Evaluation of the Factors Influencing Mortality in Patients with Stevens-Johnson Syndrome and Toxic Epidermal Necrolysis: A Multicenter Study of 166 Patients

Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are rare, life-threatening conditions that belong to the same disease spectrum. Clinically, they are characterized by widespread erythematous targetoid plaques accompanied by severe mucosal erosions [1]. Cutaneous lesions exhibit Nikolsky's sign positivity and epidermal detachment. Full-thickness epidermal necrosis is observed in pathological evaluations. The distinction between SJS and TEN is made based on the extent of body surface area involvement; SJS involves < 10% of the total body surface area, SJS-TEN overlap involves 10–30%, and TEN involves > 30% [1]. The severity-of-illness score for toxic epidermal necrolysis (SCORTEN) was first described by Bastuji-Garin et al. to assess the severity and prognosis of the disease [2]. SCORTEN is the most commonly used scale for prognostication in SJS-TEN and its effectiveness as a tool has been confirmed in several studies [3, 4]. The incidence of SJS and TEN is estimated to be 1.2–6 per million and 0.4–1.2 per million, respectively [5, 6], with mortality rates varying but averaging 1–5% for SJS and 25–35% for TEN [5, 6].

The etiopathogenesis of SJS/TEN is not fully understood, but it is primarily triggered by drugs. The most implicated drugs include sulfonamides, anticonvulsant agents, non-steroidal anti-inflammatory drugs, and allopurinol [7]. SJS/TEN is thought to be a type IV hypersensitivity reaction mediated by T cells. One of the accepted hypotheses regarding how drugs lead to the immunological reaction in SJS/TEN is the hapten/prohapten concept. According to this concept, small molecular drugs covalently bind to proteins in the serum, are recognized by certain HLA molecules, and activate T cells to produce an immune response [8, 9].

The cornerstone of treatment includes immediate discontinuation of the offending drug and multidisciplinary supportive care. Supportive care encompasses monitoring renal functions, fluid and electrolyte levels, ensuring adequate nutrition, pain control, maintaining skin integrity, and preventing infections. Treatment may need to be continued in a burn unit or intensive care unit [8]. Ensuring and maintaining enteral nutrition is crucial. Prophylactic antibiotics are not recommended, but appropriate wound care is critical in preventing secondary infections [8]. There is currently no consensus on immunomodulatory treatment for SJS-TEN. Systemic agents used in treatment include systemic steroids, intravenous immunoglobulin (IVIG), etanercept, infliximab, thalidomide, and cyclosporine [10, 11].

The primary objective of this study is to assess the factors influencing mortality among patients with SJS-TEN. Additionally, we aimed to provide a comprehensive description of the demographic and clinical characteristics of the patients included in our study.

The study was conducted with a total of 166 patients, comprising 59 (35.5%) males and 107 (64.5%) females, with ages ranging from 18 to 91 years. The average age was 50.91 ± 21.25 years.

Disease classification was TEN in 50% of cases, SJS in 33.1%, and SJS-TEN overlap in 16.9%. The SCORTEN score within the first 24 h ranged from 0 to 6, with an average of 2.44 ± 1.42 and a median of 2. The median SCORTEN scores were 2 (IQR: 1–2) for patients with SJS, 2 (IQR: 1.25–3) for SJS-TEN overlap, and 3 (IQR: 2–4) for TEN. The SCORTEN score on day 3 also ranged from 0 to 6, with an average of 2.44 ± 1.482 and a median of 2 (Table 1).

Comorbidities were present in 77.1% of cases. The most common comorbidities were coronary artery disease (CAD) with 19.3%, diabetes mellitus with 15.7%, malignancy with 15.1%, chronic renal disease with 13.3%, and epilepsy with 11.4% (Table 2).

When examining the distribution of implicated causes, the most common were antibiotics (41.6%), anti-epileptic drugs (23.5%), allopurinol (17.5%), non-steroidal anti-inflammatory drugs (12.7%), and other drugs (12.7%) (Table 3).

The time between the culprit medication use and symptom onset ranged from 2 to 45 days, with an average of 11.92 ± 8.56 days and a median of 10 days.

