Stevens-Johnson s Syndrome / Toxic Epidermal Necrolysis: An update Robert G. Micheletti, MD Assistant Professor of Dermatology and Medicine Director, Cutaneous Vasculitis Clinic, Penn Vasculitis Center Co-Director, Inpatient Consult Service University of Pennsylvania
Presentation Prodrome of fever, flu-like symptoms x 1-3 days Photophobia, conjunctival itching, dysphagia, skin tenderness Typically two or more mucous membranes Dusky atypical target lesions Denuded skin / epidermal detachment SJS < 10% BSA TEN > 30% BSA 1
Etiology and Pathogenesis Overall incidence of SJS/TEN 2-13 per million per year Risk factors include immune dysregulation: HIV infection 100-fold increase Autoimmune disease 50-fold increase in SLE patients Active malignancy (particularly hematologic) 30-to-60-fold increase J Invest Dermatol. 2016;136(7):1387-97. Br J Dermatol. 2012 Mar;166(3):575-600. Am J Clin Dermatol. 2012 Feb;13(1):49-54. 2
Etiology and Pathogenesis Genetic factors HLA-B*15:02 (carbamazepine; OR 79.84, Asian)* HLA-A*31:01 (carbamazepine; all ethnic groups)* HLA-B*58:01 (allopurinol) *Pre-screening of HLA type recommended prior to starting carbamazepine Polymorphisms in Cyt P450 and other factors associated with decreased medication clearance N Engl J Med. 2011;364(12):1126. JAMA Dermatol. 2013;149(9):1025. J Dermatol Sci. 2014 Feb;73(2):101-9. 3
Etiology and Pathogenesis Drug-specific CD8 T-cells and natural killer (NK) cells are the major inducers of keratinocyte apoptosis Drugs stimulate the immune system by binding to MHC-I and T- cell receptors, resulting in clonal expansion of cytotoxic T-cells These cells kill keratinocytes directly and indirectly via release of cytotoxic mediators: Soluble Fas ligand Perforin / granzyme Tumor necrosis factor Granulysin J Allergy Clin Immunol. 2011 Dec;128(6):1266-1276. 4
Etiology and Pathogenesis Granulysin Produced and secreted by CD8 and NK cells Highly expressed in SJS/TEN Levels in blister fluid correlate with disease severity Reproducible cytotoxic effect on keratinocytes Thought responsible for apoptosis in SJS/TEN *Being studied as a diagnostic marker and treatment target for SJS/TEN Ann Intern Med. 2009;151(7):514. J Allergy Clin Immunol. 2008;122(5):992. Curr Opin Allergy Clin Immunol. 2013 Aug;13(4):330-6. Curr Opin Allergy Clin Immunol. 2015 Aug;15(4):294-9. 5
Etiology and Pathogenesis Medications are the most common cause of SJS/TEN Typically within 1-3 weeks (avg = 14 days) Unlikely after first 8 weeks of treatment Allopurinol* Anticonvulsants Sulfa antibiotics Nevirapine Oxicam NSAIDs *Most common in EuroSCAR (17.4%) J Am Acad Dermatol. 2008;58(1):25. 6
Etiology and Pathogenesis US Derm Hospitalist Study (377 adult patients) Most common cause of SJS/TEN, in 89.7%, was medication reaction, including common culprit drugs: Class of medications N=338 Antibiotics Trimethoprim/sulfamethoxazole β-lactam antibiotics Fluoroquinolones Antiepileptic/Mood stabilizers Lamotrigine Phenytoin Carbamazepine 165 (48.8%) 89 (26.3%) 42 (12.4%) 12 (3.6%) 83 (23.7%) 32 (9.5%) 30 (8.9%) 7 (2.1%) Allopurinol v 29 (8.6%) NSAIDs 18 (5.3%) v v v Micheletti et al, Society of Dermatology Hospitalists (submitted) 7
Etiology and Pathogenesis Derm Hospitalist Study: All patients with lamotrigine-induced SJS/TEN (n = 29) survived to discharge Exposure to phenytoin increased the risk of death (RR 3.82 (0.99-14.69)) High mortality among those receiving pip/tazo (Zosyn) (73%) drug or underlying disease state? For the most part, the specific drug trigger does not contribute significantly to mortality risk in SJS/TEN 8
Etiology and Pathogenesis Mycoplasma SJS trigger common in children Significant mucosal but minimal skin involvement Tend to have better outcomes Now recognized as a distinct entity, termed Mycoplasma-induced rash and mucositis (MIRM) Derm Hospitalist Study: 2% of cases attributed to Mycoplasma; mortality among infectious cases overall 8% J Am Acad Dermatol. 2015 Feb;72(2):239-45. J Eur Acad Dermatol Venereol. 2015 Mar;29(3):595-8. 9
