Ocular Manifestations of Stevens-Johnson Syndrome and Toxic Epidermal Necrolysis in Children


To describe the acute and chronic ocular manifestations of Stevens-Johnson syndrome (SJS), toxic epidermal necrolysis (TEN), and SJS/TEN Overlap syndrome (Overlap syndrome) in children.


Retrospective case series.


Medical records of children admitted to the Hospital for Sick Children between 2001 and 2011 with SJS, TEN, and Overlap syndrome were reviewed. Demographic information, all abnormal ophthalmic findings (and median time to first diagnosis), visual acuities, and ophthalmic treatments prescribed were collected for each eye for every patient.


Thirty-six children were identified for inclusion in the study. Twenty-nine (81%) had acute ocular involvement, including all patients with TEN (n = 7). Conjunctivitis was the most common (78%) clinical sign. This, together with conjunctival membranes and subconjunctival hemorrhage, were the earliest signs, presenting by a median of 1 day. The percentage of patients and median time to occurrence of complications were as follows: for lid margin ulceration and corneal epithelial defects, 25%, 3 days; conjunctival ulceration, 39%, 3.5 days; symblepharon, 28%, 4 weeks; corneal opacification, 11%, 4 months; limbal stem cell failure, 8%, 7 months; and corneal vascularisation, 8%, 10 months after admission. Over 90% of children maintain a visual acuity of 20/40 or better in each eye at a mean follow-up of 1.4 years.


Ocular involvement in SJS, TEN, and Overlap syndrome is common and the ocular manifestations may develop many months after the initial presentation, mandating the need for long-term follow-up of these children. Despite the high frequency of sight-threatening disease, most children maintain good vision in the long term.

Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are rare and sometimes fatal mucocutaneous diseases, which can leave survivors with devastating ocular sequelae.

SJS and TEN are variants within a disease spectrum. In SJS epidermal detachment affects less than 10% of the total body surface area and in TEN it affects greater than 30%. Intermediate cases (between 10% and 30%) are called SJS/TEN Overlap syndrome (hereafter called Overlap syndrome). The incidence of SJS and TEN are 1.2–6 and 0.4–1.2 per million person-years, respectively. Adult mortality is below 5% for SJS and about 30% for TEN. SJS and TEN case series suggest the mortality is lower in children, ranging between 0% and 17%. In the most recent pediatric TEN series from our institution, the mortality rate was 0%. Antibiotics, anticonvulsants, and nonsteroidal anti-inflammatory drugs are the most commonly implicated etiologies in children.

Mucosal involvement occurs in over 90% of patients with SJS and TEN. Involvement of the ocular surface is common and may be a cause of significant morbidity in the longer term. In adult studies the incidence of ocular involvement in the acute phase of the disease is reported to be 60%–71% of SJS patients, 88%–100% of those with Overlap syndrome, and 72%–90% of TEN patients. Long-term ocular sequelae have been identified in 27%–59% of patients with SJS, TEN, and Overlap syndrome.

There is a paucity of literature reporting exclusively on children, showing an incidence of acute ocular involvement of 93% in SJS and 71%–100% in TEN. To our knowledge, there is no study reporting the incidence of acute ocular involvement in children with Overlap syndrome separately. The incidence of long-term ocular sequelae following childhood SJS or TEN is reported in 21%–29% of patients. These are all case series written with an emphasis on systemic manifestations. They have little ophthalmic clinical detail and none report the length of ophthalmic follow-up.

Given the lack of current literature focussing on the ocular manifestations of SJS, TEN, and Overlap syndrome in children, the purpose of this report is to comprehensively and systematically describe the acute and long-term ocular manifestations and visual acuity outcomes of these patients. Based on our experience, we make suggestions for the management of children with ocular manifestations of SJS, TEN, and Overlap syndrome.


Prospective Institutional Ethics Committee approval was obtained for this retrospective case series. The medical records of all patients admitted to the Hospital for Sick Children between June 1, 2001 and June 1, 2011 with SJS, TEN, or Overlap syndrome, according to the consensus definition of Bastuji-Garin and associates, were included. The records were identified using hospital admission codes, and accuracy of diagnosis was reviewed and confirmed by a senior clinician with experience in the diagnosis and management of these diseases.

