Skip to main content Skip to main content
HsiaoEye
HsiaoEye
蕭閔謙醫師 · 眼科衛教筆記
You are reading the English-mode interface. Some article body content is currently Chinese-only; full translation is in progress. Switch to Chinese
📑 RCT secondary analysis · Jacobs et al. 2026 · Am J Ophthalmol · ZEDS Taskforce
Latest Research · Keratitis · HZO

Recurrent Stromal Keratitis After HZO
Why the 3 Months After Stopping Steroids Are Riskiest

A May 2026 secondary analysis of the Zoster Eye Disease Study (ZEDS) in American Journal of Ophthalmology (Jacobs et al.). ZEDS was a US/Canada/NZ 95-center, 527-patient double-masked RCT comparing valacyclovir 1000 mg/day × 1 year vs placebo. Key findings: 105/527 (20%) developed recurrent, new or worsening stromal keratitis; 75% were diagnosed at routine surveillance visits (mostly asymptomatic); 67% had SK at or before enrollment; among those off topical steroid at SK, 38% had stopped within the prior 3 months; at SK onset, 90% were dosing ≤ once daily, evenly split between high- and low-potency; 88% of low-potency users only needed an increase in frequency rather than switch to high-potency; only 10% received open-label oral valacyclovir; vision essentially preserved (Snellen 20/25 → 20/27). Cross-referenced with Taiwan NHI §10.7.1.1 (systemic anti-herpetic agents restricted to acute use, 10-day course limit) and §14.2 (topical acyclovir restricted to V1 involvement with keratitis/ulcer). ZEDS's 1-year suppression protocol is not within Taiwan NHI's listed conditions.

⚠️ Disclaimer: General medical education, summarizing a May 2026 secondary analysis of the ZEDS SK end point (Jacobs DS et al.) in Am J Ophthalmol. This article does not replace individual evaluation by your treating ophthalmologist. Decisions about starting, stopping, tapering, or chronically maintaining any medication — particularly topical steroids and oral antivirals — should be made by your physician based on your individual case. This site does not engage in medical advertising and does not endorse any specific brand, clinic, or physician. Taiwan NHI references are drawn from current Taiwan NHI regulations; items not in that document (e.g. individual topical-steroid brands, Shingrix vaccine) are deliberately not commented on — please consult your pharmacist or the NHIA portal for those.
帶狀疱疹眼疾後反覆角膜炎：從皮疹到追蹤的敘事圖 Diagram: HZO narrative — face with V1 rash, eye cross-section with stromal inflammation, treatment timeline highlighting the 3-month post-cessation danger window, and three colored take-home cards. Highest relapse risk: the 3 months after stopping steroids Recurrent stromal keratitis after HZO — at a glance ① Rash here = high eye-risk zone Forehead Upper lid Nose tip ★ Virus travels along trigeminal V1 into the eye ★ Hutchinson sign (nose-tip rash) = high eye risk ② Inflammation is in the "stromal" middle layer Eyeball Cornea zoom (cross-section) Epithelium Stroma ← SK inflammation Endothelium Recurrent → scar → vision loss ③ Treatment timeline — 3 months post-cessation is the danger window Rash onset Day 0 Acute Rx Oral antiviral 7–10 d + topical steroid Maintenance Low-potency drop gradual taper STOP ⚠️ 3-mo danger 38% of relapses occur here Risk drops continue follow-up 3M Visit every 3 mo 75% of relapses silent you can't feel them Only the slit-lamp can ↓ Low-potency maintenance 88% don't need stronger frequency increase suffices less IOP rise / cataract risk Don't abruptly stop Within 3 mo of stopping 38% of relapses happen here Taper, don't quit Source: ZEDS trial (Jacobs et al. 2026) · n = 527 · 18-month follow-up
Recurrent stromal keratitis after HZO at a glance. ① HZO rash on the forehead, upper lid and nose tip (V1 dermatome) → virus travels along the trigeminal nerve into the eye, causing ocular complications. ② Inflammation occurs in the corneal middle layer — the "stroma" — hence stromal keratitis (SK); repeated inflammation leaves permanent scars and reduces vision. ③ Treatment timeline: acute phase = oral antiviral 7-10 days + topical steroid → maintenance = low-potency drop tapered → the 3 months after stopping is the relapse peak (ZEDS: 38% of off-steroid relapses occur here) → risk drops thereafter. Three key actions: (green) 3-month follow-up (75% of relapses are silent); (yellow) long-term low-potency maintenance (88% don't need stronger, fewer side effects); (red) don't abruptly stop steroid (38% relapse within 3 months — switch to low-dose maintenance). Source: Jacobs et al. 2026, ZEDS SK secondary analysis.

