A Challenging Case of Trimethoprim-Sulfamethoxazole Associated Drug Reaction With Eosinophilia and Systemic Symptoms With Prominent Hepatic Involvement and an Atypical Rash

Abstract

Drug reaction with eosinophilia and systemic symptoms (DRESS) is a rare, potentially life-threatening hypersensitivity reaction that manifests with a spectrum of cutaneous and systemic involvement. We present an unusual case of DRESS in a 26-year-old man with markedly elevated liver enzymes and hyperbilirubinemia but relatively limited skin activity. The patient had a recent diagnosis of human immunodeficiency virus infection and was treated with trimethoprim-sulfamethoxazole for Pneumocystis jirovecii pneumonia and bilateral pneumothoraces. This case emphasizes the importance of maintaining a high clinical suspicion for DRESS in patients presenting with elevated liver enzymes and skin rash.

INTRODUCTION

Drug reaction with eosinophilia and systemic symptoms (DRESS) is a delayed hypersensitivity reaction that presents 2 to 8 weeks after drug initiation.¹ Clinical manifestations include rash, fever, visceral involvement, and hematologic abnormalities such as eosinophilia, lymphocytosis, lymphocytopenia, atypical lymphocytes, and thrombocytopenia.^2,3 Elevated liver enzymes are the most common sign of visceral involvement.⁴ Hepatic injury varies from asymptomatic hepatitis to hepatomegaly, jaundice, and rarely fulminant liver failure.^5,6

Causative drugs include allopurinol, anticonvulsants, sulfonamides, and other antibiotics.⁴ DRESS is rare (1 in 1,000–10,000 medication exposures) and likely under-recognized.⁷ The mortality is significant, estimated from 3.8% to 10%. Long-term complications include development of autoimmune diseases such as diabetes or thyroiditis.⁸

We present a challenging case of trimethoprim-sulfamethoxazole (TMP-SMX)-induced DRESS that presented with abnormal liver enzymes, hyperbilirubinemia, atypical rash, and initially mild eosinophilia in the setting of human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS). We focus on the diagnostic dilemma of distinguishing between drug-induced liver injury (DILI) and DRESS for the hepatologist.

CASE REPORT

A 26-year-old African American man presented with a pruritic rash and flaking skin that progressed over 5 days. Six weeks prior, he presented to the emergency department with bilateral spontaneous pneumothorax. He was found to have Pneumocystis jirovecii pneumonia (PCP) and diagnosed with HIV/AIDS (CD4 count <5, HIV 166,787 copies/mL). PCP was treated with high-dose TMP-SMX. He initiated HIV antiviral therapy and continued TMP-SMX for PCP prophylaxis. No other medications were started during this time period.

On examination, he was afebrile, and vital signs were stable. Skin examination showed a generalized fine scale with desquamation involving arms, dorsal hands (Figure 1), chest, lower back, face, legs, and buttocks, with minimal erythema. There was no mucosal involvement. He had bilateral chest tubes. Abdominal examination showed bilateral inguinal lymphadenopathy but no hepatosplenomegaly.

Figure 1.

Fine scale with desquamation involving dorsal hand.

Laboratory testing revealed elevated liver enzymes: aspartate aminotransferase 411 U/L, alanine aminotransferase 391 U/L, alkaline phosphatase 1027 U/L, and total bilirubin 2.36 mg/dL. There was a mild eosinophilia (7.7%; absolute count 300 cells/µL) that peaked 4 days later (37.9%, 2,100 cells/µL). Serum viral polymerase chain reaction revealed cytomegalovirus (773 IU/mL), low-level Epstein-Barr virus (<50 IU/mL), and low viral load HIV (53 copies/mL). Serologies were negative for hepatitis A/B/C, herpes simplex virus 1/2, syphilis, and Toxoplasma. Fungal antigen detection (urine/serum) was negative for Histoplasma, Coccidioides, Blastomyces, and Cryptococcus. Blood cultures were negative. Abdominal ultrasound showed the liver was upper limit of normal size with no evidence of biliary obstruction or structural liver disease.

TMP-SMX was stopped immediately. Skin and liver biopsies were performed. The skin biopsy revealed a vacuolar interface dermatitis—a pattern often seen in drug reactions—with scattered dyskeratotic cells and a mild inflammatory infiltrate (Figure 2). No vasculitis was seen. Stains for fungal organisms and syphilis were negative. The liver biopsy showed cholestatic hepatitis with prominent eosinophils and bile duct damage (Figure 2). No fungal or acid-fast organisms were found, and cytomegalovirus and Epstein-Barr virus stains were negative.

