Key Points
Question
Are the clinical outcomes of administering plasmapheresis therapy first better than those of administering intravenous immunoglobulin (IVIG) therapy first after ineffective systemic corticosteroid therapy in patients with Stevens-Johnson syndrome or toxic epidermal necrolysis (SJS/TEN)?
Findings
This retrospective cohort study of 266 inpatients with SJS/TEN found no significant difference in mortality rates between the plasmapheresis-first and the IVIG-first groups. Patients who received plasmapheresis therapy first had longer hospitalization stays and incurred higher expenses.
Meaning
The findings of this retrospective cohort study suggest that there is no clear benefit to administering plasmapheresis before IVIG therapy to patients with SJS/TEN unresponsive to systemic corticosteroids and plasmapheresis may be associated with higher cost and longer hospital stays.
Abstract
Importance
Stevens-Johnson syndrome and toxic epidermal necrolysis (SJS/TEN) are severe cutaneous adverse reactions, and patients with SJS/TEN frequently require intensive care. However, there is limited evidence on the clinical outcomes of immunomodulating therapy, including plasmapheresis and intravenous immunoglobulin (IVIG) in patients with SJS/TEN.
Objective
To compare clinical outcomes of patients with SJS/TEN who were treated with plasmapheresis first vs IVIG first after ineffective systemic corticosteroid therapy.
Design, Setting, and Participants
This retrospective cohort study used data from a national administrative claims database in Japan that included more than 1200 hospitals and was conducted from July 2010 to March 2019. Inpatients with SJS/TEN who received plasmapheresis and/or IVIG therapy after initiation of at least 1000 mg/d of methylprednisolone equivalent systemic corticosteroid therapy within 3 days of hospitalization were included. Data were analyzed from October 2020 to May 2021.
Exposures
Patients who received IVIG or plasmapheresis therapy within 5 days after initiation of systemic corticosteroid therapy were included in the IVIG- and plasmapheresis-first groups, respectively.
Main Outcomes and Measures
In-hospital mortality, length of hospital stay, and medical costs.
Results
Of 1215 patients with SJS/TEN who had received at least 1000 mg/d of methylprednisolone equivalent within 3 days of hospitalization, 53 and 213 patients (mean [SD] age, 56.7 [20.2] years; 152 [57.1%] women) were included in the plasmapheresis- and IVIG-first groups, respectively. Propensity-score overlap weighting showed no significant difference in inpatient mortality rates between the plasmapheresis- and IVIG-first groups (18.3% vs 19.5%; odds ratio, 0.93; 95% CI, 0.38–2.23; P = .86). Compared with the IVIG-first group, the plasmapheresis-first group had a longer hospital stay (45.3 vs 32.8 days; difference, 12.5 days; 95% CI, 0.4–24.5 d; P = .04) and higher medical costs (US $34 262 vs $23 054; difference, US $11 207; 95% CI, $2789–$19 626; P = .009).
Conclusions and Relevance
This nationwide retrospective cohort study found no significant benefit to administering plasmapheresis therapy first instead of IVIG first after ineffective systemic corticosteroid treatment in patients with SJS/TEN. However, medical costs and length of hospital stay were greater for the plasmapheresis-first group.
This retrospective cohort study evaluated the outcomes of administering plasmapheresis or intravenous immunoglobulin therapy first after ineffective systemic corticosteroid treatment among inpatients with Stevens-Johnson syndrome and toxic epidermal necrolysis in Japan.
Introduction
Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are severe cutaneous adverse reactions caused by medications.1,2 Considered to be parts of a single spectrum of disease, SJS and TEN are distinguished by degree of severity based on the extent of skin detachment.3 These diseases are frequently accompanied by complications—organ failure, lung injury, and/or infection—that make treatment difficult; occasionally, SJS and TEN can lead to death. The mortality rates of SJS and TEN are approximately 10% and 30%, respectively.2,4,5 Despite the high mortality rates, no treatment strategies for SJS/TEN have been established because of their rarity and a lack of clinical evidence.6,7 Thus, further studies are needed to develop an optimal treatment for SJS/TEN.
Within the context of insufficient data, physicians consider using immunosuppressive and immunomodulatory therapies, including systemic corticosteroids, intravenous immunoglobulin (IVIG), and plasmapheresis.8,9 One choice of immunomodulatory therapy is IVIG, which is considered to inhibit widespread Fas-mediated keratinocyte apoptosis.10 The efficacy of IVIG alone is controversial.11 However, a recent multicenter observational study and meta-analyses showed that combination therapy with IVIG and systemic steroids can be more effective than monotherapy with IVIG or systemic steroids.12,13,14 Moreover, in the United States, 70% medical practitioners preferably prescribe IVIG therapy for patients with TEN.15 Another choice of immunomodulatory therapy is plasmapheresis, which has been used to clear drug metabolites and cytotoxic mediators.16,17,18,19,20,21 Although plasmapheresis is an invasive and resource-intensive intervention, some clinicians consider it to be more effective than IVIG; plasmapheresis therapy is administered before IVIG therapy.22 Limited case series and literature reviews have reported on combination therapy with plasmapheresis and systemic corticosteroids for the treatment of SJS/TEN.16,17,18,19,20,21 Some studies have reported positive effects of plasmapheresis16,17,18,19,20,21; however, a previous study has suggested that plasmapheresis is futile.23 Therefore, the effects of plasmapheresis, especially those of combination therapy with plasmapheresis and systemic steroids, are unknown.