1. Common clinic scenarios

'Doctor, five years ago I had a shingles rash on my forehead, scalp and upper eyelid, and my cornea got inflamed too. Ophthalmology gave me steroid eye drops, tapered me down, and eventually I stopped. About two or three months after stopping, this eye is getting red and a bit hazy again — is it back?'

'Doctor, I was hospitalized 10 years ago for herpes zoster ophthalmicus. Ophthalmology said to come back regularly, but I felt fine for years so I stopped going. At a recent routine check, ophthalmology found deep inflammation in my cornea — I felt absolutely nothing. Is that serious?'

'Doctor, I recently saw dermatology for shingles and finished a 7-day course of oral antiviral. The skin healed but my eye is still occasionally light-sensitive and slightly sore. Do I need to keep taking the oral antiviral? For how long?'

All three scenarios revolve around the same point: the eye component of herpes zoster ophthalmicus (HZO) is usually not "a 7-day oral antiviral course and you're done" — it tends to be a relapsing, chronic process. A May 2026 secondary analysis of the Zoster Eye Disease Study (ZEDS) SK end point by Jacobs et al. in American Journal of Ophthalmology gives us six concrete clinical lessons. This article distils them in clinic-friendly language and cross-references the relevant Taiwan NHI.

2. Common patient Q&As

Q1: What is HZO? Is it the same as stromal keratitis (SK)?
A: Not the same — it's a parent-child relationship. Herpes zoster ophthalmicus (HZO) is the umbrella term for any eye involvement when varicella-zoster virus (VZV) reactivates along the first branch (V1) of the trigeminal nerve — forehead, upper eyelid, tip of nose. HZO manifestations include: stromal keratitis (SK), dendritiform epithelial keratitis (DEK), endothelial keratitis (EK), iritis (IR), plus neuralgia, conjunctivitis, optic neuritis, acute retinal necrosis, etc. SK is the most common complication — 66% of all 160 complications in ZEDS. SK means inflammation in the middle ("stromal") layer of the cornea; it can cause corneal opacification, neovascularisation, and permanent scarring — a major cause of vision loss after HZO.
Q2: How long after the initial HZO episode can it recur?
A: It can recur 5, 10, even decades later. Ocular complications of HZO are by nature chronic, relapsing. Tran et al. 2016 in Ophthalmology showed the 5-year recurrence rate of HZO eye disease or rash is about 25%. The main ZEDS trial (Cohen et al. 2025 JAMA Ophthalmol) confirms that patients with any prior SK/DEK/EK/IR complication carry a substantial 1-year recurrence rate. The Jacobs et al. 2026 secondary analysis further showed 67% of those with recurrent SK had a prior SK history at enrollment; the virus can shed asymptomatically and maintain low-level immune activation for years. HZO should not be treated as "over once acute treatment is done" — it is a chronic relapsing disease.
Q3: Why every 3 months if I have no symptoms?
A: Because ZEDS showed 75% of SK was detected by physicians while the patient felt fine. In the Jacobs et al. 2026 data, 79 of 105 SK events (75%) were diagnosed at scheduled routine visits — not at unscheduled patient-initiated visits for redness/pain/blur. Early SK often starts as a small sectorial infiltrate (as described by Mondino 1986); the patient can't perceive it, but if untreated it spreads centrally, scars, and brings in new vessels. Slit-lamp examination with fluorescein and blue light, plus endothelial microscopy, picks up changes the patient cannot feel. "No symptoms" does not mean "no inflammation" — this is the central value of scheduled surveillance.
Q4: Why does "relapse soon after stopping topical steroids" matter so much?
A: Because the 3 months after stopping topical steroids is the highest-risk window observed in ZEDS. Of the 105 SK relapses, 52% were on topical steroids at relapse and 48% were off. Of the 48 off-steroid relapses, 38% (18) had stopped within the prior 3 months. The clinical implication: stopping doesn't mean "the eye has healed" — it means "immune suppression lifted, viral/inflammatory mechanisms re-active." A parallel single-centre review (Scott et al. 2024, Am J Ophthalmol) reaches the same conclusion — topical-steroid cessation is the strongest correlate of HZO recurrence. ZEDS therefore shifts the prescribing pattern from "taper off as soon as possible" to "low-dose long-term maintenance." But the actual maintenance dose, frequency, and whether you can truly stop — these must be decided by your ophthalmologist based on your individual exam, comorbidities (cataract, glaucoma).
Q5: My doctor said I can switch to a low-potency drop long-term — is that safe?
A: ZEDS data support this direction, but the individual decision rests with your physician. ZEDS classified topical steroids into two tiers: high-potency (prednisolone acetate 1%, difluprednate 0.05%, dexamethasone 0.1%, prednisolone phosphate 1%) and low-potency (loteprednol 0.2%, fluorometholone 0.1%/0.25%, prednisolone acetate 0.12%, rimexolone 1%). Among the 47 patients on a steroid at SK onset, the high/low split was nearly 1:1 (21 vs 26), not statistically different (p = 0.47) — high-potency didn't "protect better." And for managing recurrence, 88% (23/26) of low-potency users only needed a frequency increase, not a switch to high-potency. Only one patient in the whole series needed the most potent agent (difluprednate 4×/day). Low-potency steroids carry a lower long-term risk of IOP rise and cataract progression — but that's not zero. Any patient on long-term topical steroid still needs IOP and lens monitoring.
Q6: Do I need long-term oral valacyclovir? Will NHI cover it?
A: Most patients don't need it, and Taiwan NHI rules don't cover long-term suppression. ZEDS shows: (1) at SK relapse, only 10% (11/105) of patients had open-label oral valacyclovir added by their physician — the other 90% were controlled by topical steroid adjustment alone; (2) recurrence rate did not differ between valacyclovir and placebo groups — "not on steroid but on oral antiviral" does not substitute for the protective effect of topical steroids.

