Delayed admission to a specialist referral center for Stevens-Johnson syndrome and toxic epidermal necrolysis is associated with increased mortality: A retrospective cohort study

To the Editor: Stevens-Johnson syndrome (SJS) and Toxic Epidermal Necrolysis (TEN) are rare, severe cutaneous adverse drug reactions. The average mortality across the spectrum is 25%.1Sekula P. Dunant A. Mockenhaupt M. et al.Comprehensive survival analysis of a cohort of patients with Stevens–Johnson syndrome and toxic epidermal necrolysis.J Invest Dermatol. 2013; 133: 1197-1204Abstract Full Text Full Text PDF PubMed Scopus (254) Google Scholar Guidelines recommend that such cases be managed in reference centers.2Creamer D. Walsh S.A. Dziewulski P. et al.UK guidelines for the management of Stevens-Johnson syndrome/toxic epidermal necrolysis in adults 2016.Br J Dermatol. 2016; 174: 1194-1227Crossref PubMed Scopus (127) Google Scholar Our study aims to evaluate the impact of delayed admission to an SJS/TEN referral center on clinical outcomes. A retrospective cohort study was conducted from 2005 to 2018 at the Singapore General Hospital (SGH), the national referral center for SJS/TEN. Outcomes of interest included mortality, the incidence of bacteremia at SGH, ICU admission, and length of hospital stay. Admissions occurring within 4 days of skin blistering according to clinical records were considered early admissions and delayed admissions were defined as admissions occurring 5 days or more after skin blistering. The consensus of the 5-day cut-off was based on 1) a multi-center study in Europe,1Sekula P. Dunant A. Mockenhaupt M. et al.Comprehensive survival analysis of a cohort of patients with Stevens–Johnson syndrome and toxic epidermal necrolysis.J Invest Dermatol. 2013; 133: 1197-1204Abstract Full Text Full Text PDF PubMed Scopus (254) Google Scholar wherein the median admission delay was 5-7 days across several centers and 2) the natural history of SJS/TEN, where the disease evolves from initial blistering to hospitalization in 3 days, then progresses to maximum detachment in another 5 days.3Heng Y.K. Lee H.Y. Roujeau J.C. Epidermal necrolysis: 60 years of errors and advances.Br J Dermatol. 2015; 173: 1250-1254Crossref PubMed Scopus (45) Google Scholar A total of 123 patients with SJS/TEN were included in this study. There were 93 (76%) patients with early admissions and 30 (24%) with delayed admissions. Baseline characteristics are shown in Table I.Table IClinical characteristics for all SJS/TEN patientsClinical characteristicsAll (n = 123)Early admission (n = 93)Late admission (n = 30)P valueMale47 (38%)35 (38%)12 (41%).72Age (years)56 (19)53 (19)67 (20).11Ethnicity.30 Chinese81 (66%)58 (63%)23 (77%) Malay29 (24%)25 (27%)4 (13%) Other13 (11%)10 (11%)3 (10%)Disease.066 SJS45 (37%)39 (42%)6 (20%) SJS/TEN41 (33%)30 (32%)11 (37%) TEN37 (30%)24 (26%)13 (43%)Maximum % BSA involved23 (22)20 (19)33 (27).016SCORTEN2.1 (1.2)2.0 (1.2)2.6 (1.2).013ABCD-102.0 (1.4)1.9 (1.5)2.2 (1.2).14Transfer∗Indicates the number of patients transferred to Singapore General Hospital.57 (46%)36 (39%)21 (70%).003Inciting Drug.47 Penicillin/cephalosporin antibiotic7 (6%)5 (5%)2 (7%) Other antibiotic8 (7%)5 (5%)3 (10%) Sulfonamides10 (8%)6 (7%)4 (13%) Allopurinol19 (15%)15 (16%)4 (13%) Anticonvulsants27 (22%)24 (26%)3 (10%) NSAID5 (4%)5 (5%)0 (0%) Trimethoprim-sulfamethoxazole2 (2%)1 (1%)1 (3%) Other drug27 (22%)18 (19%)9 (30%) Unknown drug1 (0.8%)1 (1%)0 (0%) SJS/TEN not caused by drug17 (14%)13 (14%)4 (13%)Comorbidities Malignancy18 (145%)13 (14%)5 (17%).72 Liver disease4 (3%)3 (3%)1 (3%).98 Renal disease17 (14%)13 (14%)4 (13%).93 AIDS4 (3%)3 (3%)1 (3%).98 Cardiac disease25 (20%)14 (15%)11 (37%).011 Diabetes mellitus19 (15%)15 (16%)4 (13%).71 Autoimmune/connective tissue disease11 (9%)8 (9%)3 (10%).82Charlson Comorbidity Index2.70 (2.69)2.57 (2.72)3.10 (2.63).28Data are presented as count (%) or mean (SD). Bold values are statistically significant.NSAID, Non-steroidal anti-inflammatory drug.