DRESS syndrome due to first-line antitubercular therapy - A diagnostic imbroglio!

Drug reaction with eosinophilia and systemic symptoms (DRESS) syndrome after the use of first-line antitubercular drugs (ATDs) is rare and literature regarding DRESS syndrome due to ATDs is scarce in children. We report a young boy with tuberculosis who developed DRESS syndrome after exposure to isoniazid. A 9-year-old boy, diagnosed clinically as pulmonary tuberculosis, presented with fever, fast breathing, maculopapular rash, and one episode of gross hematuria. He had been on 4-drug ATD therapy (isoniazid, rifampicin, ethambutol, and pyrazinamide) for the past 4 weeks. In view of multiorgan involvement and absence of a microbiological diagnosis of tuberculosis, vasculitis was considered and he was treated with steroids. As the child recovered, both corticosteroids and ATD therapy were stopped. At 6 months of follow-up, he was presented with pneumonia. Microbiological diagnosis of tuberculosis was made and 4-drug ATD therapy was reinitiated. After 15 days, he again developed a high-grade fever and rash. On evaluation, isoniazid-induced DRESS syndrome was diagnosed. Subsequently, he received a modified regimen of ethambutol, pyrazinamide, levofloxacin, and linezolid. DRESS syndrome did not recur on these ATDs and the child became asymptomatic. Linezolid was stopped after 3 months of therapy and ethambutol, pyrazinamide, and levofloxacin are being continued. Currently, he has completed 15 months of modified ATD therapy. As a high index of suspicion is required for early diagnosis and management that are crucial to reducing morbidity and mortality, DRESS syndrome should be among the differentials in children with unexplained febrile illnesses. Copyright:

Introduction

Drug reaction with eosinophilia and systemic symptoms (DRESS) syndrome may be a life-threatening condition that is precipitated by exposure to various therapeutic agents.[1] DRESS syndrome after the use of first-line antitubercular drugs (ATDs) is rare and relevant literature in children is scarce.[2] We report a young boy with tuberculosis (TB) who developed DRESS syndrome after exposure to isoniazid. The clinical picture mimicked vasculitis causing significant clinical dilemmas and delay in diagnosis. We highlight that a high index of suspicion is required for early diagnosis of DRESS syndrome that is crucial to reducing morbidity and mortality.

Case Presentation

A 9-year-old boy presented to a local primary healthcare facility with a 3-month history of fever and cough. He was diagnosed to have TB based on Mantoux positivity (14 mm) and was initiated on 4-drug ATD therapy (isoniazid, rifampicin, ethambutol, and pyrazinamide). However, 4 weeks following ATDs, he developed fever, fast breathing, maculopapular rash, and one episode of gross hematuria. He was admitted to a local hospital for 3 weeks where he received mechanical ventilation for the initial 9 days. The treating physicians suspected vasculitis owing to multiorgan involvement. He was treated with 3 doses of intravenous methylprednisolone pulse (IVMP) therapy. Following IVMP, oral prednisolone at 2 mg/kg was initiated and ATD therapy was continued. He was subsequently referred to our institute.

On reviewing the investigations done earlier, the complete blood counts showed leucocytosis (up to 30.0 × 109/L [normal <15.0 × 109/L]) with eosinophilia (absolute count: 1.8 × 109/L [normal <0.5 × 109/L]). Urine examination showed hematuria and albuminuria. The chest radiograph showed bilateral diffuse infiltrates, which had improved significantly [Figure 1] after glucocorticoid therapy. Antinuclear (ANA) and antineutrophil cytoplasmic antibodies (ANCA) were negative. His serum IgE levels were elevated (4,064 kIU/L [normal <50]).

Figure 1

Serial chest radiographs done at the local hospital (before presenting to us) show bilateral diffuse infiltrates (right > left) that improve significantly after intravenous methylprednisolone pulse therapy and oral prednisolone. (a) Day 1 of hospitalization, (b) Day 2 of hospitalization, (c) Day 9 of hospitalization (before extubation), (d) Day 11 of hospitalization (postextubation)

On examination, he had hypertension (126/80 mmHg), cushingoid facies, maculopapular rash over the face, trunk, and extremities, and hepatomegaly. Investigations [Supplementary Table 1] showed thrombocytosis (444 × 109/L [normal <400]), microscopic hematuria, and mild elevations of urinary protein-creatinine ratio (0.38 [normal <0.2]) and alanine aminotransferase (97 U/L [normal <40]). The autoantibody profile (ANA, ANCA, anti-glomerular basement membrane antibody, and anti-dsDNA) and a detailed parasitology/infectious/immune workup were negative. Chest radiograph was unremarkable and gastric aspirates for Ziehl–Neelsen staining and mycobacterial cultures were negative. Corticosteroids were tapered and stopped over the next 6 weeks. ATD therapy was also stopped and he was followed up.