Taiwan NHI (per current Taiwan NHI regulations):
• §10.7.1.1 systemic anti-herpetic agents (acyclovir, famciclovir, valaciclovir) — restricted to: HZO/HSV involving V1 with potential corneal threat; HZO/HSV-related keratitis or corneal ulcer; acute retinal necrosis; immunocompromised/cancer/transplant patients; HZO within 3 days of rash on head/neck/genital area (5-day course); etc.
• Same section, point 3 states explicitly: "Except for the above specific conditions, treatment is limited to a 10-day course; oral, injectable and topical formulations must be used as a single agent, not combined."
• So ZEDS's "1-year suppression" use is not within Taiwan NHI's listed conditions; if your treating physician recommends long-term low-dose suppression on clinical grounds, it usually requires out-of-pocket payment.
• The standard 7-10-day acute HZO course does fall under NHI coverage.
Q7: What happens in the worst case if recurrent inflammation is not properly treated?
A: Corneal neovascularization, permanent scarring, vision loss. Long-term changes in poorly controlled HZO include: (1) corneal neovascularization — new vessels grow into the normally avascular cornea, causing opacification and reduced acuity; (2) corneal scarring — repeated stromal inflammation leaves permanent opaque areas; (3) loss of corneal sensation — the trigeminal nerve is damaged by the virus, causing neurotrophic keratopathy that heals poorly and ulcerates easily; (4) corneal thinning and structural instability — rarely progressing to ulceration with perforation. Corneal transplantation may eventually be needed, but the inflamed and denervated HZO cornea carries higher recurrence and rejection risk after transplant. The whole "regular surveillance + low-dose maintenance + don't abruptly stop steroid" strategy aims to keep the cornea in the reversible, scar-free, avascular state.

