bovine TB, avian TB and several congenital malformation or hair disorders had no supporting studies and are likely knowledge-graph artefacts.

UK Market Information

The licence record contains a single valid single-ingredient product. The BNF classification code appears to be 0502010D0 (antifungal drugs, triazole). Indication text and dosage form were not available in the data compiled for this report.

Safety Considerations

• Conditions flagged as contraindications in the knowledge graph (not label text): long QT syndrome (including familial), torsades de pointes, liver disease, gallbladder disease, potassium deficiency, hypercalcaemia, nephrocalcinosis, phosphorus metabolism disorders, impaired renal function, vision disorders and blindness, cardiomyopathy (including hypertrophic), endomyocardial fibrosis and other myocardial disorders.
• Drug interactions: 100 interactions were identified. Those shown include:
  • Major: loperamide, triamcinolone, budesonide, chlorpropamide
  • Moderate: famotidine, rabeprazole, omeprazole, hydrocortisone, dexamethasone, betamethasone, pioglitazone, alosetron, aprepitant, bisacodyl, clarithromycin
  • Minor: amphotericin B, cimetidine

Please also refer to the SmPC and BNF for authoritative safety information. Report suspected adverse reactions via the Yellow Card Scheme.

Conclusion and Next Steps

Decision: Hold

Rationale: The prediction rests on a model score alone. There are no clinical trials, and the literature concerns fungal disease occurring alongside TB, not antimycobacterial activity. No established mechanism links voriconazole to TB, and its safety profile (QT prolongation, hepatic, visual and many interaction concerns) is significant.

To proceed, the following is needed:

• Mechanism of action data and evidence of in vitro activity against M. tuberculosis, including MDR strains
• Preclinical efficacy and pharmacokinetic data at achievable drug concentrations
• Interaction assessment with rifamycin-based and second-line TB regimens, since rifampicin is a strong CYP inducer and is contraindicated with voriconazole
• UK SmPC warnings and contraindications, and approved indication text for the licensed product

  Disclaimer

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

Table of Contents

1. Vildagliptin
2. Vildagliptin: From Type 2 Diabetes Mellitus to Focal Stiff Limb Syndrome
   1. One-Sentence Summary
   2. Quick Overview
   3. Why is This Prediction Reasonable?
   4. Clinical Trial Evidence
   5. Literature Evidence
   6. UK Market Information
   7. Safety Considerations
   8. Conclusion and Next Steps
   9. Disclaimer

Vildagliptin: From Type 2 Diabetes Mellitus to Focal Stiff Limb Syndrome

One-Sentence Summary

Vildagliptin is an oral dipeptidyl peptidase-4 (DPP-4) inhibitor, originally used to manage type 2 diabetes mellitus. The TxGNN model predicts it may be effective for focal stiff limb syndrome, but this is a model prediction only, with no clinical trials and no publications currently supporting it.

Quick Overview

Why is This Prediction Reasonable?

Detailed mechanism of action data is not available in the data compiled for this report. Vildagliptin is known to act on DPP-4 and is listed as a ligand of DPP-8, DPP-9 and TRPV4. Blocking DPP-4 raises active GLP-1 and GIP levels, which improves pancreatic islet function in type 2 diabetes. Its efficacy in type 2 diabetes is established.

Focal stiff limb syndrome belongs to the stiff-person spectrum, a group of rare neurological disorders characterised by muscle stiffness and spasm. These are often thought to be autoimmune and linked to GABAergic signalling. The data compiled for this report show no mechanistic link between incretin pathway modulation and this condition. The high TxGNN score (99.88%) reflects a pattern in the knowledge graph rather than demonstrated biology. Any suggested connection, such as an immune-modulating effect of DPP-4 inhibition, is speculative and untested.

Clinical Trial Evidence

Currently no related clinical trials registered.

Literature Evidence

Currently no related literature available.

UK Market Information

This is the only authorisation recorded. It is a single-ingredient product with valid status. The number is not in the standard PL format, so it should be checked against the MHRA register. Please also consult the BNF (antidiabetic drugs section) and the current SmPC for the licensed indications.

Safety Considerations

Please refer to the SmPC and BNF for safety information. Report suspected adverse reactions via the Yellow Card Scheme.

Conclusion and Next Steps

Decision: Hold

Rationale: The prediction rests only on the TxGNN model score. There are no trials, no publications and no plausible mechanistic rationale in the data compiled for this report. This is not enough to justify further investment in focal stiff limb syndrome.

To proceed, the following is needed:

• A mechanistic hypothesis linking DPP-4 inhibition (or other vildagliptin targets) to stiff-person-spectrum pathophysiology, with preclinical support
• Detailed mechanism of action data and full UK labelling information (warnings, contraindications, interactions)
• A targeted literature and trial registry search for stiff limb syndrome and related disorders
• Review of the other TxGNN predictions for vildagliptin. Type 1 diabetes mellitus (score 99.37%) has substantially more supporting evidence in the data compiled, including a completed Phase 2 randomised trial of rapamycin plus vildagliptin and several small mechanistic studies. It is rated L2 and is a more evidence-supported research direction than this top-ranked prediction.

  Disclaimer

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