There are no standard guidelines on the use of immunomodulating therapies; their use depends on clinical experience and local guidelines.9,24 For example, the Japanese guidelines recommend systemic corticosteroids as the first choice and IVIG or plasmapheresis as the second choice of treatment for patients with SJS/TEN who have been unresponsive to systemic corticosteroids.9 However, there is a controversy regarding which therapy should be administered first when systemic corticosteroids are ineffective. Therefore, to address these knowledge gaps, we used a Japan’s nationwide inpatient database to compare clinical outcomes of patients with SJS/TEN who were treated with plasmapheresis first with the outcomes of those who were treated with IVIG first, after both groups had been unresponsive to systemic corticosteroid therapy.
Methods
The institutional review board of the University of Tokyo reviewed and approved the study protocol. Patient informed consent was waived because of the anonymous nature of the data. This study adhered to the tenets of the Declaration of Helsinki.
Data Source
This was a retrospective cohort study that used routinely collected data from the Diagnosis Procedure Combination inpatient database (complete details available elsewhere).25 This database includes administrative claims data and discharge abstracts of acute care inpatients in Japan, coded according to International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10). Data were collected from more than 1200 acute care hospitals—approximately 90% of all tertiary care hospitals in Japan. Further detailed information is described in eAppendix 1 in Supplement 1.
Patient Selection
We searched the national database for adult ( ≥18 years) hospitalized patients who had been diagnosed with SJS (ICD-10 code, L51.1) or TEN (ICD-10 code, L51.2) from July 2010 through March 2019. Among these patients, those who had received at least 1000 mg/d of methylprednisolone equivalent systemic corticosteroids within 3 days of hospitalization were included. The systemic corticosteroid initiation date was defined as the index date. We excluded patients who had received IVIG or plasmapheresis therapy before the index date and patients who did not receive either IVIG or plasmapheresis within 5 days of the index date.
We identified 5803 adult patients with SJS/TEN during the study period, of whom 1215 had received at least 1000 mg/d of methylprednisolone equivalent within 3 days of admission. We excluded 22 patients who had received IVIG or plasmapheresis before steroid pulse therapy. Of the remaining 1193 patients, 213 had received IVIG first within 5 days of the index date (IVIG-first group) and 53 had received plasmapheresis first or on the same day as IVIG administration within 5 days of the index date (plasmapheresis-first group) (Figure).
Figure. Flow Diagram of Patient Selection.
IVIG indicates intravenous immunoglobulin, and SJS/TEN, Stevens-Johnson Syndrome and Toxic Epidermal Necrolysis.
Exposure
Patients who had received IVIG earlier (or IVIG alone) were included in the IVIG-first group, and those who had received plasmapheresis earlier (or plasmapheresis alone) were included in the plasmapheresis-first group, as shown in eFigure in Supplement 1. Patients who received both IVIG and plasmapheresis on the same day were assigned to the plasmapheresis-first group because plasmapheresis eliminates most of the antibodies administered by IVIG and because IVIG is generally administered after daily plasmapheresis.
Covariates and Outcomes
On the basis of a priori knowledge, we included the covariates described in eAppendix 2 in Supplement 1.26 The primary outcome was the in-hospital mortality rate. The secondary outcomes were the length of hospital stay and total costs (USD $1 = JPY ¥110).
Statistical Analysis
Continuous variables are presented as mean (SD), whereas categorical variables are presented as frequencies and percentages. We compared the crude outcomes between the plasmapheresis- and the IVIG-first groups using the t test for continuous outcomes and the χ2 test for categorical outcomes.
To evaluate the outcomes, we used a propensity-score overlap weighting method to estimate the adjusted outcomes.27,28,29,30,31,32,33,34 The details of the overlap weighting are described in eAppendix 3 in Supplement 1. After applying the overlap weights to patients in each group, a generalized linear regression model with logit link function was used for binary outcomes to calculate the odds ratio (OR) and its 95% CI, and those with identity link function were used for continuous outcomes to calculate differences and their 95% CIs. To assess the balance of patient characteristics between the groups, we calculated the absolute standardized differences for each covariate in the weighted and unweighted populations. A standardized difference of less than 10% was considered a negligible imbalance.35,36
The 2-sided significance level for all tests was defined as P < .05. All analyses were performed from October 22, 2020, to May 12, 2021, using Stata/MP, version 16 (StataCorp).
Sensitivity Analysis
We performed 3 sensitivity analyses to confirm the robustness of the primary analysis. First, we excluded patients who received both IVIG and plasmapheresis on the same day from the plasmapheresis-first group to consider the possibility of IVIG being administered first when plasmapheresis and IVIG were administered on the same day. Second, given that the period of 5 days from systemic corticosteroid initiation in the inclusion criterion may have been arbitrary, we extended the definition of the 2 groups to within 7 days of systemic corticosteroid initiation. Third, considering the possibility of reducing the high dose of systemic corticosteroids according to the patient’s body weight, glucose tolerance, and infection risk, the inclusion criterion was changed to patients who received at least 500 mg/d of methylprednisolone equivalent systemic corticosteroid therapy within 3 days of hospitalization.
