記住我
Abstract
Background: Stevens–Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are some of the less common cutaneous adverse drug reactions with significant mortality. Objectives: This study was undertaken with the objective of studying the demographics and clinical profile of SJS/TEN and identifying parameters associated with mortality. Materials and Methods: All patients with SJS/TEN over 10 years (2010–2020) were included in the study. Data obtained from in-patient and out-patient records were analysed. Results: A total of 82 patients with SJS/TEN were admitted to our centre over a period of 10 years. Patients with SJS were significantly younger than those with TEN, with a male: female ratio >1 in SJS and <1 in TEN. The most commonly implicated drugs were antiepileptics (n = 29, 35.4%), antibiotics (n = 20, 24.4%). and Non-steroidal antiinflammatory drugs (NSAIDs) (n = 7, 8.5%). The mortality rate in the TEN group was 16% (n = 8). Certain factors such as cutaneous lesions preceding mucosal lesions at onset, high mean Body surface area (BSA) of denudation and a transfer to intensive care unit (ICU) more than 7 days after admission were significantly associated with higher mortality. There was no difference between survivors and deaths in terms of delay in hospitalisation, total disease duration, implicated drug, delay in initiation of therapy, the onset of re-epithelialisation, Severity-of-illness score for TEN (SCORTEN) and total duration of hospital stay. Conclusion: Factors significantly associated with increased mortality in TEN were cutaneous onset of lesions, mean BSA of involvement and transfer to the intensive care unit (ICU) beyond day 7 of admission.
Keywords: Cutaneous adverse drug reactions, drug rash, mortality, Stevens–Johnson syndrome, toxic epidermal necrolysis
Introduction 
Stevens–Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are among the most dreaded severe cutaneous adverse drug reactions (SCAR) on account of their high mortality rate. The most commonly implicated drugs that cause SJS/TEN are antiepileptics, antibiotics and NSAIDs, although virtually every drug known to mankind (including non-allopathic drugs) is capable of triggering this reaction.[1] The incidence of SJS and TEN is estimated to be 1.2–6/million patient-years and 0.4–1.2/million patient-years, respectively, according to global data, which corresponds to the limited data available from India.[2] Studies also show that Asians and Africans tend to have a 2–3-fold higher incidence of SJS/TEN than Caucasians.[3] This study was undertaken to look at the demographics and clinical profile of patients presenting with SJS/TEN to our hospital over the last 10 years and identify possible risk factors for poor prognosis.
Materials and Methods All patients who had been admitted at our centre for SJS, SJS-TEN overlap and TEN during 10 years from 2010 to 2020 were included in this retrospective series. Case definitions used for SJS, SJS-TEN overlap and TEN were according to the criteria proposed by Bastuji Garin.[4] After ethical clearance, in-patient and out-patient records were reviewed, and data were recorded in a study proforma. Variables recorded included patient demographics, clinical features, causative agent, re-epithelialisation time, therapy, comorbidities, complications, duration of hospitalisation and outcome. Statistical analysis was performed using Stata v. 15.0 (StataCorp LLC, Texas, USA). Mean and standard deviation/interquartile range were calculated for all normally distributed continuous variables, and sub-groups were compared using the Student's t-test. For comparison of qualitative/categorical data, the Chi-square test was used, with a P value of < 0.05 considered statistically significant, with 95% confidence intervals (CIs).
Results Clinical profile
A total of 82 patients were admitted to our centre with a diagnosis of SJS, SJS-TEN and TEN for 10 years from 2010 to 2020. Of these, 30 patients were diagnosed as SJS, 4 patients as SJS-TEN overlap and 48 patients as TEN. For analysis, the patients with SJS-TEN overlap were clubbed with TEN (n = 52). The male: female ratio was 1.14 and 0.57 for SJS and TEN, respectively. Patients in the TEN group (mean age 39.7 years, 95% CI [34.5–44.8]) were significantly older than those in the SJS group (mean age 28.0 years, 95% CI [21.7–34.3]) (P = 0.006), with the maximum number of TEN in the > 40 years age group [Table 1]. There were 16 paediatric patients (10 SJS and 6 TEN). All cases of SJS, SJS-TEN overlap and TEN were preceded by drug intake. The most common group of drugs implicated overall was antiepileptics (n = 29, 35.4%), followed by antibiotics (n = 20, 24.4%) and NSAIDs (n = 7, 8.5%) [Figure 1]. The most commonly implicated antibiotics were cephalosporins (n = 11, 55%), followed by quinolones (n = 9, 45%). Nine patients (11%) had received more than one drug before the onset of SJS/TEN.
