The present single-centre retrospective observational study was conducted in the Department of Dermatology and Venereology, Government Medical College, Thiruvananthapuram, Kerala, India. Data spanning 10 years, from January 2012 to December 2021, were retrieved from the case sheets of the central records library at this centre after a computer search using the following terms: SJS, TEN, and drug reactions. After obtaining permission from the Institutional Research and Ethical Committee (HEC No. 09/03/2022/MCT), the retrieved data were entered into a preformatted proforma and analysed over a period of six months, from November 2022 to May 2023.

Inclusion criteria: Fully documented and completed case records of diagnosed cases of SJS/TEN, including investigations were included in the study. The diagnosis of SJS/TEN entered in the case records was based on the Bastuji-Garin criteria: <10% BSA for SJS, 10-30% for SJS/TEN overlap, and >30% for TEN [3].

Exclusion criteria: Incomplete case records and those with incomplete investigations were excluded from the study.

Age and gender were the demographic variables retrieved from the case records. The data regarding co-morbidities analysed included HT, diabetes, Human Immunodeficiency Virus (HIV), atopy and malignancy. The drug variables included past history of drug reactions, whether single or multiple drugs were taken, mode of intake, duration of the drug reaction, days elapsed before being referred to this centre, total duration of stay in the hospital, and the most probable drug causing the adverse reaction.

The data regarding clinical variables retrieved from the case sheets included the presence of vesicles or bullae, erosions/crusting, presence of epidermal detachment, purpuric spots, and involvement of the eyes, oral cavity, and genital mucosa, as well as, approximate total body surface involvement. Data regarding the immediate cause of mortality were also retrieved. The mortality rate was calculated as follows: the total number of SJS/TEN patients who died during the study period divided by the total number of SJS/TEN patients in the study period, multiplied by 100.

Data regarding investigations retrieved included: blood routine tests, urine routine tests {(albumin, pus cells and Red Blood Cells (RBCs)}, Liver Function Tests (LFT) {normal range: Serum Glutamic Oxaloacetic Transaminase (SGOT): 12-38 IU/L, Serum Glutamic Pyruvate Transaminase (SGPT): 7-41 IU/L}, Renal Function Tests (RFT) (normal range: urea: 15-35 mg/dL, creatinine: 0.6-1.4 mg/dL), serum sodium (136-145 mM/L), potassium (3.6-5.4 mM/L), bicarbonate (>20 mM/L), and random blood sugar levels (<200 mg%). The reference values for the above-mentioned parameters are based on the standard Institutional laboratory values.

The data retrieved were entered into an Excel sheet and analysed using the trial version of SPSS software version 29.0. Qualitative data were expressed as proportions or percentages. Quantitative variables were expressed as means and standard deviations. Frequency distribution analysis and descriptive statistics were calculated for nominal and ratio data, respectively. Risk factors for mortality in SJS/TEN were analysed using logistic regression in both univariate and bivariate models. The Pearson’s two-sided Chi-square test was utilised to determine the statistical significance of the variables between the survival group and the mortality group, with a p-value of <0.05 considered statistically significant.

The present 10-year retrospective observational study (2012 to 2021) analysed 146 cases (N=146) of SJS/TEN during the study period. There were eight cases of mortality in the present study, accounting for a mortality rate of 5.48%. The mean±SD age in the survived group was 42.07±16.704 years, while in the mortality group, it was 67.83±7.57 years, which was statistically significant (p-value=0.003). The youngest age in the mortality group was 19 years, and the oldest was 79 years, while in the survived group, the youngest was 14 years, and the oldest was 77 years. In both groups, females outnumbered males, but this was not statistically significant [Table/Fig-1].

There were no malignancies in the mortality group, while there were 4 (2.9%) cases in the survived group. The salient demographic and drug details in the mortality group are provided in [Table/Fig-2]. The salient clinical features in the mortality group are provided in [Table/Fig-3]. The relevant investigation details in the mortality group are given in [Table/Fig-4]. The salient features in the survived group are presented in [Table/Fig-5].

Prior drug reaction history was present in 25% of the mortality group and 11.6% of the survived group (p-value=0.262). The total duration of the disease from the onset of taking the offending drug and the time spent in the hospital were not contributory to mortality (p-value=0.710) in the present study [Table/Fig-6]. Most patients were on multiple drugs when they developed the rash, making it difficult to pinpoint a particular drug as the cause. In some cases, no drugs could be implicated. Most prescriptions that led to the disease included antibiotics and analgesics. However, phenytoin (n=42) was found to be the most common drug in both the survived and mortality groups [Table/Fig-7].

