Haemolytic-Uraemic Syndrome: Neurologic Symptoms, Neuroimaging and Neurocognitive Outcome

Haemolytic-Uraemic Syndrome (HUS) was first described in 1955 by Gasser (Perez del Campo et al., 2000) and defined as a multi-systemic syndrome, due to the association of microangyopathic haemolytic anemia, thrombopenia and multiorganic aggression. HUS affects mainly kidneys and leads to acute renal failure with high levels of urea and creatinin; it often involves digestive and central nervous systems. Central nervous system (CNS) lesions, typically at the basal ganglia, may also affect cortico-subcortical areas and in so doing determines motor and neurocognitive outcome, and modify the patients’ quality of life. Incidence of HUS varies among continents, highly influenced by migration movements, and it is estimated to be around 18/100,000 in children younger than 5 years old. Some countries, like Argentina and South Africa, are considered “endemic”, with a steady and relatively high incidence of HUS during all the seasons of the year; other areas, such as Canada, most of the European countries, and the west coast of the USA, are said to be “epidemic”, with sporadic cases and a lower incidence of HUS compared to Latin America and Africa during most of the year, but with self-limited relapses during summertime (Exeni, 2001). HUS etiology is diverse and physiopathologic mechanisms are not yet well known, but infective microorganisms are frequently involved, especially Escherichia coli, serotype O15:H7; this bacteria is able to produce a toxic protein (vero-toxin –VTor Shiga-toxin –Stx), which “recognizes” the endothelial cells and provokes an endothelial lesion (Scheiring, 2010). Other bacteria seem to be involved in different cases of HUS, like Salmonella enteritidis and Streptococcus pneumoniae (De Loos et al., 2002; Prestidge & Wong, 2009). Mutations in genes coding for different components of the complement system seem to be a risk factor for HUS (Skerka et al., 2009). However, the etiologic agent remains unidentified in most patients. Clinical presentation in the acute phase includes acute renal failure (100% of patients), often high blood pressure (HBP) due to a volume surcharge (35-40% of patients), and neurological


Introduction
Haemolytic-Uraemic Syndrome (HUS) was first described in 1955 by Gasser (Pérez del Campo et al., 2000) and defined as a multi-systemic syndrome, due to the association of microangyopathic haemolytic anemia, thrombopenia and multiorganic aggression.HUS affects mainly kidneys and leads to acute renal failure with high levels of urea and creatinin; it often involves digestive and central nervous systems.Central nervous system (CNS) lesions, typically at the basal ganglia, may also affect cortico-subcortical areas and in so doing determines motor and neurocognitive outcome, and modify the patients' quality of life.Incidence of HUS varies among continents, highly influenced by migration movements, and it is estimated to be around 18/100,000 in children younger than 5 years old.Some countries, like Argentina and South Africa, are considered "endemic", with a steady and relatively high incidence of HUS during all the seasons of the year; other areas, such as Canada, most of the European countries, and the west coast of the USA, are said to be "epidemic", with sporadic cases and a lower incidence of HUS compared to Latin America and Africa during most of the year, but with self-limited relapses during summertime (Exeni, 2001).HUS etiology is diverse and physiopathologic mechanisms are not yet well known, but infective microorganisms are frequently involved, especially Escherichia coli, serotype O15:H7; this bacteria is able to produce a toxic protein (vero-toxin -VT-or Shiga-toxin -Stx), which "recognizes" the endothelial cells and provokes an endothelial lesion (Scheiring, 2010).Other bacteria seem to be involved in different cases of HUS, like Salmonella enteritidis and Streptococcus pneumoniae (De Loos et al., 2002;Prestidge & Wong, 2009).Mutations in genes coding for different components of the complement system seem to be a risk factor for HUS (Skerka et al., 2009).However, the etiologic agent remains unidentified in most patients.Clinical presentation in the acute phase includes acute renal failure (100% of patients), often high blood pressure (HBP) due to a volume surcharge (35-40% of patients), and neurological symptoms such as irritability, drowsiness, seizures, cortical blindness, hemiparesis or coma, in up to 35-50% of patients (Montoliu, 1989).These symptoms may be a consequence of different disturbances: metabolic distress (hyponatremia, hyperglycemia, acidosis, fluid imbalance), HBP itself, or CNS microangyopathy.Pancreatic failure and heart involvement are less frequent during the acute phase of HUS (2% of patients).HUS treatment is based on hydro-electrolytic management: peripheral and central venous pressure must be monitored and cardiac function must be closely controlled; renal function control is especially important, as well as caloric intake adjustment.Neurologic evaluation at the acute phase and during follow-up is crucial to diagnose CNS damage and prevent medium-and long-term sequelae.No complementary tests have yet been developed to help the clinician in establishing a medium-or long-term prognosis in patients with HUS presenting with neurologic symptoms.Although during the 1980s some authors observed a good clinical outcome in patients with microangyopathic lesions (Steinborn et al., 2004), few references have reported long-term follow-up in these patients.Over the last 20 years, some cases of posterior reversible leuko-encephalopathy syndrome of subacute onset (presenting with drowsiness, lethargy, visual disturbance or seizures) have been described in the context of HUS, sometimes not even associated with HBP (Bennett et al., 2003;Gómez-Lado et al., 2007;Kitamura et al., 2010).Prognosis factors previously described in different series of patients (Cimolai et al., 1992 ;Roche et al. 2008), including patient age, acute gastroenteritis symptoms, etiologic agent, seizures at onset, CNS images at the acute phase and neurofunctional tests performance, are reviewed below; clinical course during follow-up and long-term outcome of HUS patients with neurological symptoms are also analyzed.