Demographic information, including age at admission, sex, and length of hospital stay were collected. The day of hospital admission was designated day zero for the purposes of this paper. Clinical findings including visual acuity and clinical signs in the conjunctiva, lids, and cornea were recorded for each eye of the patient at every ophthalmic examination. The main outcome measures were incidence and the median time to first diagnosis for each abnormal clinical finding, medical and surgical therapies used, and the final visual acuity measurement.

All visual acuity records were converted to a Snellen visual acuity. Owing to the retrospective nature of the study and multiple assessments by different ophthalmologists, any acute-phase record of “conjunctival hyperaemia,” “injection,” “papillae,” “follicles,” or “red eyes” was categorized as having conjunctivitis. “Membranes” or “pseudomembranes” were both categorized as conjunctival membranes. If the record documented “clear cornea” or “cornea normal” it was taken to mean there was no corneal infiltrate, opacification, vascularisation, or ulceration. A comment on the deficiency of the tear film or tear meniscus was required for dry eye to be diagnosed. Superficial punctate epithelial erosions alone did not qualify the patient as having a dry eye. The use of any topical ophthalmic treatment or intervention (surgical and nonsurgical) was recorded. Where used, amniotic membrane grafting used cryopreserved amniotic membrane over the lid margins, palpebral conjunctiva, and ocular surface, held in place with bolstered fornix sutures, perilimbal sutures, and a conformer. A Prosthetic Replacement of the Ocular Surface Ecosystem (PROSE; Boston Foundation for Sight, Needham, Massachusetts, USA) device was used as indicated in some children. This is a gas-permeable scleral contact lens used for the symptomatic management and visual rehabilitation of patients with complex ocular surface disorders, including those with SJS.

The severity of ocular involvement during the acute phase of the disease was classified as per the schema of Power and associates. In summary, mild involvement was defined as ocular signs requiring routine eye care, which completely resolved before discharge from hospital (eg, lid edema, conjunctival injection, requiring prophylactic antibiotics or lubricants). Moderate involvement included ocular complications requiring specific treatment (eg, conjunctival membranes, corneal epithelial loss of more than 30%) and normal vision with near-complete resolution of ocular disease prior to discharge. Severe involvement included sight-threatening disease with the need for ongoing specific ophthalmic treatment after discharge (eg, symblepharon, active corneal disease at the time of discharge). Where there was asymmetry between the 2 eyes, the eye more affected was used to allocate severity.

A literature search using the MEDLINE database was performed, by combining synonyms for “SJS, TEN, or Overlap syndrome” and synonyms for “ocular” and limiting the search results to patients aged 18 years or under. The reference lists of articles thus identified were searched manually.


A total of 36 patients with SJS, TEN, or Overlap syndrome were identified from the 10-year study period, of which 20 (56%) had SJS, 7 (19%) had TEN, and 9 (25%) had Overlap syndrome ( Table 1 ). There was a predominance of male subjects, with a ratio of 1.6:1. All were consulted as in-patients and just under a third required Intensive Care Unit (ICU) admission in the acute phase of the disease (n = 11, 31%). The mean duration of acute hospital admission was 15 days and the mean duration of follow-up was 389 days. Eighteen patients (50%) had at least 1 follow-up visit to the outpatient department. Eleven patients (31%) had at least 1 year of follow-up. Fourteen patients (39%) were discharged from ophthalmic follow-up. Six (17%) were referred for ophthalmic care elsewhere, 6 (17%) are continuing their care as outpatients, and 10 children (28%) were lost to follow-up.