Rapid Diagnosis a medical emergency When SJS/TEN is clinically obvious, institute therapy even if histologic confirmation is pending However, I do always perform a biopsy for completeness In one study, 1/3 of those biopsied received an alternate diagnosis based on histology Burns. 2016 Feb 1. [Epub ahead of print]. 10
SJS Mimickers Transferred from an OSH, where he was diagnosed with SJS and treated in a burn unit Ultimate diagnosis = severe cutaneous lupus 11
SJS Mimickers Transferred from an OSH to our ICU for treatment of SJS/TEN; received IVIG Ultimate diagnosis = epidermolysis bullosa acquisita vs anti-epiligrin cicatricial pemphigoid 12
SJS Mimickers Patient with SLE on chronic immunosuppression Acute skin tenderness and widespread denuded skin concerning for SJS/TEN SSSS clinical pic Diagnosed with Staph scalded skin syndrome 13
SJS Mimickers Patient with acute leukemia 2-3 weeks s/p stem cell transplant Developed acute dusky erythema, erosions, mucositis, conjunctivitis SJS/TEN vs grade IV acute GVHD 14
SJS Mimickers Paraneoplastic pemphigus 15
Rapid Diagnosis a medical emergency When the diagnosis is in doubt, biopsy may be required to guide management Histologic appearance of SJS: Keratinocyte necrosis ranging from partial to full thickness (established lesions) Scant perivascular lymphohistiocytic inflammation 16
Rapid Diagnosis In our hospital, preliminary results of standard biopsy are available by the end of the following business day Alternative options: Frozen section biopsy Tzanck smear Jelly roll *Frozen section biopsy: Results typically available within 20-30 minutes Take the time to discuss with pathology and determine the mechanism for this in your hospital 17
Rapid Diagnosis Jelly roll The blister roof can be submitted for frozen section Reveals full-thickness epidermal necrosis 18
Data-Driven Management Published overall mortality rate for SJS/TEN is 20-25% RegiSCAR cohort 6-week mortality: SJS = 12% SJS/TEN = 29% TEN = 46% Ther Adv Drug Saf. 2011;2(3):87. J Invest Dermatol. 2013 May;133(5):1197-204. 19
Data-Driven Management How do we give our patients the best chance for survival? What management is supported by the available evidence? 20
Data-Driven Management Stop the culprit drug Earlier withdrawal = better prognosis OR of death = 0.69 for each day sooner OR of death = 4.9 for drugs with long half lives Identification of the causative agent is a non-trivial exercise Causality assessment tools (e.g., ALDEN) may help but have problems with reliability and agreement No substitute for systematic approach by an expert (dermatologist) Arch Dermatol. 2000;136(3):323. Clin Pharmacol Ther. 2017;101:S5-S99. 21
Data-Driven Management Transfer the patient to an ICU or burn unit In one study, overall mortality = 32% Mortality of those transferred after > 1 week = 51% *Patients with mild SJS or SCORTEN of 0 or 1 should still be cared for in an ICU Mild SJS is a designation best made in retrospect; patients may progress rapidly from mild presentations Published guidelines suggest initial management of SJS/TEN patients in specialized centers due to risk of rapid progression regardless of initial % BSA involved Burns. 2016;42(4):836-43. J Burn Care Rehabil. 2002;23(2):87. 22
Data-Driven Management The level of care is a surrogate for supportive care: Fluids Electrolytes Temperature Nutrition Pain control Airway maintenance Wound care Infection surveillance Ophthalmology Gynecology Urology Hematology Role of dermatology Close diagnosis, identification of culprit drug Role Cleaning, Vaseline, Topical observation; Bacteremia of dermatology lubrication, steroids, non-stick in may as team steroid vaginal 27%: gauze require May Urethritis, develop 2/3-3/4 Parkland formula quarterback drops Enteral, profound urinary not parenteral retention neutropenia Staph, Amniotic Pseudomonas, Clean dilator, prolonged Critical membranes, / sterile Vaseline intubation role in patient Enterobacteriaceae technique scleral gauze advocacy spacers Medicine (Baltimore). 2010 Jan;89(1):28-36. J Am Acad Dermatol. 2013 Aug;69(2):187.e1-187.e16. 23
