Abstract
Background
Evidence of factors associated with Stevens–Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) from population-based studies is scarce.
Objective
We aimed to identify the incidence, risk factors, and drugs that trigger the development of SJS/TEN in the general population.
Methods
A regional, population-based, longitudinal cohort with 2,398,393 Japanese individuals was analyzed using the Shizuoka Kokuho Database from 2012 to 2020.
Results
Among 1,909,570 individuals, 223 (0.01%, 2.3 cases/100,000 person-years) patients were diagnosed with SJS/TEN during the observational period of a maximum of 7.5 years. In a multivariable analysis, the risks of SJS/TEN were an older age, and the presence of type 2 diabetes, peripheral vascular disease, and systemic autoimmune diseases. The administration of drugs, such as immune checkpoint inhibitors, insulin, and type 2 diabetes agents, triggered the onset of SJS/TEN.
Limitations
The results may apply only to the Japanese population.
Conclusion
In this cohort population from a database representing the general population, the risks of developing SJS/TEN were old age and a history of type 2 diabetes, peripheral vascular disease, and systemic autoimmune disease. Furthermore, in addition to previously reported drugs, the administration of immune checkpoint inhibitors, insulin, and type 2 diabetes agents, may trigger the development of SJS/TEN.
Key words: claims database, drug eruption, population-based cohort study, pharmacoepidemiology, risk factor, safety, Stevens–Johnson syndrome, toxic epidermal necrolysis
Abbreviations used: CI, confidential interval; HR, hazard ratio; ICD-10, International Classification of Disease, 10th revision; SJS, Stevens–Johnson syndrome; SKDB, Shizuoka Kokuho Database; TEN, toxic epidermal necrolysis
Capsule Summary.
-
•
Limited information regarding Stevens–Johnson syndrome or toxic epidermal necrolysis is available.
-
•
Physicians may wish to consider risk factors, such as older age and diabetes, that may be associated with the development of Stevens–Johnson syndrome or toxic epidermal necrolysis, and drugs that trigger this disease such as immune-checkpoint inhibitors.
Introduction
Stevens–Johnson syndrome and toxic epidermal necrolysis (SJS/TEN) represent a spectrum of severe cutaneous adverse reactions,1,2 and are clinically characterized by erythema, blistering, and erosions. SJS/TEN may also induce lung and liver damage,3, 4, 5 causing life-threatening events and death in the worst-case scenario. Nevertheless, there are only limited data related to the incidence of SJS/TEN because of few longitudinal cohort studies. These studies have shown an incidence of 1 to 7 cases/million/year for SJS and 0.4 to 1.5 for TEN.6,7 Previous studies have reported risk factors associated with SJS/TEN, such as skin disease, liver disease,8 diabetes, psoriasis, systemic lupus erythematosus, and previous drug allergies.9 However, these previous studies did not examine the risk factors using a cohort in the general population as a riskset.
Previous case series studies showed potential agents that cause the onset of SJS/TEN, such as anti-inflammatory analgesics, antibacterial agents, anticonvulsants, antigout agents, and anticancer agents (eg, immune checkpoint inhibitors).10, 11, 12, 13, 14, 15 However, the risk of drug administration for the onset of SJS/TEN has not been investigated using causal inference in cohort studies.
Therefore, we investigated the risk factors and drugs that trigger the onset of SJS/TEN using a large, population-based, longitudinal cohort from an insurance claims database (Shizuoka Kokuho Database [SKDB]).
Methods
Data resource
In this study, we used the SKDB with already cleaned data.16 The data of all enrollees were anonymized.16 The SKDB can be considered a regional, population-based, longitudinal cohort of 2,398,393 Japanese people (women, n = 1,303,667, 54.4%) living in Shizuoka Prefecture, near the center of Japan (population of approximately 3.6 million).16 The SKDB has been used as a data source in several studies.17, 18, 19 Comprehensive, personally linked data were collected, and everyone was assigned a unique identifier. This dataset includes basic information from the subscriber list (eg, sex, age, zip code, observation period, and reason for disenrollment including death) and claims from public health insurance organizations (<75 years old for the National Health Insurance system and >75 years old for the Latter-stage Elderly Medical Care System).
The SKDB is a suitable database for real-world studies because it contains accurate information on deaths and loss to follow-up from the Basic Resident Registration System. In addition, in this database, Anatomical Therapeutic Chemical classification codes and International Classification of Diseases, 10th Revision (ICD-10) codes are included for searching for drugs and diagnoses.
Population and observational period
We used a dataset of the SKDB that comprised 8.5 years of longitudinal data from April 2012 to September 2020. All enrollees were investigated using individually linked databases between their annual health checkups and medical claims. To clarify the risk factors for new-onset SJS/TEN, we excluded patients who had already been diagnosed with drug eruption, such as SJS/TEN and drug-induced hypersensitivity syndrome/drug reaction with eosinophilia and systemic symptoms, before the index date. The study schema is shown in Fig 1. Each enrollee’s observation period was defined as the time from the date of insurance registration or 1 April 2012, whichever came earlier, to the date of insurance withdrawal or 30 September 2020, whichever came later.
Fig 1.
Study schema. Cohort entry is the date of registration with the health insurance provider or April 1, 2012, whichever occurred later. The index date is the initial date with ≥1-year continuous subscribership to the health insurance system. The follow-up period is the interval between the index date and (1) the end of the study (September 30, 2020), or (2) the withdrawal date from the health insurance system, whichever occurred first. SJS/TEN, Stevens–Johnson syndrome and toxic epidermal necrolysis; SKDB, Shizuoka Kokuho Database.
