Stevens-Johnson Syndrome and Toxic Epidermal Necrolysis: 11-year Demographic Clinical and Prognostic Characteristics.

Background: Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are acute, life-threatening, severe drug reactions. Randomized studies on these diseases are difficult to perform. Aims and
Objectives: The purpose of this study was to summarize the demographic and clinical characteristics of patients with SJS and TEN in a tertiary hospital in Turkey. Materials and
Methods: We evaluated the records of 33 patients with SJS and TEN who were followed in our clinic or examined between January 2008 and June 2019, retrospectively. Age, sex, time of admission to hospital, causative drug, presence of concomitant disease, skin findings, mucosal involvement, the severity-of-illness score for TEN, the medication used, antibiotic use, transfer to intensive care, development of complications, and death or discharge status were noted.
Results: Of the 33 patients, 11 (33.3%) had SJS, 3 (9.1%) had SJS/TEN overlap, and 19 (57.6%) had TEN. The majority (60.6%) of the patients were female. Nineteen (57.6%) patients had one, and 13 (39.4%) had more than one suspected drug exposure in their history. The most commonly suspected drugs were antibiotics. Twelve (36.4%) patients had intensive care unit hospitalization. Ten (30.3%) patients died.
Conclusion: The demographic data of our study were consistent with the literature. Similar to the literature, antibiotics were the most common reaction-causing drugs. However, antiepileptic drugs, which were more frequently reported in other studies, were identified as suspicious in only one patient. We believe that our study will contribute to the determination of characteristics of this rare disease with real-life data. Copyright:

Introduction

Stevens–Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are acute, life-threatening, severe drug reactions. The disease progresses rapidly and diffuse keratinocyte necrosis results in dermo-epidermal and mucocutaneous separation. SJS (separation <10% of body surface area [BSA]), SJS/TEN overlap (separation between 10% and 30% of BSA) and TEN (separation >30% of BSA) are named according to the ratio of the separated area. Prompt diagnosis is critical for immediate discontinuation of the causative drug and initiation of appropriate treatment, which can be lifesaving.[12]

The identified risk factors are as follows: having a slow acetylator genotype, the presence of immunosuppression, use of anticonvulsants concurrent with radiotherapy, and having specific human leukocyte antigen (HLA) alleles such as HLA-B × 15: 02, HLA-A × 31: 01, HLA-B × 58: 01.[1] However, the pathogenesis is still not fully understood. Exposing the suspected drug in the presence of risk factors, the immune response occurs when the drug metabolites react with the tissues of the recipient. CD8 +cytotoxic T cells and natural killer (NK) cell-mediated cytotoxicity are triggered, and this reaction causes keratinocyte apoptosis by Fas/FasL, granulysin, perforin/granzyme B, and keratinocyte necroptosis by annexin A1.[12]

The most common drugs responsible for SJS/TEN are the sulfonamide group, allopurinol, carbamazepine, lamotrigine, nonsteroidal anti-inflammatory drugs (NSAID) containing oxicam, nevirapine, and sulfasalazine, and antibiotics such as cephalosporin, fluoroquinolone, aminopenicillin, doxycycline, vancomycin, ciprofloxacin, and erythromycin.[34]

In SJS and TEN, the interval between the drug exposure and the onset of the cutaneous reaction is usually 4–28 days. Prodromal symptoms such as headache, rhinitis, cough, weakness, sore throat, and a burning sensation in the eyes are generally observed.[2] Skin pain is an alerting finding for epidermal necrosis. Skin involvement usually starts from the trunk, face, and proximal extremities. Initial cutaneous lesions are atypical target lesions, usually seen as dark red purpuric macules. Nikolsky's sign, an indicator of epidermal necrosis, may be positive. Flaccid bullae, erythema, and peeled skin can be seen.[3] The severity of illness score for TEN (SCORTEN), a prognostic scoring system, should be performed within the first 24 hours of the patient's admission to the hospital [Table 1].[5] Mortality rates can be predicted according to SCORTEN scores [Table 2].[5]

Table 1

Features used in the SCORTEN calculation and their corresponding scores[5]

Features	Points
Age ≥40	1
Concomitant malignancy	1
BUN >28 mg/dL	1
Glucose >252 mg/dL	1
Pulse ≥120/min	1
Serum bicarbonate level <20 mmol/L	1
Body surface area involved >10%	1

SCORTEN: Severity-of-illness score for toxic epidermal necrolysis.

