ZOSTER-GUARD

Paper

ZOSTER-GUARD: Herpes Zoster Reactivation Risk Stratification Before JAK Inhibitor or Biologic Therapy in Rheumatic and Autoimmune Disease

Authors: Dr. Erick Zamora-Tehozol, DNAI, RheumaAI
ORCID: 0000-0002-7888-3961

Abstract

Herpes zoster (HZ) causes substantial morbidity in autoimmune disease, especially when Janus kinase inhibitors (JAKi), glucocorticoids, and combination immunosuppression are used. Clinicians face a practical problem before therapy initiation: how urgent is recombinant zoster vaccine review, and how intensive should early surveillance be if treatment cannot be delayed? We present ZOSTER-GUARD, a transparent 10-domain weighted clinical score that estimates near-term HZ reactivation risk in rheumatic and autoimmune disease. The model integrates diagnosis group, therapy class, steroid intensity, age, prior zoster, lymphopenia, additional immunosuppressants, diabetes, chronic kidney disease, and completed recombinant zoster vaccination. The implementation is executable as standalone Python using only the standard library and adds Monte Carlo uncertainty estimation to reflect measurement variability. Three demo scenarios show clinically coherent stratification: vaccinated low-intensity rheumatoid arthritis (LOW 5.8), rheumatoid arthritis starting a JAK inhibitor with prior zoster (HIGH 37.4), and severe systemic lupus erythematosus on high steroids plus JAK inhibitor (HIGH 45.6). This score is intended for vaccination-planning and monitoring support, not diagnosis of active infection, antiviral prophylaxis decisions, or replacement of guideline-based care. Limitations include expert-derived weights, absence of prospective validation, and incomplete capture of all host immune determinants.

Keywords: herpes zoster, JAK inhibitor, vaccination, rheumatology, autoimmune disease, risk stratification, Monte Carlo, DeSci

1. Clinical problem

Autoimmune inflammatory diseases already confer excess HZ risk versus the general population. That baseline vulnerability is amplified by immunomodulatory therapy, especially JAK inhibitors and glucocorticoids. Clinicians often must choose between delaying treatment for vaccination, vaccinating during disease control windows, or proceeding urgently with stronger patient education and follow-up.

The practical question is not merely whether HZ can occur; it is whether a given patient's profile makes pre-treatment zoster prevention urgent. Existing guidance strongly supports recombinant zoster vaccination in at-risk patients with rheumatic disease, but bedside estimation of relative urgency remains qualitative. ZOSTER-GUARD addresses that gap with a transparent, auditable score.

2. Methodology

2.1 Design principles

The score was designed around five clinical principles:

1. Therapy intensity matters, with JAK inhibitors weighted highest because multiple rheumatology sources report comparatively elevated HZ incidence.
2. Host susceptibility matters, including age, prior zoster, lymphopenia, diabetes, and chronic kidney disease.
3. Combination risk matters, because steroids and multiple immunosuppressants amplify infection vulnerability.
4. Vaccination is protective but not absolute, so completed recombinant zoster vaccination reduces but does not erase estimated risk.
5. Transparency is mandatory, so every domain contributes an explicit weighted value visible to the user.

2.2 Domains and weights

Each domain is scored on a clinically interpretable raw scale and multiplied by its weight. The final composite score is clamped to 0-100.

2.3 Uncertainty estimation

To avoid false precision, the implementation performs 5,000 Monte Carlo simulations with moderate perturbation of steroid dose, duration, age, lymphocyte count, and immunosuppressant burden. This generates an approximate 95% interval around the composite score.

3. Executable skill

3.1 Python code

#!/usr/bin/env python3
from pathlib import Path
print(Path('skills/zoster-guard/zoster_guard.py').read_text())

The full executable implementation is included in the submission content and locally at skills/zoster-guard/zoster_guard.py.

3.2 Demo output

=== Scenario 1 — Younger RA patient on methotrexate with vaccination complete ===
Composite score: 5.8/100
Risk category: LOW

=== Scenario 2 — RA starting JAK inhibitor with prior zoster and moderate lymphopenia ===
Composite score: 37.4/100
Risk category: HIGH

=== Scenario 3 — SLE flare on high steroids plus JAK inhibitor without vaccination ===
Composite score: 45.6/100
Risk category: HIGH

4. Interpretation

• LOW (<15): routine vaccine review remains appropriate, but urgent therapy delay for HZ risk alone is usually unnecessary.
• INTERMEDIATE (15-29.9): review vaccine status and early symptom counseling.
• HIGH (30-49.9): vaccination review should be strongly prioritized before escalation when feasible.
• VERY HIGH (≥50): urgent preventive planning and close early follow-up if treatment cannot wait.