∗ Indicates the number of patients transferred to Singapore General Hospital. Open table in a new tab Data are presented as count (%) or mean (SD). Bold values are statistically significant. NSAID, Non-steroidal anti-inflammatory drug. Univariate analyses demonstrated a significant difference in outcomes between early and late admission patients for mortality (14% vs 37%; P = .007), incidence of bacteremia (20% vs 52%; P = .001), rate of ICU admission (14% vs 41%; P = .002), and total length of stay at SGH (16 ± 22 vs 28 ± 23; P = .01). A dose-response relationship was demonstrated between mortality and admission delay, with mortality increasing from 15% in those admitted within 0-3 days, to 19% in those admitted within 4-5 days, and to 33% in those admitted within 6 days. A logistic regression model adjusting for disease classification and transfer status showed similar outcomes after adjustment (predictors SCORTEN and ABCD-10 were collinear with disease classification and excluded from the final model) (Table II). Similar findings were seen during a secondary analysis of patients directly admitted to SGH from the community (n = 66), suggesting that these differences in outcomes were unlikely to be influenced by transfer status.Table IIEffect of admission delay on clinical outcomes adjusted for disease classification and transfer statusOutcomeEffect of admission delayP valueEffect of disease classification∗The reference category was Overlap and TEN for the reporting of OR or Beta coefficient.P valueEffect of transferP valueMortality2.9 (1.0, 8.3).053.4 (1.6, 7.1).0010.64 (0.2, 1.9).44Bacteremia3.7 (1.2, 11.2).026.6 (2.9, 15.1)<.000111 (0.4, 3.4).843ICU admission3.3 (1.1, 9.8).035.8 (2.4, 13.9)<.00010.9 (0.3, 2.9).85Length of stay6.9 (-3.8, 17.7).2011.7 (6.1, 17.2)<.0001-0.3 (-9.7, 9.2).96For length of stay outcome, beta coefficient (95% CI) is reported, all other outcomes are reported as OR (95% CI). Only survivors are included in the analysis for length of hospital stay. Bold values are statistically significant.∗ The reference category was Overlap and TEN for the reporting of OR or Beta coefficient. Open table in a new tab For length of stay outcome, beta coefficient (95% CI) is reported, all other outcomes are reported as OR (95% CI). Only survivors are included in the analysis for length of hospital stay. Bold values are statistically significant. Our findings suggest that early admission to reference centers is associated with better outcomes. Early admission and management may mitigate acute metabolic and septic complications; allow for early identification and withdrawal of culprit medications, which improves outcomes4Garcia-Doval I. LeCleach L. Bocquet H. Otero X.L. Roujeau J.C. Toxic epidermal necrolysis and Stevens-Johnson syndrome: does early withdrawal of causative drugs decrease the risk of death?.Arch Dermatol. 2000; 136: 323-327Crossref PubMed Google Scholar; and permits earlier initiation of immunomodulatory therapy which is most effective during the progressive phase of the disease. Additionally, outcomes have been shown to be superior in centers with higher volumes of SJS/TEN cases.5Traikia C. Hua C. Le Cleach L. et al.Individual-and hospital-level factors associated with epidermal necrolysis mortality: a nationwide multilevel study, France, 2012-2016.Br J Dermatol. 2020; 182: 900-906Crossref PubMed Scopus (5) Google Scholar Limitations, including referral bias, potential confounders of transfer status and disease classification must be acknowledged. However, we adjusted for these in our multi-variate model and a separate analysis was performed, limiting the data to nontransfer cases, which yielded similar outcomes. Establishing a network of reference centers has been advocated.5Traikia C. Hua C. Le Cleach L. et al.Individual-and hospital-level factors associated with epidermal necrolysis mortality: a nationwide multilevel study, France, 2012-2016.Br J Dermatol. 2020; 182: 900-906Crossref PubMed Scopus (5) Google Scholar Our findings provide contemporary evidence supporting the expeditious transfer and management of cases in reference centers. None disclosed.