1042 Haemolytic-uraemic syndrome SiR-Moake’s review (Feb 12, p 393) of haemolytic-uraemic syndrome (HUS) provides a compelling argument for the direct role of shiga or shiga-like toxins in the pathophysiology of diarrhoea associated D+HUS. Moake implicates these bacterial exotoxins in the endothelial dysfunction and von Willebrand factor mediated platelet consumption associated with the disease. It seems unlikely, however, that these potent toxins fully explain the inflammatory response and more widespread extra-renal endothelial damage that seem to be central to the multisystem dysfunction. 1.2 Although the accentuating effects of co-factors such as lipopolysaccharide (LPS), tumour necrosis factor, and interleukin-1 are alluded to, we contend that LPS has a more important role than Moake suggests. Several groups have detected specific antibodies to the LPS of verocytotoxin-producing organisms in patients with HUS .3-5 We have investigated the acute humoral immune response to the highly conserved, inner-core glycolipid region of LPS in 20 children with moderate (mean age 4-0 years, range 0-7-11) and 14 children with severe (3-4 years, range 1-6-11-8) D+HUS. D+HUS severity was classified into moderate (neutrophil count < 10 x 109jL) or severe.2 21 patients had haemorrhagic colitis, 8 of the severely affected patients had severe gastrointestinal disease, and 5 died as a result of their illness. Serum IgG and IgM endotoxin-core antibodies were measured by ELISA. Results are recorded here in median units (MU), the median for normal blood donors being set at 100 MU. The median antibody concentrations in acute phase D+HUS were 64-9 (95% CI 32’1-94,69) for IGM and 53 1 (0-54-71-55) for IgG, and were lower than those seen in adult blood donors (figure). Significantly lower IgG values were . Moderately severe o Severe Normal range Figure: Endotoxin core concentrations (median and 95% CO In 20 children with moderate and 14 with severe D + HUS in acute phase *p=0 0 12, **p< 0 007. Results calculated in relation to standard reference serum and expressed as median units (MU). Normal median (broken line) and ranges (10th-90th percentiles) were determined for 1000 healthy blood donors. Levels normally reach adult range by end of first year of life (unpublished). seen in patients with severe HUS compared with those with moderate disease (median 32-9 [0-57’7] vs 59 4 [501-7624] MU; U test p < 0-007). Although specific IgM values were also lower in the severely affected group (median 37 5 [22-9-157 4] vs 75-1 [52-98’5] MU) the difference was not significant (p=012). Low or falling IgM and IgG concentrations were recorded in the patients who died. Endotoxin core-antibodies constitute an important component of host defence against gram-negative bacteria, are capable of opsonisation, and are thought to act as "master keys" in the initial phases of infection before more specific antibodies develop. We propose that in severe D+HUS, mucosal inflammation results in the dissemination of LPS from gram-negative bacteria into the systemic circulation, resulting in the consumption of endotoxin core antibodies, activation of inflammatory cells, and disruption of endothelial function. These findings further implicate LPS in the systemic manifestations of D + HU S, and suggest that LPS- core antibodies may be important in host defence. Further studies are required to confirm these possibilities and to explore the therapeutic benefit of immunotherapy with endotoxin and LPS-core antibodies in severe disease. Robert S Heyderman, Margaret M Fitzpatrick, G Robin Barclay Paediatric Infectious Diseases Unit, St Mary’s Hospital Medical School, London W2 1NY, UK; Department of Paediatrics, St James’s University Hospital, Leeds; and South East Scotland Regional Blood Transfusion Centre, Edinburgh 1 Kaplan BS, Cleary BS, Obrig TG. Recent advances in understanding the pathogenesis of haemolytic uraemic syndromes. Pediatr Nephrol 1990; 4: 276-83. 2 Fitzpatrick MM, Shah V, Trompeter RS, Dillon MJ, Barratt TM. Interleukin-8 and polymorphoneutrophil leukocyte activation in haemolytic uraemic syndrome of childhood. Kidney Int 1992; 42: 951-56. 3 Chart H, Scotland SM, Rowe B. Serum antibodies to Escherichia coli 0157:H7 in patients with hemolytic uremic syndrome. J Clin Microbiol 1989; 27: 285-90. 4 Bitzan M, Moeblus E, Ludwig K, Muller-Wiefel DE, Heesemann J, Karch H. High incidence of serum antibodies to Escherichia coli lipopolysaccharide in children with hemolytic-uremic syndrome. J Pediatr 1991; 119: 380-85. 5 Caprioli A, Luzzi I, Rosmini F, et al. Hemolytic-uremic syndrome and vero-cytotoxin-producing Escherichia coli infection in Italy. J Infect Dis 1992; 166: 154-58 SiR-In their reviews, Moake and Neild mention shigella-like toxins in diarrhoea-associated HUS. Verotoxin-producing Escherichia coli (VTEC) is known to play a major part in diarrhoea-associated HUS,1 but the isolation of VTEC in blood or the association of VTEC with pyelonephritis is not reported. We record a well-documented HUS complicating the course of a VTEC urinary infection. A 10-month-old girl with fever for 2 days and red urine for 6 h was admitted to our department. 10 days before admission the patient and her mother had mild non-bloody diarrhoea. Laboratory findings were: white cell count 23-5 x 109/L with 13% band forms, haemoglobin 92 g/L, platelets 179 x 10’/L, C-reactive protein 311 mg/L, creatinine 83 &micro;mol/L. Urinalysis showed large numbers of erythrocytes, many bacteria, 4-8 leucocytes per centrifuged high-power field, and proteinuria ( + + + ). Intravenous amoxycillin and netilmicin were started. On the next day both urine (> 108 colony forming units per L) and blood cultures were positive for E coli. Within 2 days she developed worsened anaemia (haemoglobin 5 g/L), red blood cell fragmentocytosis, thrombocytopenia (11 x 109/L), and anuric renal failure. Peritoneal dialysis was instituted for 14 days. Renal failure subsequently resolved. Voiding cystourethrogram showed severe vesicoureteric reflux on the left side. The E coli strains isolated in blood and urine were serotyped: they both belonged to the serovar 017:H18 and harboured