Does vitiligo come back after stopping treatment?
Maintenance and recurrence risk assessment

At-a-glance — three numbers you'll want

Why does vitiligo come back after stopping treatment?

Vitiligo is fundamentally an autoimmune attack on one's own melanocytes. Even when no new lesions are visible, tissue-resident memory T cells (TRMs) can lie dormant in previously affected skin.

When treatment intensity is reduced (stopping topical agents or phototherapy), previously suppressed TRMs reactivate, releasing IFN-γ and CXCL9/10/11 chemokines that recruit cytotoxic CD8+ T cells to attack melanocytes again. Clinically this presents as gradual lesion expansion or new depigmentation in previously repigmented areas.

A key challenge is that relapse is usually silent — no pain, itch, or redness. It's often only detected at the next clinic visit via Wood's lamp or serial photography. This is why maintenance therapy is increasingly emphasised: low-intensity continuous suppression of TRMs is preferable to retreating a full-blown relapse.

Who's at higher risk? The three predictors

(1) Acral involvement — aHR 1.56

The strongest phenotypic predictor. Acral sites — fingers, palms, toes, soles — are 'glabrous' regions with little hair-follicle melanocyte reservoir. Patients with acral involvement had a 1-year cumulative recurrence of 33.3% vs 22.9% in those without (a 10-percentage-point gap).

Mechanism: hair-follicle melanocytes serve as a reservoir for repigmentation. Hair-bearing regions (scalp, trunk, proximal limbs) can recover via melanocyte migration from the outer root sheath. Acral sites are glabrous and lack this reservoir, so depigmentation tends to be more persistent ('colour scarring').

(2) Thyroid disease — aHR 1.42

In the CGMH cohort, 22.1% had concurrent thyroid disease — vs ~1% in the general population. The two most common are autoimmune thyroiditis (Hashimoto) and Graves disease, both autoimmune attacks on the thyroid gland.

Why do these two diseases cluster? Vitiligo and autoimmune thyroid disease share a common dysregulated immune network — IFN-γ signalling, CXCL chemokine pathways, and shared genetic susceptibility loci (e.g., PTPN22, HLA region). Patients with thyroid disease tend to have a more 'autoreactive' baseline, making vitiligo more relapse-prone.

✦ Clinical tip: Order TSH, free T4, and anti-TPO antibodies at vitiligo presentation. Even if initially normal, repeat every 1–2 years — thyroid disease may develop years after vitiligo onset.

(3) Other autoimmune comorbidities — aHR 1.73

This category covers non-thyroid autoimmune conditions: alopecia areata (AA), rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), type 1 diabetes, pernicious anaemia, Sjögren's syndrome, etc. Prevalence in the cohort was 17.8%, consistent with prior literature.

Interestingly, the adjusted HR (1.73) was even higher than thyroid disease (1.42), possibly reflecting that systemic autoimmune dysregulation has a stronger impact on vitiligo persistence than organ-restricted autoimmunity. Clinically, vitiligo patients with any prior autoimmune diagnosis should be treated as harder-to-stabilise cases.

High-risk subgroup: acral + thyroid = 2.5× hazard

This is the study's most clinically actionable finding. The team stratified patients into four groups (neither / acral only / thyroid only / both) and reported:

✦ Quick self-assessment questions:

• (1) Do I have vitiligo patches on my fingers, palms, toes, or soles?
• (2) Do I have Hashimoto thyroiditis, Graves disease, or any thyroid abnormality? Are my TSH / anti-TPO labs abnormal?
• (3) Do I have any other autoimmune disease (alopecia areata, RA, SLE, type 1 diabetes…)?

Any 'yes' → you fall into the 'don't stop cold-turkey, extend maintenance' category. Both ① + ② → highest-risk subgroup; discuss a long-term maintenance plan with your dermatologist.

If I'm high-risk, what should I do?

The goal of maintenance is not 'cure' — it's recurrence suppression. In the cohort, 25.9% of relapsers had been completely off treatment for >6 months, vs only 6.3% of non-relapsers. Translation: those who maintained treatment had clearly lower relapse rates.