Results
Patient Sample and Baseline Characteristics
The unweighted and weighted characteristics of the 266 eligible patients (mean [SD] age, 56.7 [20.2] years; 114 [42.9%] men and 152 [57.1%] women) are described in Table 1. In the plasmapheresis-first group, 56.6% (30 of 53) patients received IVIG later, and in the IVIG-first group, 6.6% (14 of 213) patients received plasmapheresis later. In the plasmapheresis-first group, the mean (SD) period from the index date to receiving plasmapheresis and IVIG was 1.8 (1.3) and 6.3 (6.1) days, respectively. In the IVIG-first group, mean (SD) period from the index date to receiving IVIG and plasmapheresis was 1.5 (1.5) days and 5.8 (3.1) days, respectively. The C statistic was 0.851 (95% CI, 0.797–0.904). After overlap weighting, the baseline characteristics were well balanced between the 2 groups.
Table 1. Patient Characteristics Before and After Overlap Weighting in the Plasmapheresis-First Group (n = 53) vs Intravenous Immunoglobulin–First (n = 213) Group.
| Characteristic | Before overlap weighting | After overlap weightinga | |||
|---|---|---|---|---|---|
| Group, No. (%) | ASD | Group, No (%) | |||
| Plasmapheresis first | IVIG first | Plasmapheresis first | IVIG first | ||
| Age and sex | |||||
| Mean, y | 56.7 | 55.3 | NA | 55.9 | 55.5 |
| Age category | |||||
| ≤49 y | 18 (34.0) | 88 (41.3) | 15.2 | 37.3 | 37.3 |
| 50-59 y | 8 (15.1) | 28 (13.1) | 5.6 | 12.7 | 12.7 |
| 60-69 y | 11 (20.8) | 29 (13.6) | 19.0 | 21.7 | 21.7 |
| 70-79 y | 10 (18.9) | 40 (18.8) | 0.2 | 19.0 | 19.0 |
| ≥80 y | 6 (11.3) | 28 (13.1) | 5.6 | 9.3 | 9.3 |
| Female | 28 (52.8) | 124 (58.2) | 10.8 | 53.4 | 53.4 |
| Male | 25 (47.2) | 89 (41.8) | 10.8 | 46.6 | 46.6 |
| BMI | |||||
| ≤18.5 | 6 (11.3) | 28 (13.1) | 5.6 | 12.4 | 12.4 |
| 18.5-24.9 | 23 (43.4) | 131 (61.5) | 36.9 | 46.9 | 46.9 |
| 25.0-29.9 | 9 (17.0) | 32 (15.0) | 5.3 | 20.4 | 20.4 |
| ≥30.0 | 5 (9.4) | 5 (2.3) | 30.4 | 4.9 | 4.9 |
| Missing data | 10 (18.9) | 17 (8.0) | 32.3 | 15.4 | 15.4 |
| Smoking history | |||||
| Nonsmoker | 24 (45.3) | 127 (59.6) | 29.0 | 51.0 | 51.0 |
| Current/past | 20 (37.7) | 43 (20.2) | 39.4 | 32.8 | 32.8 |
| Missing data | 9 (17.0) | 43 (20.2) | 8.3 | 16.2 | 16.2 |
| Charlson comorbidity index | |||||
| 0 | 29 (54.7) | 121 (56.8) | 4.2 | 56.2 | 56.2 |
| 1 | 12 (22.6) | 49 (23.0) | 0.9 | 18.8 | 18.8 |
| 2 | 4 (7.5) | 28 (13.1) | 18.5 | 10.3 | 10.3 |
| 3 | 4 (7.5) | 7 (3.3) | 18.9 | 8.3 | 8.3 |
| ≥4 | 4 (7.5) | 8 (3.8) | 16.5 | 6.4 | 6.4 |
| Hospital volume (cases/study period) and type | |||||
| Low (1) | 20 (37.7) | 83 (39.0) | 2.5 | 36.7 | 36.7 |
| Medium-low (2) | 6 (11.3) | 38 (17.8) | 18.6 | 14.5 | 14.5 |
| Medium-high (3-4 ) | 19 (35.8) | 36 (16.9) | 44.0 | 28.7 | 28.7 |
| High (≥5) | 8 (15.1) | 56 (26.3) | 27.9 | 20.0 | 20.0 |
| Teaching hospital | 51 (96.2) | 205 (96.2) | 0.1 | 95.4 | 95.4 |
| ICU admission | 23 (43.4) | 48 (22.5) | 45.2 | 37.0 | 37.0 |
| Treatment ≤5 d from index date | |||||
| Mechanical ventilation | 6 (11.3) | 21 (9.9) | 4.7 | 11.7 | 11.7 |
| Kidney replacement therapy | 10 (18.9) | 9 (4.2) | 46.7 | 11.0 | 11.0 |
| Gross wound area treated/d, cm2 | |||||
| <3000 | 6 (11.3) | 44 (20.7) | 25.7 | 4.9 | 4.9 |
| 3000-6000 | 8 (15.1) | 21 (9.9) | 15.9 | 12.6 | 12.6 |
| ≥6000 | 18 (34.0) | 48 (22.5) | 25.6 | 29.6 | 29.6 |
| Missing data | 21 (39.6) | 100 (46.9) | 14.8 | 41.1 | 41.1 |
| Antibiotic ophthalmic solution/ointment | 24 (45.3) | 84 (39.4) | 11.8 | 45.1 | 45.1 |
| Corticosteroid ophthalmic solution/ointment | 27 (50.9) | 109 (51.2) | 0.5 | 50.2 | 50.2 |
| Corticosteroid body ointment | 26 (49.1) | 116 (54.5) | 10.8 | 52.0 | 52.0 |
| Antibiotic body ointment | 19 (35.8) | 53 (24.9) | 23.9 | 39.5 | 39.5 |