Table 2

Estimated mortality rates based on SCORTEN scores[5]

SCORTEN	Mortality %
0-1	3.2
2	12.1
3	35.3
4	58.3
≥5	90

SCORTEN: Severity-of-illness score for toxic epidermal necrolysis.

A multidisciplinary approach is necessary, especially in the acute phase of treatment.[12] There is no definite accepted treatment because no randomized controlled trial is possible for SJS/TEN. The most important step in the treatment is the rapid detection and discontinuation of the drug and the initiation of supportive therapy. Intensive care follow-up is required in patients with more than 10% of BSA detachment. In addition to skin care, patients should be monitored and treated in terms of accompanying thermoregulatory dysfunction, fluid loss, electrolyte imbalance, nutrition, skin, and systemic infection. Immunomodulatory treatment options include systemic steroids, cyclosporine, intravenous immunoglobulin (IVIG), and tumor necrosis factor-alpha (TNF-alpha) inhibitors.[134]

Materials and Methods

We retrospectively evaluated the records of 33 patients with SJS and TEN who were examined or hospitalized in our clinic between January 2008 and June 2019. Age, sex, the presence of additional disease, time of admission to our hospital, presence of suspicious drugs, skin findings, mucosal and ocular involvement, SCORTEN scores in patients with TEN, corticosteroid (CS), IVIG, and antibiotic use, history of transfer to the intensive care unit (ICU), complications, and time of death or discharge were noted.

The Local Ethics Committee approved the study (protocol 19-7T/64). Written informed consent was obtained from the patients for use of the images. Statistical analyses were performed using the IBM SPSS 25.0 software. Descriptive continuous variables are expressed using mean, standard deviation, maximum and minimum, and categorical variables are expressed as frequency and percent distribution values. Survival time and functions were estimated using Kaplan–Meier survival analysis. Multivariate analyse was performed with Cox regression analysis. Differences between the groups were evaluated using the log-rank test. Multivariate analyse was performed with cox regression analysis.

Results

Of the 33 patients, 11 (33.3%) had SJS, 3 (9.1%) had SJS/TEN overlap, and 19 (57.6%) had TEN. Twenty (60.6%) patients were female. The mean age was 60.9 ± 21 (range, 9–91) years. There was no significant relationship between sex and survival (P = 0.391). Twenty-seven patients were aged over 40 years. No significant relationship was found between age above or below 40 years and survival (P = 0.850). The most common comorbidities are shown in Table 3. The relationships between frequently accompanying diseases, presence of malignancy, and immunosuppression and survival, were not significant [Table 3].

Table 3

Frequency and P values of concomitant diseases

Concomitant diseases	Number (n)	Percent (%)	Relationship with survival (P)
Hypertension	11	33.3	0.173
Diabetes mellitus	8	24.2	0.234
Malignancy (breast, bladder, ovarian, prostate carcinoma, lymphoma, and glioblastoma multiforme)	6	18.2	0.154
Thyroid dysfunction	4	12.1	0.907
Chronic renal failure	4	12.1	0.083
Immunosuppression	4	12.1	0.512
Gout	3	9.1
Chronic heart failure	2	6.1
Chronic arterial disease	2	6.1
Pneumonia	2	6.1
Epilepsy	2	6.1
Depression	2	6.1

Nineteen (57.6%) patients had one, and 13 patients (39.4%) had more than one suspected drug in their medical history. Antibiotics were the most common reaction-causing drugs, followed by allopurinol. The most suspected drugs that caused SJS and TEN in the entire group are shown in Table 4. Nivolumab, methenolone, phenytoin sodium, esomeprazole, cyclophosphamide, ganciclovir, and bicalutamide were the other causative agents. The relationship between survival and the use of the most common causative drugs was not significant [Table 4].

Table 4

Distribution and P values of suspicious drugs

Suspicious drug	Number (n)	Percentage (%)	Relationship with survival (P)
Antibiotics	15	45.5	0.161
Cephalosporin	5	15.2
Amoxicillin clavulonate	4	12.1
Quinolone	3	9.1
Macrolides	2	6.1
Rifampicin	1	3
Doxycycline	1	3
Gentamicin	1	3
Allopurinol	7	21.2	0.478
NSAIDs (naproxen, dexketoprofen, lornoxicam, diclofenac)	6	18.2	0.675

NSAIDs=nonsteroidal anti-inflammatory drugs

After the onset of symptoms, the mean interval to admission to the hospital was 5 ± 3.6 (range, 1–15) days. The relationship between prolonged interval to admission and survival was not significant (P = 0.187). The mean length of hospital stay (LOS) was 17.3 ± 9.5 days. There was no significant relationship between LOS and survival (P = 0.158).