Supplementary Table 1

Laboratory investigations at first presentation to us

Parameter	Result
Hemoglobin (g/L)	11.8 (11.5-15.5)
Total leukocyte count (×109/L)	8.62 (5.0-14.5)
Differential leukocyte count	N40L46M14E0B0
Platelets (×109/L)	444 (150-400)
Serum urea (mg/dL)	23 (10-50)
Serum creatinine (mg/dL)	0.13 (0.5-1.2)
Total serum bilirubin (mg/dL)	0.6 (<1.0)
Alanine aminotransferase (U/L)	97 (5-45)
Aspartate aminotransferase (U/L)	37 (15-50)
Alkaline phosphatase (U/L)	118 (145-420)
Total serum protein (mg/dL)	6.6 (6.4-8.3)
C-reactive protein (mg/L)	0.058 (0-3.5)
Total serum albumin (mg/dL)	3.8 (3.5-5.6)
Urine routine microscopy	120-125 RBCs/HPF
Repeat urine microscopy	Normal
Urinary spot protein creatinine ratio	0.378 (<0.2)
Prothrombin time (seconds)	10.2 (10-14.6)
International normalized ratio	0.84
Activated partial thromboplastin time (seconds)	26.5 (26-39)
Plasma fibrinogen (g/L)	1.75 (1.89-4.75)
Serum ferritin (ng/mL)	38.74 (30-40)
Serum N-terminal pro-B type natriuretic peptide (pg/mL)	183.4 (0-125)
Hepatitis B surface antigen	Negative
Antihepatitis C virus IgM antibody	Negative
Human immunodeficiency virus serology	Negative
Antistreptolysin O (IU/mL)	44.2 (<200)
Gastric lavage for acid-fast bacilli	Negative
Gastric lavage for Mycobacterial tuberculosis CB-NAAT	Negative
Filaria, Toxocara, and Toxoplasma serology	All negative
Antinuclear antibody (IIF)	Negative
Antineutrophil cytoplasmic antibody (ELISA for MPO/PR3)	Both negative
Antiglomerular basement membrane antibody	Negative
Antidouble stranded DNA antibody (IU/mL)	18 (<60)
Antitissue tranglutaminase IgA antibody (EliA U/mL)	<0.1 (<7)
Gastric lavage for hemosiderin laden macrophages	Negative
Complement C3 (mg/dL)	125.7 (78.9-178.9)
Complement C4 (mg/dL)	16.2 (14.5-61.6)
CD3+ T cells [% of lymphocytes]	75.7 (60-76)
CD19+ B cells [% of lymphocytes]	15.7 (13-27)
CD16+56+ NK cells [% of lymphocytes]	2.72 (4-17)
Absolute no. of CD3+ T cells (× 109/L)	2.474 (1.200-2.600)
Absolute no. of CD19+ B cells (× 109/L)	0.514 (0.270-0.860)
Absolute no. of CD16+CD56+ NK cells (×109/L)	0.089 (0.100-0.480)
IgG (g/L)	14.04 (7.72-17.71)
IgA (g/L)	1.10 (0.73-2.09)
IgE (kIU/L)	572 (0-78)

Abnormal values are highlighted in bold. CB-NAAT, cartridge-based nucleic acid amplification test; ELISA, enzyme-linked immunosorbent assay; HPF, high-power field; IIF, indirect immunofluorescence; MPO, myeloperoxidase; RBCs, red blood cells; PR3, proteinase 3

After 6 months, he again presented to us with fever, cough, hemoptysis, and acute onset stridor with progressive respiratory distress. On examination, he was pale, tachypneic, with decreased oxygen saturation on pulse oximetry (SpO2: 82%). He had supraglottic and intercostal chest retractions, and auscultation revealed bronchial breathing on the left side of the chest. The child required mechanical ventilation for the management of hypoxemia. No hepatomegaly or rash was noted. Investigations [Supplementary Table ] showed anemia, thrombocytosis, and elevated erythrocyte sedimentation rate. His liver transaminases and urine microscopy were normal. Chest imaging showed large areas of consolidation in the left upper, lingular, and lower lobes [Figure 2]. The laryngoscopic examination was suggestive of tubercular laryngitis. Repeat gastric lavage showed positivity for acid-fast bacilli (on multiple occasions) and the cartridge-based nucleic acid amplification test was positive for Mycobacterium tuberculosis, which was sensitive to isoniazid and rifampicin. Lung aspiration showed necrotic epithelioid granulomas with positivity for acid-fast bacilli. A diagnosis of TB was made and he was commenced on 4-drug ATD therapy (isoniazid, rifampicin, ethambutol, and pyrazinamide). After 15 days of ATD therapy, he developed a maculopapular, pruritic rash distributed over the face, trunk, and extremities [Figure 3]. It was also accompanied by a continuous high-grade fever without any other specific focus.