3. Key takeaways: 8 things you really need to know

30-second takeaways

  • HZO is chronic-relapsing: 5-yr recurrence ~25%; 67% of recurrent-SK patients already had SK history
  • SK is the dominant complication: 66% of all ZEDS complications; 20% (105/527) of trial patients had recurrent/new/worsening SK
  • 75% are asymptomatic: scheduled 3-month follow-up is the main diagnostic route — don't wait until you "feel" something
  • 3 months post-cessation is the danger window: 38% of off-steroid relapses had stopped within the prior 3 months
  • Low-potency suffices: high vs low potency showed no statistical difference in prevention or treatment; 88% of low-potency users only needed frequency increase
  • Oral antiviral rarely needed: 90% of SK relapses controlled by steroid adjustment alone; only 10% had open-label oral valacyclovir added
  • Vision is preserved: enrollment logMAR 0.10 (20/25) → 12-month 0.13 (20/27), not statistically different
  • Taiwan NHI restricts to acute use: §10.7.1.1 limits systemic anti-herpetic course to 10 days; 1-year suppression is not within NHI's listed conditions

4. Background: why does shingles "hit the eye"?

The pathogen is varicella-zoster virus (VZV). After childhood chickenpox, VZV is not cleared from the body but stays latent in sensory ganglia (dorsal root and trigeminal). Aging, immune decline, stress, illness, chemotherapy etc. can reactivate VZV along a single ganglion, producing a rash along that dermatome — this is shingles (zoster).

When reactivation occurs in the first division of the trigeminal nerve (the ophthalmic nerve, V1), the rash appears on the forehead, upper eyelid, tip and side of the nose — this is herpes zoster ophthalmicus (HZO). Because V1 also supplies sensation to the cornea, uveal tract and retina, VZV can travel along the nerve to involve the eye itself. A key clinical clue is Hutchinson's sign — rash extending to the tip of the nose indicates involvement of the nasociliary branch of V1 and a much higher risk of eye involvement.

Old vs current view: historically, HZO ocular complications were considered "purely inflammatory, no remaining virus," so steroids were the only treatment. The past two decades of evidence (Wilson 1992, Cohrs 2008, Pavan-Langston 1995, Hu 2010, Li 2018 etc.) show that VZV can shed asymptomatically and maintain low-grade immune activation for years; recurrent SK has both an infectious and an inflammatory component. This shift in understanding underpinned the design of the ZEDS trial — if infection contributes, then long-term oral antiviral suppression ought to reduce recurrence.

5. What is ZEDS, and why does this dataset matter?

The Zoster Eye Disease Study (ZEDS) is the largest, most rigorously designed RCT for recurrent HZO eye disease (ClinicalTrials.gov: NCT03134196). Basic design:

ItemDetail
Participants527 patients with prior HZO; stratified by age (< 60 / ≥ 60) and disease phase (recent / chronic)
AllocationDouble-masked randomization: valacyclovir 1000 mg/day × 1 year vs placebo
Follow-up18 months, with scheduled ophthalmology visits every 3 months (slit-lamp, visual acuity, IOP, etc.)
Primary end pointsRecurrent, new or worsening stromal keratitis (SK), SK with ulceration (SKU), endothelial keratitis (EK), iritis (IR), dendritiform epithelial keratitis (DEK)
Scale95 sites (US 85 + Canada 9 + NZ 1), Nov 2017 – Jun 2024
Main result (Cohen 2025, JAMA Ophthalmology)1-year valacyclovir suppression reduced SK/DEK/IR/EK recurrence at 18 months; 12-month primary endpoint not statistically significant, 18-month secondary endpoint significant

Jacobs et al. 2026 is the "SK secondary analysis" of ZEDS — unpacking the 105 SK relapse cases: who, when, on what medication, how was the relapse managed, and what was the visual outcome. This is currently the most complete real-world dataset on SK management after HZO.

6. The 6 clinical lessons from ZEDS

LessonZEDS dataClinical implication
1. SK is usually silent 79/105 (75%) diagnosed at scheduled visits HZO patients can't wait until symptoms appear; 3-monthly follow-up is the baseline
2. Topical steroid alone is enough in most cases Only 11/105 (10%) needed open-label oral valacyclovir; 90% controlled by steroid alone No need to add oral antiviral for every relapse; first reassess whether steroid dose/frequency is adequate
3. Low-potency is sufficient High vs low potency nearly 1:1 (p = 0.47); 88% of low-potency users only needed frequency increase Long-term low-potency maintenance, lower IOP rise and cataract progression risk
4. Don't abruptly stop steroid Of off-steroid relapses, 38% had stopped within prior 3 months Rather than full discontinuation, switch to a low-frequency maintenance (e.g. a few times per week)
5. High-potency doesn't prevent recurrence better No statistical difference between high vs low potency in preventing SK recurrence "Stay on the strongest" is not safer; only adds side effects
6. Vision can be preserved Enrollment logMAR 0.10 (20/25) → 12-mo 0.13 (20/27), not statistically different Active surveillance + early low-dose intervention preserves vision in most cases