Figure 2.

Histopathologic findings (H&E stain). (A) Skin punch biopsy shows vacuolar interface dermatitis with mild superficial perivascular infiltrate (×40). (B) Higher magnification reveals basal cell vacuolization at the dermoepidermal junction, scattered dyskeratotic cells (arrowheads), pigment incontinence, and eosinophils (arrows) within the infiltrate (×200). (C) Liver core biopsy shows portal tracts expanded by inflammatory infiltrate with mild interface activity and lobular disarray with sinusoidal inflammation (×100). (D) Mixed portal inflammation includes lymphocytes, histiocytes, and eosinophils; interlobular bile duct (arrowhead) shows epithelial damage and inflammatory infiltration (×200).

The patient was given supportive treatment for suspected DRESS. Dermatology prescribed triamcinolone topical steroids for rash. Liver enzymes decreased after drug cessation, and follow-up labs 6 weeks later showed aspartate aminotransferase 30 U/L, alanine aminotransferase 26 U/L, alkaline phosphatase 275 U/L, and total bilirubin 1.3 mg/dL, consistent with resolving liver injury. PCP prophylaxis was changed to atovaquone. He was discharged in stable condition.

DISCUSSION

The working diagnosis was DRESS due to TMP-SMX. Although the incidence of DRESS is rare, antibiotics are a frequent culprit (up to 74% in 1 American electronic health record review), and 3% to 9% of antibiotic-induced cases are attributed to sulfonamides.^9,10 Although the overall incidence of DRESS is estimated at 1 in 1,000 to 1 in 10,000 exposures to high-risk medications, published studies do not provide a precise incidence rate specifically for TMP-SMX. However, multiple case series consistently identify TMP-SMX as one of the most frequently implicated antibiotic triggers of DRESS. The elevated liver enzymes, rash, eosinophilia, and lymphadenopathy all supported the diagnosis. Our patient's skin findings were atypical due to the minimal erythema and scaly appearance resembling xerosis, as can sometimes occur in dark skinned individuals.¹¹ The predominantly cholestatic pattern of liver enzymes was also unusual, since hepatocellular injury is more common in DRESS.¹² Other hematologic abnormalities were difficult to interpret in the setting of newly diagnosed HIV/AIDS, given his preexisting lymphopenia and anemia. Using the international Registry of Severe Cutaneous Adverse Reactions diagnostic criteria, our patient's score of 5 indicates a probable case of DRESS (Table 1).^3,7 Similarly, Japanese SCAR diagnostic criteria pertain to drug-induced hypersensitivity syndrome, which likely represents the same disease spectrum despite differences in nomenclature.²

Table 1.

RegiSCAR scoring system used to diagnose DIHS/DRESS syndrome in the patient

Criteria	Patient		Score
	No/Yes	Score	−1	0	+1	+2
Fever ≥ 38.5°C	No	−1	No/U	Yes
Enlarged lymph nodesa	Yes	+1		No/U	Yes
Eosinophils (cells/µL)	Yes				700–1,499	≥1,500
Eosinophils, if leukocytes <4,000		+2			10%–19.9%	≥20%
Atypical lymphocytes	U	0		No/U	Yes
Extensive rash (>50% TBSA)	Yes	+1		No/U	Yes
Rash suggesting DRESSb	U	0	No	U	Yes
Biopsy suggesting DRESS	Yes	0	No	Yes/U
Hepatic impairment	Yes	+1		No/U	Yes
Other organ impairmentc	No	0		No/U	Yes
Resolution in ≥15 d	Yes	0	No/U	Yes
Evaluation of other potential causesd	Yes	+1			Yes
Total score		5

Total score <2 (negative case); 2 to 3 (possible case); 4 to 5 (probable case); >5 (definitive case).

DIHS, drug-induced hypersensitivity syndrome; DRESS, drug reaction with eosinophilia and systemic symptoms; RegiSCAR, Registry of Severe Cutaneous Adverse Reactions; TBSA, total body surface area; U, unknown.

^aEnlarged lymph nodes: >1 cm and ≥2 different areas.

^bRash is suggestive for DRESS if ≥ 2 criteria: scaling/desquamation, facial edema, purpura, infiltrating.

^cScore 1 for each organ, maximum 2 points for internal organ impairment (renal, lung, muscle, heart, pancreas).