Our study included 166 patients with a female/male ratio of 1.81, aligning with the reported female predominance in the literature [12]. The average age was 50.91 ± 21.25 years, similar to Micheletti et al.'s study, which reported an average age of 49 ± 19.2 years [13]. The SCORTEN scores within the first 24 h and on day 3 were similar in our study with a median of 2 but a statistically significant increase in the maximum percentage of skin detachment during follow-up was observed.

One hundred twenty-eight (77.1%) of our patients had comorbidities, with the most common being CAD, diabetes mellitus, malignancy, chronic renal disease, and epilepsy. Manvi et al.'s study found seizure disorder, hyperuricemia/gout, and HIV infection as the most common comorbidities [14]. Schroeder et al. evaluated 28 patients with SJS/TEN in a retrospective study of 11 years of experience in a high-complexity tertiary care hospital; 89.2% of 28 patients were reported to have comorbidities, and cardiovascular comorbidities were more common [15]. Taken together, the data highlight a remarkable proportion of patients to suffer from co-existent disorders, which also necessitate the use of multiple medications.

In our study, antibiotics (41.6%), antiepileptic drugs (23.5%), allopurinol (17.5%), and non-steroidal anti-inflammatory drugs (12.7%) were the most frequently implicated causes. Abulatan et al.'s study compiled drugs associated with SJS-TEN, highlighting antibiotics, especially sulfonamides, as the most commonly associated [12]. Hsu et al. identified antiepileptics, particularly carbamazepine, as most frequently associated with SJS-TEN [16].

Due to the rarity of the disease, most studies on SJS-TEN are retrospective and involve a small number of patients. Only a few prospective studies analyzing the efficacy of specific immunomodulatory treatments have been conducted [4, 6, 17]. The absence of large randomized controlled trials prevents the establishment of a standard pharmacological treatment for SJS-TEN. However, there are publications suggesting that immunomodulatory treatments and combinations could be beneficial because of the immunological nature of the disease [18‐21]. Whether the observed outcomes are attributable to these treatments or the cessation of the triggering drug and supportive care leads to remission remains uncertain.

Nearly all of our patients received supportive care. The most frequently used systemic immunomodulatory treatments were systemic steroids in 84.3% of patients, IVIG in 49.3%, and cyclosporine in 38.6%. Numerous systematic reviews have attempted to standardize immunomodulatory treatments for SJS-TEN or to find the ideal treatment for SJS-TEN [22, 23]. Zimmermann et al.'s meta-analysis in 2017 compiled 96 studies (covering 3248 patients) and evaluated the immunomodulatory and supportive treatments administered. The treatments assessed included supportive care, glucocorticoids, intravenous immunoglobulin, cyclosporine, plasmapheresis, thalidomide, cyclophosphamide, hemoperfusion, tumor necrosis factor inhibitors, and granulocyte colony-stimulating factors. Their analysis reported that only glucocorticoids and cyclosporine appeared promising in improving survival [22]. Tsai et al.'s recent systematic review and network meta-analysis evaluated the effects of systemic immunomodulatory treatments on mortality in SJS-TEN overlap and TEN. Covering 2079 patients across 67 studies, the analysis of 10 studies found that immunomodulatory treatments did not demonstrate superiority in reducing mortality compared to supportive care, and thalidomide increased mortality. Analysis of 11 studies suggested that the combination of corticosteroids and IVIG could be beneficial for survival, with a SCORTEN-based standardized mortality ratio (SMR) of 0.53 (95% CI, 0.31–0.93). Cyclosporine and etanercept were mentioned as promising treatments, but further research is needed [23]. As can be seen, the results obtained from these meta-analyses are not in harmony with each other. Due to the rarity of SJS-TEN and their high morbidity and mortality rates, conducting randomized controlled clinical trials is challenging. This makes it hard to determine a definitive "gold standard" treatment approach [24]. In general, systemic steroids, IVIG, and cyclosporine were used as first-line treatments in our study. The selection of these therapies was guided by the patients' comorbidities (such as diabetes mellitus, renal disease, immune suppression leading to increased risk of infection, chronic hepatitis, and hypertension) as well as laboratory findings (such as positive blood cultures, renal function results, and electrolyte levels). In our comparative analysis of survivors and deceased patients, we found that systemic steroids, IVIG, and cyclosporine treatments did not have an impact on mortality. Contrary to our findings, some studies in the literature have reported positive effects of cyclosporine on mortality [11, 17, 25]. Gilbert et al.'s review on the efficacy and safety of cyclosporine treatment in SJS-TEN suggested that using a daily dose of 3 mg/kg could have positive effects on mortality [11]. Kirchhoff et al., in their retrospective study of 64 patients with SJS/TEN/overlap, reported that cyclosporine might have a positive effect on survival compared to IVIG [25]. Valeyrie-Allanore et al.'s phase II open prospective study on 29 patients with SJS/TEN/overlap treated with cyclosporine found a lower death rate than predicted by SCORTEN, suggesting that cyclosporine could improve survival [17].