| Other kind of body ointment | 40 (75.5) | 147 (69.0) | 14.4 | 71.9 | 71.9 |
| Vasopressor | 6 (11.3) | 17 (8.0) | 11.3 | 11.4 | 11.4 |
| Opioid analgesics | 21 (39.6 | 59 (27.7) | 39.6 | 32.7 | 32.7 |
| Nonopioid analgesics | 21 (39.6) | 46 (21.6) | 25.3 | 35.3 | 35.3 |
| Proton-pump inhibitor use | 32 (60.4) | 137 (64.3) | 8.1 | 61.5 | 61.5 |
| H2 receptor antagonist use | 25 (47.2) | 76 (35.7) | 23.3 | 47.6 | 47.6 |
| Antibiotics (oral/intravenous) | |||||
| Penicillin, no antipseudomonal | 3 (5.7) | 12 (5.6) | 0.1 | 5.8 | 5.8 |
| Penicillin with antipseudomonal | 7 (13.2) | 11 (5.2) | 27.9 | 5.9 | 5.9 |
| Cephalosporin, no antipseudomonal | 9 (17.0) | 35 (16.4) | 1.5 | 18.5 | 18.5 |
| Cephalosporin with antipseudomonal | 3 (5.7) | 6 (2.8) | 14.0 | 6.5 | 6.5 |
| Macrolide | 4 (7.5) | 8 (3.8) | 16.4 | 7.9 | 7.9 |
| Sulfamethoxazole-trimethoprim | 8 (15.1) | 16 (7.5) | 0.1 | 14.2 | 14.2 |
| Fluoroquinolone | 5 (9.4) | 21 (9.9) | 1.4 | 10.3 | 10.3 |
| Carbapenem | 8 (15.1) | 21 (9.9) | 15.8 | 12.6 | 12.6 |
| Anti-MRSA drug | 9 (17.0) | 16 (7.5) | 29.0 | 13.1 | 13.1 |
| Antifungal drug | 6 (11.3) | 23 (10.8) | 1.7 | 13.4 | 13.4 |
| Red blood cell transfusion | 6 (11.3) | 16 (7.5) | 13.0 | 11.0 | 11.0 |
| Platelet concentrates transfusion | 5 (9.4) | 7 (3.3) | 25.2 | 6.7 | 6.7 |
Abbreviations: ASD, absolute standardized difference; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); H2, histamine type-2; ICU, intensive care unit; IVIG, intravenous immunoglobulin; NA, not applicable; MRSA, methicillin-resistant Staphylococcus aureus.
All ASDs after overlap weighting were zero.
Table 2 shows the results of crude outcome comparisons between the groups. The in-hospital mortality rates were 14.1% (30 patients of 213) and 18.9% (10 patients of 53) in the IVIG-first and plasmapheresis-first groups, respectively. The mean (SD) length of hospital stay was 34.7 (26.5) and 44.8 (32.0) days in the IVIG- and plasmapheresis-first groups, respectively. The total (SD) costs were US $23 584 ($19 002) and $34 527 ($25 498) in the IVIG- and plasmapheresis-first groups, respectively.
Table 2. Crude Outcomes, by Plasmapheresis-First Group (n = 53) vs Intravenous Immunoglobulin–First (n = 213) Group.
| Outcome | Treatment group, mean (SD) | |
|---|---|---|
| Plasmapheresis first | IVIG first | |
| In-hospital mortality, No. (%) | 10 (18.9) | 30 (14.1) |
| Length of stay, d | 44.8 (32.0) | 34.8 (26.5) |
| Total costs, US dollars | 34 537 (25 498) | 23 584 (19 002) |
Abbreviation: IVIG, intravenous immunoglobulin.
The adjusted outcomes obtained using overlap weighting are presented in Table 3. No significant difference in the in-hospital mortality rate was observed between the plasmapheresis- and IVIG-first groups (18.3% vs 19.5%; OR, 0.93; 95% CI, 0.38–2.23; P = .86). The plasmapheresis-first group showed a longer average length of hospital stay than the IVIG-first group (45.3 vs 32.8 days; difference, 12.5 days; 95% CI, 0.4–24.5 days; P = .04) and higher average total costs (US $34 262 vs $23 054; difference, $11 207; 95% CI, $2789–$19 626; P = .009).
Table 3. Outcome Analyses After Propensity-Score Overlap Weighting, by Plasmapheresis-First Group (n = 53) vs Intravenous Immunoglobulin–First (n = 213) Group.
| Outcome | Treatment group, mean (SD) | OR/difference (95% CI) | P value | |
|---|---|---|---|---|
| Plasmapheresis first | IVIG first | |||
| In-hospital mortality, % | 18.2 | 19.8 | 0.90 (0.37-2.17) | .81 |
| Length of stay, d | 45.3 (32.6) | 32.7 (26.7) | 12.5 (0.42-24.7) | .04 |
| Total cost, US $ | 34 177 (27 178) | 23 029 (17 978) | 11 148 (2801-19 492) | .009 |
Abbreviations: IVIG, intravenous immunoglobulin; OR, odds ratio.