SCORTEN was calculated in 15 of 19 patients with TEN. The SCORTEN scores were two in five (33.33%) patients, three in four (26.6%) patients, four in five (33.3%) patients, and five in one (6.6%) patient. Data of skin findings were present in 32 patients. The prevalence of skin findings and oral mucosa and ocular involvement are summarized in Table 5. Skin lesions of patients with SJS and TEN were shown in Figures 1 and 2. The relationship between having particular lesions, mucosal and ocular involvement, and survival was not significant [Table 5].

Table 5

Frequency and P values of skin lesions and mucosal involvement

Cutaneous and mucous membrane findings	Number (n)	Percentage (%)	Relationship with survival (P)
Maculopapular lesions	23	69.7	0.530
Exulceration	21	63.6	0.695
Vesicle-blister	19	57.6	0.526
Target lesion	5	15.2	0.739
Purpura	4	12.1	0.497
Oral mucosa involvement	29	87.9	0.421
Ocular involvement	18	54.5	0.164

Figure 1

(a) Purpura, hemorrhagic crusts and (b) detached skin in two patients with TEN. (TEN: toxic epidermal necrolysis)

Figure 2

(a) Hemorrhagic crust on the upper lip, atypical target lesions, vesicle and blister, (b) atypical target lesions, (c-d) maculopapular lesions and exulceration in patients with SJS (SJS: Stevens–Johnson syndrome)

Management protocol for SJS and TEN is shown as a chart [Figure 3]. All patients had skin dressing with gauze containing mupirocin or fusidic acid ointment. Systemic CSs were started in 32 patients. The dose and duration data of CS in two patients were missing. One patient did not receive CS due to concomitant active pneumonia. The duration of CS use ranged from 1 to 53 days; the mean duration was 19.03 ± 13.2.. Multivariate analyse of the relationship between survival and SCORTEN, having maculopapular skin lesion, having exulceration, ocular or oral mucosal involvement, admission time to hospital and treatment type (CS or CS+IVIG) was performed with cox regression analysis. The relationship between survival and SCORTEN, having maculopapular skin lesion, having exulceration, ocular or oral mucosal involvement, admission time to hospital and treatment type (CS or CS+IVIG) was not significant (p=1.00).

Figure 3

A chart for management procedure for TEN. (TEN: toxic epidermal necrolysis, CSs: corticosteroids, IVIG: Intravenous immunoglobulin)

The relationship between survival and SCORTEN, having maculopapular skin lesion, having exulceration, ocular or oral mucosal involvement, admission time to hospital and treatment type (CS or CS+IVIG) was not significant (p=1.00).

Sixteen of 22 patients with SJS/TEN overlap and TEN had IVIG treatment. One patient's dose and duration data were missing. The other 15 patients had IVIG within the first 6 days of hospitalization. Eight patients received 3 g/kg, six patients received 2 g/kg, and one patient received 1 g/kg IVIG. Of the 15 patients for whom SCORTEN was calculated, 10 patients received IVIG and five did not. Mortality estimated in the non-IVIG group according to SCORTEN value was 1.764; none of the patients died. The estimated mortality value in the IVIG group was 4.073, and seven patients died.

During the follow-up, 24 (72.7%) patients received antibiotics. There was no significant relationship between antibiotic use and survival (P = 0.502). Twelve (36.4%) patients had intensive care hospitalization, which was found to have a negative effect on survival (P = 0.003).

Twenty-two (66.7%) patients had complications during follow-up. There was no significant relationship between complication development and survival (P = 0.280). The most common complications were sepsis (n = 7), impaired renal function tests and electrolyte imbalance (n = 7), pneumonia and respiratory disorders (n = 3), and blood glucose elevation (n = 3). Ten (30.3%) patients died during the follow-up, eight of whom had TEN and two had SJS. The relationship between having SJS, SJS/TEN overlap, and TEN and survival was not significant.

