Introduction

Thrombotic antiphospholipid syndrome, or APS, has a high risk of recurrent arterial and venous thrombosis. Some events occur while patients appear to be receiving adequate anticoagulant or antiplatelet treatment. Clinicians often describe these cases as drug resistance, but this term combines several different problems: nonadherence, drug access, complications, inadequate dosing, drug interactions, altered absorption, unreliable laboratory monitoring, high residual platelet reactivity, and true pharmacologic nonresponse.

Pharmacogenomic variation may explain part of this residual risk. Variants affecting drug activation, metabolism, transport, and molecular targets can alter the response to vitamin K antagonists, clopidogrel, cilostazol, aspirin, and possibly direct oral anticoagulants. Most available pharmacogenomic models rely heavily on European populations. Their performance in Mexican Mestizo and Indigenous-enriched populations is uncertain.

STORM-APS would apply the STORM framework to antithrombotic therapy. It would combine ancestry-adjusted pharmacogenomics, measured drug exposure, functional coagulation or platelet assays, APS phenotype, and clinical

Primary hypothesis

As we already know, among mexican patients with thrombotic or haemorragic antiphospholipid syndrome, pharmacogenetic variants that affect drug activation, metabolism, transport, or target sensitivity are associated with inadequate drug exposure or high residual coagulation or platelet activity, besides drug and patient care access. These variants increase the risk of objectively confirmed recurrent thrombosis during treatment with direct oral anticoagulants, vitamin K antagonists, aspirin, clopidogrel, or cilostazol, independent of adherence, prescribed dose, renal function, drug interactions, antiphospholipid antibody profile, and traditional cardiovascular risk factors. There are a lot of existing publications in cardiology and neurology addressing this situation but we don’t have anything about this data In rheumatology.

secondary hypotheses

1. Variants in ABCB1, ABCG2, CYP3A4/5, and CES1 contribute to altered DOAC exposure and may explain a subset of apparent DOAC treatment failures, being independently associated with autoimmunity.
2. Variants in CYP2C9, VKORC1, and CYP4F2 are associated with unstable anticoagulation, reduced time in therapeutic range, or unusually high vitamin K antagonist dose requirements, something that won’t be related to diet or patient adherence.
3. CYP2C19 loss-of-function (LoF) alleles Are associated with high on-treatment platelet reactivity during clopidogrel therapy. CYP2C19 and CYP3A5 variants may also modify cilostazol exposure, although this association should initially be considered exploratory. Some Aspirin resistant theories are based on this.
4. Variants affecting the thromboxane and platelet activation pathways may contribute to high on-aspirin platelet reactivity, but aspirin nonresponse will require functional confirmation because no single genetic marker has sufficient predictive value. Aspirin also needs more time to onset the therapeutic effect and is the most complicated drug for halting for surgery or bleeding.
5. The frequency and effect of relevant variants differ according to mexican regional ancestry. A population-adjusted model will therefore predict treatment response more accurately than models derived mainly from European populations.
6. A model combining pharmacogenomics, measured drug exposure, platelet or coagulation function, adherence, and SAAF phenotype will predict recurrent thrombosis better than clinical variables alone

A STORM-based pharmacogenomic model that integrates ancestry-adjusted allele frequencies with drug-specific functional measurements will identify Mexican patients with SAAF who remain at increased thrombotic risk despite apparently adequate antithrombotic treatment.

Main objective

To develop and prospectively validate an ancestry-adjusted model that identifies pharmacogenomically mediated nonresponse to:

• direct oral anticoagulants
• vitamin K antagonists
• aspirin
• clopidogrel
• cilostazol

Methodology

Phase 1: Evidence synthesis and panel development

Conduct a structured review of PubMed, Embase, LILACS, PharmGKB, CPIC, and regulatory pharmacogenomic sources.

Candidate genes would include:

• DOACs:’ABCB1`, `ABCG2`, `CYP3A4`, `CYP3A5`, and `CES1`.
• Vitamin K antagonists: `CYP2C9`, `VKORC1`, and `CYP4F2`.
• Clopidogrel: `CYP2C19` and `ABCB1`.
• Cilostazol`CYP2C19` and `CYP3A5`.
• Aspirin, exploratory: `PTGS1`, `PEAR1`, `TBXAS1`, and selected platelet-receptor genes.

Each association would be classified as robust, suggestive, exploratory, or lacking population-specific evidence. Holm-Bonferroni and Benjamini-Hochberg corrections would control multiple testing.

Phase 2: Pilot cohort

Design: Prospective, multicenter observational study.

Target: 100 adults with clinician-confirmed thrombotic APS.

Participants would remain on the treatment selected by their physicians. The protocol would not assign high-risk patients to DOACs or change treatment based on exploratory results.