The results of each sensitivity analyses are reported in eTables 1 to 3 in Supplement 1. The results of the sensitivity analysis were similar to the original results.
Discussion
Using a nationwide inpatient database and the overlap weighting method, we investigated whether IVIG or plasmapheresis should be performed first in patients with SJS/TEN who do not respond to steroid pulse therapy. In-hospital mortality did not differ significantly between the 2 groups. However, longer in-hospital stays and higher hospitalization costs were observed in patients who received plasmapheresis therapy first than in those who received IVIG first.
The population in this study comprised inpatients with severe SJS/TEN with poor response to systemic corticosteroids. Patient mortality in that study was comparable with that reported in recent studies.5,12,13,37 The present study findings suggest that plasmapheresis is not superior to IVIG in terms of mortality. There are several possible reasons for this finding. First, repeated plasmapheresis may deplete patient immunoglobulins, possibly affecting susceptibility to sepsis.38 Sepsis is a major cause of death in patients with SJS/TEN, and patients with significant epidermal necrosis are also at a high risk of bacteremia.39,40,41 Immunoglobulin deficiency may offset the benefit of plasmapheresis, including the removal of inflammatory mediators.16,17,18,19,20,21 Second, patients in the plasmapheresis-first group who received IVIG later may have missed the appropriate timing of IVIG administration. Patients with SJS/TEN maintain a high level of Fas-ligand for 5 days from disease onset.42 Because IVIG inhibits the Fas and Fas-ligand interaction, cell apoptosis and inhibition of epidermal necrosis follow.10
On the basis of these findings, a previous study suggested that the earlier administration of IVIG may be associated with better ocular outcomes.43 However, in the present study, more than 50% of patients in the plasmapheresis-first group received IVIG administration within an average of 4 to 5 days from plasmapheresis. Considering whether to administer IVIG on the basis of the response to plasmapheresis may have occluded the appropriate timing for IVIG administration, and the patients who received IVIG later may have already had advanced epidermal necrosis; therefore, they may not have received the potential benefit of IVIG. In the IVIG-first group, as well as the plasmapheresis-first group, there were 4 to 5 days from prior treatment to another treatment; however, fewer than 10% of patients in the IVIG-first group received plasmapheresis. This difference may have been associated with the longer length of hospital stay. Furthermore, the relatively higher cost of plasmapheresis vs IVIG cost may have increased hospitalization costs.
The strengths of the current study are its relatively large sample size of 266 patients with severe SJS/TEN who did not respond to systemic corticosteroids; the use of an appropriate control group using overlap weighting, and analyses that adjusted for a large number of potential confounding factors. Given that SJS/TEN is a rare and severe disease, conducting a large-scale prospective cohort study or randomized clinical trial would be difficult. Moreover, the invasiveness of plasmapheresis makes it difficult to conduct any randomized clinical trials. To our knowledge, only a few small case series studies have reported the benefits of plasmapheresis16,17,18,19,20,21 and because these studies had no control group, the results may have been subject to significant biases (eg, selection bias).
We used propensity-score overlap weighting because patient characteristics across the treatments were substantially different. Cautious interpretation is required when using this method because patients who had a high probability of receiving either treatment contributed less to the findings. In other words, the main target of inference was the overlap population, a reasonable clinical equipoise for each treatment decision.31 These study findings suggest that plasmapheresis does not have an evident advantage over IVIG. However, the effects of plasmapheresis are unknown for patients with profoundly severe symptoms who nearly always receive plasmapheresis in response to their medical condition. Further studies should evaluate the effects of plasmapheresis on this patient population and the effects after both steroid pulse therapy and IVIG administration.
Limitations
This study had a number of limitations. First, although we tried to adjust for the potential confounding factors by using overlap weighting, unmeasured confounding factors may have biased the results, including the percentage of skin detachment, pulse rate, and several laboratory data contained in SCORETEN (a scoring system of prognosis in patients with SJS/TEN),44,45 despite selecting several treatments as surrogate covariates for the unmeasured confounding factors. Second, the current study focused on which treatment was given first, but combination therapy may have led to residual confounding. Third, we diagnosed SJS/TEN on the basis of ICD-10 codes. Several diagnostic ICD-10 codes in the database have been well validated (eg, heart failure and liver disease codes).46 However, the accuracy of the ICD-10 code for SJS/TEN itself has not been directly validated. To address this problem, we combined the diagnostic and procedure codes, which are well validated and highly accurate,46 in the inclusion criteria. Given that a high dose (≥1000 mg/d of methylprednisolone) of systemic corticosteroids is rarely administered to patients with dermatological diseases that should be differentiated from SJS/TEN, —eg, acute generalized exanthematous pustulosis or drug reactions with eosinophilia and systemic symptoms—it is reasonable to consider that the accuracy of SJS/TEN diagnosis in our study should be high. Fourth, we did not investigate long-term life prognosis, functional prognosis, and patient quality of life, including ocular and vulvovaginal sequelae, because this national database does not contain these data. Fifth, generalizability to other countries is limited because the study population was predominantly composed of individuals of Japanese origin.
