Early high-dose intravenous corticosteroids rapidly arrest Stevens Johnson syndrome and drug reaction with eosinophilia and systemic symptoms recurrence on drug re-exposure

Stevens Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN), collectively called SJS/TEN, form a severe cutaneous adverse drug reaction (SCAR) spectrum characterized by skin and mucosal denudation. SJS affects ≤10% body surface area, and TEN ≥30%, with SJS/TEN overlap lying in between.¹ Drug reaction with eosinophilia and systemic symptoms (DRESS) is another SCAR characterized by rash, hematologic abnormalities (particularly eosinophilia), lymphadenopathy, and internal organ involvement.²

Withdrawal of the drug and multidisciplinary supportive care are the mainstay of treatment. Currently, there is no consensus on efficacious pharmacologic treatment of SJS/TEN to halt progression and reduce mortality. Disease rarity, absence of randomized controlled trials, lack of good case-control studies, small study sample sizes, inclusion of noncases, inconsistent definition severity, and therapeutic response are the main reasons.³ Pharmacologic interventions that have been used include thalidomide, systemic corticosteroids, intravenous immunoglobulins (IVIG), cyclosporine, and TNF-alpha blockers.¹ For DRESS, corticosteroids are often used in the absence of an accepted gold standard and suffer the same flaws as in SJS/TEN.²

Corticosteroids are commonly used to improve outcomes in SJS/TEN due to familiarity with their use and cost-effectiveness. A survey of American clinicians reported the use of systemic corticosteroids as monotherapy by 43%, 30%, and 20% for SJS, SJS/TEN, and TEN, respectively, whereas in combination with IVIG by 34%, 59%, and 70%, respectively.⁴ However, their efficacy in SJS/TEN is unclear. Systematic reviews of best studies using systemic corticosteroids in SJS/TEN are inconclusive.^1,5 Supporters of corticosteroid suggest that they are efficacious when used early in high doses, highlighting that most studies used low doses late in the disease evolution.^1,5,6 Concerns regarding corticosteroid use in SJS/TEN relate to higher mortality, delaying of re-epithelialization, risk of sepsis, and protein catabolism.^5,6 Knowledge gaps on their use for SJS/TEN include cutoff times for potential benefits, best administration route, safest most effective corticosteroid, optimum dose, duration of use, and adverse events.^1,6

Rechallenge with the implicated and structurally related drugs can cause a life-threatening recurrence in severe SCAR phenotypes and should not be attempted in the presence of effective alternative drugs. Exceptions exist in low- and middle-income countries where diseases of high global burden, such as HIV and tuberculosis, demand complex treatment regimens and where these reactions may significantly restrict treatment options. Our high HIV prevalence and tuberculosis incidence results in disproportionately high SCAR incidence during multidrug treatment for tuberculosis. Because of paucity of affordable, effective antituberculosis drugs and resource constraints, we developed a rechallenge protocol for patients with antituberculosis drug–associated SCAR (see Figure E1 in this article’s Online Repository at www.jaci-inpractice.org). We rechallenge with all 4 drugs to identify and remove the offender from the regimen as we have previously described.⁷ Briefly, after the patient returns to normal, we rechallenge them sequentially and additively in the order isoniazid, rifampicin, pyrazinamide, and ethambutol. The usual interval between individual drug rechallenges was 4 days, as >95% of rechallenge reactions (RR) develop with 72 hours. We previously described features we used to identify RR⁷ (see Table E1 in this article’s Online Repository at www.jaci-inpractice.org). In case of RR, the culprit drug is permanently stopped and rechallenge with the next drug is delayed until clinical and laboratory parameters return to normal. Three-quarters of our patients develop RRs to first-line antituberculosis drugs that range from mild to severe.⁷ In a previous study of 97 patients rechallenged with first-line antituberculosis drugs, we found 1 RR to prolong the rechallenge process and hospitalization by a median of 10 days (interquartile range, 5-17), significantly increasing personal and institutional costs. Delay in normalization of clinical and laboratory parameters after RR was the biggest contributor to prolonged hospitalization.⁷

In an attempt to reduce severity of RRs, shorten hospitalization, and prevent sequelae, we decided to empirically initiate intravenous corticosteroids with first clinical signs of RR we adjudged as rapid and severe. Before rechallenge, we obtained renal function, peripheral IV access, and performed a cardiac examination to facilitate rapid treatment with systemic corticosteroids if indicated. We postulated that because of our unique situation of being able to initiate corticosteroids very early compared with published studies, a single comparatively lower dose of 100 mg of methylprednisolone would suffice.¹ However, the dose and frequency of administration could be adjusted at the discretion of the treating physician. Our selection of methylprednisolone was based on institutional availability and its relatively longer half-life compared with other infusible corticosteroids.

We retrospectively report outcomes of early methylprednisolone infusion in 6 consecutive cases of SJS/TEN and DRESS who were rechallenged with first-line antituberculosis drugs at a tertiary center in Cape Town, South Africa.