5. Clinical limitations

1. The weights are evidence-informed but not regression-derived from a prospective cohort.
2. The score estimates relative urgency, not absolute incidence probability.
3. Vaccine effectiveness varies with timing, host factors, and immunosuppressive regimen.
4. The model does not include all immune determinants such as VZV-specific cellular immunity, frailty, or detailed biologic exposure history.
5. The score is not validated for pediatric patients, transplant populations, or oncology regimens.
6. The score must not replace guideline-based vaccination decisions or prompt treatment of suspected active zoster.

6. Why this solves a real problem

Rheumatology clinicians routinely face a narrow window between recognizing the need for JAK inhibitor or biologic therapy and deciding whether zoster vaccination or extra follow-up should occur first. A transparent score can make that conversation explicit, especially in settings where vaccination uptake is poor and risk communication is inconsistent.

References

1. Winthrop KL, Harigai M, Genovese MC, et al. Prevention and management of herpes zoster in patients with rheumatoid arthritis and psoriatic arthritis: a clinical review. Clin Exp Rheumatol. 2022;40(1):162-172. DOI: 10.55563/clinexprheumatol/cpu6r9
2. Nash P, Kerschbaumer A, Dörner T, et al. JAK inhibitors and infections risk: focus on herpes zoster. Ther Adv Musculoskelet Dis. 2020;12:1759720X20936059. DOI: 10.1177/1759720X20936059
3. Cito A, Venerito V, Stefanizzi P, et al. Turning the Tide against Herpes Zoster in Rheumatoid Arthritis Patients Treated with JAK Inhibitors. J Clin Med. 2024;13(15):4423. DOI: 10.3390/jcm13154423
4. Venerito V, Stefanizzi P, Cantarini L, et al. Immunogenicity and Safety of Adjuvanted Recombinant Zoster Vaccine in Rheumatoid Arthritis Patients on Anti-Cellular Biologic Agents or JAK Inhibitors: A Prospective Observational Study. Int J Mol Sci. 2023;24(8):6967. DOI: 10.3390/ijms24086967
5. Bass AR, Chakravarty E, Akl EA, et al. 2022 American College of Rheumatology Guideline for Vaccinations in Patients With Rheumatic and Musculoskeletal Diseases. Arthritis Care Res (Hoboken). 2023;75(3):449-464. DOI: 10.1002/acr.25045

Executable Python code

#!/usr/bin/env python3
"""
ZOSTER-GUARD: Herpes Zoster Reactivation Risk Stratification Before
JAK Inhibitor or Biologic Therapy in Rheumatic and Autoimmune Disease

Authors: Dr. Erick Zamora-Tehozol, DNAI, RheumaAI
ORCID: 0000-0002-7888-3961
License: MIT

Clinical purpose:
Estimate the near-term risk of clinically relevant herpes zoster (HZ)
reactivation when immunomodulatory therapy is being started or intensified in
rheumatic and autoimmune disease. The score supports vaccination planning,
monitoring intensity, and risk communication. It does NOT diagnose active HZ,
replace vaccination guidance, or supersede infectious-disease consultation.

Key references:
- Winthrop KL et al. Clin Exp Rheumatol. 2022. DOI: 10.55563/clinexprheumatol/cpu6r9
- Nash P et al. Ther Adv Musculoskelet Dis. 2020. DOI: 10.1177/1759720X20936059
- Cito A et al. J Clin Med. 2024. DOI: 10.3390/jcm13154423
- Venerito V et al. Int J Mol Sci. 2023. DOI: 10.3390/ijms24086967
- Bass AR et al. Arthritis Care Res. 2023. DOI: 10.1002/acr.25045
"""

from __future__ import annotations

import random
from dataclasses import dataclass, field
from typing import List


@dataclass
class ZosterPatient:
    diagnosis_group: str = "other_aiird"
    therapy_class: str = "csdmard"
    prednisone_mg_day: float = 0.0
    steroid_weeks: int = 0
    age: int = 50
    prior_herpes_zoster: bool = False
    lymphocytes_per_ul: int = 1500
    multiple_immunosuppressants: int = 0
    diabetes: bool = False
    chronic_kidney_disease: bool = False
    rcv_zoster_vaccinated: bool = False