Five patients (6.1%) gave a history of a previous adverse drug reaction. The morphology of the previous reaction was not suggestive of SJS/TEN in any of them; however, the implicated drug for the previous rash was the same as the implicated drug for the present rash in three patients.
Comorbid conditions (including diabetes, hypertension, renal insufficiency, arthritis, hypothyroidism, hyperuricemia, asthma, cardiac disease, cerebrovascular disease, psychiatric disease, malignancy, HIV and tuberculosis) were present in 13.3% (n = 4) of the SJS group and 30.8% (n = 16) of the TEN group.
Prodromal symptoms (fever/myalgia/burning sensation) were present in 76.7% (n = 23) of SJS and 80.8% (n = 42) of TEN patients. Concurrent infections were present in 34.2% (n = 28) of all patients at the time of presentation, including respiratory, gastrointestinal and urinary infections.
Mucosal involvement was present in all patients; oral mucosal involvement was the most common, followed by ocular mucosa. Ocular mucosal involvement was in the form of conjunctivitis, cicatricial conjunctivitis, keratitis, corneal epithelial defect and dry eye. Mucosal sequelae included symblepharon (n = 6), ankyloblepharon (n = 1) and genital synechiae (n = 1).
The clinical morphology of TEN was 'TEN with spots' (spots ± atypical targets) in 27 (51.9%) and 'TEN without spots' (diffuse erythema, no spots or targets) in 25 (48.1%) patients. The mean maximum BSA of denudation in TEN was 42.0% (95% CI [30–55]). The mean SCORTEN calculated at admission was 2.2 (95% CI [1.9–2.5]).
Secondary bacterial infection of the cutaneous erosions and bloodstream infection/sepsis were seen exclusively in TEN patients, in 32.7% (n = 17) and 39.6% (n = 19), respectively. The most common pathogens were Staphylococcus aureus, methicillin-resistant S. aureus (MRSA), Enterococcus and non-fermenting gram-negative bacilli. Apart from infections, other complications seen included biochemical abnormalities (including electrolyte imbalance, metabolic acidosis and hyperglycaemia), haematological abnormalities (including anaemia, leukopaenia, pancytopaenia, thrombocytopaenia and coagulopathy), drug-induced liver injury and acute kidney injury (AKI).
Treatment
All patients received supportive treatment in the form of strict barrier nursing, wound care with saline/potassium permanganate compresses and paraffin gauze dressings, ophthalmic care, analgesia, fluid replacement, enteral protein supplementation, followed by intensive monitoring of vitals and output and periodic wound swabs/blood cultures. Patients with SJS were treated with either steroids or supportive treatment alone. Those with TEN (n = 52) received steroids (mostly as daily intravenous dexamethasone in doses of 8–12 mg/day) (n = 29), cyclosporine (n = 10), intravenous immunoglobulin (IVIG) (n = 3) or supportive treatment alone (n = 11) [Table 2].
Outcome
In the TEN group (n = 52), two patients were discharged against medical advice, and eight deaths occurred (mortality rate of 16%); there were no deaths in the SJS group (n = 30). The maximum BSA of denudation was significantly associated with mortality (P = 0.003). There was a clinical subset of patients who developed cutaneous lesions before or concurrent with mucosal erosions, and these patients had significantly higher mortality (P < 0.001), compared to those who had mucosal lesions preceding skin lesions. There was no significant correlation between mortality and age, gender, implicated drug, SCORTEN, time delay in the initiation of therapy and choice of therapy [Table 3].