Supplementary Table 2

Laboratory investigations at 6 months of follow-up

Parameter	Result
Hemoglobin (g/L)	9.9 (11.5-15.5)
Total leukocyte count (×109/L)	10.7 (5.0-14.5)
Differential leukocyte count	N92L5M3E0B0
Platelets (×109/L)	447 (150-400)
Serum urea (mg/dL)	15 (10-50)
Serum creatinine (mg/dL)	0.29 (0.5-1.2)
Total serum bilirubin (mg/dL)	0.54 (<1.0)
Alanine aminotransferase (U/L)	8 (5-45)
Aspartate aminotransferase (U/L)	20 (15-50)
Total serum protein (mg/dL)	7.3 (6.4-8.3)
Total serum albumin (mg/dL)	3.3 (3.5-5.6)
Urine routine microscopy	Normal
Repeat urine microscopy	Normal
Urinary spot protein creatinine ratio	0.1 (<0.2)
Prothrombin time (seconds)	12.4 (10-14.6)
International normalized ratio	1.02
Activated partial thromboplastin time (seconds)	38.2 (26-39)
Plasma fibrinogen (g/L)	4.51 (1.89-4.75)
Gastric lavage for acid-fast bacilli	Positive: 10-15/HPF
Gastric lavage for Mycobacterial tuberculosis CB-NAAT	Positive
Gastric lavage for mycobacterial culture	M. tuberculosis
Urine for acid-fast bacilli	Negative
Urine for Mycobacterial tuberculosis CB-NAAT	Negative
Urine for mycobacterial culture	Negative
Transthoracic echocardiography	Normal

Abnormal values are highlighted in bold. CB - NAAT, cartridge-based nucleic acid amplification test; HPF, high-power field

Figure 2

Chest radiograph (a) and computed tomography (b-d) done at the time of the second admission (at our center) show large areas of consolidation involving the left upper, lingular, and lower lobes

Figure 3

Appearance of erythematous maculopapular pruritic skin rash over face (a), trunk (b), and extremities (c, d) after initiation of isoniazid

Repeat laboratory evaluation [Supplementary Table 3] showed eosinophilia (12.0%, absolute count: 1.666 × 109/L), elevated serum alanine and aspartate aminotransferases of 240 and 279 U/L, respectively, which increased to 307 and 397, respectively, over the next 4 days. The presence of fever, rash, systemic manifestations and eosinophilia aroused the suspicion of DRESS syndrome. Skin biopsy was consistent with DRESS syndrome [Supplementary Figure 1] and evaluation for other potential causes was negative [Supplementary Table 3]. The ATDs were discontinued and tapering doses of oral prednisolone were administered (initially at 2 mg/kg/day) resulting in clinical improvement. To identify the culprit, ATDs were reintroduced one by one. He again developed a fever and maculopapular rash once isoniazid was initiated. Subsequently, a modified regimen of ethambutol, pyrazinamide, levofloxacin, and linezolid was initiated gradually. DRESS syndrome did not recur on these ATDs and he became completely asymptomatic. Linezolid was stopped after 3 months of therapy and ethambutol, pyrazinamide, and levofloxacin were continued. Although the index child continued to remain asymptomatic, follow-up imaging showed a large pneumatocele in the left upper and lingular lobes resulting in compression-collapse of adjacent lung parenchyma. The pneumatocele was drained successfully via a percutaneous pigtail catheter placed under radiological guidance [Supplementary Figure 2]. Currently, he has completed 15 months of modified ATD-therapy and is doing well.