The baseline characteristics of the 6 patients, their RR features, offending drugs, doses of methylprednisolone used, and clinical responses are summarized in Table I. In summary, their median age was 27 years, 4 were female, 5 were HIV infected with a median CD4 count of 178 cells/mm³, and HIV-uninfected patient had systemic lupus erythematosus; 3 patients presented with DRESS, and 3 with SJS/TEN. All 4 drugs caused at least 1 RR. The median interval (range) between initiating rechallenge and first clinically detectable RR features was 45 minutes (20-150 minutes). The median interval (range) between first RR features and initiating methylprednisolone infusion was 25 minutes (20-180 minutes). All 6 received a single infusion of methylprednisolone: three received 100 mg, one 120 mg, and two 200 mg. The median time (range) to patient-reported clinically significant improvement was 45 minutes (25-150 minutes). None reported a recurrence of the symptoms that had defined their RR during the following 72 hours. All 6 tolerated the other antituberculosis drugs to treatment completion (range, 6-9 months).

TABLE I.

Baseline characteristics and clinical features of 6 patients treated with intravenous methylprednisolone after developing a rechallenge reaction to an antituberculosis drug

Variable	Case 1	Case 2	Case 3	Case 4	Case 5	Case 6
Gender	Female	Female	Female	Male	Male	Female
Age (y)	25	24	29	31	37	48
HIV-infected	Yes	Yes	Yes	Yes	Yes	No
CD4 count (cells/mm3)	66	202	181	36	178	N/A
Phenotype of the original drug reaction	DRESS	SJS	DRESS	TEN	DRESS	SJS/TEN overlap
Offending drug identified by rechallenge	H	Z	H	R	E	H
Interval between index reaction and initiation of rechallenge (d)	24	15	65	23	39	44
Interval between rechallenge and first symptoms (min)	45	30	60	45	150	20
Interval between symptoms and IVI corticosteroid (min)	10	20	120	20	30	180
Dose of methylprednisolone given (mg)	100	100	100	200	200	120
Time to clinical improvement (min)	120	25	45	30	150	45
Features of the rechallenge reaction
Rigors	No	No	Yes	No	No	Yes
Fever	No	No	Yes	No	No	Yes
Tachycardia	Yes	Yes	Yes	No	No	Yes
Increase in blood pressure	No	No	No	No	Yes	Yes
Pruritus	Yes	Yes	Yes	Yes	No	Yes
Painful skin	No	Yes	No	Yes	No	Yes
Nausea	No	No	Yes	No	No	No
Vomiting	No	No	No	No	No	No
Headache	No	No	No	No	No	No
Photophobia	No	Yes	No	No	No	Yes
Mucosal erythema	No	No	No	No	No	Yes
Skin erythema	Yes	Yes	Yes	Yes	Yes	Yes
Mucosal necrosis	No	No	No	No	No	No
Skin necrosis	No	No	No	No	No	No
Morbilliform rash	No	No	Yes	No	No	Yes
Eosinophilia	No	No	Yes	No	No	No
Elevated transaminases	No	No	Yes	No	No	No

DRESS, Drug reaction with eosinophilia and systemic symptoms; E, ethambutol; H, isoniazid; IVI, intravenous injection; R, rifampicin; SJS, Stevens Johnson syndrome; TEN, toxic epidermal necrolysis; Z, pyrazinamide.

The speed (<60 minutes) and nature (frequent systemic symptoms) of these RRs illustrate the resting state of immune activation after SCAR and the importance of preformed or rapidly available inflammatory molecules, for example, cytokines and arachidonic acid derivatives resulting in RR. In addition, the rapid onset of symptom relief from intravenous methylprednisolone highlights the importance of nongenomic anti-inflammatory corticosteroid effects, which may include immune cell membrane stabilization, inhibition of arachidonic acid release, or T-cell receptor signaling.^8,9 Further research is required to understand these effects especially in the context of HIV immune dysregulation. These findings suggest that if given early enough, at even lower doses than previously suggested, a single intravenous dose may be sufficient to prevent disease worsening after drug provocation in SCAR. If proven, these findings may improve safety of rechallenging in situations where there are no alternative drugs; open a window for an in vivo model to study earliest immunological, cytokine, and cellular responses in SCAR for a better understanding of pathogenesis and more targeted therapies; the role of transient drug metabolites that are thought to play a role in these reactions; and the mechanism by which corticosteroids ameliorate SCAR. Our findings need to be proven in larger prospective studies to determine if they apply to all drugs, all SCAR, the safest and most efficacious corticosteroid and optimal dosing, side effects, and optimizing patient selection.

Clinical Implications.

• Single early infusion of corticosteroids rapidly and sustainably reverses cutaneous and systemic features of Stevens Johnson syndrome/toxic epidermal necrolysis and drug reaction with eosinophilia and systemic symptoms developing on re-exposure to the offending drug. These findings support the role of corticosteroids in these reactions and potentially improve the safety of drug rechallenge protocols.