Five (3%) of our patients received plasmapheresis treatment. On average, treatment commenced on day 13 ± 6.67. Of the five patients who underwent plasmapheresis, three had a SCORTEN score of 2, one had a score of 4, and one had a score of 5. Two of the three patients with a SCORTEN score of 2 died. Plasmapheresis is a potential treatment option in SJS-TEN therapy, but its efficacy is uncertain. The rationale behind plasmapheresis is its ability to remove toxins, including drugs, drug metabolites, and other cytotoxic agents. However, plasmapheresis has potential disadvantages, such as depleting immunoglobulin levels, which could potentially increase the risk of sepsis [26]. While some studies do not find the use of plasmapheresis in TEN beneficial [27, 28], others support its efficacy [29]. Senda et al.'s study analyzed outcomes of 38 patients treated with plasma exchange (PE) within the first 24 h of hospital admission and 218 patients who did not receive PE, finding no benefit of PE in reducing in-hospital mortality or length of stay [27]. In the study by Krajewski et al., 21 patients who were followed up in a burn center for TEN were treated with plasmapheresis and IVIG. Mortality was observed in 52% of patients. Severe concomitant diseases were observed in the group with mortality. They concluded that plasmapheresis and IVIG treatment should be used only in the group without severe comorbidities; otherwise, it may increase mortality [28]. Han et al.'s prospective observational study randomly applied plasmapheresis to 13 of 28 patients followed for SJS-TEN overlap and TEN, with the remaining 15 receiving systemic treatments other than plasmapheresis. Of the 13 patients who received plasmapheresis treatment, 7 were also receiving corticosteroid or IVIG treatment. This study suggested that plasmapheresis might have advantages over conventional treatments like IVIG and corticosteroids, and its use as monotherapy could be more beneficial [29]. Although the number of the patients receiving plasmapheresis is low in our study, our results do not justify the use of this treatment modality.

In our study, the complication rate was 51.8%, with the most common complications being sepsis (14.5%), intubation (13.9%), acute renal failure (12.7%), urinary tract infection (11.4%), and bacteremia. Micheletti et al.'s comprehensive retrospective study of 377 patients identified significant in-hospital complications of SJS/TEN, including acute renal failure (34.5%), intubation (23.6%), pneumonia (15.1%), urinary tract infection (14.6%), bacteremia (13.3%), sepsis (12.7%), major thromboembolic event/disseminated intravascular coagulation (8.2%), and skin infection (8.0%) [13]. Schroeder et al.'s study found systemic infections as the most common complication, with septic shock occurring in 10.7% of cases [15]. Although 60.8% of our patients used antibiotics, infection was the leading cause of mortality, which underlines the importance of supportive care and avoidance of unnecessary antibiotic use.

Forty (24.1%) of our patients died in hospital. Schroeder et al.'s study found a mortality rate of 17.8% [15]. In our study, the average hospital stay for survivors was 19.98 ± 14.18 days, with a median of 18 days. The hospital stay for those who died ranged from 3 to 76 days, with an average of 21.80 ± 17.87 days and a median of 15 days. Similarly, Micheletti et al.'s study reported an average hospital stay of 16.2 (SD = 16.1) days for discharged patients. The overall hospital stay, including survivors, was 21.9 (SD = 79.9) days [13].