@dataclass
class DomainScore:
    name: str
    score: float
    weight: float
    weighted: float
    detail: str


@dataclass
class ZosterResult:
    composite_score: float
    risk_category: str
    vaccination_recommendation: str
    monitoring_comment: str
    ci_lower: float
    ci_upper: float
    domains: List[dict]
    notes: List[str] = field(default_factory=list)


WEIGHTS = {
    "diagnosis": 0.10,
    "therapy": 0.22,
    "steroids": 0.14,
    "age": 0.10,
    "history": 0.12,
    "lymphopenia": 0.10,
    "combo": 0.08,
    "diabetes": 0.05,
    "ckd": 0.05,
    "vaccination": 0.04,
}


def score_diagnosis(group: str):
    mapping = {
        "ra": (30, "Rheumatoid arthritis carries a baseline HZ excess versus the general population"),
        "sle": (40, "SLE combines immune dysregulation with frequent steroid exposure"),
        "aav": (45, "AAV often requires intense remission-induction immunosuppression"),
        "psa": (22, "Psoriatic arthritis baseline HZ risk is elevated but lower than SLE/AAV"),
        "iim": (34, "Inflammatory myopathy often requires higher steroid exposure"),
        "ssc": (20, "Systemic sclerosis risk depends more on treatment intensity than diagnosis alone"),
        "other_aiird": (25, "Other autoimmune inflammatory rheumatic disease"),
    }
    return mapping.get(group, (25, f"Unmapped diagnosis group: {group}"))


def score_therapy(therapy_class: str):
    mapping = {
        "csdmard": (8, "Conventional synthetic DMARD only"),
        "tnfi": (24, "TNF inhibitor"),
        "other_biologic": (26, "Non-TNF biologic"),
        "rituximab": (35, "B-cell depletion may increase infection vulnerability"),
        "jak inhibitor": (68, "JAK inhibitors show the clearest modern HZ signal in rheumatology cohorts"),
    }
    return mapping.get(therapy_class.lower(), (20, f"Unmapped therapy class: {therapy_class}"))


def score_steroids(mg_day: float, weeks: int):
    if mg_day <= 0 or weeks <= 0:
        return 0, "No current glucocorticoid exposure"
    if mg_day < 7.5:
        return 10, f"Low-dose steroids ({mg_day:.1f} mg/day for {weeks} weeks)"
    if mg_day < 15:
        return 24, f"Moderate steroids ({mg_day:.1f} mg/day for {weeks} weeks)"
    if mg_day < 30:
        return 48, f"High steroids ({mg_day:.1f} mg/day for {weeks} weeks)"
    return 78, f"Very high steroids ({mg_day:.1f} mg/day for {weeks} weeks)"


def score_age(age: int):
    if age < 40:
        return 0, f"Age {age}"
    if age < 50:
        return 12, f"Age {age}"
    if age < 65:
        return 28, f"Age {age}"
    if age < 75:
        return 48, f"Age {age}"
    return 62, f"Age {age}"


def score_history(prior_herpes_zoster: bool):
    return (55, "Prior herpes zoster episode") if prior_herpes_zoster else (0, "No prior herpes zoster")


def score_lymphopenia(lymphocytes: int):
    if lymphocytes >= 1200:
        return 0, f"lymphocytes={lymphocytes}/uL"
    if lymphocytes >= 800:
        return 18, f"lymphocytes={lymphocytes}/uL"
    if lymphocytes >= 500:
        return 40, f"lymphocytes={lymphocytes}/uL"
    return 68, f"lymphocytes={lymphocytes}/uL"


def score_combo(count: int):
    if count <= 0:
        return 0, "No additional immunosuppressants"
    if count == 1:
        return 18, "One additional immunosuppressant"
    if count == 2:
        return 34, "Two additional immunosuppressants"
    return 52, f"{count} additional immunosuppressants"


def score_diabetes(diabetes: bool):
    return (24, "Diabetes present") if diabetes else (0, "No diabetes")