Material and methods
Over the last 30 years

205
-Medium-and long-term outcome (2-18 years).Neurologic evaluation was performed by a pediatric neurologist when abnormalities at the initial neurological examination or complementary tests were identified.Neurological sequelae were considered "medium-term" when they were present between 4 weeks and 12 months after clinical onset; complications were considered "long-term" when they persisted for more than 1 year after admission.Neurocognitive evaluation was performed whe n m e d i u m o r l o n g t e r m s e q u e l a e w e r e identified.In these patients, physiotherapy and neurocognitive intervention were started as soon as possible after hospital discharge and continued during the school years.
-Pathology data of the exitus are also summarized.Follow-up was maintained until clinical normalization or at least 2 years after admission.

Results
The following tables summarize the patients' characteristics (sex, age at onset), causative agent, clinical presentation, diagnostic tests and clinical course of the 25 patients with HUS and neurological symptoms at onset.Median age at presentation was 2 years 8 months (range 7 months-7 years old).As shown in Table 1, sex distribution in HUS patients with neurologic symptoms reveals a higher proportion of girls (64%), with a boy/girl rate of 1:1.7; the rate among patients without neurologic symptoms was 1:1.2, slightly more frequent in girls.Recent history of acute gastroenteritis (D+) was present in 24/25 patients with HUS and neurological symptoms at onset, although etiologic agent was only found in blood in 4/25 (two Salmonella enteritidis and two E. coli).One of these patients presented E coli both in blood and urine, and another had Salmonella in blood and E. coli in urine; E. coli was also present in urine in another patient.The most frequent neurologic sign at onset was drowsiness alone (40%) or together with irritability (16%), while irritability alone was present in 10%.Nine patients suffered seizures at onset (generalized tonic-clonic, tonic or myoclonic seizures), which stands for 14% of all HUS patients and 36% of neurologic HUS patients.All nine patients survived without important long term sequelae.However, patients presenting seizure recurrence (patient 6) or myoclonic seizures during the acute phase (patient 9) developed medium term sequelae.
One patient presented orolingual dystonia shortly after clinical onset with irritability and drowsiness; no other patients showed abnormal movements at the acute phase or during follow up.
Eleven children had some neurological complementary test performed: eye funduscopy showed fovea erythrosis in patient 4 (exitus), and patient 21 presented delayed and disorganized VEP with normal BEAR.Video EEG was abnormal for all the patients who underwent it (5/9), with slow background activity; transfontanellar US, BMA Doppler US brain CT and MRI (both in the acute phase and during follow-up) findings are summarized in Table 2. Brain MRI findings of patients 6 (Figures 1 and 2) and 21 (Figures 3 and 4) are consistent with vasculitic lesions due to diffuse hypoxic-ischemic aggression, with cortico-subcortical and basal ganglia distribution.Despite these findings, which were persistent along followup, both patients presented a favorable course without important long-term sequelae.Five of the 25 patients with neurologic symptoms at the acute phase showed one or more medium-term neurological deficits (Table 3): 1/5 hemiparesia, 4/5 mild cognitive dysfunction and 2/5 visuo-perception and construction deficits, which almost normalized during long-term follow up.Nineteen of the 25 presented normal neurological examination at hospital discharge, and one year later.Patient 4 died within the first 15 hours after admission, after a rapidly progressive neurologic deterioration and respiratory arrest.He presented lower limb myoclonias after life rescue.Thorax x-ray revealed right inferior lobe (RIL) pneumonia.Abdomen and transfontanellar US were normal.Pathology studies confirmed RIL pneumonia, severe segmentary glomerular and tubular nephropathy, acute pancreatitis, lung and heart interstitial inflammation, diffuse alveolar damage, intracapillary thrombi in lungs and kidneys, brain cortical necrosis with edema and cerebellar granular necrosis.Steptococcus pneumoniae was not identified.This represents a mortality of 1.5% of the HUS patients and 4% of the patients with HUS and neurologic symptoms at onset.