Common maintenance options (low → high intensity):

• Intermittent topical： Apply calcineurin inhibitor (e.g., tacrolimus 0.1%) or low-potency topical corticosteroid twice weekly to previously affected areas.
• Reduced-interval phototherapy： NB-UVB tapered from 2–3×/week (induction) down to once every 1–2 weeks (maintenance).
• Low-dose systemic immunomodulation： A minority with highly active, widespread vitiligo, or with multiple autoimmune comorbidities, may receive ongoing low-dose oral immunomodulators under specialist supervision.

The most common patient mistake is 'stop completely once skin looks fine' — which is exactly the 'off treatment >6 months' bucket that contained the largest fraction of relapsers. Plan a step-down with your dermatologist instead of a cold-turkey stop.

FAQ (patient-facing)

Q1: I have vitiligo on hands and feet — do I need lifelong treatment?

Not literally lifelong — but expect 'long-term, low-intensity, intermittent.' Acral sites have limited melanocyte reservoir, slow repigmentation, and higher relapse risk, so complete withdrawal is harder. Maintenance can be as light as twice-weekly topical — minimal lifestyle impact.

Q2: I have hypothyroidism / Hashimoto — does this make vitiligo harder to treat?

Yes. Thyroid-positive patients had 1-year recurrence of 36.9% vs 24.8%, adjusted hazard ~1.42×. This doesn't mean treatment is futile — it means maintenance must be more careful. Controlling the thyroid itself (compliant medication, regular TSH monitoring) also helps overall immune stability.

Q3: When am I most at risk of relapse after stopping?

The study followed patients ≥1 year after achieving ≥6 months stability. One-year cumulative recurrence was ~27–28% for low-risk, 33–40% for high-risk patients. So 'the first year after stopping' is the key observation window — especially the first 6 months. Any subtle expansion under Wood's lamp warrants prompt re-evaluation.

Q4: Are anti-TPO and ANA tests useful?

Anti-TPO and ANA positivity were significant in univariable analysis but attenuated to non-significance after multivariable adjustment, because they overlap with clinically diagnosed autoimmunity (r = 0.25–0.34). Still worth checking — treat them as useful clues rather than direct predictors.

Q5: Does family history affect relapse rate?

In this cohort, family history (vitiligo / thyroid / other autoimmune) was non-significant in both univariable and multivariable models, but all showed a mildly elevated trend (HR ~1.1–1.2). This may reflect already-high baseline genetic susceptibility in this Taiwanese cohort. Still worth disclosing — it informs broader immune-status assessment.

Q6: Can vitiligo self-resolve? Can I choose 'wait and see'?

Possible, but verify two things first: (1) Are you in a rapidly progressive phase (visible expansion over the past 6–12 months)? (2) Do you have any of the high-risk factors? If both answers are 'no,' observation + sun protection is reasonable. If either is 'yes,' early treatment is preferred — 'earlier intervention preserves more melanocytes' is consistent across the vitiligo literature (including this study's introduction).

Background and clinical question

Even after induction-phase stabilisation and repigmentation in non-segmental vitiligo (NSV), the post-discontinuation relapse rate remains a long-standing clinical challenge. BAD 2021, TDA 2024, and the 2023 global recommendations all emphasise maintenance therapy, but evidence on who needs it most, what regimen, and for how long has been limited.

Aim: identify phenotypic + autoimmune predictors of NSV recurrence/relapse after treatment stabilisation, in a large single-centre retrospective cohort, and build a clinically usable risk-stratification framework. Primary outcome = time from first documented stability (≥6 months without new/expanding lesions) to first recurrence/relapse (new or expanding lesion under Wood's lamp).

Study design and cohort

Source: CGRD, 2015/01/01–2024/10/31. Inclusion: (1) NSV diagnosed by a board-certified dermatologist (ICD-10-CM L80); (2) guideline-based induction (NB-UVB ± topical corticosteroid ± calcineurin inhibitor ± short course (≤3 mo) systemic corticosteroid); (3) ≥6 months documented stability; (4) ≥1 year post-stability follow-up.