7. Topical steroid tiers (ZEDS definitions)

ZEDS applied the following tiers (per the trial protocol; clinically difluprednate is generally regarded as twice as potent as prednisolone acetate 1%, but ZEDS grouped it with other high-potency agents):

PotencyAgents listed by ZEDS
High-potency prednisolone acetate 1%, difluprednate 0.05%, dexamethasone 0.1%, prednisolone phosphate 1%
Low-potency loteprednol 0.2%, fluorometholone 0.1%/0.25%, prednisolone acetate 0.12%, rimexolone 1%

About Taiwan prescribing: whether each agent is available, NHI-covered, or its self-pay cost in Taiwan — please consult your prescribing physician, pharmacist, or the latest NHIA announcements. This article does not make brand-specific coverage claims. The point here is the "potency-tier concept," not a brand recommendation.

8. Taiwan NHI antiviral coverage conditions

The following summary covers the wording of Taiwan NHI's antiviral coverage conditions. Items or scenarios not explicitly covered are not commented on here (e.g. individual topical steroid brands, Shingrix vaccine, etc.). For actual coverage, duration, and applicability, refer to your prescriber, pharmacist, or the latest NHIA announcements.

✅ §10.7.1.1 Systemic anti-herpetic agents (oral / IV acyclovir, famciclovir, valaciclovir)

Coverage conditions (use limited to):

  • Herpetic encephalitis
  • HZO/HSV involving V1 with potential corneal threat
  • HZO/HSV involving sacral S2 dermatome affecting excretion
  • Immunocompromised / cancer / transplant patients with HZO or HSV
  • Varicella in neonates or immunocompromised patients
  • HZO/HSV-related keratitis or corneal ulcer
  • Acute retinal necrosis
  • HZO within 3 days of rash on head/neck/genital area — up to 5 days oral or topical

Duration cap (§10.7.1.1 point 3): "Except for the above specific conditions, treatment is limited to a 10-day course; oral, injectable and topical formulations must be used as a single agent — not combined."

Also: acyclovir 200 mg (Deherp), 400 mg, 800 mg (Virless) oral tablets used outside the above conditions are limited to a 7-day course.

✅ §14.2 Topical acyclovir (restricted to)

  • HZO/HSV involving V1 with potential corneal threat
  • HZO/HSV-related keratitis or corneal ulcer
  • Acute retinal necrosis
  • Not to be combined with oral, injectable or other topical antivirals (topical acyclovir ointment cannot be co-prescribed with oral antiviral)

⚠️ Key gap: ZEDS protocol vs Taiwan NHI rules

  • ZEDS uses valacyclovir 1000 mg/day × 1 year as chronic HZO suppression
  • Taiwan NHI §10.7.1.1 lists only acute indications (keratitis, ulcer, V1 involvement, within 3 days of rash, etc.), with a 10-day course cap
  • So "1-year suppression" is NOT within Taiwan NHI's listed conditions. If your physician recommends it on individualised clinical grounds, it is typically out-of-pocket
  • This article does not decide for your physician whether you need long-term oral antiviral. Discuss with your ophthalmologist whether low-dose topical steroid maintenance is sufficient for your individual case, or whether oral suppression is still indicated

9. When to see your doctor immediately

ZEDS shows SK is usually silent — that's why scheduled 3-monthly follow-up is the baseline. But if you develop any of the following, see your doctor same-day, not at the next scheduled visit:

🚨 Red flags for HZO patients

  • Sudden decline or blurring of vision (any degree)
  • Red, painful, light-sensitive, watery eye (especially within 3 months of stopping topical steroid)
  • Sudden floaters, flashes, visual field defect (possible acute retinal necrosis — rare but serious HZO complication)
  • New rash on the same forehead/upper eyelid/nose tip side as the original HZO — possible recurrence requiring immediate treatment
  • Corneal trauma / scratch history with pain or vision loss — denervated HZO corneas heal poorly

If any of the above, head to ophthalmology or general ER same-day. This article does not provide acute-management SOPs — let the on-site emergency clinician assess.

