Previous studies have shown that age at onset of primary haemolytic ur
aemic syndrome (HUS) is a feature of prognostic significance, the dise
ase being of much better outcome in paediatric patients younger than 3
years than in older children. In an attempt to find an explanation fo
r such a difference, we analysed the clinical and pathological feature
s of 42 children over 3 years of age who presented with HUS between 19
55 and 1990 in our department. On the basis of the presence of a prodr
omal diarrhoea, we divided our patients into two groups: 21 children p
resented with the diarrhoea-associated (typical or D+) form of HUS, wh
ereas 21 had the non-diarrhoea-associated (atypical or D-) form. Of th
e 42 children, 20 (47.5%) progressed to end-stage renal failure. Howev
er our study shows that age at onset of HUS is not a prognostic featur
e per se. The difference in outcome between children and infants is mo
st likely related to the high incidence of the atypical subset of HUS
in children over 3 years, a subset that is very uncommon in infants. T
he ominous features which characterise this form of the disease are: (
1) the absence of a diarrhoeal prodrome, (2) normal urine output, (3)
marked proteinuria, (4) hypertension, (5) the occurrence of relapses o
r recurrences and (6) the presence of widespread and severe arteriolar
changes on renal biopsy. The poor prognosis of the atypical form of H
US warrants the use of fresh-frozen plasma infusions and/or plasma exc
hange as early as possible in the course of the disease.