def score_ckd(ckd: bool):
    return (28, "Chronic kidney disease present") if ckd else (0, "No chronic kidney disease")


def score_vaccination(vaccinated: bool):
    return (-40, "Completed recombinant zoster vaccine series") if vaccinated else (0, "No documented recombinant zoster vaccination")


def compute_zoster_risk(patient: ZosterPatient, n_simulations: int = 5000, seed: int = 42) -> ZosterResult:
    items = [
        ("diagnosis", score_diagnosis(patient.diagnosis_group)),
        ("therapy", score_therapy(patient.therapy_class)),
        ("steroids", score_steroids(patient.prednisone_mg_day, patient.steroid_weeks)),
        ("age", score_age(patient.age)),
        ("history", score_history(patient.prior_herpes_zoster)),
        ("lymphopenia", score_lymphopenia(patient.lymphocytes_per_ul)),
        ("combo", score_combo(patient.multiple_immunosuppressants)),
        ("diabetes", score_diabetes(patient.diabetes)),
        ("ckd", score_ckd(patient.chronic_kidney_disease)),
        ("vaccination", score_vaccination(patient.rcv_zoster_vaccinated)),
    ]

    domains: List[DomainScore] = []
    composite = 0.0
    for name, (raw, detail) in items:
        weight = WEIGHTS[name]
        weighted = raw * weight
        composite += weighted
        domains.append(DomainScore(name, round(raw, 1), weight, round(weighted, 1), detail))

    composite = round(max(0.0, min(composite, 100.0)), 1)

    rng = random.Random(seed)
    sims: List[float] = []
    for _ in range(n_simulations):
        noisy = ZosterPatient(
            diagnosis_group=patient.diagnosis_group,
            therapy_class=patient.therapy_class,
            prednisone_mg_day=max(0.0, patient.prednisone_mg_day * (1 + rng.gauss(0, 0.12))),
            steroid_weeks=max(0, int(round(patient.steroid_weeks * (1 + rng.gauss(0, 0.10))))),
            age=max(18, int(round(patient.age + rng.gauss(0, 1.5)))),
            prior_herpes_zoster=patient.prior_herpes_zoster,
            lymphocytes_per_ul=max(50, int(round(patient.lymphocytes_per_ul * (1 + rng.gauss(0, 0.15))))),
            multiple_immunosuppressants=max(0, patient.multiple_immunosuppressants + int(round(rng.gauss(0, 0.25)))),
            diabetes=patient.diabetes,
            chronic_kidney_disease=patient.chronic_kidney_disease,
            rcv_zoster_vaccinated=patient.rcv_zoster_vaccinated,
        )
        total = 0.0
        noisy_items = [
            ("diagnosis", score_diagnosis(noisy.diagnosis_group)),
            ("therapy", score_therapy(noisy.therapy_class)),
            ("steroids", score_steroids(noisy.prednisone_mg_day, noisy.steroid_weeks)),
            ("age", score_age(noisy.age)),
            ("history", score_history(noisy.prior_herpes_zoster)),
            ("lymphopenia", score_lymphopenia(noisy.lymphocytes_per_ul)),
            ("combo", score_combo(noisy.multiple_immunosuppressants)),
            ("diabetes", score_diabetes(noisy.diabetes)),
            ("ckd", score_ckd(noisy.chronic_kidney_disease)),
            ("vaccination", score_vaccination(noisy.rcv_zoster_vaccinated)),
        ]
        for name, (score, _) in noisy_items:
            total += score * WEIGHTS[name]
        sims.append(max(0.0, min(total, 100.0)))

    sims.sort()
    ci_lower = round(sims[int(0.025 * n_simulations)], 1)
    ci_upper = round(sims[int(0.975 * n_simulations)], 1)

    if composite < 15:
        category = "LOW"
        vaccination_recommendation = "Routine zoster vaccination counseling remains appropriate, but therapy delay for HZ risk alone is usually unnecessary."
        monitoring_comment = "Usual clinical surveillance; reassess if therapy intensifies or steroids increase."
    elif composite < 30:
        category = "INTERMEDIATE"
        vaccination_recommendation = "Strongly review recombinant zoster vaccine status before escalation, especially if age ≥50 or prior zoster exists."
        monitoring_comment = "Counsel early rash recognition and reassess risk after 1-3 months of therapy."
    elif composite < 50:
        category = "HIGH"
        vaccination_recommendation = "Complete recombinant zoster vaccination if feasible before JAK inhibitor/biologic initiation or intensification."
        monitoring_comment = "Lower threshold for prompt antiviral treatment if dermatomal pain or vesicular rash occurs."
    else:
        category = "VERY HIGH"
        vaccination_recommendation = "Vaccination review is urgent. If therapy cannot wait, explicitly document HZ risk, educate the patient, and plan close follow-up."
        monitoring_comment = "Close early follow-up is favored, especially during the first months after JAK inhibitor start or steroid escalation."

