We describe a patient similar to those with quinine-induced pancytopenia and coagulopathy described by Maguire and colleagues [1].

A 74-year-old white woman was hospitalized with acute-onset fever and chills on 13 May 1991. The physical examination was unrevealing. Laboratory results included a serum creatinine level of 140 µmol/L (1.6 mg/dL), a lactate dehydrogenase level of 298 U/L, pancytopenia (hemoglobin concentration, 103 g/L; platelet count, 22 x 10⁹/L; and leukocyte count, 1.7 x 10⁹/L), and coagulopathy (prothrombin time, 14.8 seconds; partial thromboplastin time, 43.3 seconds; and fibrin degradation products >0.04 g/L). She recovered without specific therapy, and two subsequent hospital admissions for a similar clinical syndrome were marked by extensive but fruitless evaluations. When a computed tomographic scan of the abdomen showed cholelithiasis, biliary sepsis was suspected to be the cause; the patient had a laparoscopic cholecystectomy on 18 June 1991.

A subsequent recurrence of symptoms led to a referral to our institution on 16 July 1991. Because she looked well, she was observed and discharged. However, she returned on 2 September 1991 with fever, mild renal insufficiency, and pancytopenia. After repeated questioning, she admitted to ingesting quinine sulfate from her niece's supply to relieve nocturnal leg cramps. The patient recovered spontaneously and with her consent, a drug challenge was started.

On 27 September 1991, one 260-mg tablet of quinine sulfate from the niece's supply was administered to the patient after prechallenge blood samples were drawn. Within 90 minutes, her temperature increased to 38.5 °C, and chills, nausea, vomiting, and hypotension developed. A completed blood count 2 hours after drug administration showed decreases from challenge values in the platelet count (from 214 x 10⁹/L to 53 x 10⁹/L) and leukocyte count (from 7.8 x 10⁹/L to 1.1 x 10⁹/L). Despite the parenteral administration of corticosteroids, the hematologic effects worsened within 72 hours to include anemia (hemoglobin concentration decreased from 118 g/L to 73 g/L) and more severe leukopenia and thrombocytopenia. With supportive care, the patient's hematologic abnormalities resolved and she recovered with no sequelae.

In contrast to the leukopenia seen in the patients described by Maquire and colleagues [1], our patient had a more profound lymphocytopenia (absolute count, 0.068 x 10⁹/L in the initial episode), with relative sparing of the granulocytic cell line. The pathophysiologic mechanism responsible for this finding is unclear, but quinine-induced antibodies such as those found by Stroncek and colleagues [2] to react with platelets, red blood cells, neutrophils, and T lymphocytes might be involved.

It is clear that this newly described phenomenon represents a dangerous side effect of a compound found in many beverages and over-the-counter preparations. We agree with the authors that such preparations should be labeled with strong warnings, and urge complete and repetitive history-taking, with attention to all preparations ingested, when evaluating patients with recurring cryptogenic illnesses.