10. Conclusion: HZO is not a 7-day disease — it's a 7-year disease

HZO is easily mistaken for a "dermatology, 7-day antiviral and done" acute illness. The biggest message from ZEDS is the opposite: HZO ocular complications are chronic, relapsing, and mostly asymptomatic. The Jacobs et al. 2026 SK secondary analysis offers a practical prescribing framework:

  • 3-month surveillance — most SK is silent, only routine visits catch it early
  • Low-potency, low-frequency long-term maintenance — don't abruptly stop; the 3 months after cessation is the highest-risk window
  • Oral antiviral is rarely needed — 90% of SK relapses are controlled by steroid adjustment alone; long-term suppression is not within Taiwan NHI's listed conditions
  • Vision can be preserved — with active surveillance and avoiding abrupt cessation, most patients maintain near-20/25 acuity

If you or a family member has a history of HZO, even "everything seems fine now," please keep up with your ophthalmology follow-ups. That is the single most important reminder from the 527-patient, 18-month ZEDS dataset.

📚 HsiaoEye Related Articles — Cornea / Emergencies

References

  1. Jacobs DS, Lee T, Asbell P, Shen J, Choulakian M, Holland GN, Baratz KH, Prescott CR, Colby K, Hochman JS, Troxel AB, Cohen E, Jeng BH. Stromal Keratitis in the Zoster Eye Disease Study: Lessons Learned. Am J Ophthalmol. 2026;285:144–151. doi:10.1016/j.ajo.2026.01.034.
  2. Cohen EJ, Troxel AB, Liu M, et al. Low-dose valacyclovir in herpes zoster ophthalmicus: the Zoster Eye Disease randomized clinical trial. JAMA Ophthalmol. 2025;143(4):269–276.
  3. Liesegang TJ. Herpes zoster ophthalmicus natural history risk factors, clinical presentation, and morbidity. Ophthalmology. 2008;115(2 Suppl):S3–12.
  4. Miserocchi E, Fogliato G, Bianchi I, Bandello F, Modorati G. Clinical features of ocular herpetic infection in an Italian referral center. Cornea. 2014;33(6):565–570.
  5. Sy A, McLeod SD, Cohen EJ, et al. Practice patterns and opinions in the management of recurrent or chronic herpes zoster ophthalmicus. Cornea. 2012;31(7):786–790.
  6. Lo DM, Jeng BH, Gillespie C, Wu M, Cohen EJ. Current practice patterns and opinions on the management of recent-onset or chronic herpes zoster ophthalmicus of Zoster Eye Disease Study investigators. Cornea. 2019;38(1):13–17.
  7. Cobo LM, Foulks GN, Liesegang T, et al. Oral acyclovir in the treatment of acute herpes zoster ophthalmicus. Ophthalmology. 1986;93(6):763–770.
  8. Hoang-Xuan T, Büchi ER, Herbort CP, et al. Oral acyclovir for herpes zoster ophthalmicus. Ophthalmology. 1992;99(7):1062–1070.
  9. Severson EA, Baratz KH, Hodge DO, Burke JP. Herpes zoster ophthalmicus in Olmsted County, Minnesota: have systemic antivirals made a difference? Arch Ophthalmol. 2003;121(3):386–390.
  10. Tran KD, Falcone MM, Choi DS, et al. Epidemiology of herpes zoster ophthalmicus: recurrence and chronicity. Ophthalmology. 2016;123(7):1469–1475.
  11. Liesegang TJ. Corneal complications from herpes zoster ophthalmicus. Ophthalmology. 1985;92(3):316–324.
  12. Mondino BJ, Farley MK, Aizuss DH. Sectorial corneal infiltrates and pannus in herpes zoster ophthalmicus. Graefes Arch Clin Exp Ophthalmol. 1986;224(3):313–316.
  13. Wilson A, Sharp M, Koropchak CM, Ting SF, Arvin AM. Subclinical varicella-zoster virus viremia, herpes zoster, and T-lymphocyte immunity to varicella-zoster viral antigens after bone marrow transplantation. J Infect Dis. 1992;165(1):119–126.
  14. Cohrs RJ, Mehta SK, Schmid DS, Gilden DH, Pierson DL. Asymptomatic reactivation and shed of infectious varicella-zoster virus in astronauts. J Med Virol. 2008;80(6):1116–1122.
  15. Pavan-Langston D, Yamamoto S, Dunkel EC. Delayed herpes zoster pseudodendrites. Polymerase chain reaction detection of viral DNA and a role for antiviral therapy. Arch Ophthalmol. 1995;113(11):1381–1385.
  16. Hu AYH, Strauss ED, Holland GN, et al. Late varicella-zoster virus dendriform keratitis in patients with histories of herpes zoster ophthalmicus. Am J Ophthalmol. 2010;149(2):214–220.
  17. Li JY. Herpes zoster ophthalmicus: acute keratitis. Curr Opin Ophthalmol. 2018;29(4):328–333.
  18. Scott DAR, Liu K, Danesh-Meyer HV, Niederer RL. Herpes zoster ophthalmicus recurrence: risk factors and long-term clinical outcomes. Am J Ophthalmol. 2024;268:1–9.

