Introduction

Steven Johnson Syndrome (SJS) and Toxic Epidermal Necrolysis (TEN) are grouped together under the terminology of epidermal necrolysis (EN), which is a life-threatening dermatologic condition characterized by a dusky morbilliform eruption with mucosal involvement, that rapidly progresses to vesicles, blisters, and skin sloughing.

There is a well-established need for evidence-based studies to define best supportive care practices and short-term treatment in EN [1]. Most existing studies examining these treatment methods are retrospective, but several randomized clinical trials have been completed. A lack of standardization and validation for existing endpoints has been identified as a key barrier when comparing therapies, hindering the development of evidenced-based treatment [1]. The purpose of this review is to systematically evaluate instruments used for outcome measures in prospective studies in the acute phase of EN.

Methods

This systematic review followed PRISMA guidelines and was registered in PROSPERO (CRD42023405479). PubMed and EMBASE were utilized to identify prospective studies evaluating systemic and/or topical treatments for EN, including dressings and ocular treatments. We included studies which used one or more acute, EN-specific therapies, reported clear outcome measurements, were prospective in nature, and were available in English. Studies were excluded if they were not primary literature, presented outcome measurements for EN that could not be separated from other conditions, or assessed outcome measures related to EN secondary complications (e.g. xerophthalmia).

Data collected from these studies included study type, outcome measures, and primary or secondary endpoints if defined. The Cochrane Risk of Bias 2 (RoB 2) tool was used to assess the risk of bias in randomized control trials (RCTs) [2]. Each study was reviewed and scored according to the recommended protocol given by the authors of the tool.

Results

A total of 13 studies from 1997 to 2018 were included in the analysis (Fig. 1). Three RCTs, 5 case series, and 5 retrospective cohort studies were identified. Outcomes measures were divided into four categories: mortality assessment, cutaneous outcomes, mucosal outcomes, and non-cutaneous outcomes (Table 1).

Table 1 Outcome measurements in treatment of acute EN
Full size table
Fig. 1
figure 1

PRISMA flow diagram for identifying studies assessing treatment of acute EN

Full size image

Mortality assessment

Seven studies assessed patient mortality. The most commonly used metric to assess mortality was the standardized mortality ratio (n = 4), which is defined as the ratio of observed deaths compared to the expected deaths using the severity of illness score for TEN (SCORTEN) mortality predictor and is a frequently utilized endpoint in studies of epidermal necrolysis [1]. Other methods used in mortality assessment included comparison of treatment group mortality to an alternate treatment group (n = 1), placebo group (n = 3), or historical control group from a retrospective database (n = 1, Table 1).

Cutaneous outcomes

Assessment of cutaneous outcomes primarily focused on skin detachment and reepithelialization. Studies that quantified cutaneous involvement (n = 8) used classic burn tables or palm measurements to provide an estimated body surface area. All cutaneous outcomes were clinician reported outcome measurements (ClinROM) and there were no validated methods or instruments utilized.

Mucosal outcomes

Outcomes of oral and/or ocular mucosal disease were assessed in 2 of 13 studies (Table 2). Measured ocular outcomes included adverse events and complications, changes in tear break up time (TBUT), and visual acuity. All mucosal outcomes were ClinROMs and were unvalidated.

Table 2 Risk of bias in randomized control trials using the cochrane risk of bias 2 (RoB 2 tool)
Full size table

Non-cutaneous clinical outcomes

Three studies collectively assessed non-cutaneous adverse events and complications, including hospitalization duration. Adverse events, complications, and general state of health, following treatment were non-standardized descriptive reports by authors.

Scoring tools used included the Severity of Illness Score (n = 1) and Simplified Acute Physiology Score (SAPS, n = 1). SAPS is a validated critical care scoring tool used to assess prognosis in critically ill patients and severity of disease in clinical trials. Scores were calculated and compared at different time points to quantify treatment effectiveness. In contrast, the Severity of Illness Score used by Aihara et al. was designed from a retrospective survey assessing ophthalmic lesions, lip/oral lesions, cutaneous lesions, and general condition [3]. Other objective outcomes included hospitalization time and fever duration. All overall clinical outcome measures were ClinROMs.

Risk of bias in randomized control trials

The Cochrane RoB 2 tool was created to provide an assessment of the risk of bias in RCTs, where bias is defined as systematic deviation from a theoretically ideal RCT. All three RCTs included in this study had some concerns for risk of bias overall (Table 2). All studies had low risk of bias in the randomization process and two had low risk of bias due to missing outcome data. In contrast, one study had high rates of patient death resulting in a high risk of bias due to missing outcome data. There were concerns noted in every other domain for all studies.

Discussion

Measured outcomes used to assess treatment effectiveness in the acute phase of EN varied across studies. Most data came from prospective case reports and cohort studies representing the lack of available randomized clinical trial data available in EN.

The search did not reveal any EN-specific validated measures or scoring tools used to assess disease progression and outcomes. Adverse events and complications were assessed in multiple studies; however, they were not a priori defined, and it was unclear if these events were the result of the EN disease process, therapeutics, or another cause.

The standardized mortality ratio was widely used across studies. As an outcome measurement, the standardized mortality ratio likely requires additional calibration as SCORTEN is thought to underestimate mortality for low scores and overestimate mortality for high scores [1]. Regardless, the standardized mortality ratio provides a consistent (albeit unvalidated) method of mortality assessment across studies.

Measures of EN-specific cutaneous outcomes across studies varied. Many studies assessed skin detachment or reepithelialization, however their specific endpoints were nuanced. Some studies took a “time-to-event” approach, which measures the time to desired outcome, such as completion or commencement of skin reepithelialization. This measurement does not include data from patients who died, and therefore must be considered in light of mortality measurements. Further, measurement of epithelial damage and healing has not been standardized as an outcome measure across clinicians.

In addition to the time-to-event approach, several studies compared cutaneous outcomes (skin detachment, reepithelialization) at predetermined timepoints. While EN is known to take weeks to resolve, there was no apparent strategic rationale for selecting these timepoints in relation to the SJS/TEN disease process. Further, there is no validated metric to quantitate morphology and distribution of cutaneous lesions at baseline for adequate comparison.

Interestingly, resolution of oral mucosal and ocular involvement were monitored in one study each [4, 5]. All outcomes measured were ClinROMs and there were no patient reported outcome measures (PROMs). There are currently no studies which assess the validity of ClinROMs or PROMs as outcome measurements for EN.

Across all RCTs included in this review, there was some concern for risk of bias. It is worth considering that the high mortality rate of patients in Wolkenstein et al. contributed to a high risk of bias from missing outcome data [5]. This highlights a unique challenge in minimizing bias for RCTs looking at treatments for EN, especially at tertiary centers which accept patients with increased disease severity.

Limitations of this study include the exclusion of non-prospective literature and studies in languages other than English.

With little consensus about management and treatment of EN, development of validated outcome measures, with representation of the oral and genital mucosa, is essential for future studies including RTCs. One way of accomplishing this is the establishment of a Core Outcome Set for EN, which identifies consensus-based outcomes to be measured and reported across all clinical trials.