One of the key objectives of our study was to analyze the factors affecting mortality in SJS-TEN. Therefore, we compared the characteristics of cases that resulted in exitus with those who survived. In the multivariate (adjusted) regression analysis, that beyond SCORTEN 2, the risk of mortality increased exponentially with each increment in SCORTEN, reaching an 84-fold increase in mortality at SCORTEN 5–6. These results affirm SCORTEN as a reliable tool for determining disease prognosis and a primary determinant of mortality. Similar to our study, Krajewski et al. reported SCORTEN as an essential element in stratifying the therapeutic process, allowing for the distinction between patients with a high risk of death at the outset and those expected to have a better prognosis [28]. Our findings also seem to confirm the conclusions of a recent review that an underestimation of mortality was found for SCORTEN values ≤ 3 and the opposite for those > 3 (SCORTEN range: 0–7) [3]. Multivariate analysis also revealed that plasmapheresis was associated with a 22-fold higher risk of mortality. Although plasmapheresis is generally implemented as a relatively late-stage rescue therapy, with only the most severe and unresponsive cases selected for this treatment, that was not the case in our study. As this is a retrospective study, we do not have comprehensive information on the criteria used for selecting patients for plasmapheresis. The fact that plasmapheresis was applied to patients with a lower SCORTEN in our study may be attributed to variations in physicians' treatment approaches. However, data from our regression analysis indicate that plasmapheresis is associated with an increased risk of mortality, even independent of SCORTEN. We also consider the possibility that the increased mortality that we determined associated with plasmapheresis could be related to the timing of its administration. Similar to our study, several studies have analyzed mortality risk factors in SJS-TEN [30‐32]. A recent national study by Wasuwanich et al. in the US evaluated potential factors causing an increase in mortality for SJS-TEN [30]. In their univariable model, increased age, diabetes mellitus, congenital urea cycle metabolism disorders, chronic kidney disease, pneumonia, sepsis, and malignancy presence were significantly associated with increased mortality. Being non-Hispanic white race/ethnicity and having conjunctivitis were associated with a decreased risk of mortality. In the multivariate model, increased age, chronic kidney disease, pneumonia, sepsis, and malignant neoplasia were associated with increased mortality, while being non-Hispanic white was associated with decreased mortality risk [30]. Sunaga et al.'s nationwide study in Japan, evaluating data from 489 patients with SJS-TEN, reported higher mortality rates in patients using systemic steroids. The rate of sepsis in the steroid-using group was higher compared to the non-steroid group, though this increase was not statistically significant. The study found that SJS-TEN patients who developed sepsis had a higher mortality rate in later follow-ups compared to those without sepsis [31]. Sekula et al.'s large-scale, multinational study followed 460 patients diagnosed with SJS/TEN for 1 year [32]. The 6-week mortality rate within this cohort was 23% (95% CI 19–27), with mortality continuing to increase after the 6th week, reaching an overall mortality of 34% (95% CI 30–39) at the end of 1 year. Various factors were found to influence mortality: age, severity of the reaction, recent malignancy, pre-existing severe kidney or liver disorders, and recent infections. Among these, severe liver and kidney disorders were identified as independent risk factors for mortality. Additionally, recent infection was also determined to be an independent risk factor for mortality [32].

The primary limitation of our study is its retrospective design. The majority of treatments being used in combination and the fact that almost all patients received supportive therapy alongside make it difficult to evaluate treatment responses, especially in groups with few patients (e.g., etanercept). There might be biases in the selection of different treatment options, and not all subjects received the same doses of systemic treatments, which complicates comparisons. Additionally, the potential for missing or incorrect data in our study, which compiled data over a lengthy period of 10 years, is a concern.

Our study found no impact of systemic corticosteroids, IVIG, and cyclosporine treatments on mortality. We determined that higher SCORTEN (> 2) within the first 24 h and the use of plasmapheresis was related to increased mortality. We believe the data obtained from 12 tertiary care hospitals over a 10-year experience will contribute to the literature. The lack of large prospective, randomized controlled trials regarding SJS-TEN still hinders the establishment of a standard treatment protocol. Until effective treatment protocols are developed and integrated into daily practice, it is recommended to meticulously provide supportive care and to determine the most suitable individualized option for each patient based on the physicians’ experience.