Table 1

Demographic Details of Study Population Including Children With Diagnoses of Stevens-Johnson Syndrome, Toxic Epidermal Necrolysis, and Stevens-Johnson Syndrome/Toxic Epidermal Necrolysis Overlap According to the Bastuji-Garin and Associates Consensus Definition

Patient Characteristics
Total number 36
Male, n (%) 22 (61%)
Female, n (%) 14 (39%)
Age at admission (range) 8.8 years (0.7–15 years)
Number (%) of patients admitted to the Intensive Care Unit 11 (31%)
Mean duration of admission (range) 15 days (2–87 days)
Mean duration of follow-up (range) 389 days (0 days–9 years)
Diagnosis, n (%)
SJS 20 (56%)
Overlap 9 (25%)
TEN 7 (19%)

Overlap = Stevens-Johnson syndrome/toxic epidermal necrolysis overlap; SJS = Stevens-Johnson syndrome; TEN = toxic epidermal necrolysis.

Twenty-nine patients (81%) had acute ocular involvement affecting 15 (15/20; 75%) of those with SJS, 7 of 7 patients with TEN, and 7 of 9 among those with Overlap syndrome ( Figure 1 ). Severity of ocular involvement was asymmetrical in only 3 patients, and in each of these, the 2 severity categories were adjacent to one another. Overall, 11 children (31%) had mild, 13 (36%) had moderate, and 5 (14%) had severe ocular involvement in the acute phase of the disease.

Figure 1

The severity and frequency of ocular involvement during the acute phase in children with Stevens-Johnson syndrome (SJS), Stevens-Johnson syndrome/toxic epidermal necrolysis overlap (Overlap), and toxic epidermal necrolysis (TEN). The severity classification has been previously described. Ocular involvement was asymmetrical in 3 patients and in these cases the more severely affected eye was used to allocate severity.

Clinical features of children with SJS, TEN, and Overlap syndrome are shown in Table 2 . Conjunctivitis (78%), conjunctival membranes (28%), and subconjunctival hemorrhage (33%) were the earliest of signs, all presenting within the first day. Bulbar (39%) and tarsal (33%) conjunctival ulceration and corneal epithelial degradation including superficial punctate epithelial erosions (50%) and epithelial defects (25%) were common and appeared at a median of day 3. Symblepharon (28%) and ankyloblepharon (11%) presented during the fourth week, followed by trichiasis (8%), anterior blepharitis (6%), and punctual auto-occlusion (8%) in the fifth and sixth weeks. Subconjunctival scarring presented late in the seventh week. Chronic lid changes such as lid margin keratinization (22%), meibomian gland disease (25%) and entropion (8%), and corneal opacification (11%) presented between the third and fourth months. Dry eye (28%), distichiasis (11%), limbal stem cell failure (8%), and corneal vascularization (8%) presented between 5 and 11 months after initial presentation. Illustrative examples of the clinical signs identified in this cohort are shown in Figure 2 .

Table 2

Frequency and Mean Day of Diagnosis of Ocular Clinical Signs in Patients With Stevens-Johnson Syndrome, Stevens-Johnson/Toxic Epidermal Necrolysis Overlap, and Toxic Epidermal Necrolysis

Clinical Sign Number (%) of Patients Affected Range (Days) Median (Days) Median (Equivalent Week or Month)
Conjunctivitis 28 (78%) 0–9 0.5
Conjunctival membranes 10 (28%) 0–5 0.5
Subconjunctival hemorrhage 12 (33%) 0–30 1
Bulbar conjunctival ulceration 14 (39%) 0–18 2
Tarsal conjunctival ulceration 12 (33%) 0–20 3.5
Symblepharon 10 (28%) 1–207 23.5 Week 4
Ankyloblepharon 4 (11%) 13–539 26 Week 4
Subconjunctival scarring 5 (14%) 20–339 47 Week 7
Lid edema 14 (39%) 0–44 2.5
Lid margin ulceration 9 (25%) 0–26 3
Trichiasis 3 (8%) 10–31 30 Week 5
Blepharitis (Anterior) 2 (6%) 22–37 34.5 Week 5
Punctal auto-occlusion 3 (8%) 36–906 39 Week 6
Lid margin keratinization 8 (22%) 14–642 80.5 Month 3
Entropion 3 (8%) 30–388 100 Month 4
Meibomian gland disease 9 (25%) 15–359 114 Month 4
Distichiasis 4 (11%) 96–717 170.5 Month 6
Superficial punctate epithelial erosions 18 (50%) 0–9 2
Corneal epithelial defect 9 (25%) 0–242 3
Corneal opacification 4 (11%) 40–149 116 Month 4
Dry eye 10 (28%) 60–509 135 Month 5
Limbal stem cell failure 3 (8%) 114–1150 193 Month 7
Corneal vascularization 3 (8%) 114–674 333 Month 10