Conclusions
The findings of this retrospective cohort study suggest that there is no significant difference in mortality rates between inpatients with SJS/TEN who received plasmapheresis first vs IVIG therapy first after ineffective systemic corticosteroid therapy. Moreover, the findings suggest that patients with SJS/TEN who receive plasmapheresis therapy first may have longer hospitalization stays and incur higher expenses. Thus, IVIG administration first, before plasmapheresis, may be the preferred course of treatment when corticosteroids have been ineffective in patients with SJS/TEN.
eAppendix 1. Details of Diagnosis Procedure Combination (DPC) database
eAppendix 2. Details of covariates
eAppendix 3. Details of overlap weighting and calculating propensity score
eFigure. Examples of patients who were assigned to or excluded from each group
eTable 1. Results of the overlap weighting analyses after excluding patients who received IVIG and plasmapheresis on the same day
eTable 2. Results of the overlap weighting analyses after extending the definition of the 2 groups to within 7 days of steroid administration.
eTable 3. Results of the overlap weighting analyses after including patients who received 500–1000 mg/day of methylprednisolone in addition to the original cohort
eReferences.
Data Sharing Statement
References
- 1.Mockenhaupt M. The current understanding of Stevens-Johnson syndrome and toxic epidermal necrolysis. Expert Rev Clin Immunol. 2011;7(6):803-813. doi: 10.1586/eci.11.66 [DOI] [PubMed] [Google Scholar]
- 2.Auquier-Dunant A, Mockenhaupt M, Naldi L, Correia O, Schröder W, Roujeau JC; SCAR Study Group. Severe Cutaneous Adverse Reactions . Correlations between clinical patterns and causes of erythema multiforme majus, Stevens-Johnson syndrome, and toxic epidermal necrolysis: results of an international prospective study. Arch Dermatol. 2002;138(8):1019-1024. doi: 10.1001/archderm.138.8.1019 [DOI] [PubMed] [Google Scholar]
- 3.Bastuji-Garin S, Rzany B, Stern RS, Shear NH, Naldi L, Roujeau JC. Clinical classification of cases of toxic epidermal necrolysis, Stevens-Johnson syndrome, and erythema multiforme. Arch Dermatol. 1993;129(1):92-96. doi: 10.1001/archderm.1993.01680220104023 [DOI] [PubMed] [Google Scholar]
- 4.Roujeau JC, Stern RS. Severe adverse cutaneous reactions to drugs. N Engl J Med. 1994;331(19):1272-1285. doi: 10.1056/NEJM199411103311906 [DOI] [PubMed] [Google Scholar]
- 5.Mahar PD, Wasiak J, Hii B, et al. A systematic review of the management and outcome of toxic epidermal necrolysis treated in burns centres. Burns. 2014;40(7):1245-1254. doi: 10.1016/j.burns.2014.02.006 [DOI] [PubMed] [Google Scholar]
- 6.Dodiuk-Gad RP, Olteanu C, Chung WH, Shear NH. Drug Safety: the 9th International Congress on Cutaneous Adverse Drug Reactions at the 23rd World Congress of Dermatology in Vancouver. 2016;39(3):271-276. doi: 10.1007/s40264-015-0390-9 [DOI] [PubMed] [Google Scholar]
- 7.Dodiuk-Gad RP, Olteanu C, Jeschke MG, Cartotto R, Fish J, Shear NH. Treatment of toxic epidermal necrolysis in North America. J Am Acad Dermatol. 2015;73(5):876-7.e2. doi: 10.1016/j.jaad.2015.08.008 [DOI] [PubMed] [Google Scholar]
- 8.Creamer D, Walsh SA, Dziewulski P, et al. U.K. guidelines for the management of Stevens-Johnson syndrome/toxic epidermal necrolysis in adults 2016. Br J Dermatol. 2016;174(6):1194-1227. doi: 10.1111/bjd.14530 [DOI] [PubMed] [Google Scholar]
- 9.Kinoshita Y, Saeki H. A review of toxic epidermal necrolysis management in Japan. Allergol Int. 2017;66(1):36-41. doi: 10.1016/j.alit.2016.06.001 [DOI] [PubMed] [Google Scholar]
- 10.Viard I, Wehrli P, Bullani R, et al. Inhibition of toxic epidermal necrolysis by blockade of CD95 with human intravenous immunoglobulin. Science. 1998;282(5388):490-493. doi: 10.1126/science.282.5388.490 [DOI] [PubMed] [Google Scholar]
- 11.Schneck J, Fagot JP, Sekula P, Sassolas B, Roujeau JC, Mockenhaupt M. Effects of treatments on the mortality of Stevens-Johnson syndrome and toxic epidermal necrolysis: a retrospective study on patients included in the prospective EuroSCAR Study. J Am Acad Dermatol. 2008;58(1):33-40. doi: 10.1016/j.jaad.2007.08.039 [DOI] [PubMed] [Google Scholar]
- 12.Micheletti RG, Chiesa-Fuxench Z, Noe MH, et al. Stevens-Johnson syndrome/toxic epidermal necrolysis: a multicenter retrospective study of 377 adult patients from the United States. J Invest Dermatol. 2018;138(11):2315-2321. doi: 10.1016/j.jid.2018.04.027 [DOI] [PubMed] [Google Scholar]