Data-Driven Management Despite this, there is wide variability in supportive and systemic care: Survey of 102 burn centers showed variability in consultation of other services: Ophthalmology (66%) Dermatology (47%) Gynecology (13%) Burns. 2016 Jun;42(4):830-5. 24
Data-Driven Management ICU versus Burn Center: Burn centers are highly skilled in wound care and fluid management, but do they have a dermatologist? Burn center may be less familiar with management of complex medical comorbidities Patients who are post-transplant, have autoimmune disease, etc. require specialized knowledge and care Dermatology hospitalists may be best positioned to provide the differential diagnosis, wound care, and advocacy patients need Both are better than non-specialized ward, but supportive care should be more homogeneous and studied prospectively J Am Acad Dermatol. 2014 Jul;71(1):195-6. 25
Latest Evidence for Pharmacotherapies Besides supportive care, published literature has not consistently shown benefit from the use of any particular systemic therapy for SJS/TEN Case series and retrospective cohorts are limited by confounding factors, low numbers, and inherent biases that make drawing conclusions from existing data difficult 26
Latest Evidence for Pharmacotherapies Systemic Corticosteroids: Early observational studies indicated a higher complication and mortality rate EuroSCAR cohort suggested possible benefit to early steroids, though not statistically significant RegiSCAR cohort and systematic review: no mortality advantage over supportive care alone J Am Coll Surg. 1995;180(3):273 Ther Adv Drug Saf. 2011;2(3):87. J Am Acad Dermatol. 2008;58(1):33. J Invest Dermatol. 2013 May;133(5):1197-204. 27
Latest Evidence for Pharmacotherapies Systemic Corticosteroids: More recently, a meta-analysis of 96 studies in JAMA Dermatology, suggested there is a survival benefit While these results had limitations and were only marginally statistically significant, corticosteroids were felt to be promising On balance, despite concern for increased infection, early use of systemic steroids may be beneficial JAMA Dermatol. 2017;153(6):514-22. 28
Latest Evidence for Pharmacotherapies Intravenous Immunoglobulin: Data are limited and conflicting Thought to inhibit Fas FasL binding, but this is no longer considered the primary mediator of SJS Drugs. 2005;65(15):2085. Science. 1998;282(5388):490. Arch Dermatol. 2003;139(1):26. Arch Dermatol. 2003;139(1):39. Br J Dermatol. 2012 Aug;167(2):424-32. 29
Latest Evidence for Pharmacotherapies Intravenous Immunoglobulin: Case series and a systematic review suggested highdose, early IVIG (>2g/kg total) improves survival However, a number of reviews, including the recent meta analysis from 2017, have shown no mortality benefit Drugs. 2005;65(15):2085. Science. 1998;282(5388):490. Arch Dermatol. 2003;139(1):26. Arch Dermatol. 2003;139(1):39. Br J Dermatol. 2012 Aug;167(2):424-32. 30
Latest Evidence for Pharmacotherapies Intravenous Immunoglobulin: For these reasons, IVIG has fallen out of favor in Europe However, it continues to be an agent of choice for SJS/TEN among US dermatology hospitalists Br J Dermatol. 2015 Nov;173(5):1250-4. 31
Latest Evidence for Pharmacotherapies A survey of 131 US providers (academic dermatologists and burn centers) reported that: Majority do not use systemic steroids More likely to use IVIG if more severe disease majority use IVIG for SJS/TEN overlap and TEN Those who see SJS/TEN more frequently are more likely to use IVIG regardless of severity J Am Acad Dermatol. 2016 Feb;74(2):379-80. 32
US Derm Hospitalist Study (377 adult patients) In our cohort managed by inpatient dermatologists, IVIG was used extensively, alone or in combination (55.6% of patients) Patients receiving IVIG had more severe disease at presentation: Higher median BSA involvement (30% vs 12%, p-value<0.01) Higher rate of TEN (28.0% vs 18.7%, p-value=0.01) Higher rates of severe ocular, oral, and genitourinary involvement (all p-values <0.05) compared to other treatment groups Micheletti et al, Society of Dermatology Hospitalists (submitted) 33