Comorbidities as candidate risk factors associated with onset of SJS/TEN
Patients with SJS/TEN were defined as those with a medical claim with the disease code, which is the code used for billing medical fees for SJS/TEN (SJS: 6951003, TEN: 8845586). In this study, the candidate risk factors were age, sex, and the presence of comorbidities using the Elixhauser comorbidity index20 and previously reported risk factors and candidate risk factors8,9,21,22 (as defined using the ICD-10, Supplementary Table I, available via Mendeley at https://data.mendeley.com/datasets/yd63zxsgxy/1).
Drugs as candidate risk factors associated with onset of SJS/TEN
To identify drugs that trigger the onset of SJS/TEN, we created a list of causative drugs for SJS/TEN and drug reaction with eosinophilia and systemic symptoms, analogous drugs with caution for drug eruptions in medication package inserts, and potential causative drugs, such as immune checkpoint inhibitors, insulin, and type 2 diabetes agents (Supplementary Table II, available via Mendeley at https://data.mendeley.com/datasets/yd63zxsgxy/1).
Statistical analysis
The data are shown as the mean and standard deviation for continuous variables and the frequency and percentage for categorical variables. The Wilcoxon rank-sum test for continuous variables and the chi-square test for categorical variables were performed for comparison between 2 groups.
Univariate and multivariable Cox regression analyses were performed to examine the factors associated with the onset of SJS/TEN. We calculated hazard ratios (HRs), 95% confidence intervals (CIs), and P values in the Wald test. The reported risk factors (eg, age) and potential risk factors (eg, critical etiological and epidemiological factors) were used in the regression analysis. One (hypertension) of 2 variables with a high correlation was not used in the multivariable model owing to multicollinearity, based on the criterion of an absolute Spearman correlation coefficient >0.4. Because missing covariates did not occur completely at random among all patients, simple missing data imputation was not carried out. The variables with ≥10 cases of onset of SJS/TEN shown in Table I and a P value < .05 were selected as factors to be included in the multivariable regression model.
Table I.
Characteristics of patients with onset of SJS/TEN
| Variable | Category or unit | SJS/TEN |
Others |
P value |
|---|---|---|---|---|
| (n = 223) | (n = 1,909,347) | |||
| Age (y) | 1 | 69.46 ± 14.78 | 57.46 ± 23.42 | <.001 |
| Age (y) | 0 to <40 | 14 (6.3) | 450,724 (23.6) | <.001 |
| 40 to <50 | 7 (3.1) | 142,411 (7.5) | ||
| 50 to <60 | 8 (3.6) | 155,583 (8.1) | ||
| 60 to <70 | 56 (25.1) | 456,473 (23.9) | ||
| 70 to <80 | 98 (43.9) | 429,628 (22.5) | ||
| ≥80 | 40 (17.9) | 274,528 (14.4) | ||
| Sex | Women | 116 (52.0) | 1,031,695 (54.0) | .546 |
| Allergy disease | Presence | 69 (30.9) | 502,906 (26.3) | .128 |
| Dermatological disease | Presence | 79 (35.4) | 487,342 (25.5) | .001 |
| Kidney disease | Presence | 38 (17.0) | 138,766 (7.3) | <.001 |
| Liver disease | Presence | 40 (17.9) | 184,542 (9.7) | <.001 |
| Gout | Presence | 32 (14.3) | 140,851 (7.4) | <.001 |
| Mycoplasma infection | Presence | 0 (0.0) | 4019 (0.2) | >.999 |
| HHV infection | Presence | 6 (2.7) | 25,811 (1.4) | .132 |
| Psoriasis | Presence | 2 (0.9) | 10,903 (0.6) | .364 |
| Drug allergy | Presence | 0 (0.0) | 0 (0.0) | NA |
| Type-1 diabetes | Presence | 0 (0.0) | 2544 (0.1) | >.999 |
| Type-2 diabetes | Presence | 28 (12.6) | 99,036 (5.2) | <.001 |
| AIDS/HIV | Presence | 1 (0.4) | 332 (0.0) | .038 |
| Congestive heart failure | Presence | 40 (17.9) | 160,246 (8.4) | <.001 |
| Peripheral vascular disease | Presence | 44 (19.7) | 128,233 (6.7) | <.001 |
| Chronic pulmonary disease | Presence | 51 (22.9) | 335,373 (17.6) | .043 |
| Hemiplegia or paraplegia | Presence | 2 (0.9) | 15,950 (0.8) | .771 |
| Metastatic solid tumor | Presence | 5 (2.2) | 16,594 (0.9) | .046 |
| Weight loss | Presence | 0 (0.0) | 5944 (0.3) | >.999 |
| Fluid and electrolyte disorders | Presence | 23 (10.3) | 141,972 (7.4) | .123 |
| Deficiency anemia | Presence | 24 (10.8) | 110,284 (5.8) | .004 |
| Alcohol abuse | Presence | 0 (0.0) | 8614 (0.5) | .631 |
| Drug abuse | Presence | 0 (0.0) | 632 (0.0) | >.999 |
| Depression | Presence | 18 (8.1) | 92,603 (4.8) | .040 |
| Psychoses | Presence | 7 (3.1) | 46,934 (2.5) | .510 |
| Cardiac arrythmia | Presence | 36 (16.1) | 165,450 (8.7) | <.001 |
| Valvular disease | Presence | 12 (5.4) | 56,709 (3.0) | .045 |