Table 3: Comparison of survivors vs. deaths (TEN only, n=50, DAMA=2 excluded) Discussion Clinical profile
Our series saw a difference in the mean age of SJS and TEN patients, with SJS patients being significantly younger by about a decade [Table 1]. Previous series have also found the most common age group for SJS to comprise children and young adults, whereas TEN is more common among older adults.[5],[6] Gender differences are also well described in SJS/TEN, with a male: female ratio >1 in SJS and <1 in TEN, as seen in our series as well.[5],[7] Although our SJS and TEN patients were more or less similar in terms of demographics (with the exception of age), we are in agreement with Rokea et al., that SJS and TEN are vastly different entities, despite being on the same spectrum, and should be treated as such.[5]
A formal causality scoring was not performed in all cases, with the diagnosis of SJS/TEN made clinically by an experienced dermatologist, and any drug received within an exposure window of 4–28 days before the rash was considered potentially causative for the rash.[7] The most commonly implicated drugs for SJS/TEN across series reported from India include antiepileptics, antibiotics and NSAIDs, which is similar to what we have seen in our series.[8],[9],[10],[11] One patient experienced SJS due to modafinil, a drug prescribed for the treatment of narcolepsy. There are no published reports of SJS/TEN to modafinil, although the risk of SJS is mentioned on the product label, on account of a rash seen during clinical trials, and a single case of SJS to its congener-armodafinil has been published recently.[12] Adverse cutaneous drug reactions to indigenous systems of medicine are a fairly common occurrence in India, and this series included one patient each, with SJS and TEN, potentially to Ayurvedic medicines.
Treatment
Clear evidence-based therapeutic guidelines TEN are lacking, because of the paucity of data, the rarity of the disease and the practical and ethical difficulties of conducting prospective randomised trials on small numbers of patients over several years. According to the available evidence in the form of meta-analyses and systematic reviews, the most robust evidence is in favour of cyclosporine and steroid-IVIg combination, in terms of a favourable effect on mortality.[8],[13] Until recently, steroids were the mainstay of therapy for TEN at our centre, used as intravenous dexamethasone 8–12 mg/day in divided doses until the arrest of disease progression, for a mean duration of 7.1 ± 3.9 days in this series. The use of steroids has been a subject of controversy in the management of TEN, with some studies showing that they are beneficial when used early in the disease, in high doses (e.g. pulse therapy), and in specific subgroups, whereas others have demonstrated higher infection/complication rates and poor outcomes in steroid-treated patients.[13],[14] The highest mortality occurred in the steroid group in our series (n = 6, 21.4%), followed by the supportive care group (n = 2, 18.2%). There was no significant difference in the rates of cutaneous re-epithelialisation or duration of hospital stay in the four treatment groups in this series. The most robust data currently available supports the use of cyclosporine, which is consistent with our experience over the last few years, during which we have increasingly chosen cyclosporine as the first-line treatment of TEN at our centre. Data on IVIg shows a trend towards reduced mortality, with a better prognosis in children; however, most often in our setting, the cost of the drug is a major limiting factor in its widespread use.[15] We did not observe any mortalities in the cyclosporine (n = 10) and IVIg (n = 3) groups; however, the numbers were small in these groups. We did not institute steroids or any immunomodulatory therapy in patients who presented late in the course of the disease, and/or had no active/progressing areas of cutaneous necrosis. These patients were managed with supportive care alone (n = 11).