- 13.Tsai TY, Huang IH, Chao YC, et al. Treating toxic epidermal necrolysis with systemic immunomodulating therapies: a systematic review and network meta-analysis. J Am Acad Dermatol. 2021;84(2):390-397. doi: 10.1016/j.jaad.2020.08.122 [DOI] [PubMed] [Google Scholar]
- 14.Ye LP, Zhang C, Zhu QX. The effect of intravenous immunoglobulin combined with corticosteroid on the progression of Stevens-Johnson syndrome and toxic epidermal necrolysis: a meta-analysis. PLoS One. 2016;11(11):e0167120. doi: 10.1371/journal.pone.0167120 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Curtis JA, Christensen LC, Paine AR, et al. Stevens-Johnson syndrome and toxic epidermal necrolysis treatments: an internet survey. J Am Acad Dermatol. 2016;74(2):379-380. doi: 10.1016/j.jaad.2015.08.033 [DOI] [PubMed] [Google Scholar]
- 16.Yamada H, Takamori K. Status of plasmapheresis for the treatment of toxic epidermal necrolysis in Japan. Ther Apher Dial. 2008;12(5):355-359. doi: 10.1111/j.1744-9987.2008.00609.x [DOI] [PubMed] [Google Scholar]
- 17.Lissia M, Figus A, Rubino C. Intravenous immunoglobulins and plasmapheresis combined treatment in patients with severe toxic epidermal necrolysis: preliminary report. Br J Plast Surg. 2005;58(4):504-510. doi: 10.1016/j.bjps.2004.12.007 [DOI] [PubMed] [Google Scholar]
- 18.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]
- 19.Kamanabroo D, Schmitz-Landgraf W, Czarnetzki BM. Plasmapheresis in severe drug-induced toxic epidermal necrolysis. Arch Dermatol. 1985;121(12):1548-1549. doi: 10.1001/archderm.1985.01660120074023 [DOI] [PubMed] [Google Scholar]
- 20.Bamichas G, Natse T, Christidou F, et al. Plasma exchange in patients with toxic epidermal necrolysis. Ther Apher. 2002;6(3):225-228. doi: 10.1046/j.1526-0968.2002.00409.x [DOI] [PubMed] [Google Scholar]
- 21.Szczeklik W, Nowak I, Seczynska B, Sega A, Krolikowski W, Musial J. Beneficial therapeutic effect of plasmapheresis after unsuccessful treatment with corticosteroids in two patients with severe toxic epidermal necrolysis. Ther Apher Dial. 2010;14(3):354-357. doi: 10.1111/j.1744-9987.2009.00800.x [DOI] [PubMed] [Google Scholar]
- 22.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]
- 23.Furubacke A, Berlin G, Anderson C, Sjöberg F. Lack of significant treatment effect of plasma exchange in the treatment of drug-induced toxic epidermal necrolysis? Intensive Care Med. 1999;25(11):1307-1310. doi: 10.1007/s001340051063 [DOI] [PubMed] [Google Scholar]
- 24.Kridin K, Brüggen MC, Chua SL, et al. Assessment of treatment approaches and outcomes in Stevens-Johnson syndrome and toxic epidermal necrolysis: insights from a pan-European multicenter study. JAMA Dermatol. 2021;157(10):1182-1190. doi: 10.1001/jamadermatol.2021.3154 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Yasunaga H. Real world data in Japan: chapter II the diagnosis procedure combination database. Ann Clin Epidemiol. 2019:1(3):76-79. doi: 10.37737/ace.1.3_76 [DOI] [Google Scholar]
- 26.Morita K, Matsui H, Michihata N, Fushimi K, Yasunaga H. Association of early systemic corticosteroid therapy with mortality in patients with Stevens-Johnson syndrome or toxic epidermal necrolysis: a retrospective cohort study using a nationwide claims database. Am J Clin Dermatol. 2019;20(4):579-592. doi: 10.1007/s40257-019-00443-9 [DOI] [PubMed] [Google Scholar]
- 27.Li F, Thomas LE, Li F. Addressing extreme propensity scores via the overlap weights. Am J Epidemiol. 2019;188(1):250-257. doi: 10.1093/aje/kwy201 [DOI] [PubMed] [Google Scholar]
- 28.Desai RJ, Franklin JM. Alternative approaches for confounding adjustment in observational studies using weighting based on the propensity score: a primer for practitioners. BMJ. 2019;367:l5657. doi: 10.1136/bmj.l5657 [DOI] [PubMed] [Google Scholar]
- 29.Li F, Morgan KL, Zaslavsky AM. Balancing covariates via propensity score weighting. J Am Stat Assoc. 2018;113(521):390-400. doi: 10.1080/01621459.2016.126046629930437 [DOI] [Google Scholar]
- 30.Mehta N, Kalra A, Nowacki AS, et al. Association of use of angiotensin-converting enzyme inhibitors and angiotensin II receptor blockers with testing positive for coronavirus disease 2019 (COVID-19). JAMA Cardiol. 2020;5(9):1020-1026. doi: 10.1001/jamacardio.2020.1855 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Thomas LE, Li F, Pencina MJ. Overlap weighting: a propensity score method that mimics attributes of a randomized clinical trial. JAMA. 2020;323(23):2417-2418. doi: 10.1001/jama.2020.7819 [DOI] [PubMed] [Google Scholar]
