Hepatitis E virus infection with nonimmune hemolytic anemia.

INTRODUCTION Anemia frequently occurs in association with acute and chronic hepatitis. Aplastic anemia and autoimmune hemolytic anemia occur in various viral infections including hepatitis A, B and C (1-5). Nonimmune hemolytic anemia is well described in Wilson disease (6,7). Anemia with hepatitis E infection has only been reported in patients with glucose-6-phosphate dehydrogenase (G6PD) deficiency (8,9). We report a case of Coombs-negative hemolytic anemia in association with acute hepatitis E virus infection that resulted in a decrease of hemoglobin from 10.7 to 3.6 g/dL. In this case no underlying hemolytic disorder was initially apparent but the patient was subsequently confirmed to be G6PD deficient. CASE REPORT A 10-year-old girl of Pakistani origin presented to her local hospital with a 4-day history of vomiting, malaise, abdominal pain, headache, pyrexia and jaundice. She was transferred to our center for evaluation. Three weeks previously she had returned from a 4-week holiday in Pakistan during which she had been well. She had been diagnosed with autoimmune thyroiditis 18 months previously and was receiving thyroxine. She was not taking any other medication and had not taken antimalarial chemoprophylaxis. She had received hepatitis A vaccination before travel. She had no previous history of jaundice, pallor or abdominal pain. She was the second child born to nonconsanguinous parents and there was no significant family history of anemia or jaundice. On physical examination she was afebrile but was pale and jaundiced. There were no stigmata of chronic liver disease and no lymphadenopathy. She had tender hepatomegaly palpable 3 cm below the subcostal margin and a palpable spleen tip but no ascites. She was not encephalopathic. Cardiovascular, respiratory, musculo-skeletal and neurologic examinations were normal. No Kayser-Fleisher rings were seen on slit lamp examination of the eyes. Laboratory investigations (Table 1) revealed an initial hemoglobin of 10.7 g/dL (normal range, 12-16 g/dL), which subsequently decreased to 3.6 g/dL by day 4, leukocyte count of 23 × 109/L (normal range, 4 to 11 × 109/L), neutrophils 15 × 109/L (normal range, 2 to 7.5 × 109/L) and platelets of 440 × 109/L (normal range, 140 to 380 × 109/L). The absolute reticulocyte count was 170 × 109/L (normal range, 20 to 80 × 109/L), and the reticulocyte percentage was 14.2% (0.8% to 1.8%). Peripheral blood smear showed a leucoerythroblastic blood picture with polychromasia, some spherocytes and irregular contracted cells. There was no morphologic evidence of red blood cell membrane abnormality. No malarial parasites were seen.TABLE 1: Laboratory valuesUrine hemosiderin was positive, direct antiglobulin test (direct Coombs test) was negative and haptoglobin was 0.09 g/L (normal range, 0.4 to 1.6 g/L), all consistent with nonimmune hemolysis. Glucose-6-phosphate dehydrogenase (G6PD) level was 6.6 IU/g hemoglobin (normal range, 4.6 to 13.5 IU/g hemoglobin). Pyruvate kinase analysis was not performed. Biochemical indicators of liver dysfunction peaked as follows: serum bilirubin 665 m mol/L (3 to 15 m mol/L), alanine aminotransferase 3140 IU/L (0 to 35 IU/L), alkaline phosphatase 877 IU/L (70 to 300 IU/L) and gamma glutamyl transferase 71 IU/L (8 to 48 IU/L). Serum albumin was 36 g/L (37 to 49 g/L) and the international normalized ratio was 1.3 (0.9 to 1.3). Serum cholesterol and renal function were normal and C- reactive protein was not increased. Blood and urine cultures were negative and viral markers for hepatitis A, B and C, Epstein-Barr virus, cytomegalovirus, herpes simplex, measles, varicella zoster and parvovirus were all negative. Paul-Bunnel test and blood smear for malarial parasites were negative. Hepatitis E immunoglobulin (Ig)M (by ELISA: "Genelab", Singapore) was strongly positive at presentation and weakly positive after 6 weeks of convalescence. Hepatitis E IgG was positive on both occasions. Antinuclear antibodies, antimitochondrial antibodies and anti-LKM antibodies were negative. Serum immunoglobulins were marginally increased with IgG, IgM and IgA of 16.5 g/L (5.4 to 16.1 g/L), 4.55 g/L (0.50 to 1.80 g/L) and 3.5 g/L (0.70 to 2.50 g/L), respectively. Serum ceruloplasmin and 24-hour