The co-morbidity parameters with significance in both the survived and mortality groups are provided in [Table/Fig-8]. Two cases (25%) in the mortality group and 25 cases (18.11%) in the survived group had diabetes. In the survived group 24 (17.39%) cases and 5 (62.5%) cases in the mortality group had hypertension (p-value=0.002).

Investigative parameters with significance are presented in [Table/Fig-9]. Only deranged urine microscopy was significantly associated with mortality (p-value=0.04). The most common probable drug in the mortality group and the survived group was phenytoin, with 3 (33.3%) cases in the mortality group and 36 (26.09%) cases in the survived group, respectively. In the survived group, 113 (81.88%) cases had eye involvement, 124 (89.85%) cases had oral involvement, and 69 (50%) cases had genital involvement, while in the mortality group, eye involvement was seen in 6 (75%) cases, oral involvement in 7 (87.5%) cases and genital involvement in 50% (4 cases) (p-value=0.774, 0.404, 0.467, respectively). Vesicles and bullae, along with phenytoin and TEN, were significantly associated with mortality (p-value <0.05) [Table/Fig-10]. However, epidermal detachment, purpuric spots, and erosion/crusting were not associated with mortality (p-value >0.05) [Table/Fig-11].

All the patients were treated with systemic steroids (betamethasone), antibiotics, and supportive measures. In the mortality group, 7 (87.5%) cases and 134 (97.8%) in the survival group were managed in the general wards. Four patients died of septicaemia; two died of septicaemia and renal failure, one died due to ARDS and another died due to renal failure.

The present retrospective case record-based observational study demonstrated a mortality rate of 5.48% in cases of SJS/TEN, one of the lowest in similar studies worldwide. Mortality figures in foreign studies range from 15.4-48%, while in Indian studies, they range from 8.10-28.20% [4-7]. Moreover, the majority of the survivors (96.37%) in the present study were managed in general wards rather than in the Intensive Care Unit (ICU). This contrasts with most advanced centres in the world, where SJS/TEN is managed in burn wards and ICUs, yet they have a much higher mortality rate compared to the present study and other Indian studies [8,9]. The exact cause for this discrepancy cannot be explained, other than by racial and biological factors. A study in the United States indicated greater mortality in the Hispanic population compared to the non hispanic white population (p-value=0.05), suggesting racial and biological influences [4]. However, in the present study, the management in the ICU or general ward did not determine the final clinical outcome regarding survival or mortality (p-value=0.147). In cases of TEN, where there is epidermal detachment with erosions and crusting, skin barrier function is lost; thus, it is preferable to manage these patients in ICU units or burn wards to ensure better barrier nursing and patient care [8]. This may not be feasible in most centres in India due to limited resources, as patients from the Medicine Department are often given priority in these units.

The mean age in the mortality group was higher than in the survival group, and this difference was significant (p-value=0.003). This aligns with the study by Noe MH et al., which found that increasing age significantly contributes to mortality [8]. Age is a contributing factor in mortality for SJS/TEN and is one parameter in Severity-of-illness Score for Toxic Epidermal Necrolysis (SCORTEN), where an age greater than 40 years receives 1 point [3]. Age is also one of the parameters in the ABCD-10 model (A-age, B-bicarbonate, C-cancer, D-dialysis, 10-10% BSA) for predicting mortality in SJS/TEN [8]. Older individuals may have co-morbidities such as diabetes, hypertension and cardiac problems, which contribute to mortality [9].

Females were the majority in both the survival and mortality groups, but this was not significant (p-value=0.361). This finding is consistent with other studies [10,11]. The exact factors responsible for the increased frequency of SJS/TEN and mortality in females have not been elucidated, but they may involve biological factors.

The past history of drug reactions in the index patient was not contributory to mortality in the present study (p-value=0.262). The total duration of the disease from the onset of taking the offending drug and the time spent in the hospital was not contributory to mortality in this study (p-value=0.710). This observation is consistent with the studies by Thakur V et al., and Kanagarajan A et al., in India [12,13]. However, prolonged hospital stay contributing to mortality has been reported in other studies [12-14]. Prolonged hospital immobilisation may be complicated by pneumonia and septicaemia in SJS/TEN, where the barrier functions of the skin are already compromised, leading to mortality [15].

An interesting finding in the present study is the association of HT as a co-morbidity contributing to mortality, which was significant (p-value=0.002). We could not find any study with a similar finding. Other studies have linked co-morbidities such as DM, HIV, tuberculosis and cutaneous autoimmune diseases as risk factors for mortality [16,17]. DM was not a significant co-morbidity in the current study (p-value=0.961).