Variables and statistical methods

Covariates: age, sex, family history (vitiligo / thyroid / autoimmune), thyroid disease, other autoimmune disease, acral involvement, facial involvement, baseline BSA (<5%, 5–10%, >10%), anti-TPO (>5.6 IU mL⁻¹), ANA (≥1:80). Outcome = first recurrence/relapse, defined as expansion of previously stable lesions OR new lesions, confirmed with Wood's lamp + serial photography.

Statistics: Cox proportional hazards regression (Efron method). Univariable → prespecified multivariable model (all covariates retained a priori, no data-driven selection). Multiplicative interaction model for acral × thyroid + 4-cell joint-exposure analysis. PH assumption verified by standard diagnostics. Kaplan–Meier 12-month RFS with Greenwood CIs, log-rank test for group comparisons.

Primary results — Baseline + 1-year cumulative recurrence

Cohort baseline: N = 809, mean age 44.0 ± 15.8, female 57.1%. Recurrence/relapse 266 (32.9%).

✦ Note: anti-TPO/ANA were significant in KM and univariable but attenuated under multivariable adjustment (below); BSA, family history, and facial involvement were non-significant at every level; age and sex were non-significant throughout.

Multivariable Cox — three independent predictors

Acral × thyroid — joint-exposure analysis

Authors performed two interaction analyses:

• Multiplicative interaction model： Acral × thyroid HRint 1.68 (95% CI 0.98–2.93, p = 0.061) — did not reach multiplicative significance, but borderline.
• 4-cell joint-exposure model： Reference = neither; acral-only aHR 1.35 (0.99–1.83, p = 0.051); thyroid-only aHR 1.07 (0.69–1.62, p = 0.755); both aHR 2.44 (1.64–3.59, p < 0.001) — strong additive effect.

Statistically interesting: each factor alone fails to reach significance after adjustment, but joint exposure does — suggesting that co-occurrence flags a deeper systemic immune dysregulation than either alone. Clinically, the 'acral + thyroid' group should be stratified as its own category rather than treating each factor separately.

Sub-threshold predictors — anti-TPO, ANA, family history, BSA

Anti-TPO+ / ANA+: univariable HR 1.43 / 1.53; attenuated under multivariable adjustment because they correlate with clinically diagnosed autoimmune / thyroid disease (r = 0.25–0.34) and act as surrogate markers. Still warrant routine checking — useful for follow-up immune-status assessment.

Family history: vitiligo / thyroid / autoimmune family hx all NS (p = 0.176 / 0.787 / 0.203), but point estimates all > 1.0, suggesting underlying genetic susceptibility may still play a role — but this cohort's already-high baseline susceptibility limits stratification.

BSA: 3-level (<5% / 5–10% / >10%) log-rank p = 0.434, NS. The >10% group had 33.3% vs 26.5% in <5% — numerically higher but statistically underpowered (n = 69 vs 593). Clinically, widespread vitiligo is still considered harder to stabilise.

Facial involvement: 27.9% vs 25.8%, p = 0.677. Mechanism: facial sites have a rich melanocyte reservoir (dense hair follicles + thicker epidermis), so prior facial involvement does not increase relapse risk — a striking contrast to acral sites.

Maintenance treatment patterns — distribution of 5 modalities

The authors categorised treatment exposure at the time of event (or during follow-up) into 5 groups: (1) off treatment (>6 months no Tx); (2) topical only (≥6 months); (3) phototherapy (NB-UVB ± topical ≥6 months); (4) systemic tapering (within past 6 months); (5) systemic maintenance (stable low-dose systemic, no active taper).

✦ Important caveat: this is descriptive, not a prespecified outcome; treatment is time-varying and clinician-determined, with risks of immortal-time bias and confounding by indication. We cannot conclude causally that 'systemic maintenance reduced relapse 6.9-fold.' But two patterns are clinically notable:

• Off treatment >6 months: 25.9% of relapsers vs 6.3% of non-relapsers — even with confounding, the ~4× difference suggests 'complete discontinuation' is over-represented in relapsers.
• Systemic tapering: 37.6% of relapsers — suggesting rapid systemic taper is a high-risk transition window; BAD 2021 recommends slow taper to maintenance dose rather than abrupt cessation.

