Background

Drug rash with eosinophilia and systemic symptoms (DRESS) is a rare but serious idiosyncratic drug reaction with an incidence of 0.1 to 1 per 1000 individuals taking high-risk medications and an associated mortality rate of 10% to 20%.^1,2 Medications most commonly implicated with DRESS include allopurinol (18%); sulfonamides (12%); antibiotics (11%); and aromatic antiepileptics (35%), such as carbamazepine (20%), phenytoin (7%), lamotrigine (7%), and phenobarbital (1.7%).¹ The diagnosis of DRESS is difficult due to the overlapping symptoms between DRESS and other serious conditions, such as serotonin syndrome and neuroleptic malignant syndrome (NMS).^1,3,4 The clinical presentation of DRESS is characterized by fever, rash, lymphadenopathy, internal organ involvement (eg, liver), and hematologic abnormalities (eg, eosinophilia).^1,2 A biopsy can help confirm the diagnosis; however, the histopathologic features of the rash often show variable inflammatory patterns.⁵ Drug rash with eosinophilia and systemic symptoms typically occurs 2 to 8 weeks following initiation of the culprit medication.¹ The pathogenesis is not well understood but thought to be related to one of the following mechanisms: a delayed immunologic reaction, a transient state of immune suppression, and/ or reactivation of latent herpes virus infections. Discontinuation of the causative agent is the primary treatment.² Following an episode of DRESS, choosing alternative pharmacological therapies must be done cautiously as there are limited data available regarding cross-reactivity between various medications. In this case report, we discuss a patient who was initiated on carbamazepine after the development of DRESS following recent exposure to divalproex sodium, lithium, and olanzapine.

Case Report

Hospital Course

On admission, the patient's home dose of divalproex was continued (Table 1). From hospital day (HD) 1 to HD 10, he received as needed oral doses of olanzapine every other day and received one dose of intramuscular olanzapine for acute agitation. On HD 10, olanzapine 5 mg oral daily was scheduled and increased to 10 mg by HD 12. On HD 15, CPK was elevated, and his olanzapine was decreased to 2.5 mg daily. Creatinine phosphatase kinase continued to rise (3642 IU/L on HD 22), and olanzapine was discontinued as a precaution based on history of possible NMS with antipsychotics despite absence of other symptoms of NMS. This elevation in CPK was attributed to an episode of agitation requiring physical restraints and a dose of intramuscular olanzapine. He continued to appear manic and was subsequently started on lithium and lorazepam. He remained clinically stable, and his CPK down-trended to 718 IU/L. On HD 28, he developed a fever, lymphadenopathy, diarrhea, diffuse abdominal pain, and an episode of dark emesis, and his CPK increased to 1435 IU/mL. Additionally, liver function enzymes, white blood count, and serum creatinine were significantly elevated from baseline. He became increasingly tired and confused, and all psychotropic medications aside from lorazepam were discontinued due to concern for NMS. He was transferred to general medicine where he developed a morbilliform rash on his chest, trunk, arms, and legs. Because serotonin syndrome was not able to be ruled out, he was started on cyproheptadine and continued until the punch biopsy results returned consistent with drug eruption (perivascular lymphocytic inflammation with mild dermal hemorrhage). Dermatology, psychiatry, and hematology consult services agreed on a suspected diagnosis of DRESS, and he was initiated on a prednisone taper. Given the anticipated prolonged corticosteroid taper, sulfamethoxazole-trimethoprim was added for prophylaxis against Pneumocystis pneumonia based on a statement by the American Thoracic Society.⁶ His workup was negative for infection, hepatitis virus, and antinuclear antibody. As he improved medically, he began to exhibit grandiosity and mood lability concerning for reemergence of manic symptoms. Following initiation of carbamazepine on HD 34, the medication was slowly titrated, and his mood returned to baseline by discharge on HD 59. He was briefly rehospitalized 2 days after discharge due to residual symptoms in the context of psychosocial stressors, but no medication changes were made. During an outpatient follow-up, he continued to appear euthymic and all DRESS-related symptoms were resolved.