Discussion

In this retrospective study, epidemiologic data, causative drugs, complications, and survival rates of patients with SJS/TEN were investigated. Thirty-three patients were included in the study. The TEN, SJS/TEN overlap, and SJS ratios were 57.6%, 9.1%, and 33.3%, respectively. In the literature, the TEN ratio ranges between 54% and 80.3%, the SJS/TEN overlap ratio ranges between 2.4% and 39%, and the SJS ratio ranges between 7% and 36.5%.[6-12]

A female predominance (60.6%) was detected in this study. In the literature, studies mostly report female predominance, ranging between 52.7% and 62%[678101113141516] fewer studies report a male predominance.[91217] Lim et al.[11] reported that female patients had worse prognoses and higher mortality rates. We did not find any significant relationship between sex and survival.

The mean age of our patients was 60.9 years. In other studies, the mean age ranged from 35.9 to 66 years.[678910111213141718] This difference is thought to be due to the presence or absence of pediatric patients. We had one patient aged under 18 years.

SCORTEN scores of our patients ranged between 2 and 5 and the mean value was 3.1. The mean SCORTEN was reported as 2.4,[7] 1.9,[8] 2.52,[19] and 2.9[20] (in the TEN group) in the other studies. The reason for the higher SCORTEN value in our study may be because SCORTEN was calculated only in patients with TEN and that the patients in our study group had more severe diseases.

About 39.4% of patients in this study had more than one suspected drug in their history. Medieros et al.[6] reported 47.5% of their patients had more than one suspected drug. We found antibiotics as suspicious drugs in 45% of the patients. This rate was reported between 39% and 56.7% in other studies.[6781013] Antiepileptic drugs (AEDs) were second in four studies,[67818] whereas we detected only one patient exposed to an AED. Firuz et al.[17] reported that more patients who developed TEN due to allopurinol died than those with TEN caused by other drugs. We had seven patients whose suspected drug was allopurinol, and only one patient died in the follow-up. The overall mortality rate was 30.3% (10/33); 42.1% (8/19) in the TEN group, 18.2% (2/11) in the SJS group, and 0 in the SJS/TEN overlap group in our study. Among the two deceased patients in the SJS group, one had SJS while he was hospitalized for bladder carcinoma, underwent routine hemodialysis, and had inotropic support due to hypotension. He had a poor general condition and underwent concomitant radiotherapy while he had the diagnosis of SJS. The other patient underwent liver transplantation 8 months ago, was receiving immunosuppressive therapy, and had bicytopenia on arrival.

Mortality rates in the literature ranged between 9.5 and 38%.[78910111217182021] Mortality rates in TEN groups ranged between 16.7% and 71.4% [Table 6].[67131519] SJS/TEN mortality is known to be lower in children than in adults.[22] We think the factors that affect the variability of mortality rates in the literature are due to the nature of the disease; the studies can only be performed retrospectively; some studies are conducted in burns centers, whereas others not; variability in the numbers of patients; and variability in the distribution of pediatric and adult patients.

Table 6

Studies evaluating SJS/TEN mortality in the literature and their features

Studies	Mortality %	TEN mortality %	No	Mean age	Burn center	Country	Treatment	Duration (years)	Mean SCORTEN
Our study	30.3	42.1	33 (1 ped)	60.9	−	Turkey	ST/CS/IVIG	11	3.1
Nizamoglu et al.[20]	9.5 (no ped death)	NA	42 (19 ped)	37.3	+	UK	Aggressive wound care/IVIG	12	3.1
McCullough et al.[23]	10	NA	40	45.9	+	USA	ST/IVIG	15	2.06
Zhang et al.[9]	12.5	21.4	48 (6 ped)	49.2	±	USA	ST/IVIG/CYC-A	10	NA
Carrasquillo et al.[10]	13.8	NA	30 (6 ped)	35.9	−	Porto Rico	ST/CS/IVIG	11	1.9
Chan et al.[11]	19	NA	42	56	+	Australia	ST/CS/IVIG	10	NA
Lim et al.[13]	23.7	24.5	76	54.3	+	USA	ST/CS/IVIG	11	NA
Firoz et al.[19]	29.2	NA	82	45.1	+	USA	ST/IVIG/CYC-A	5	NA
Cartotto et al.[14]	29.5	NA	61	51	+	Canada	ST/debridement/IVIG	11	NA
Weinand et al.[12]	38	NA	72	63	+	Germany	ST/blister debridement	27	NA
Medeiros et al.[8]	NA	16.7	41	43	−	Brazil	ST/CS/IVIG/CYC-A	10
Lalosevic et al.[17]	NA	17.6	38 (8 ped)	42.1	−	Serbia	ST/CS/IVIG	20	NA
Krajewski et al.[21]	NA	52	21	53.2	+	Poland	Plasmapheresis/IVIG	10	2.5
Barrerio et al.[15]	NA	71.4	21	66	+	Portugal	ST/IVIG	12	NA
Papp et al.[22]	20.9	35	67	55.8	±	Canada	ST/CS/IVIG/CYC-A	10	2.9 (in TEN)