Supplementary Table 3

Laboratory investigations at the time of appearance of rash after initiation of isoniazid

Parameter	Result
Hemoglobin (g/L)	9.1 (11.5-15.5)
Total leukocyte count (×109/L)	13.88 (5.0-14.5)
Differential leukocyte count	N72L16M0E12B0
Absolute eosinophil count (×109/L)	1.666 (0.05-0.50)
Platelets (×109/L)	402 (150-400)
Total serum bilirubin (mg/dL)	0.7 (<1.0)
Alanine aminotransferase (U/L)	240 (5-45)
Aspartate aminotransferase (U/L)	279 (15-50)
Total serum protein (mg/dL)	6.3
Total serum albumin (mg/dL)	2.5 (3.5-5.6)
Urine routine microscopy	Normal
Urinary spot protein creatinine ratio	0.15 (<0.2)
Prothrombin time (seconds)	13.9 (12-14)
International Normalized Ratio	0.93
Activated partial thromboplastin time (seconds)	36 (28-40)
Hepatitis B surface antigen	Negative
Antihepatitis C virus IgM antibody	Negative
Human immunodeficiency virus serology	Negative
Anti-Epstein-Barr viral capsid antigen (IgM)	Negative
Parvovirus serology	Negative
Antinuclear antibody (IIF)	Negative
Antidouble stranded DNA antibody	43 (<60)
Antineutrophil cytoplasmic antibody (ELISA for MPO/PR3)	Both negative
Complement C3 (mg/dL)	141.0 (78.9-178.9)
Complement C4 (mg/dL)	20.7 (14.5-61.6)
Ultrasonography of the abdomen	Enlarged liver, normal echotexture

Abnormal values are highlighted in bold. ELISA, enzyme-linked immunosorbent assay; IIF, indirect immunofluorescence; MPO, myeloperoxidase; PR3, proteinase 3

Discussion

DRESS syndrome due to ATDs is rare and only a few reports are available in the pediatric literature.[12] In a systematic review, Metterle et al. showed that the leading therapeutic agents causing DRESS syndrome were anticonvulsant drugs (e.g., carbamazepine, phenytoin), antimicrobials, and nonsteroidal anti-inflammatory agents (e.g., aspirin, diclofenac, ibuprofen).[3] The mean duration of drug exposure at symptom onset was 24 days.[3] Among ATDs, rifampicin is the most common agent implicated for DRESS syndrome. However, isoniazid, ethambutol, pyrazinamide, and fluoroquinolones have also been implicated.[4] It may be difficult to identify the culprit drug at the beginning as first-line ATDs are mostly prescribed in combination. The common clinical manifestation of DRESS includes skin rash (morbilliform), fever, lymphadenopathy, liver dysfunction, leukocytosis, and eosinophilia. However, the difficulty arises when systemic symptoms are present without any cutaneous manifestations.[5] The RegiScar or the Japanese criteria help establish the diagnosis.[67] The total RegiScar score in our patient was more than 5, fulfilling the criteria for a definitive case of DRESS syndrome [Supplementary Table 4]. As our child had systemic symptoms each time after isoniazid was reintroduced, we considered it to be the culprit drug. Rifampicin was avoided as being the most common ATD implicated to cause DRESS syndrome. There was no adverse reaction to ethambutol, pyrazinamide, levofloxacin, and linezolid in the index child.

Supplementary Table 4

RegiScar score of the index patient

Parameters	Designated Score	Score initially	Score at diagnosis
Fever ≥38.5 °C	No/Unkown=−1 Yes=0	0	0
Enlarged lymph nodes	No/Unkown=0 Yes=1	0	0
Eosinophilia Absolute eosinophil count (×109/L) or eosinophil percentage if total leukocyte count <4.0×109/L	No=0 0.70-1.499 or 10-19.9%=1 >1.500 or >20%=2	2	2
Atypical lymphocytes	No/Unkown=0 Yes=1	0	0
Skin rash >50% of body surface area	No/Unkown=0 Yes=1	0	1
Skin rash suggesting DRESS	No=−1 Unknown=0 Yes=1	1	1
Skin biopsy DRESS	No/Unkown=−1 Yes=0	−1	0
Organ involvement	No/Unkown=0 One organ=1 Two or more organs=2	2	1
Resolution ≥15 days	No/Unkown=−1 Yes=0	0	0
Evaluation of other potential causes	Yes=1	1	1
Total score	Minimum=−2 Maximum=9	5	6

Note: Final score <2, no case; final score 2-3, possible case; final score 4-5, probable case; final score >5, definite case

Although identification of the culprit drug is based on the temporal correlation of symptoms with drug exposure, patch and lymphocytic transformation tests may be valuable adjunctive tools.[8] Inability to perform these tests (due to availability constraints) may be an important limitation in our case; however, the strong temporal correlation (on multiple occasions) asserted the diagnosis of isoniazid-induced DRESS syndrome in our case.

Ethical approval and informed consent

As this manuscript pertains only to a case report, specific ethics approval is not mandated. Informed consent was taken from the parents of the child before inclusion into the manuscript.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.