To the Editor: A 78-year-old man was admitted to our Department feeling pain in the upper right quadrant for the last 3 days and fever of 6-h duration. The patient suffered from autoimmune cholangitis, and thus, he had been treated with ursodeoxycholic acid 1000 mg/day for the last 3 years. Apart from the above, he had been suffering from chronic lymphocytic leukemia (CLL) for the last 18 months. The patient had an unremarkable course of his liver and hematological diseases. For the latter, he had not received any chemotherapy until his present admission. During physical examination, the patient had a high fever reaching to 38.5°C, with hepatosplenomegaly and lymphadenopathy. The lymph nodes were firm, rubbery, and nontender. Admission laboratory findings included the following: Ht: 33%, Hb: 10.8 gr/dL, erythrocyte sedimentation rate (ESR): 110 mm per first hour, C-reactive protein: 40 mg/dL, white blood cell: 12,300 per cubic millimeter (lymphocytes: 75%), aspartate aminotransferase (AST): 45 IU/L [upper normal level (UNL) <40], alanine aminotransferase (ALT): 52 IU/L (UNL < 40), alkaline phosphatase (ALP) (UNL < 290), γ-glestamyl transpeptidase (γGT): 230 IU/L (UNL < 48), bilirubin: 2.3 mg/dL, and γ-globulin: 6.5 g/dL (without monoclonal protein). Abdominal computer tomography revealed enlargement of the liver and spleen and abdominal lymphadenopathy. There was evidence of CLL progression in our patient's clinical and laboratory blood tests, but his liver tests had deteriorated, compared with those 2 months ago. Thus, with the possible diagnosis of biliary sepsis, piperacillin/tazobactam (Tazocin) 4.5 gr qid was administered. Three days after the initiation of the antibiotic regimen, the patient showed a mild improvement in his clinical condition, yet he developed painful nonpruritic erythematous skin plaques (1-2 cm in diameter) on both legs (Fig. 1). Skin biopsy showed perivascular and interstitital infiltrate with predominately neutrophils in the upper dermis, without any vasculitic changes or fibrinoid necrosis (Fig. 2). These lesions were compatible with Sweet syndrome. When piperacillin/tazobactam was substituted by imipenem 1gr tid, the patient's fever reduced 3 days later, and the skin lesions were gradually disappeared, without the need of any corticosteroid therapy. Eventually, the patient was discharged in a good condition 10 days after his admission. He was followed up in the clinic, without any recurrence of the skin lesions.FIGURE 1: Multiple painful nonpruritic erythematous skin plaques 1-2 cm in diameter on both legs of the patient after piperacillin/tazobactam administration.FIGURE 2: Patient's skin biopsy showing perivascular and interstitital infiltrate with predominate neutrophils in the upper dermis, compatible with the diagnosis of Sweet syndrome (eosin and hematoxylin stain, ×40).Since the original report in 1964, multiple cases of Sweet syndrome have been reported.1 Although the exact pathogenesis of Sweet syndrome has not yet been established, dysregulation of cytokines may play an important role.1 Different diseases, including malignancies (eg, CLL), infections, and autoimmune disorders, and, less frequently, drugs have been reported to be associated with the syndrome.2 Although several drugs, such as granulocyte colony-stimulating factor, carbamazepine, and antibiotics (eg, ofloxacin, tetracycline, and trimethoprim-sulfamethoxazole) have been reported to cause Sweet syndrome,3 in most of the cases, it is difficult to confirm whether Sweet syndrome is drug induced or not, due to concurrent underlying disease. The diagnostic criteria that have been proposed for the determination of drug-induced Sweet syndrome are the following: (1) painful erytematous plaques, (2) characteristic lesions in skin biopsy, (3) pyrexia, (4) relation with the drug administration or recurrence after drug rechallenge, and (5) resolution after drug withdrawal or treatment with corticosteroids.3 Although our patient was suffering from CLL, Sweet syndrome usually coincides with the diagnosis or when there is a deterioration of malignancy.1,2,4 However, our patient had no evidence of CLL progression. In addition, he had not received any medical treatment, apart from ursodeoxycholic acid and piperacillin/tazobactam, but the former had been administered for the last 3 years. Therapeutic intervention based on corticosteroids is usually necessary, but recession after treatment of the underlying disease (eg, malignancy) or withdrawal of the causative drug is reported.1 The skin lesions of our patient were improved after discontinuity of piperacillin/tazobactam and without any corticosteroid therapy. Thus, our patient fulfills all the diagnostic criteria for piperacillin/tazobactam-induced Sweet syndrome. In addition, the use of the Naranjo probability scale indicated that Sweet syndrome was possibly due to piperacillin/tazobactam administration (total score 3).3,5 To our knowledge, this is the first report of Sweet syndrome associated with piperacillin/tazobactam treatment. We believe that a high proportion of awareness is needed for prompt diagnosis of Sweet syndrome, particularly when drugs are implicated in the etiology of this rare entity. In this way, the responsible drug can be withdrawn, and an appropriate treatment can be administered whenever necessary. Evangelos Cholongitas, MD* Chrysoula Pipili, MD* Maria Dasenaki, MD* Department of Internal Medicine General Hospital of Sitia Sitia, Greece Loukas Kaklamanis, MD† †Department of Histopathology Onasio Hospital Athens, Greece