- 32.Otaka S, Aso S, Matsui H, Fushimi K, Yasunaga H. Associations between early parenteral nutrition and in-hospital outcomes in underweight patients with gastrointestinal surgery. Clin Nutr ESPEN. 2021;43:464-470. doi: 10.1016/j.clnesp.2021.03.005 [DOI] [PubMed] [Google Scholar]
- 33.Yasunaga H. Introduction to applied statistics—chapter 1 propensity score analysis. Ann Clin Epidemiol. 2020;2(2):33-37. doi: 10.37737/ace.2.2_33 [DOI] [Google Scholar]
- 34.Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika. 1983;70(1):41-55. doi: 10.1093/biomet/70.1.41 [DOI] [Google Scholar]
- 35.Rosenbaum PR, Rubin DB. Constructing a control group using multivariate matched sampling methods that incorporate the propensity score. Am Stat. 1985;39(1):33-38. doi: 10.1080/00031305.1985.10479383 [DOI] [Google Scholar]
- 36.Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med. 2009;28(25):3083-3107. doi: 10.1002/sim.3697 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Zimmermann S, Sekula P, Venhoff M, et al. Systemic immunomodulating therapies for Stevens-Johnson syndrome and toxic epidermal necrolysis: a systematic review and meta-analysis. JAMA Dermatol. 2017;153(6):514-522. doi: 10.1001/jamadermatol.2016.5668 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Wing EJ, Bruns FJ, Fraley DS, Segel DP, Adler S. Infectious complications with plasmapheresis in rapidly progressive glomerulonephritis. JAMA. 1980;244(21):2423-2426. doi: 10.1001/jama.1980.03310210025020 [DOI] [PubMed] [Google Scholar]
- 39.Knight L, Muloiwa R, Dlamini S, Lehloenya RJ. Factors associated with increased mortality in a predominantly HIV-infected population with Stevens Johnson syndrome and toxic epidermal necrolysis. PLoS One. 2014;9(4):e93543. doi: 10.1371/journal.pone.0093543 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.de Prost N, Ingen-Housz-Oro S, Duong TA, et al. Bacteremia in Stevens-Johnson syndrome and toxic epidermal necrolysis: epidemiology, risk factors, and predictive value of skin cultures. Medicine (Baltimore). 2010;89(1):28-36. doi: 10.1097/MD.0b013e3181ca4290 [DOI] [PubMed] [Google Scholar]
- 41.Koh HK, Chai ZT, Tay HW, et al. Risk factors and diagnostic markers of bacteremia in Stevens-Johnson syndrome and toxic epidermal necrolysis: a cohort study of 176 patients. J Am Acad Dermatol. 2019;81(3):686-693. doi: 10.1016/j.jaad.2019.05.096 [DOI] [PubMed] [Google Scholar]
- 42.Murata J, Abe R, Shimizu H. Increased soluble Fas ligand levels in patients with Stevens-Johnson syndrome and toxic epidermal necrolysis preceding skin detachment. J Allergy Clin Immunol. 2008;122(5):992-1000. doi: 10.1016/j.jaci.2008.06.013 [DOI] [PubMed] [Google Scholar]
- 43.Kim KH, Park SW, Kim MK, Wee WR. Effect of age and early intervention with a systemic steroid, intravenous immunoglobulin or amniotic membrane transplantation on the ocular outcomes of patients with Stevens-Johnson syndrome. Korean J Ophthalmol. 2013;27(5):331-340. doi: 10.3341/kjo.2013.27.5.331 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Guégan S, Bastuji-Garin S, Poszepczynska-Guigné E, Roujeau JC, Revuz J. Performance of the SCORTEN during the first five days of hospitalization to predict the prognosis of epidermal necrolysis. J Invest Dermatol. 2006;126(2):272-276. doi: 10.1038/sj.jid.5700068 [DOI] [PubMed] [Google Scholar]
- 45.Bastuji-Garin S, Fouchard N, Bertocchi M, Roujeau JC, Revuz J, Wolkenstein P. SCORTEN: a severity-of-illness score for toxic epidermal necrolysis. J Invest Dermatol. 2000;115(2):149-153. doi: 10.1046/j.1523-1747.2000.00061.x [DOI] [PubMed] [Google Scholar]
- 46.Yamana H, Moriwaki M, Horiguchi H, Kodan M, Fushimi K, Yasunaga H. Validity of diagnoses, procedures, and laboratory data in Japanese administrative data. J Epidemiol. 2017;27(10):476-482. doi: 10.1016/j.je.2016.09.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eAppendix 1. Details of Diagnosis Procedure Combination (DPC) database
eAppendix 2. Details of covariates
eAppendix 3. Details of overlap weighting and calculating propensity score
eFigure. Examples of patients who were assigned to or excluded from each group
eTable 1. Results of the overlap weighting analyses after excluding patients who received IVIG and plasmapheresis on the same day
eTable 2. Results of the overlap weighting analyses after extending the definition of the 2 groups to within 7 days of steroid administration.
eTable 3. Results of the overlap weighting analyses after including patients who received 500–1000 mg/day of methylprednisolone in addition to the original cohort
eReferences.
Data Sharing Statement