Latest Evidence for Pharmacotherapies Intravenous Immunoglobulin: In summary, no high quality evidence supports the use of IVIG in SJS/TEN However, those who receive it appear to be sicker to start with, and the effects of this and other confounders may not be completely accounted for by SCORTEN-predicted mortality Early in the disease course, a dose of 1g/kg/day x 3-4 days IVIG can be considered Potential side effects include clot / hyperviscosity Side effect profile more favorable than high-dose steroids Br J Dermatol. 2015 Nov;173(5):1250-4. 34
Latest Evidence for Pharmacotherapies Cyclosporine: Anti-apoptotic, inhibits T-cells, including CD8 T-cells Dose 3 to 5 mg/kg/day, tapered over one month Supported mostly by small, retrospective, and uncontrolled case series J Trauma. 2000;48(3):473. Cutis. 2011 Jan;87(1):24-9. J Am Acad Dermatol 2012;67(4):630-5. 35
Latest Evidence for Pharmacotherapies Cyclosporine: Retrospective review of 64 patients Outcome of 15 who received cyclosporine (SMR = 0.43) better than 50 who received IVIG (SMR = 1.43) IVIG dose and timing varied widely Retrospective review of 44 patients Among 24 who received cyclosporine, SMR = 0.42 Open, Phase II trial of 29 patients: SCORTEN predicted mortality 2.75; actual = 0 Br J Dermatol. 2010;163(4):847. J Am Acad Dermatol. 2014;71(5):941. J Am Acad Dermatol. 2017;76(1):106-113 36
Latest Evidence for Pharmacotherapies Cyclosporine: More recently, two 2017 studies suggested a mortality benefit of cyclosporine via meta-analysis and retrospective case series (49 patients) Increasing favor in Europe, but the actual number of treated patients remains small, and the drug has important side effects (renal, immunosuppression) JAMA Dermatol. 2017;153(6):514-22. J Invest Dermatol. 2017 Oct;137(10):2092-100. 37
Latest Evidence for Pharmacotherapies Cyclosporine: Letter-to-the-editor: Existing studies either exclude patients with renal insufficiency and other known SJS/TEN mortality risk factors altogether or do not report them In our cohort, renal failure was the greatest mortality RF Recent retrospective cohort study of 174 patients: No significant benefit from cyclosporine in 95 patients (versus 79 supportive care only) Acute renal failure more common in treatment group J Invest Dermatol. 2018 Jan 21 [Epub ahead of print] 38
Latest Evidence for Pharmacotherapies Cyclosporine: In summary, cyclosporine is certainly deserving of further study, but there are reasons to temper enthusiasm Early use of cyclosporine in young, otherwise healthy patients at a dose of 3-5mg/kg/day is a reasonable treatment option, but we do not have enough data to recommend this in sicker patients JAMA Dermatol. 2017;153(6):514-22. 39
Latest Evidence for Pharmacotherapies Tumor Necrosis Factor (TNF) inhibitors: Thalidomide: Only double-blind, randomized, placebo-controlled TEN study Trial stopped early due to excess mortality (10/12 died, vs 3/10 in placebo) Suggested reason for caution with respect to TNF inhibitors for SJS/TEN Lancet. 1998;352(9140):1586. 40
Latest Evidence for Pharmacotherapies Tumor Necrosis Factor (TNF) inhibitors: Etanercept: Case series in 8/2014 JAAD One 50mg injection 10 patients with good outcomes; no controls Infliximab: One 5mg/kg infusion Case reports only Pediatr Dermatol. 2014;31(4):532. J Am Acad Dermatol. 2014;71(2):278. 41
Latest Evidence for Pharmacotherapies Tumor Necrosis Factor (TNF) inhibitors: 96-patient randomized trial of etanercept versus corticosteroids SMR = 0.47 among those receiving etanercept (0.80 with corticosteroids) Decreased time to skin healing and lower rates of complications compared to corticosteroid group J Clin Invest. 2018 Feb 5. [Epub ahead of print] 42
Latest Evidence for Pharmacotherapies Tumor Necrosis Factor (TNF) inhibitors: Recent experience with etanercept is promising 25mg or 50mg etanercept twice weekly until skin lesions healed J Clin Invest. 2018 Feb 5. [Epub ahead of print] 43