Study groups would reflect actual treatment:

• DOAC
• vitamin K antagonist;
• aspirin
• clopidogrel
• cilostazol
• anticoagulant plus antiplatelet combinations.

Because clopidogrel and cilostazol use may be uncommon in APS, recruitment would require several rheumatology, hematology, neurology, and vascular-medicine centers.

- direct oral anticoagulants;

- vitamin K antagonists;

- aspirin;

- clopidogrel;

- cilostazol.

- direct oral anticoagulants;

- vitamin K antagonists;

- aspirin;

- clopidogrel;

- cilostazol.

Baseline information

1. arterial or venous index event
2. single, double, or triple aPL positivity
3. lupus anticoagulant
4. associated systemic lupus erythematosus
5. renal and hepatic function
6. platelet count and hematocrit
7. diabetes, smoking, hypertension, and dyslipidemia
8. prescribed dose and treatment duration
9. adherence
10. interacting medications
11. regional ancestry and molecular ancestry markers when available.

Pharmacogenomic testing

Use a targeted MassARRAY, TaqMan, or equivalent panel covering approximately 20 to 30 variants. A subset of samples should undergo duplicate testing for quality control.

The model would estimate allele frequencies using Beta distributions and run Monte Carlo simulations to represent uncertainty across Mexican ancestry groups.

Outcomes

The primary pharmacodynamic outcome would be reproducible treatment nonresponse on two separate measurements after confirming adherence and correct sampling time.

The primary clinical outcome would be objectively confirmed recurrent arterial, venous, or microvascular thrombosis during 24 months.

Safety outcomes would include:

• ISTH major bleeding
• clinically relevant non-major bleeding
• treatment discontinuation because of adverse effects.

Statistical analysis

The analysis would:

1. test each drug class separately
2. avoid combining all treatments into one resistance category
3. examine genotype and ancestry interactions
4. adjust for APS phenotype, adherence, renal function, dose, interactions, and cardiovascular risk
5. use nested cross-validation to limit overfitting;
6. compare clinical-only, pharmacogenomic-only, and integrated models
7. report discrimination, calibration, and decision-curve analysis.

The pilot would estimate prevalence, assay feasibility, variance, and effect sizes. Those results would determine the sample size for definitive validation. A fixed efficacy claim should not be made from the first 100 participants.

Estimated cost

These are planning estimates in USD, not vendor quotations.

Phase 1 and pilot cohort, 100 participants

1. Protocol, ethics, database, and governance: $8, 000
2. Panel design and targeted genotyping: $15, 000
3. Platelet-function and anticoagulant assays: $15,000
4. Sample collection, processing, and storage: $5,000
5. Research coordination for 12 months: $35,000
6. Biostatistics and model development: $7,000
7. Follow-up and blinded event adjudication: $5,000
8. Contingency, 15%: $9,600

**Estimated pilot total: $99,600**

Expanded validation cohort, approximately 300 participants

Estimated total: $187,175 including two years of coordination, repeated functional testing, event adjudication, biostatistics, and a 15% contingency.

Mexican institutional quotations could move these estimates by 30% to 50%. Existing laboratory equipment, donated assays, and institutional personnel could lower the cash requirement.

Return on investment

Scientific ROI

The pilot would produce:

1. the first integrated APS pharmacogenomic and functional-response dataset in a Mexican cohort
2. one methods or evidence-synthesis paper
3. one prospective cohort paper
4. an ancestry-adjusted STORM-APS model
5. preliminary data for national and international grant applications
6. foundation for a clinically validated panel.

Commercial scenario

This is a planning model, not a revenue forecast.

Assume:

• test price: $250
• variable laboratory and reporting cost: $100
• contribution per test: $150

Under those assumptions:

1. the $99,600 pilo would break even after approximately 491 tests;
2. the $187, 175 validation program would break even after approximately 1, 108 tests.

An institutional model could charge $25,000 per center per year for testing support, software, quality control, and reporting. Four centers would generate $100,000 in annual revenue. Assuming 30% delivery and support costs, annual contribution would be approximately $70,000.

That implies an estimated payback period of:

1. 1.05 yearsfor the pilot investment;
2. 2.37 years for the expanded validation program.

Clinical ROI

Clinical savings cannot yet be claimed. The study must first measure:

• recurrent thromboses avoided
• bleeding events
• hospitalizations
• unnecessary treatment changes
• cost of repeated INR and platelet testing
• cost per correctly reclassified patient
• incremental cost per thrombosis avoided
• quality-adjusted life-years if a health-economic analysis is planned.

The initial product should remain research-only. Genotype-guided treatment recommendations should not enter clinical care until prospective validation shows that they improve decisions and outcomes.