^dIf at least 3 investigations performed to exclude alternative diagnosis and found negative: antinuclear antibody, blood culture, serology for hepatitis A/B/C, Chlamydia/Mycoplasma pneumoniae, other serology/polymerase chain reaction/cultures.

The 5-week time course of symptom onset after initiating TMP-SMX is typical of DRESS.² His rapid improvement after stopping the drug further supported the diagnosis and highlights the variable clinical course of DRESS. Biopsies helped to confirm the diagnosis and rule out other causes such as vasculitis and DILI. Exclusion of alternative etiologies is particularly important in the setting of HIV/AIDS, and the patient underwent extensive infectious disease workup. Autoimmune serologies were not performed, which is a limitation. The skin findings of vacuolar interface dermatitis, although not definitive, were suggestive of DRESS. Although the liver biopsy findings of cholestatic hepatitis with eosinophilia support a diagnosis of DRESS, there is notable histologic overlap with DILI. Antiretroviral therapy was initiated during the same time period, and rarely antiretroviral therapy medications (eg, nevirapine, raltegravir) have been associated with DRESS.¹³ However, the prompt resolution of symptoms after cessation of TMP-SMX favored the sulfonamide as the offending agent.

The differential diagnosis includes DILI, which has been reported in 0.091% of hospitalized patients prescribed TMP-SMX.¹⁴ Distinguishing DRESS from DILI is clinically relevant because these conditions differ in prognosis and management. Liver manifestations of both DILI and DRESS range widely from mild elevations in liver enzymes to fulminant hepatic failure. Diagnosis is particularly challenging in atypical presentations of DRESS with subtle skin findings, which could be misinterpreted as pruritus in the setting of hyperbilirubinemia.¹¹ A review of the U.S. DILI Network (DILIN) showed that 6% of participants had probable DRESS using modified Registry of Severe Cutaneous Adverse Reactions criteria.¹⁵ Compared to DILI alone, patients with DRESS were younger (mean age 42 vs 51 years), more likely Black race (26% vs 12%), had higher frequency of rash (100% vs 13%), and more severe liver injury (45% vs 21.5%). DRESS was associated with higher mortality than DILI alone, so early recognition with drug cessation is critical.¹¹ Systemic corticosteroids may be given in severe cases of DRESS to prevent progression to organ dysfunction.

Treatment of DRESS relies on early recognition and immediate withdrawal of the causative drug. In mild cases without visceral involvement or severe rash, management is supportive and aimed at relieving symptoms with topical steroids, antihistamines, and careful monitoring.² Systemic corticosteroids are generally recommended in severe cases of DRESS, although evidence from large clinical trials and clinical guidelines are lacking.^2,5 For steroid refractory and/or relapsed severe DRESS, second-line or steroid sparing agents include cyclosporine, mycophenolate, rituximab, cyclophosphamide, intravenous immunoglobulin, and plasmapheresis. In this case, steroids were not necessary because the liver enzymes decreased after drug cessation and the skin findings were mild.

Although TMP-SMX is effective in treating PCP and other bacterial infections, clinicians and hepatologists should be aware of the potential risk of adverse events, including DRESS. A meta-analysis showed that TMP-SMX is associated with increased adverse events compared to other antibiotics.¹⁶ Current treatment guidelines for PCP recommend using high-dose TMP 15–20 mg/kg/d and SMX 75–100 mg/kg/d for 2–3 weeks. However, some studies have suggested that intermediate or low-dose TMP-SMX therapy may be equally effective for PCP with fewer adverse drug events.¹⁷ This case highlights the importance of early recognition of DRESS with subtle skin findings, and prompt discontinuation of the drug to prevent morbidity and mortality associated with this syndrome.

DISCLOSURES

Author contributions: A Baddam and Z Ahmed jointly contributed to the conceptualization and writing of the initial manuscript. DR Fels Elliott and PP Chow composed figures, reviewed, and edited the manuscript. Z. Ahmed is the article guarantor.

Financial disclosure: None to report.

Informed consent was obtained for this case report.

ABBREVIATIONS:

AIDS

Acquired immunodeficiency syndrome

DIHS

drug-induced hypersensitivity syndrome

DRESS

drug reaction with eosinophilia and systemic symptoms

HIV

human immunodeficiency virus

PCP

Pneumocystis jirovecii pneumonia

RegiSCAR

Registry of Severe Cutaneous Adverse Reactions

TMP-SMX

trimethoprim-sulfamethoxazole