| Pulmonary circulation disorders | Presence | 1 (0.4) | 3915 (0.2) | .367 |
| Hypertension, uncomplicated | Presence | 124 (55.6) | 659,825 (34.6) | <.001 |
| Hypertension, complicated | Presence | 2 (0.9) | 12,204 (0.6) | .657 |
| Other neurological disorders | Presence | 13 (5.8) | 72,441 (3.8) | .113 |
| Diabetes, no complications | Presence | 9 (4.0) | 24,726 (1.3) | .003 |
| Diabetes, complications | Presence | 20 (9.0) | 75,544 (4.0) | .001 |
| Hypothyroidism | Presence | 6 (2.7) | 29,320 (1.5) | .163 |
| Renal failure | Presence | 11 (4.9) | 45,353 (2.4) | .023 |
| Liver disease | Presence | 44 (19.7) | 195,870 (10.3) | <.001 |
| Peptic ulcer disease | Presence | 56 (25.1) | 237,060 (12.4) | <.001 |
| Lymphoma | Presence | 2 (0.9) | 5906 (0.3) | .152 |
| Solid tumor without metastasis | Presence | 28 (12.6) | 116,459 (6.1) | <.001 |
| Systemic autoimmune diseases | Presence | 16 (7.2) | 49,041 (2.6) | <.001 |
| Coagulopathy | Presence | 6 (2.7) | 13,908 (0.7) | .006 |
| Obesity | Presence | 2 (0.9) | 5804 (0.3) | .148 |
| Blood loss anemia | Presence | 2 (0.9) | 2876 (0.2) | .045 |
Data are shown as the mean ± standard deviation or n (%).
AIDS, Acquired immunodeficiency syndrome; HHV, human herpesvirus; HIV, human immunodeficiency virus; SJS/TEN, Stevens–Johnson syndrome and toxic epidermal necrolysis.
A risk classification (low, middle, and high) was constructed using the individual predictive value estimated from the multivariable Cox model. To examine the causal relationship between the administration of drugs and the subsequent onset of SJS/TEN, a Cox regression analysis was performed with the time to onset of SJS/TEN within 4 months8,23 from the date of the first administration of each drug (or the index date if not administered). The onset of SJS/TEN can occur within 1 month after taking the triggered drugs, but to identify the triggered drugs based on causal inference, the maximum possible period until the onset was set at 4 months.
A P-value of < .05 was considered statistically significant. All analyses were performed using SAS version 9.4 (SAS Institute) and EZR version 1.55 (https://www.jichi.ac.jp/saitama-sct/SaitamaHP.files/statmedEN.html).24
Results
Characteristics of patients with SJS/TEN
The procedure for extracting the first diagnosed SJS/TEN among all cases in the SKDB is shown in Fig 2. Among 1,909,570 individuals, the number of SJS/TEN cases that developed during the observation period was 223. The median (minimum-maximum) observation period was 6.0 (0-7.5) years, and the incidence of SJS/TEN onset was 2.3 cases/100,000 person-years.
Fig 2.
Flow chart of the patients. The population as an analysis set of this study excluded patients with a baseline period of <1 year and patients diagnosed with drug eruption, such as SJS/TEN and DIHS/DRESS, before the start date of the health examination from all registrants from April 2012 to September 2020. DIHS, Drug-induced hypersensitivity syndrome; DRESS, drug reaction with eosinophilia and systemic symptoms; SJS/TEN, Stevens–Johnson syndrome and toxic epidermal necrolysis.
The characteristics of patients with SJS/TEN at the time of diagnosis and other participants without SJS/TEN are shown in Table I. The patients who developed SJS/TEN were more likely to have an older age (>60 years) and a history of kidney disease, liver disease, gout, type 2 diabetes, congestive heart failure, peripheral vascular disease, chronic pulmonary disease, deficiency anemia, depression, cardiac arrythmias, valvular disease, hypertension (uncomplicated), diabetes (complications), renal failure, peptic ulcer disease, solid tumor without metastasis, or systemic autoimmune disease than other individuals.
Identification of risk factors for the development of SJS/TEN
The results of univariable and multivariable regression analyses in the analysis set are shown in Supplementary Table III, available via Mendeley at https://data.mendeley.com/datasets/yd63zxsgxy/1 and Table II. Hypertension was excluded because of its high correlation with age. Other variables with P ≤ .05 were entered into the multivariable model. The multivariable analysis showed that age (60-<70 vs 0-40 years as a reference: HR: 1.94 [95% CI: 1.04-3.51], 70-80: 2.91 [1.57-5.23], >80: 1.91 [1.01-3.62]), the presence of type 2 diabetes (1.53 [1.01-2.32]), peripheral vascular disease (1.76 [1.24-2.51]), and systemic autoimmune disease (1.80 [1.07-3.03]) increased the risk of developing SJS/TEN.
Table II.