Figure 2

Ocular manifestation of Stevens-Johnson syndrome and toxic epidermal necrolysis (Top left) bulbar conjunctival ulceration, (Top right) desquamation of the ocular surface and lid margin, (Bottom left) symblepharon, and (Bottom right) lid margin and conjunctival keratinization, with a prosthetic replacement of the ocular surface ecosystem (PROSE device; Boston Foundation for Sight, Needham, Massachusetts, USA) in situ.

The characteristics of children with at least 1 year of follow-up and those with less than 1 year of follow-up were compared. All children with severe ocular involvement in the acute phase of the disease were followed for at least 1 year. Early (within the first 7 days) conjunctival membranes, subconjunctival haemorrhage, bulbar conjunctival ulceration, and superficial punctate epithelial erosions were much more common in the cohort that went on to have follow-up for more than 1 year than in the cohort with less than 1 year of follow-up. All children without acute ocular involvement, 10 of those with mild involvement (10/11; 91%), and 8 of those with moderate involvement (8/13; 62%) had less than 1 year of follow-up.

The clinical signs were compared across children with mild, moderate, and severe ocular involvement ( Table 3 ). Conjunctivitis is common across all 3 severity categories. All children with severe ocular involvement had signs of subconjunctival hemorrhage, bulbar and tarsal conjunctival ulceration, symblepharon, lid edema, lid margin ulceration and keratinization, superficial punctate epithelial erosions, and dry eye.

Table 3

Frequency of Clinical Signs in Children With Mild, Moderate, and Severe Ocular Involvement in the Acute Phase of the Disease

Clinical Sign Mild, n (%) (N = 11) Moderate, n (%) (N = 13) Severe, n (%) (N = 5)
Conjunctivitis 10 (91%) 13 (100%) 5 (100%)
Conjunctival membranes 0 6 (46%) 4 (80%)
Subconjunctival hemorrhage 2 (18%) 5 (38%) 5 (100%)
Bulbar conjunctival ulceration 0 9 (69%) 5 (100%)
Tarsal conjunctival ulceration 0 7 (54%) 5 (100%)
Symblepharon 1 (9%) 4 (31%) 5 (100%)
Ankyloblepharon 0 0 4 (80%)
Subconjunctival scarring 1 (9%) 2 (15%) 2 (40%)
Lid edema 2 (18%) 7 (54%) 5 (100%)
Lid margin ulceration 1 (9%) 3 (23%) 5 (100%)
Trichiasis 0 0 3 (60%)
Blepharitis (anterior) 0 1 (8%) 1 (20%)
Punctal auto-occlusion 0 1 (8%) 2 (40%)
Lid margin keratinization 1 (9%) 2 (15%) 5 (100%)
Entropion 0 0 3 (60%)
Meibomian gland disease 1 (9%) 5 (38%) 3 (60%)
Distichiasis 0 1 (8%) 3 (60%)
Superficial punctate epithelial erosions 5 (45%) 8 (62%) 5 (100%)
Corneal epithelial defect 0 4 (31%) 3 (60%)
Corneal opacification 0 0 4 (80%)
Dry eye 1 (9%) 4 (31%) 5 (100%)
Limbal stem cell failure 0 0 3 (60%)
Corneal vascularization 0 0 3 (60%)

Only gold members can continue reading. Log In or Register to continue

Jan 6, 2017 | Posted by in OPHTHALMOLOGY | Comments Off on Ocular Manifestations of Stevens-Johnson Syndrome and Toxic Epidermal Necrolysis in Children
Premium Wordpress Themes by UFO Themes