ST: supportive treatment, ped: pediatric; CS: corticosteroid; IVIG: intravenous immunoglobulin; TEN: toxic epidermal necrolysis; SJS: Stevens-Johnson syndrome; CYC-A: Cyclosporine A

The most common comorbidities in our study were hypertension (HT), diabetes mellitus (DM), and renal failure. In the other studies, the most reported comorbidities were HT, DM, asthma, epilepsy, respiratory distress, and renal failure.[8913] HIV infection, which is a risk factor for SJS and TEN, was not present in our study. Malignancy, defined as another risk factor, was seen in 18.2% of patients (n = 6). Four of these patients died (TEN n = 3, SJS n = 1). In the study of Chan et al.,[9] the ratio of patients with active malignancy was 3%.

The most common complications were sepsis (21.2%) and impaired renal function (21.2%). In the study of Barreiro et al.,[13] sepsis (53%) was most common, followed by renal failure (23%). The sepsis development rate was found as 14.6% by Zhang et al.,[7] 35.3% by Weinand et al.,[10] and 47.1% by Carrasquillo et al.[8]

In our study, 12 patients (36.4%) had intensive care hospitalization, which was found to have a negative effect on survival. Intensive care unit transferals were reported as 43.3%[10] and 64% in the literature.[9]

The efficacy of immunomodulatory treatments in the treatment of TEN is still controversial. Different studies claim that CSs have positive and negative effects on survival in the treatment of SJS/TEN.[23] There are contradictory views about whether early CS onset inhibits inflammation and increases the risk of sepsis.[4] In our study, 97% of the patients received CS treatment. The mechanism of action of IVIG in the treatment of SJS/TEN is inhibition of Fas-Fas ligand interaction and apoptosis. Studies have shown that IVIG alone or in combination with CS is effective[8924] and ineffective in the treatment of TEN.[1317] Also, increased mortality was reported in patients receiving IVIG.[20] In a meta-analysis performed in 2017, the authors stated that the findings did not support the use of IVIG in the treatment of SJS/TEN.[25] Almost half (48.5%) of our patients received IVIG treatment. Other studies in the literature reported IVIG use rates as 28%,[17] 56.7%,[8] and 70%.[9]

The characteristics of the study were summarized in Table 7. In conclusion, antibiotics were the most common cause of SJS and TEN in our study, similar to the literature. However, AEDs were more commonly identified as causative drugs in SJS and TEN in previous studies, but we found AEDs as suspicious in only one patient. ICU stay negatively affected survival. This may be due to patients having a worse general condition being monitored in ICUs wherein monitoring increases the risk of hospital infections. The rate of accompanying malignancy was higher in our patients than in the literature. Our hospital is a regional tertiary healthcare facility that receives referrals for patients with malignancy, which accounts for the high malignancy rate in our study group. This had a negative effect on mortality. Unfortunately, due to the nature of the disease, it is not possible to conduct randomized controlled trials, thus contradictory results regarding immunomodulatory treatments in SJS and TEN have been reported. However, there is a consensus that immediate discontinuation of the suspected drug, proper skin care, and timely supportive treatment is lifesaving. Therefore, a multidisciplinary approach is required after rapid diagnosis.

Table 7

Summary of characteristics of the study

Characteristics of the study		P
Female ratio	60.6%	0.391
Mean age	60.9	-
The most accompanying disease	HT	0.173
The most causative drug	Antibiotics	0.161
The most seen lesion type	Maculopapular lesions	0.530
The most complications	Sepsis, impaired renal function test, and electrolyte imbalance	-
Negative correlation on survival	ICU stay	0.003

HT: hypertension, ICU: Intensive care unit

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.