Multivariable Cox regression analysis of the onset of SJS/TEN
| Variable (reference) | Category | Multivariable model |
||
|---|---|---|---|---|
| HR | 95% CI | P value | ||
| Age (0 to <40 y) | 40 to <50 y | 1.16 | 0.47-2.88 | .749 |
| 50 to <60 y | 1.00 | 0.41-2.38 | .999 | |
| 60 to <70 y | 1.94 | 1.04-3.51 | .030 | |
| 70 to <80 y | 2.91 | 1.57-5.23 | <.001 | |
| ≥80 y | 1.91 | 1.01-3.62 | .046 | |
| Sex (women) | Men | 0.87 | 0.66-1.15 | .319 |
| Dermatological disease (absence) | Presence | 1.22 | 0.92-1.62 | .173 |
| Kidney disease (absence) | Presence | 1.41 | 0.96-2.07 | .077 |
| Liver disease (absence) | Presence | 1.17 | 0.82-1.67 | .376 |
| Gout (absence) | Presence | 1.15 | 0.77-1.73 | .498 |
| Type-2 diabetes (absence) | Presence | 1.53 | 1.01-2.32 | .043 |
| Congestive heart failure (absence) | Presence | 1.29 | 0.88-1.92 | .192 |
| Peripheral vascular disease (absence) | Presence | 1.76 | 1.24-2.51 | .002 |
| Deficiency anemia (absence) | Presence | 1.15 | 0.73-1.80 | .551 |
| Depression (absence) | Presence | 1.35 | 0.83-2.20 | .231 |
| Cardiac arrythmia (absence) | Presence | 1.04 | 0.71-1.54 | .827 |
| Peptic ulcer disease (absence) | Presence | 1.27 | 0.91-1.75 | .156 |
| Solid tumor without metastasis (absence) | Presence | 1.39 | 0.92-2.09 | .122 |
| Systemic autoimmune disease (absence) | Presence | 1.80 | 1.07-3.03 | .027 |
Bold type indicates statistical significance.
CI, Confidence interval; HR, hazard ratio.
Identification of drugs that trigger the development of SJS/TEN
The frequency of potential drug use during the 4 months before SJS/TEN was determined by classifying them into risk categories (low, middle, and high, Supplementary Table IV, available via Mendeley at https://data.mendeley.com/datasets/yd63zxsgxy/1). We found that patients at a high risk of developing SJS/TEN had a higher rate of using anticonvulsants, antiarrhythmics drugs, sulfa drugs, antigout agents, immune checkpoint inhibitors, antifungals, mucolytic agents, anti-cancer drugs, antipsychotics, insulin, and type-2 diabetes agents than those at a low risk, but not antibacterials and non-steroidal anti-inflammatory drugs.
The results of univariable and multivariable regression analyses on the causal relationship between drugs and the onset of SJS/TEN are shown in Table III. In the multivariable analysis, the administration of drugs other than antiarrhythmics increased the risk of SJS/TEN onset. In particular, the drugs with an HR of >50 were anticonvulsants (HR: 66.97 [95% CI: 29.99-149.50]), immune checkpoint inhibitors (486.70 [175.70-1348.00]), anticancer drugs (50.02 [21.77-115.00]), and insulin (73.36 [33.80-159.20]).
Table III.
Cox regression analysis of the incidence of SJS/TEN by a drug taken within 4 months
| Drug taken within 4 mo | Incidence of SJS/TEN within 4 mo of taking the drug (%) | Incidence of SJS/TEN within 4 mo of not taking the drug (%) | Univariable model |
Multivariable model |
||||
|---|---|---|---|---|---|---|---|---|
| HR | 95% CI | P value | HR | 95% CI | P value | |||
| Anticonvulsant | 0.036 | <0.001 | 76.95 | 35.03-169.00 | <.001 | 66.97 | 29.99-149.50 | <.001 |
| Antiarrhythmic | 0.000 | 0.001 | N.E. | N.E. | N.E. | N.E. | N.E. | N.E. |
| Sulfa drug | 0.077 | 0.001 | 7.00 | 1.60-22.60 | <.001 | 4.83 | 1.09-21.54 | <.001 |
| Antigout agent | 0.027 | 0.001 | 46.54 | 17.21-125.80 | <.001 | 35.04 | 12.79-96.03 | <.001 |
| Antibacterial | 0.004 | <0.001 | 7.66 | 3.16-18.54 | <.001 | 7.12 | 2.93-17.32 | <.001 |
| NSAID | 0.003 | <0.001 | 7.77 | 2.96-20.36 | <.001 | 8.10 | 3.08-21.31 | <.001 |
| Immune checkpoint inhibitor | 0.258 | 0.001 | 537.80 | 195.30-1481.00 | <.001 | 486.70 | 175.70-1348.00 | <.001 |
| Antifungal | 0.012 | 0.001 | 20.24 | 6.44-63.56 | <.001 | 15.41 | 4.85-49.02 | <.001 |
| Anti-peptic ulcer agent | 0.006 | <0.001 | 13.72 | 5.79-2.51 | <.001 | 14.69 | 6.17-35.02 | <.001 |
| Mucolytic agent | 0.002 | 0.001 | 3.12 | 1.26-7.76 | .014 | 2.83 | 1.13-7.07 | .026 |
| Anti-cancer drug | 0.031 | 0.001 | 58.66 | 25.72-133.80 | <.001 | 50.02 | 21.77-115.00 | <.001 |
| Antipsychotic | 0.007 | <0.001 | 15.83 | 6.71-37.35 | <.001 | 13.66 | 5.61-33.25 | <.001 |
| Insulin | 0.040 | 0.001 | 76.71 | 37.80-155.70 | <.001 | 73.36 | 33.80-159.20 | <.001 |
| Type-2 diabetes agent | 0.010 | 0.001 | 16.86 | 6.54-43.49 | <.001 | 13.91 | 5.35-36.17 | <.001 |
Bold type indicates statistical significance.