Risk-stratified maintenance intensity — clinical translation

The 3 independent predictors map to a 3-tier risk-stratification framework, each tier corresponding to a maintenance-intensity recommendation:

Limitations

• Retrospective design： Selection / documentation / misclassification biases. While CGRD is a large multi-centre database, missing data and inter-physician variability can affect outcome ascertainment.
• Cannot distinguish recurrence vs relapse： The authors combined both as a unified outcome. Strictly: recurrence = depigmentation in previously stable areas; relapse = new lesions. Their underlying immunology may differ (TRM reactivation vs systemic immune flare). Future prospective studies need to separate them.
• Treatment not in regression model： Maintenance modalities were descriptive only, not included as outcome predictors. Confounding by indication is severe — milder cases discontinue more readily AND maintain stability more easily. Hence the maintenance-type proportions cannot be directly interpreted as efficacy.
• Inconsistent documentation of autoimmune diagnoses： 'Other autoimmune' is a composite — not subclassified into AA / RA / SLE / T1DM, etc. Disease-specific signals may be diluted.
• Single-centre Taiwanese cohort： Generalisability to other populations (e.g., predominantly Fitzpatrick I-II Western cohorts) is uncertain. Clinical utility for Taiwan is high.

Context with prior literature

• Xu 2023 J Cosmet Dermatol single-centre real-life study of 'cured' vitiligo patients reported ~34.2% recurrence — closely matching this study's 32.9%, providing mutual corroboration.
• Gill 2016 J Am Acad Dermatol US cross-sectional study reported autoimmune comorbidity prevalence ~20% in vitiligo — comparable to this cohort's 17.8%.
• Chivu 2022 J Pers Med systematic review highlighting the bidirectional vitiligo / autoimmune thyroid link — this study's 22.1% prevalence aligns with their meta-estimate.
• Ezzedine 2025 J Clin Invest provides mechanistic context (IFN-γ / CXCL9-10-11 / CD8+ TRM) for maintenance therapy necessity — this study's phenotypic findings + Ezzedine's mechanistic framework together form a 'wet evidence + bench evidence' alignment.
• Esmat 2012 JEADV examined why acral lesions resist photochemotherapy, proposing the melanocyte-reservoir hypothesis. This study's acral signal is the phenotypic correlate.
• BAD 2021 + TDA 2024 Both guidelines recommend maintenance but lack explicit criteria for who needs it most — this study provides cohort-level data to refine those criteria.

Clinical take-aways

1. Even with ≥6 months stability, ~33% of NSV patients recur within 1–2 years. Patients should be clearly counselled: 'repigmentation ≠ permanent resolution.'
2. At first visit, proactively screen thyroid (TSH, free T4, anti-TPO) and ask about other autoimmune history. Reassess every 1–2 years if initially normal.
3. Acral involvement should be treated as a dual signal: limited melanocyte reservoir + high relapse risk. Weight it more heavily when planning maintenance.
4. The 'acral + thyroid/autoimmune' subgroup is highest-risk (aHR 2.44). Recommendations: (1) do not stop completely, (2) maintenance ≥12 months, (3) co-care with endocrinology / rheumatology, (4) follow-up every 3 months.
5. Systemic taper is a high-risk transition window — follow BAD 2021 / TDA 2024 guidance for slow taper to maintenance dose, avoid abrupt cessation; bridge with topical / phototherapy during taper.
6. Anti-TPO / ANA, though not independent predictors, are useful baseline immune markers — routinely test; positive results warrant closer follow-up and vigilance for future thyroid abnormalities.