Frequently asked questions

Can herpes zoster ophthalmicus come back after treatment is finished?
Yes. The ocular complications of herpes zoster ophthalmicus are chronic and relapsing, and can flare again 5, 10 or more years later. The 5-year recurrence rate of eye disease or rash is about 25%, and in the ZEDS secondary analysis about 67% of patients with recurrent stromal keratitis already had a prior history at enrollment, so it is not an acute illness that ends once treatment is done.
If my eye feels fine, do I still need regular ophthalmology follow-up?
Yes. The ZEDS secondary analysis showed that about 75% of recurrent stromal keratitis was detected by the physician at routine slit-lamp follow-up while the patient felt nothing. Early inflammation often begins as a small sectorial infiltrate that the patient cannot perceive, but if untreated it spreads and leaves permanent scarring. No symptoms does not mean no inflammation, so scheduled follow-up is what catches it early.
Why is the relapse risk highest in the first 3 months after stopping steroid eye drops?
Because stopping the drop lifts immune suppression and lets the viral and inflammatory mechanisms reactivate — it does not mean the eye has fully healed. Among ZEDS relapse patients who were off steroid, about 38% had stopped within the prior 3 months, making that window the highest-risk period observed. This is why abrupt cessation is discouraged in favour of low-dose long-term maintenance, though the exact tapering and timing must be decided by your ophthalmologist.
Is a low-potency steroid eye drop less effective for long-term use?
ZEDS data support low-potency long-term maintenance. Among patients on a steroid at relapse onset, high- and low-potency users were split roughly evenly with no statistical difference, meaning high-potency did not protect better. When managing recurrence, about 88% of low-potency users only needed an increase in dosing frequency rather than a switch to high-potency, and low-potency carries a lower long-term risk of raised eye pressure and cataract. Still, anyone on long-term steroid needs regular eye-pressure and lens monitoring.
Do I need long-term oral antiviral, and will Taiwan NHI cover it?
Most people do not. ZEDS showed that at stromal keratitis relapse only about 10% of patients had oral antiviral added by their physician, while roughly 90% were controlled by adjusting topical steroid alone. Under Taiwan NHI, systemic anti-herpetic agents (§10.7.1.1) are listed only for acute indications and treatment is limited to a 10-day course, so the 1-year suppression used in ZEDS is not within the listed conditions; if your physician recommends long-term use on individual grounds it is usually out-of-pocket.
What is the worst outcome if recurrent inflammation is not well controlled?
Long-term recurrent inflammation can cause new blood vessels to grow into the cornea, leave permanent scarring, and damage corneal sensation leading to neurotrophic keratopathy; the cornea may thin and become structurally unstable, and in rare cases progress to ulceration with perforation, eventually possibly requiring a corneal transplant. Conversely, with active surveillance, low-dose maintenance and avoiding abrupt steroid cessation, most ZEDS patients preserved vision near 20/25 — the reversible state this long-term strategy aims to protect.