Outcome
The overall mortality for TEN in this series was 16%, which is lower than the rates reported in most series (ranging from 10% for SJS to 50% for TEN), irrespective of treatment modality or geographic location.[16],[17] There were no deaths in our SJS patients, in contrast to some published series that have reported mortalities in SJS patients, although they have not characterised the cause of death in these patients.[18] A number of factors associated with poor outcome have been analysed by various studies, with most being in agreement that advancing age, severity of illness (extent of denudation), high SCORTEN, medical comorbidities and malignancy are important factors that worsen outcomes in TEN.[11],[17],[19] Interestingly, in our series, younger patients (< 40 years) had higher overall mortality (21.4%) compared to older patients (9.1%), and the majority of the deaths (7 out of 8; 87.5%) were in patients of the female gender, although neither of these parameters achieved statistical significance. We do not have an explanation for this; however, the experience at our centre has been that young women with TEN tend to do worse than males and older patients. This may be important from the point of view of prognostication and counselling at the onset of illness because many of the other parameters such as SCORTEN have not been found to be reliable predictors of mortality, in previous studies as well.[11],[20] It is also our observation that patients who presented with cutaneous necrosis before mucosal lesions at the onset of illness did significantly worse overall (mortality = 31.8%) than those who presented with prominent mucosal lesions before cutaneous necrosis (mortality = 3.6%). Previously, Canavan et al. made the observation that patients on the erythema multiforme (EM)/SJS spectrum who presented with prominent mucositis with sparse skin lesions had a milder disease course and should be re-classified as MIRM (Mycoplasma-induced rash and mucositis) rather than EM or SJS.[21] Our observations indicate that there could potentially be a similar subset among TEN patients, that is, those with a cutaneous onset (and severe course) vs. those with a mucosal onset (and milder course). Patients who presented with the clinical subtype of “TEN without spots” also had higher mortality (24% vs. 7.4%) than those who presented with “TEN with spots,” although this was not statistically significant.
We found that a higher maximum BSA of denudation and the development of sepsis were significantly associated with poorer outcomes but did not help practically in prognostication at the onset of illness because these events usually occurred later in the course of the illness. Patients who developed sepsis and/or had unstable vitals were shifted to the ICU; a total of 22 patients required ICU care during the course of their hospital stay. All eight mortalities in this series occurred in an ICU setting, with the cause of death being sepsis with multiorgan dysfunction in all patients. With regard to early transfer to an ICU setting, previous studies have found that patients who were transferred early (within 6 days of admission) to a burns ward/ICU setting tended to do better overall.[22] In our series as well, patients who were directly admitted to the ICU at the time of initial hospitalisation (n = 9; mortality = 18.2%) and those who were transferred to an ICU setting within 7 days of hospitalisation (n = 14; mortality = 7.1%) did significantly better than those who were transferred to ICU after the first 7 days of hospitalisation (n = 7; mortality = 85.7%). This makes a strong case for early aggressive supportive care, preferably in an ICU/burns ward setting, which many burns wards across the world have found to be as effective as immunomodulatory therapy, if not more.[18],[23] Unfortunately, this is not always feasible in a resource-poor setting, where both the financial constraints of the patient and the availability/cost of ICU beds play a major role in decision-making. Because wound infection and sepsis appear to be predictors of a poor outcome, frequent monitoring of the TEN patient to pick up early signs of local and systemic infections in terms of wound swabs and blood/urine cultures (every 48–72 h) should be a part of the TEN protocol in any centre that manages these patients. In terms of parameters such as a delay in hospitalisation, total disease duration, implicated drug, delay in initiation of therapy, the onset of re-epithelialisation, SCORTEN and total duration of hospital stay, there was no difference between survivors and deaths, across treatment groups.
Conclusion Patients with SJS and TEN were largely similar in terms of demographics, with the exception of age and gender differences, and a better prognosis with no mortalities in the SJS group. The factors significantly associated with a poor prognosis in the TEN group were cutaneous lesions preceding mucosal lesions at onset, high mean BSA of denudation and a transfer to ICU >7 days after admission.
Limitations
Although this is the largest single-centre series of SJS/TEN from India, differences in treatment efficacy and outcome between subgroups could not be compared because treatments were individualised and not protocol-based, the study was retrospective in nature, and the numbers in each group were too small to compare statistically.
Ethics committee/institutional review board's permission
We also declare that the study was assessed and approved by the Institutional Ethics Committee/Institutional Review Board and that the letter of approval is available with us for examination (IEC no. 191/2017).
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References 
留言 (0)