References

1. Primary source
2. Shih HY, Lin HW, Tsai TF, Lai YJ, Ng CY. Recurrence and relapse in nonsegmental vitiligo: phenotypic and autoimmune predictors from a large retrospective cohort. Br J Dermatol. 2026;194(5):854-861. [Source]
3. Vitiligo guidelines and consensus
4. Eleftheriadou V, Atkar R, Batchelor J, et al. British Association of Dermatologists guidelines for the management of people with vitiligo 2021. Br J Dermatol. 2022;186(1):18-29. [Source]
5. Taiwan Dermatological Association.Vitiligo Clinical Treatment Consensus. 2024.
6. Ring J. Global recommendations for vitiligo management. J Eur Acad Dermatol Venereol. 2023;37(11):2157-2158.
7. Passeron T. Medical and maintenance treatments for vitiligo. Dermatol Clin. 2017;35(2):163-170. PMID: 28317527.
8. Goulden V, Ling TC, Babakinejad P, et al. British Association of Dermatologists and British Photodermatology Group guidelines for narrowband ultraviolet B phototherapy 2022. Br J Dermatol. 2022;187(3):295-308. [Source]
9. Mechanism
10. Ezzedine K, Tannous R, Pearson TF, Harris JE. Recent clinical and mechanistic insights into vitiligo offer new treatment options for cell-specific autoimmunity. J Clin Invest. 2025;135:e185785.
11. Ivashkiv LB. IFN-γ: signalling, epigenetics and roles in immunity, metabolism, disease and cancer immunotherapy. Nat Rev Immunol. 2018;18(9):545-558.
12. Esmat SM, El-Tawdy AM, Hafez GA, et al. Acral lesions of vitiligo: why are they resistant to photochemotherapy? J Eur Acad Dermatol Venereol. 2012;26(9):1097-1104.
13. Mecinska-Jundzill K, Tadrowski T, Jundzill A, et al. Evaluation of polymorphisms and expression of PTPN22, NLRP1 and TYR genes in vitiligo patients. Postepy Dermatol Alergol. 2023;40(2):225-233.
14. Comorbidity epidemiology
15. Gill L, Zarbo A, Isedeh P, et al. Comorbid autoimmune diseases in patients with vitiligo: a cross-sectional study. J Am Acad Dermatol. 2016;74(2):295-302.
16. van Geel N, Speeckaert M, Brochez L, et al. Clinical profile of generalized vitiligo patients with associated autoimmune/autoinflammatory diseases. J Eur Acad Dermatol Venereol. 2014;28(6):741-746.
17. Chivu AM, Balasescu E, Pandia LD, et al. Vitiligo-thyroid disease association: when, in whom, and why should it be suspected? A systematic review. J Pers Med. 2022;12(12):2048.
18. Ma Z, Cai M, Yang K, et al. Predicting the risk of autoimmune thyroid disease in patients with vitiligo: development and assessment of a new predictive nomogram. Front Endocrinol (Lausanne). 2023;14:1109925.
19. Colucci R, Dragoni F, Moretti S. Oxidative stress and immune system in vitiligo and thyroid diseases. Oxid Med Cell Longev. 2015;2015:631927.
20. Sandru F, Carsote M, Albu SE, et al. Vitiligo and chronic autoimmune thyroiditis. J Med Life. 2021;14(2):127-130.
21. Recurrence / real-world data
22. Xu W, Qiu Z, Li C, et al. Recurrence and risk factors in cured patients with vitiligo: A real-life single-center retrospective study. J Cosmet Dermatol. 2023;22(5):1680-1684.
23. Oh J, Lee RW, Lee HR, et al. Classification of facial and truncal segmental vitiligo and its clinical courses including recurrence rate and patterns: a retrospective review of 956 patients. Br J Dermatol. 2021;184(4):750-753.
24. Delbaere L, Duponselle J, Herbelet S, et al. Predictive value of disease activity signs in vitiligo: An observational study. Exp Dermatol. 2024;33(2):e15167.
25. Kayal A, Gupta LK, Khare AK, et al. Pattern of childhood onset vitiligo at a tertiary care centre in South-West Rajasthan. Indian J Dermatol. 2015;60(5):520.
26. Al-Shobaili HA. Update on the genetics characterization of vitiligo. Int J Health Sci (Qassim). 2011;5(2):167-179.