    notes = [
        "This score is intended for vaccination and monitoring planning, not antiviral prophylaxis decisions.",
        "Breakthrough zoster can still occur after vaccination; the vaccine domain lowers but does not eliminate risk.",
    ]
    if patient.therapy_class.lower() == "jak inhibitor":
        notes.append("JAK inhibitor exposure is weighted heavily because multiple rheumatology datasets report comparatively higher HZ incidence.")
    if patient.rcv_zoster_vaccinated:
        notes.append("Vaccination effect is modeled conservatively because effectiveness varies by host factors and timing.")

    return ZosterResult(
        composite_score=composite,
        risk_category=category,
        vaccination_recommendation=vaccination_recommendation,
        monitoring_comment=monitoring_comment,
        ci_lower=ci_lower,
        ci_upper=ci_upper,
        domains=[d.__dict__ for d in domains],
        notes=notes,
    )


def print_case(label: str, patient: ZosterPatient):
    result = compute_zoster_risk(patient)
    print(f"\n=== {label} ===")
    print(f"Composite score: {result.composite_score}/100")
    print(f"Risk category: {result.risk_category}")
    print(f"95% CI: [{result.ci_lower}, {result.ci_upper}]")
    print(f"Vaccination recommendation: {result.vaccination_recommendation}")
    print(f"Monitoring: {result.monitoring_comment}")
    if result.notes:
        print("Notes:")
        for note in result.notes:
            print(f"- {note}")
    print("Top domains:")
    for domain in sorted(result.domains, key=lambda d: d['weighted'], reverse=True)[:5]:
        print(f"- {domain['name']}: raw {domain['score']} × w {domain['weight']} = {domain['weighted']} ({domain['detail']})")


if __name__ == "__main__":
    print_case(
        "Scenario 1 — Younger RA patient on methotrexate with vaccination complete",
        ZosterPatient(
            diagnosis_group="ra",
            therapy_class="csdmard",
            prednisone_mg_day=5,
            steroid_weeks=2,
            age=42,
            prior_herpes_zoster=False,
            lymphocytes_per_ul=1550,
            multiple_immunosuppressants=0,
            diabetes=False,
            chronic_kidney_disease=False,
            rcv_zoster_vaccinated=True,
        ),
    )
    print_case(
        "Scenario 2 — RA starting JAK inhibitor with prior zoster and moderate lymphopenia",
        ZosterPatient(
            diagnosis_group="ra",
            therapy_class="jak inhibitor",
            prednisone_mg_day=10,
            steroid_weeks=6,
            age=61,
            prior_herpes_zoster=True,
            lymphocytes_per_ul=760,
            multiple_immunosuppressants=1,
            diabetes=True,
            chronic_kidney_disease=False,
            rcv_zoster_vaccinated=False,
        ),
    )
    print_case(
        "Scenario 3 — SLE flare on high steroids plus JAK inhibitor without vaccination",
        ZosterPatient(
            diagnosis_group="sle",
            therapy_class="jak inhibitor",
            prednisone_mg_day=30,
            steroid_weeks=8,
            age=68,
            prior_herpes_zoster=False,
            lymphocytes_per_ul=420,
            multiple_immunosuppressants=2,
            diabetes=False,
            chronic_kidney_disease=True,
            rcv_zoster_vaccinated=False,
        ),
    )

Demo output

=== Scenario 1 — Younger RA patient on methotrexate with vaccination complete ===
Composite score: 5.8/100
Risk category: LOW

=== Scenario 2 — RA starting JAK inhibitor with prior zoster and moderate lymphopenia ===
Composite score: 37.4/100
Risk category: HIGH

=== Scenario 3 — SLE flare on high steroids plus JAK inhibitor without vaccination ===
Composite score: 45.6/100
Risk category: HIGH