CI, Confidence interval; HR, hazard ratio; N.E., not evaluable; NSAID, nonsteroidal anti-inflammatory drug.
Discussion
This large-scale cohort study clarified the risk factors for SJS/TEN. We found that the incidence of SJS/TEN was 2.3 cases/100,000 person-years. The risk factors for the onset of SJS/TEN were an old age (>60 years) and a history of type 2 diabetes, peripheral vascular disease, and systemic autoimmune disease. Additionally, the drugs that triggered the development of SJS/TEN were anticonvulsants, sulfa drugs, antigout agents, antibacterial agents, nonsteroidal anti-inflammatory drugs, antifungals, mucolytic agents, anticancer drugs, antipsychotics, immune checkpoint inhibitors, insulin, and type 2 diabetes agents. These results could contribute to the early diagnosis and prevention of the onset of SJS/TEN in the future, and such information was not available from previous reports.
SJS and TEN, which are differently defined on the basis of affected skin areas, represent a spectrum of severe cutaneous adverse reactions characterized by extended epidermal necrosis. There are several key factors involved in the pathogenesis of SJS/TEN, such as drug-reactive cytotoxic CD8+ T-lymphocytes25 and soluble biomolecules, including Fas-ligand,26 granulysin,26 and annexin-A1,27 and neutrophil extracellular traps.28 Although the mechanism of SJS/TEN has been recently determined,25, 26, 27, 28 the appropriate early therapeutic intervention is unknown. Furthermore, the mortality of SJS/TEN has worsened in the last decade.6 Therefore, identifying risk factors for developing SJS/TEN is important or early therapeutic intervention.
Older people tend to use various medicines for a number of diseases that they suffer from. Additionally, such a situation for older people may promote desynchronized T-cell activation under a microenvironment with development of cytotoxicity by senescence,29,30 which could be associated with a high incidence of SJS/TEN.
In this study, we showed that 3 underlying diseases were associated with the development of SJS/TEN, namely systemic autoimmune disease and type 2 diabetes, as previously shown,9 and peripheral vascular disease. In autoimmune disease, disturbed regulatory immunity augments cytotoxic reactions, which are associated with the induction of SJS/TEN.31,32 Diabetes often causes internal organ damage, such as that to the kidney and liver, leading to increased drug concentrations by impaired metabolizing activity, and this might be associated with the augmentation of hypersensitivity reactions. Peripheral vascular disease (eg, ICD-10 codes I70, I71, I731, I738, I739, I771, I790, I792, K551, K558, K559, Z958, and Z959) is an ambiguous category including intermittent claudication, peripheral angiopathy (not otherwise specified), and spasm of an artery. Therefore, the reason for this association remains unclear and it requires further investigation.
Surprisingly, we found that the use of immune checkpoint inhibitors had an extremely high HR (>400) in the incidence of SJS/TEN compared with nonusers. This HR was the highest among the investigated drugs, and the presence of malignancy was not associated with the incidence of SJS/TEN (Table II), which suggested a high risk of immune checkpoint inhibitor use. These agents abrogate regulatory immunity and augment antitumor immunity.33, 34, 35, 36, 37 Therefore, these agents may promote SJS/TEN. Indeed, many cases of SJS/TEN in patients treated with immune checkpoint inhibitors have been reported previously.15,38, 39, 40, 41 Because the number of patients with cancer treated with immune checkpoint inhibitors has dramatically increased because of their remarkable efficacy, even at the advanced stage, physicians should be aware of the emergence of SJS/TEN.
We also showed that the administration of insulin and type 2 diabetes drugs was associated with a high risk of developing SJS/TEN, which should be interpreted carefully because we also detected type 2 diabetes as a risk. Whether having diabetes or using insulin and type 2 diabetes drugs is a risk of developing SJS/TEN, remains inconclusive, because this finding may be due to reverse causality. Alternatively, because insulin and type 2 diabetes drugs may be proxy variables for type 2 diabetes, this disease may be a potent risk factor for developing SJS/TEN.
Strengths and limitations
The strength of this study is the identification of risk factors and drugs that trigger the development of SJS/TEN using incidence-related information over a recent 7.5-year period. However, this study has several limitations. First, no genetic information was available for the individuals included in this study. Recent epidemiological studies have shown that several genetic factors specific to drugs have a risk of causing SJS/TEN.42, 43, 44 Additionally, in-silico biomolecular studies of the distinctive combinations between human leukocyte antigen alleles and drugs have shown exquisite molecular interactions,45, 46, 47, 48 which are caused by the enhancement of cytotoxic T-cell reactions, such as allorecognition via the “pharmacological-interaction” concept.47,49 Therefore, the diseases that were risk factors in our study might be associated with genetic factors, such as human leukocyte antigen alleles, predisposing to SJS/TEN. However, in the general population of this study, the probability that individuals had such genetic information is rare.
Second, information regarding the disease severity, skin histology of SJS/TEN, and the associated complications were unavailable. Third, all of the drugs and their interactions with the comorbidities that trigger SJS/TEN could not be examined in this study. To comprehensively search for drugs and the interactions that trigger SJS/TEN, high-throughput screening studies using the same method as that used in this study are required in the future. Fourth, because the database used in this study was based on National Health Insurance claim data in Shizuoka Prefecture, which does not cover health insurance claims of the Social Assurance Union, there is likely to have been selection bias. Fifth, over-the-counter drugs as causative drugs could not be assessed because the SKDB does not have claim data of such drugs. Nonsteroidal anti-inflammatory drugs such as oxicam, which is strongly associated with SJS/TEN,50 are not marketed as over-the-counter drugs, excluding topical piroxicam in Japan. Finally, the results may not apply to other ethnic groups because the present population was exclusively Japanese.