urinary copper before and after penicillamine administration were normal. Abdominal ultrasound demonstrated hepatosplenomegaly with no intrahepatic duct dilatation, normal common bile duct and patent vessels with forward flow. The gall bladder was thick walled and contained sludge. There were no ascites. She received a blood transfusion, oral fat-soluble vitamins and intravenous fluids. She made a gradual recovery and was discharged home 1 week after initial presentation. At review 6 weeks later she was well with normal liver function tests and hemoglobin and no further evidence of hemolysis. When tested 1 year later, G6PD deficiency was apparent. The red blood cell G6PD levels (Sigma diagnostic) were 0.8 IU (normal range, 7.0 to 11.1) and her hemoglobin was 12.1 g/dL with reticulocyte percent-age of 1.1. Family screening for G6PD deficiency is in progress. DISCUSSION Hepatitis E virus (HEV) is a non-enveloped RNA virus responsible for large epidemics of acute hepatitis and sporadic hepatitis cases in southeast and central Asia (10). Sporadic cases have been reported in developed countries, including the United Kingdom, usually in travelers but also in patients with no history of travel abroad or other risk factors (11). HEV is transmitted by the fecal-oral route and there is very low person to person transmission rate. The incubation period ranges from 2 to 10 weeks. The illness is usually self limiting lasting 1 to 4 weeks. As the illness subsides, serum aminotransferases and bilirubin decrease, returning to normal after 6 weeks in most patients. In acute HEV hepatitis, IgM appears during the early phase of clinical symptoms, preceding IgG by several days, and disappears over 4 to 5 months. IgG persists for several years after infection (10). Our patient had visited Pakistan, where HEV infection is endemic, before becoming unwell, and the diagnosis was established by the clinical presentation and detection of anti-HEV IgM and IgG. Although HEV RNA was not assayed, the serological profile was conclusive of recent HEV infection. Our patient had evidence of intravascular hemolysis with a sudden decrease in hemoglobin and an associated increase in bilirubin, low serum haptoglobin and urine positive for hemosiderin but negative direct Coombs test. Nonimmune hemolytic anemia (Coombs test negative) may occur in liver disease and is well described in Wilson disease (6,7). This diagnosis was not supported by laboratory investigations and further excluded by her spontaneous recovery. Previous case reports have identified other hepatitic viruses as the cause of aplastic anemia (1) and autoimmune haemolytic anemia in association with acute hepatitis. Reported viruses include cytomegalovirus (12) Epstein-Barr virus (13) hepatitis A (2) and hepatitis B (3,4). These infections and hepatitis C, varicella zoster, herpes simplex, measles and parvovirus infection were excluded in our patient by negative serology. Although our patient had autoimmune thyroiditis, there was no evidence to support a diagnosis of autoimmune hepatitis. Furthermore, her anemia was Coombs negative. The blood smear showed leucoerythroblastic blood picture with some contracted cells and spherocytes. There was no evidence of red blood cell membrane abnormalities. Severe hemolysis has been documented in children with G6PD deficiecy who acquire HEV infection (8,9). In our patient G6PD assay was normal during her acute presentation. On retesting at 1 year, however, she was confirmed to be G6PD deficient. Although elevation of G6PD level into the normal range may occur as a result of reticulocytosis, this diagnosis was initially considered to be unlikely. Her blood smear was not typical of G6PD deficiency, and the presentation would more typically be in a younger child. As there was no history to support a diagnosis of drug-induced hemolysis, and as the presence of Coombs negative hemolysis in our patient occurred and resolved simultaneously with acute hepatitis E infection, it appears most likely that this was a related phenomenon. In conclusion, this case presents further evidence that hepatitis E may trigger hemolysis in children with G6PD deficiency and also may lead to the first clinical presentation of the disorder. G6PD testing should be performed when acute hemolysis has resolved.