TABLE 1 Patient summary

View Fullscreen

Discussion

The differential diagnosis in this case included DRESS, serotonin syndrome, sepsis, and NMS. Based on the punch biopsy results, eosinophilia, and improvement following corticosteroid administration, it was determined that DRESS was the most definitive explanation for this presentation. Additionally, he scored a 6 of 9 on the DRESS scoring system developed by Karduan et al,¹ which indicates a definite case of DRESS. However, the patient also met the diagnostic criteria for serotonin syndrome and NMS (Table 2). Divalproex, lithium, and olanzapine were subsequently discontinued. The patient was treated with corticosteroids based on treatment success in prior reports.^1,2

TABLE 2 Differential diagnosis^a

View Fullscreen

Determining the causative agent was challenging as all 3 medications have been associated with DRESS. Divalproex has been implicated as a cause of DRESS and has the most published reports of DRESS among the medications that the patient was taking.² However, he previously tolerated divalproex monotherapy and historically only had adverse reactions when it was combined with antipsychotics or lithium. Lithium has been associated with DRESS in a limited number of case reports.^7,8 However, the patient was only taking lithium for 6 days prior to the onset of DRESS, making it an unlikely culprit. Symptoms began roughly 21 days after the first dose of olanzapine, which is consistent with the onset of DRESS reported in the literature.^1,2 Olanzapine is not an agent that has traditionally been associated with DRESS. However, in 2016, the US Food and Drug Administration⁹ published a warning regarding the risk of DRESS associated with olanzapine following 23 cases of olanzapine-induced DRESS.

In addition to discontinuing the agent(s) that may have induced DRESS, it is recommended to avoid agents with similar structures because cross-reactivity between aromatic anticonvulsants may be as high as 80%.^2,10,11 Although less common, there are case reports of DRESS caused by nonaromatic psychotropic medications.² As valproate and lithium are not aromatic and structurally dissimilar to the aromatic anticonvulsant carbamazepine, cross-reactivity was unlikely.¹² There are some structural similarities between carbamazepine and olanzapine. In fact, 1 case report exists describing the cross-reactivity between these agents in a patient who developed DRESS.¹³ Case reports^10,14-16 also describe secondary neosensitization to certain medications following DRESS induced by structurally different agents. Neosensitization is thought to be caused by a transient state of immunosuppression induced during the first episode of DRESS that may trigger a nonspecific immune system response leading to the inability to tolerate other drugs present at the time. In several reports of neosensitization, the agent that was related to the second hypersensitivity episode was initiated during the first episode of DRESS,^15,16 but this was not consistent in every case.^10,14,16 Although a positive history of DRESS is a major risk factor for future episodes, our patient was left with limited treatment options for his bipolar mania.²

The American Psychiatric Association (APA)¹⁷ recommends lithium, valproate, or a second-generation antipsychotic as the first-line treatment of acute mania. Because the patient was receiving all of these medications prior to the development of DRESS and none could be excluded as the causative agent, it was difficult to determine the most appropriate agent for the treatment of his bipolar disorder. Alternative treatment options recommended by the American Psychiatric Association include carbamazepine and ECT. Based on the patient's preference, ECT was not considered at this time. A MEDLINE search revealed no published case reports of the successful use of carbamazepine in a patient with a history of DRESS from a nonaromatic medication. Although carbamazepine carries a relatively high risk of DRESS, it was determined to be the safest option. Additionally, sulfamethoxazole-trimethoprim (which has a relatively high rate of DRESS) was successfully used.

Conclusion

When patients have severe drug reactions, it is important to minimize risk of recurrence. Previous reports have demonstrated the cross-reactivity of certain psychotropic medications in causing DRESS. Ideally, alternative medications following an episode of DRESS should be structurally dissimilar and associated with a low risk of reaction. In this case, we were successfully able to use both carbamazepine and sulfamethoxazole-trimethoprim even though both have a relatively high rate of DRESS.