Conclusion
In a cohort population representing the general population, the risk factors for developing SJS/TEN were old age and a history of type 2 diabetes, peripheral vascular disease, and systemic autoimmune disease. Furthermore, in addition to previously reported drugs, the administration of immune checkpoint inhibitors, insulin, and type 2 diabetes drugs may trigger the development of SJS/TEN.
Conflicts of interest
Nanako Ubukata is an employee of Sato Pharmaceutical Co, Ltd.
Acknowledgments
We thank Ellen Knapp, PhD, from Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript.
Footnotes
Drs Ubukata and Nakatani contributed equally to this work.
Funding sources: The Shizuoka Graduate University of Public Health conducts contract research projects for public health in Shizuoka Prefecture with funding from Shizuoka Prefecture, including the current study.
IRB approval status: The ethics committee of Shizuoka Graduate University of Public Health approved the study (#SGUPH_2021_001_027).
Patient consent: Not applicable.
References
- 1.Stevens A.M., Johnson F.C. A new eruptive fever associated with stomatitis and ophthalmia: report of two cases in children. Am J Dis Child. 1922;24(6):526–533. [Google Scholar]
- 2.Lyell A. Toxic epidermal necrolysis: an eruption resembling scalding of the skin. Br J Dermatol. 1956;68(11):355–361. doi: 10.1111/j.1365-2133.1956.tb12766.x. [DOI] [PubMed] [Google Scholar]
- 3.Powell N., Munro J.M., Rowbotham D. Colonic involvement in Stevens-Johnson syndrome. Postgrad Med J. 2006;82(968) doi: 10.1136/pgmj.2005.042952. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.de Prost N., Mekontso-Dessap A., Valeyrie-Allanore L., et al. Acute respiratory failure in patients with toxic epidermal necrolysis: clinical features and factors associated with mechanical ventilation. Crit Care Med. 2014;42(1):118–128. doi: 10.1097/CCM.0b013e31829eb94f. [DOI] [PubMed] [Google Scholar]
- 5.Pradka S.P., Smith J.R., Garrett M.T., Fidler P.E. Mesenteric ischemia secondary to toxic epidermal necrolysis. J Burn Care Res. 2014;35(5):e346–e352. doi: 10.1097/bcr.0000000000000006. [DOI] [PubMed] [Google Scholar]
- 6.Sunaga Y., Kurosawa M., Ochiai H., et al. The nationwide epidemiological survey of Stevens-Johnson syndrome and toxic epidermal necrolysis in Japan, 2016-2018. J Dermatol Sci. 2020;100(3):175–182. doi: 10.1016/j.jdermsci.2020.09.009. [DOI] [PubMed] [Google Scholar]
- 7.Yang M.S., Lee J.Y., Kim J., et al. Incidence of Stevens-Johnson syndrome and toxic epidermal necrolysis: a nationwide population-based study using National Health Insurance Database in Korea. PLoS One. 2016;11(11) doi: 10.1371/journal.pone.0165933. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Chung S.J., Ahn K.M., Oh J.H., Shim J.S., Park H.W. Incidence rates of severe cutaneous adverse reactions due to antiseizure medication: a nationwide study using health claims data in Korea. Epilepsia. 2021;62(1):250–257. doi: 10.1111/epi.16751. [DOI] [PubMed] [Google Scholar]
- 9.Gronich N., Maman D., Stein N., Saliba W. Culprit medications and risk factors associated with Stevens-Johnson syndrome and toxic epidermal necrolysis: population-based Nested case-control study. Am J Clin Dermatol. 2022;23(2):257–266. doi: 10.1007/s40257-021-00661-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Roujeau J.C., Kelly J.P., Naldi L., et al. Medication use and the risk of Stevens–Johnson syndrome or toxic epidermal necrolysis. N Engl J Med. 1995;333(24):1600–1608. doi: 10.1056/NEJM199512143332404. [DOI] [PubMed] [Google Scholar]
- 11.Mockenhaupt M., Viboud C., Dunant A., et al. Stevens–Johnson syndrome and toxic epidermal necrolysis: assessment of medication risks with emphasis on recently marketed drugs. The EuroSCAR-study. J Invest Dermatol. 2008;128(1):35–44. doi: 10.1038/sj.jid.5701033. [DOI] [PubMed] [Google Scholar]
- 12.Roujeau J.C., Stern R.S. Severe adverse cutaneous reactions to drugs. N Engl J Med. 1994;331(19):1272–1285. doi: 10.1056/NEJM199411103311906. [DOI] [PubMed] [Google Scholar]
- 13.Letko E., Papaliodis D.N., Papaliodis G.N., Daoud Y.J., Ahmed A.R., Foster C.S. Stevens-Johnson syndrome and toxic epidermal necrolysis: a review of the literature. Ann Allergy Asthma Immunol. 2005;94(4):419–436. doi: 10.1016/S1081-1206(10)61112-X. quiz 436-438, 456. [DOI] [PubMed] [Google Scholar]
- 14.Sueki H. Epidemiology of SJS/TEN. MB Derma. 2021;307:47–52. Japanese. [Google Scholar]
