Streptococcus Pneumoniae-Associated Hemolytic Uremic Syndrome

Hemolytic uremic syndrome (HUS) is the most common cause of acute renal failure in many Western countries.1 Most cases are caused by Shiga-like toxin-producing strains of Escherichia coli and are associated with a prodromal illness of bloody diarrhea. Those cases not associated with a diarrhea prodrome have diverse etiologies including inherited abnormalities of the complement system, other familial forms and various infections.2 Infection with Streptococcus pneumoniae has been reported as a cause of HUS not associated with a diarrhea prodrome, and this may not be as rare as commonly believed.3 Because early recognition of the disease may have important implications for treatment and prognosis, it is essential that clinicians be familiar with the disease.4 Two cases that illustrate important clinical and laboratory features of the disease are presented. Case reports.Case 1. A 2 1/2-year-old girl was admitted to another hospital with left lower lobe pneumonia. She was given iv erythromycin because of a history of penicillin allergy. After 48 h of treatment she remained febrile with clinical signs of left lower lobe consolidation and associated empyema, confirmed by radiography. A limited thoracotomy with insertion of two chest drains was performed, and the antibiotic was changed to ceftriaxone. Pus from the empyema cultured S. pneumoniae resistant to erythromycin but susceptible to penicillin and ceftriaxone. Eight days after admission she deteriorated rapidly with decreasing urine output, increasing respiratory distress and decreasing arterial oxygen saturation. Her hemoglobin concentration fell from 12.4 g/dl on admission to 7.7 g/dl. She was given a transfusion of 10 ml/kg of red blood cell concentrate after which her condition deteriorated dramatically. After endotracheal intubation and initiation of positive pressure ventilation, she was transferred to the Red Cross War Memorial Children's hospital. On admission she had clinical and radiologic evidence of left lower lobe consolidation and pulmonary edema. The liver was enlarged to 6 cm below the costal margin. Her blood pressure was 150/70 mm Hg. Plasma sodium was 121 mmol/l, potassium 5.5 mmol/l, urea 32.3 mmol/l, creatinine 435 μmol/l, alanine aminotransferase (ALT) 193 units/l and lactate dehydrogenase 3207 units/l. The hemoglobin concentration was 10.1 g/dl falling to 8.8 g/dl the following day, the leukocyte count was 26.2 × 109/l and the platelet count was 78 × 109/l. The blood smear was positive for schistocytes. Erythrocyte T cryptantigen activation was positive by the peanut lectin assay. She required assisted respiration for 3 days and peritoneal dialysis for 18 days. She received two further transfusions of washed red blood cell concentrate. Three years after discharge she is normotensive but has persistent 2+ proteinuria on dipsticks. Her urine protein: creatinine ratio is 87 mg/mmol (normal, <20) and her glomerular filtration rate measured by 51Cr-labeled diaminotetraacetic acid (51Cr-EDTA) clearance is 78 ml/min/1.73 m2 (normal, 80 to 120). Case 2. This 7-month-old infant was referred with pneumonia and empyema, anemia, thrombocytopenia and hepatomegaly. He had been unwell for 1 week and had been admitted to the hospital the day before transfer. On examination he was pale and in distress. He had right-sided pneumonic consolidation with an associated effusion. His liver was enlarged 8 cm below the right costal margin, and the spleen was enlarged 3 cm below the left. His blood pressure was 140/70 mm Hg. His plasma sodium was 125 mmol/l, potassium 6.8 mmol/l, urea 22.5 mmol/l, creatinine 353 μmol/l, bilirubin 46 μmol/l, ALT 152 units/l and lactate dehydrogenase 2850 units/l. The leukocyte count was 32.2 × 109/l, hemoglobin 5.7 g/dl and platelets 14 × 109/l. T cryptantigen was strongly positive, and the direct Coombs test was positive. Although the initial blood smear was reported as showing no schistocytes, there was marked red blood cell agglutination and a repeat smear in warm conditions revealed numerous fragmented red blood cells. A penicillin-susceptible S. pneumoniae was cultured from pus aspirated from the pleural space. He required 16 days of peritoneal dialysis. One year later he is normotensive but has persistent proteinuria (protein:creatinine ratio, 100 mg/mmol) and glomerular filtration rate measured by 51Cr-EDTA clearance is reduced at 67 ml/min/1.73 m2. Discussion. The Thomsen-Friedenreich (T) antigen is part of a normal structure on the surface of erythrocytes and glomerular endothelial cells which is usually hidden by neuraminic acid. Neuraminidase, secreted by S. pneumoniae, removes neuraminic acid, exposing the antigen. Most humans beyond early infancy have IgM antibodies against the T antigen in their serum. These antibodies react with exposed (activated) T antigen on the surface of cells, causing agglutination and lysis of erythrocytes and damage to glomerular endothelial cells, resulting in HUS.5 Neuraminidase activity has been demonstrated in the sera of patients with pneumococcus-induced HUS.6 Circulating neuraminidase has not been demonstrated in neonates with necrotizing enterocolitis, who have T antigen activation by Clostridium. It appears that erythrocyte T antigen activation occurs in a regional manner in necrotizing enterocolitis and that renal endothelial cells are not involved. Several authors have suggested that S. pneumoniae is an extremely uncommon cause of HUS.7, 8 However, a recent review of HUS patients in Queensland, Australia, found that 3 of 55 (5%) had pneumococcal disease with T antigen activation.3 In our hospital HUS was uncommon with about 1 case per year requiring dialysis until the onset of a Shigella dysenteriae epidemic in early 1997 (unpublished data). The situation is similar in other hospitals in Southern Africa.9 These 2 cases of S. pneumoniae-associated HUS have therefore constituted the major cause of HUS during that period. The apparent rarity may simply be a result of lack of reporting, and possibly some patients have been mistakenly diagnosed with pneumococcal glomerulonephritis or other causes of renal failure. Both our patients had hepatomegaly with deranged liver functions. Given that both patients were fluid-overloaded at the time of admission, some of this may have been related to hepatic congestion, but this may have been related to liver involvement in the disease process as has recently been documented.10 Liver involvement has a benign course and extensive evaluation of the liver disease is not justified unless it persists beyond 2 to 3 weeks. The mechanism of liver disease is not established, but might also be caused by T antigen activation on hepatocytes. Early recognition of HUS caused by S. pneumoniae infection is important because use of plasma or plasma containing blood products may aggravate the disorder because of the addition of donor IgM specific for the T antigen.4 There is an additional danger of fluid overload in patients with unrecognized renal failure. The possibility of T antigen activation should be considered in any child with severe pneumococcal disease, whether pneumonia or meningitis, who deteriorates while receiving treatment, particularly if anemia is present. Those cases recognized early in their course have an excellent prognosis for survival, whereas unrecognized cases have an extremely high mortality.4 Where T antigen activation is confirmed, blood should be given only in the form of washed red blood cells. If fresh frozen plasma is essential for coagulation support, only low titer anti-T plasma should be used. Early diagnosis of renal impairment allows the clinician to avoid errors in fluid and electrolyte therapy. There is little information in the medical literature on the long term prognosis of this illness. Both of our patients has abnormal glomerular filtration rates and proteinuria, and the possibility of eventual progression to end stage renal failure cannot be excluded. Rodney D. Gilbert, M.Med., F.C.P., D.C.H. Andrew C. Argent, M.Med., F.C.P., D.C.H.