Another important highlight of the present study was that the presentation of TEN with vesicles and bulla predominantly contributed to mortality (p-value=0.002), rather than epidermal detachment with charring (p-value=0.424) and purpuric spots (p-value=0.775). This contrasts with other studies where epidermal detachment had a worse prognosis in TEN [14,15,18,19]. The vesicles and bulla in SJS/TEN are rich in electrolytes such as sodium, potassium, bicarbonate and chloride. The subsequent rupture of these blisters results in the loss of these vital electrolytes, leading to an electrolyte imbalance, which is an important factor contributing to mortality [20].

Another highlight of the present study was that there was no significant association between the extent of BSA involvement and mortality (p-value=0.140). This is in contrast to most foreign studies where total BSA involvement was associated with an increased frequency of mortality [8,20-22]. Vaishampayan SS et al., in an Indian study also related BSA to mortality [23]. However, studies conducted by Thakur V et al., and Kanagarajan A et al., in India demonstrated no significant association between BSA and mortality (p-value >0.05) [12,13].

The presence of erosions or crusting was not significantly associated with mortality in this study (p-value=0.745), while some foreign studies have demonstrated that the presence of erosions or crusting contributes to mortality, as these lesions are a good nidus for bacterial growth, especially MRSA [20-22]. Similar Indian studies have not factored erosions or crusting into their parameters to assess mortality [2,5,12].

The presence or absence of eye, oral, and genital lesions in the surviving group and the mortality group was not statistically significant in the present study (p-value=0.774, 0.404, 0.467, respectively). Most other studies have demonstrated similar findings [5,12,24]. However, involvement of the eyes and oral cavity contributed significantly to morbidity and prolonged hospital stays due to the lengthy healing process in these areas, and there may be severe ophthalmological sequelae post-discharge [25].

Phenytoin was the most commonly implicated drug in both the surviving group and the mortality group, and it was statistically more likely to cause mortality than other drugs (p-value=0.004). However, in several North Indian and South Indian studies, carbamazepine was the most commonly implicated drug [26-28]. In India, these two antiepileptic drugs are widely used due to their efficacy and low cost, which could explain the high frequency of SJS/TEN associated with these drugs. This is in contrast to foreign studies, where beta-lactam and sulphonamide antibiotics are the most common culprits [24,29]. In the present study, mortality was significantly higher in TEN than in SJS (p-value=0.001). This is consistent with similar Indian and foreign studies [24-28].

Conditions SJS/TEN is complicated by systemic organ involvement, such as acute renal tubular necrosis, drug-induced hepatitis, hepatic failure and respiratory tract issues like ARDS and pneumonia, all contributing to mortality. Deranged LFT, RFT, and blood haemogram are serological markers used to detect early organ involvement. However, in this study, deranged LFT, RFT and blood haemogram were not significantly associated with mortality (p-value=0.784, 0.660, 0.736, respectively). Nevertheless, Indian and foreign studies have demonstrated a relationship between deranged LFT and RFT and mortality [11,27,28]. The present study demonstrated that urine microscopy findings of albumin, RBCs and pus cells were significantly associated with mortality (p-value=0.040). This finding has not been reported in previous studies. Interestingly, out of the five cases in the mortality group that exhibited urine microscopy abnormalities, three cases had normal serum urea and creatinine levels. This indicates that urine microscopy abnormalities may be the earliest manifestation of acute renal tubular necrosis rather than serological RFT derangement. Therefore, it is warranted to perform urine microscopy in the emergency laboratory, in addition to electrolytes, RFT, and LFT, as soon as, SJS/TEN cases are admitted [29].

In the present study, data extracted from the case sheets indicated that all cases of SJS/TEN were treated uniformly with systemic steroids, antibiotics and supportive measures. Intravenous Immunoglobulin (IVIG) and ciclosporin were not used; hence, treatment outcomes regarding survival and mortality rates could not be assessed. However, systemic steroids and prophylactic antibiotics could be determining factors in the very low mortality rate observed in the present study compared to foreign studies. Future multicentric studies that factor in both SCORTEN and ABCD-10 are required to determine the mortality factors in SJS/TEN.

The present retrospective case record-based study was conducted at a single-centre. In the present study, it was not possible to correlate the SCORTEN score with the mortality rate, as it is a 10-year retrospective analysis and all SCORTEN parameters were available only in the recent case sheets, not in the older ones. Additionally, only the probable drug could be identified in cases where multiple drugs were administered.

The present study showed one of the lowest mortality rates for SJS/TEN in the world. Co-morbidities such as HT were significant risk factors for mortality. Phenytoin was the most common drug associated with SJS/TEN in both the survival and mortality groups and was significantly more likely to cause mortality than other drugs. Presentation with vesicles and bullae, rather than epidermal detachment with charring, was significantly associated with mortality. Urine microscopic abnormalities were also significantly associated with mortality. Patients with TEN were more likely to die than those with SJS. The final clinical outcome of survival or mortality did not depend on whether the patient was managed in the ICU or in general wards.