- 15.Geisler A.N., Phillips G.S., Barrios D.M., et al. Immune checkpoint inhibitor-related dermatologic adverse events. J Am Acad Dermatol. 2020;83(5):1255–1268. doi: 10.1016/j.jaad.2020.03.132. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Nakatani E., Tabara Y., Sato Y., Tsuchiya A., Miyachi Y. Data resource profile of Shizuoka Kokuho Database (SKDB) using integrated health- and care-insurance claims and health checkups: the Shizuoka study. J Epidemiol. 2021;32(8):391–400. doi: 10.2188/jea.JE20200480. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Goto H., Nakatani E., Yagi H., Moriki M., Sano Y., Miyachi Y. Late-onset development of psoriasis in Japan: a population-based cohort study. JAAD Int. 2021;2:51–61. doi: 10.1016/j.jdin.2020.10.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Nishimura S., Kumamaru H., Shoji S., et al. Assessment of coding-based frailty algorithms for long-term outcome prediction among older people in community settings: a cohort study from the Shizuoka Kokuho Database. Age Ageing. 2022;51(3) doi: 10.1093/ageing/afac009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Hashizume H., Nakatani E., Sato Y., Goto H., Yagi H., Miyachi Y. A new susceptibility index to predict the risk of severe herpes zoster-associated pain: a Japanese regional population-based cohort study, the Shizuoka study. J Dermatol Sci. 2022;105(3):170–175. doi: 10.1016/j.jdermsci.2022.02.006. [DOI] [PubMed] [Google Scholar]
- 20.Quan H., Sundararajan V., Halfon P., et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005;43(11):1130–1139. doi: 10.1097/01.mlr.0000182534.19832.83. [DOI] [PubMed] [Google Scholar]
- 21.Fukasawa T., Takahashi H., Takahashi K., Tanemura N., Amagai M., Urushihara H. Risk of Stevens-Johnson syndrome and toxic epidermal necrolysis associated with anticonvulsants in a Japanese population: matched case–control and cohort studies. Allergol Int. 2021;70(3):335–342. doi: 10.1016/j.alit.2021.01.004. [DOI] [PubMed] [Google Scholar]
- 22.Downey A., Jackson C., Harun N., Cooper A. Toxic epidermal necrolysis: review of pathogenesis and management. J Am Acad Dermatol. 2012;66(6):995–1003. doi: 10.1016/j.jaad.2011.09.029. [DOI] [PubMed] [Google Scholar]
- 23.Ordoñez L., Salgueiro E., Jimeno F.J., Manso G. Spontaneous reporting of Stevens-Johnson syndrome and toxic epidermal necrolysis associated with antiepileptic drugs. Eur Rev Med Pharmacol Sci. 2015;19(14):2732–2737. [PubMed] [Google Scholar]
- 24.Kanda Y. Investigation of the freely available easy-to-use software ‘EZR’ for medical statistics. Bone Marrow Transplant. 2013;48(3):452–458. doi: 10.1038/bmt.2012.244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Duong T.A., Valeyrie-Allanore L., Wolkenstein P., Chosidow O. Severe cutaneous adverse reactions to drugs. Lancet. 2017;390(10106):1996–2011. doi: 10.1016/S0140-6736(16)30378-6. [DOI] [PubMed] [Google Scholar]
- 26.Chung W.H., Hung S.I., Yang J.Y., et al. Granulysin is a key mediator for disseminated keratinocyte death in Stevens-Johnson syndrome and toxic epidermal necrolysis. Nat Med. 2008;14(12):1343–1350. doi: 10.1038/nm.1884. [DOI] [PubMed] [Google Scholar]
- 27.Saito N., Qiao H., Yanagi T., et al. An annexin A1-FPR1 interaction contributes to necroptosis of keratinocytes in severe cutaneous adverse drug reactions. Sci Transl Med. 2014;6(245) doi: 10.1126/scitranslmed.3008227. [DOI] [PubMed] [Google Scholar]
- 28.Kinoshita M., Ogawa Y., Hama N., et al. Neutrophils initiate and exacerbate Stevens-Johnson syndrome and toxic epidermal necrolysis. Sci Transl Med. 2021;13(600) doi: 10.1126/scitranslmed.aax2398. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Guo Z., Wang G., Wu B., et al. DCAF1 regulates Treg senescence via the ROS axis during immunological aging. J Clin Invest. 2020;130(11):5893–5908. doi: 10.1172/JCI136466. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Tsaknaridis L., Spencer L., Culbertson N., et al. Functional assay for human CD4+CD25+ Treg cells reveals an age-dependent loss of suppressive activity. J Neurosci Res. 2003;74(2):296–308. doi: 10.1002/jnr.10766. [DOI] [PubMed] [Google Scholar]
- 31.Watanabe Y., Yamaguchi Y. Drug allergy and autoimmune diseases. Allergol Int. 2022;71(2):179–184. doi: 10.1016/j.alit.2022.02.001. [DOI] [PubMed] [Google Scholar]
- 32.Sueki H., Kitami Y., Watanabe H. Severe cutaneous adverse reactions during tapering of high-dose systemic steroid therapy for autoimmune diseases: implications for non-HIV immune reconstitution inflammatory syndrome. Case Rep Dermatol. 2019;11(2):166–174. doi: 10.1159/000500945. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Rausch J., Lopez P.A., Bialojan A., et al. Combined immunotherapy: CTLA-4 blockade potentiates anti-tumor response induced by transcutaneous immunization. J Dermatol Sci. 2017;87(3):300–306. doi: 10.1016/j.jdermsci.2017.06.013. [DOI] [PubMed] [Google Scholar]