US Derm Hospitalist Study (377 adult patients) Overall and subgroup mortality less than predicted by SCORTEN Predicted mortality for the overall cohort was 21.1%, vs 14.7% actual Adjusting for SCORTEN, mortality was lower in the steroid + IVIG group compared to IVIG or steroids alone and to supportive care only *Bias toward using IVIG to treat patients with more severe SJS/TEN *Only 5 patients treated with cyclosporine or etanercept In-hospital mortality SCORTEN predicted mortality, N (%) Overall N=368 IVIG only N=92 Steroid only N=116 IVIG + Steroid N=54 Supportive care N=117 77.7 (21.1) 21.6 (23.5) 20.8 (17.8) 11.6 (21.2) 22.7 (19.4) Observed mortality 54 (14.7) 17 (18.5) 15 (12.9) 6 (10.7) 16 (13.7) SMR (95%CI) 0.70 (0.58, 0.79) 0.79 (0.55, 0.92) 0.72 (0.48, 0.89) 0.52 (0.21, 0.79) 0.70 (0.47, 0.87) Micheletti et al, Society of Dermatology Hospitalists (submitted) 44
US Derm Hospitalist Study (377 adult patients) One of the largest existing SJS/TEN cohorts (largest in N America) Overall survival rate better than reported in the literature and better than predicted by SCORTEN: Published mortality rate 20-25% in our cohort, 14.7% Large systematic review reported SMR 0.82-0.92 in our cohort, the SMR was 0.70 (0.58, 0.79) Ther Adv Drug Saf. 2011;2(3):87. Micheletti et al, Society of Dermatology Hospitalists (submitted). 45
US Derm Hospitalist Study (377 adult patients) Unclear if improved survival reflects excellent supportive care in tertiary centers, the presence of a consulting dermatologist, or inadequacy of SCORTEN as a predictive tool The optimal pharmacologic regimen remains uncertain Some small published series support steroids + IVIG Cyclosporine, etanercept, corticosteroids all reasonable Need better ways to account and adjust for prescriber practices, particularly with respect to IVIG severity bias Additional analysis of the cohort is ongoing J Burn Care Res. 2012 Nov;33(6):e295-308. Indian J Dermatol Venereol Leprol. 2013;79(4):506. Micheletti et al, Society of Dermatology Hospitalists (submitted). 46
Latest Evidence for Pharmacotherapies At our institution, we favor: Admission of all SJS/TEN patients to the MICU At least 1:1 nursing care Daily dermatology and universal ophthalmology and gynecology consultation Early high-dose IVIG (1g/kg/day) x 3-4 days with a goal to try to stop progression (actively re-evaluating) Given as soon as the diagnosis is made (regardless of severity); early Rx may be the key with any treatment 47
Latest Evidence for Pharmacotherapies All retrospective studies and literature reviews are limited by heterogeneous populations, dosing, and confounding factors In general, a lack of consensus regarding the appropriate pharmacologic management of SJS/TEN persists, even among experienced providers 48
Next Steps Research priorities: Better understanding of the genetic and immunologic basis of disease with prospective sample collection Improved risk assessment so the reaction can be avoided Investigation of novel treatments to inhibit granulysin Br J Dermatol. 2015 Nov;173(5):1250-4. 49
Next Steps Research priorities: Systematic evaluation of SJS/TEN sequelae and quality of life Survey study Establish expert opinion standard for supportive care Delphi effort Br J Dermatol. 2016 Aug;175(2):422-4. Br J Dermatol. 2015 Nov;173(5):1250-4. 50
Next Steps Research priorities: Revision of SCORTEN, which may overestimate mortality Improved prognostic model (SMR 0.99) Utilize existing networks to conduct a prospective SJS/TEN study and, ultimately, a randomized controlled trial J Invest Dermatol. 2017;137(5S):S37 51
Summary SJS/TEN is a severe cutaneous adverse drug reactions with published mortality of 20-25% Dermatologists play a critical role in diagnosing this disease, differentiating mimickers, identifying and stopping culprit drugs, and coordinating treatment efforts Much progress still to make with respect to pathogenesis, prevention, and management dermatologists can and should be at the forefront of these efforts 52
The Dermatology Foundation has supported & advanced my career.
Thank you Robert.Micheletti@uphs.upenn.edu University of Pennsylvania