- 34.Dodagatta-Marri E., Meyer D.S., Reeves M.Q., et al. α-PD-1 therapy elevates Treg/Th balance and increases tumor cell pSmad3 that are both targeted by α-TGFβ antibody to promote durable rejection and immunity in squamous cell carcinomas. J Immunother Cancer. 2019;7(1):62. doi: 10.1186/s40425-018-0493-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Saleh R., Elkord E. Treg-mediated acquired resistance to immune checkpoint inhibitors. Cancer Lett. 2019;457:168–179. doi: 10.1016/j.canlet.2019.05.003. [DOI] [PubMed] [Google Scholar]
- 36.Gan X., Shan Q., Li H., et al. An anti-CTLA-4 heavy chain-only antibody with enhanced Treg depletion shows excellent preclinical efficacy and safety profile. Proc Natl Acad Sci U S A. 2022;119(32) doi: 10.1073/pnas.2200879119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Kumar P., Saini S., Prabhakar B.S. Cancer immunotherapy with check point inhibitor can cause autoimmune adverse events due to loss of Treg homeostasis. Semin Cancer Biol. 2020;64:29–35. doi: 10.1016/j.semcancer.2019.01.006. [DOI] [PubMed] [Google Scholar]
- 38.Quach H.T., Johnson D.B., LeBoeuf N.R., Zwerner J.P., Dewan A.K. Cutaneous adverse events caused by immune checkpoint inhibitors. J Am Acad Dermatol. 2021;85(4):956–966. doi: 10.1016/j.jaad.2020.09.054. [DOI] [PubMed] [Google Scholar]
- 39.He C., Si X., Zhang L., Xu Y., Qu T. Immune checkpoint inhibitor–related epidermal necrolysis: a rare condition with poor prognosis. Eur J Cancer. 2021;145:194–196. doi: 10.1016/j.ejca.2020.12.018. [DOI] [PubMed] [Google Scholar]
- 40.Logan I.T., Zaman S., Hussein L., Perrett C.M. Combination therapy of ipilimumab and Nivolumab-associated toxic epidermal necrolysis (TEN) in a patient with metastatic melanoma: a case report and literature review. J Immunother. 2020;43(3):89–92. doi: 10.1097/CJI.0000000000000302. [DOI] [PubMed] [Google Scholar]
- 41.Coleman E.L., Olamiju B., Leventhal J.S. The life-threatening eruptions of immune checkpoint inhibitor therapy. Clin Dermatol. 2020;38(1):94–104. doi: 10.1016/j.clindermatol.2019.10.015. [DOI] [PubMed] [Google Scholar]
- 42.Tohkin M., Kaniwa N., Saito Y., et al. A whole-genome association study of major determinants for allopurinol-related Stevens-Johnson syndrome and toxic epidermal necrolysis in Japanese patients. Pharmacogenomics J. 2013;13(1):60–69. doi: 10.1038/tpj.2011.41. [DOI] [PubMed] [Google Scholar]
- 43.Kaniwa N., Sugiyama E., Saito Y., et al. Specific HLA types are associated with antiepileptic drug-induced Stevens-Johnson syndrome and toxic epidermal necrolysis in Japanese subjects. Pharmacogenomics. 2013;14(15):1821–1831. doi: 10.2217/pgs.13.180. [DOI] [PubMed] [Google Scholar]
- 44.Kaniwa N., Saito Y., Aihara M., et al. HLA-B∗1511 is a risk factor for carbamazepine-induced Stevens-Johnson syndrome and toxic epidermal necrolysis in Japanese patients. Epilepsia. 2010;51(12):2461–2465. doi: 10.1111/j.1528-1167.2010.02766.x. [DOI] [PubMed] [Google Scholar]
- 45.Jaruthamsophon K., Tipmanee V., Sangiemchoey A., Sukasem C., Limprasert P. HLA-B∗15:21 and carbamazepine-induced Stevens-Johnson syndrome: pooled-data and in silico analysis. Sci Rep. 2017;7:45553. doi: 10.1038/srep45553. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Illing P.T., Vivian J.P., Dudek N.L., et al. Immune self-reactivity triggered by drug-modified HLA-peptide repertoire. Nature. 2012;486(7404):554–558. doi: 10.1038/nature11147. [DOI] [PubMed] [Google Scholar]
- 47.Pichler W.J., Beeler A., Keller M., et al. Pharmacological interaction of drugs with immune receptors: the p-i concept. Allergol Int. 2006;55(1):17–25. doi: 10.2332/allergolint.55.17. [DOI] [PubMed] [Google Scholar]
- 48.Chung W.H., Hung S.I., Hong H.S., et al. A marker for Stevens–Johnson syndrome. Nature. 2004;428(6982):486. doi: 10.1038/428486a. [DOI] [PubMed] [Google Scholar]
- 49.Pichler W.J., Watkins S., Yerly D. Risk assessment in drug hypersensitivity: detecting small molecules which outsmart the immune system. Front Allergy. 2022;3:827893. doi: 10.3389/falgy.2022.827893. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Mockenhaupt M., Kelly J.P., Kaufman D., Stern R.S., SCAR Study Group The risk of Stevens-Johnson syndrome and toxic epidermal necrolysis associated with nonsteroidal antiinflammatory drugs: a multinational perspective. J Rheumatol. 2003;30(10):2234–2240. [PubMed] [Google Scholar]