Comments
HUS is a multisystemic entity; its incidence in Europe has been sporadic in the past, although recent migration movements have facilitated a relapse of cases in several countries.I n g e n e r a l , o l d e r p a t i e n t s t e n d t o s h o w m i l d e r n e u r o l o g i c s y m p t o m s a t o n s e t , l i k e drowsiness or irritability, while younger patients, especially under 18 months, tend to present seizures during the acute phase.Physiopathology is not yet well understood, but experimental and in vivo studies (Ren et al., 1999;Carter, 1986;Cimolai, 1896) have proved that Escherichia coli VT induces thrombopenia through consuming, kidnapping, aggregation and platelet dysfunction mechanisms; plasminogen inhibitor activity is also enhanced, and therefore fibrinolysis is inhibited.Released factors such as TNF, IL, FvW monomers, free radicals, thromboxane, etc., provoke endothelial lesions and vasculitic events in several organs, especially kidneys, digestive system and brain (Seth et al., 1896;Miller & Kin, 1987;Montoliu, 1989;Hahn et al., 1989;Erikson et al., 2001;Steinborn et al., 2004;Rivero et al., 2004).VT receptors are present in various troncoencephalic nuclei, the amygdala and the hippocampus, and in the posterior root neurons of the ganglia.This suggests VT may induce primary neuronal damage as well as a vasculitic lesion (Hahn et al., 1989;Hamano et al., 1993;Rivero et al., 2004).This probably happens also at the basal ganglia, especially at putamen nucleae, the most frequent localization of CNS lesions (Nakamura et al., 2003).The vasculitic damage (due to the diffuse hypoxic-ischemic aggression) observed in our patients was mainly localized at cortico-subcortical areas and the basal ganglia, as described in previous reports (Ren et al., 1999;Akasaka et al., 1999;Garel et al., 2004).Clinical course of patients with these lesions was favorable, and MRI lesions became smaller on follow-up controls.Brain MRI sequenced controls of patients 6 and 21 reinforce the hypothesis of vasculitic lesion as the main cause of tissue damage, although direct neuronal toxicity could not be disclosed (Hahn et al., 1989).
In contrast with previous reports (Theobald et al., 2001), basal ganglia necrosis has not proved to be a bad prognosis factor in our series: our patients did not present extrapyramidal signs, as other authors have reported (Di Mario et al., 1987;Barnett et al., 1995).Unfavorable neurologic outcome was formerly correlated with seizures at onset of symptoms and plasmapheresis (unnecessary for our patients) at diagnosis (Cimolai et al., 1986).The nine patients who presented seizures at onset survived without long-term sequelae, whereas the only exitus presented initial drowsiness and rapidly progressive neurologic deterioration, without seizures (myoclonias happened after resuscitation maneuvers).This was the only patient with neurologic symptoms, acute pancreatitis, endocarditis and RIL pneumonia; Streptococcus pneumoniae was not detected (De Loos et al., 2002;Prestidge & Wong, 2009).Neurologic evaluation and follow-up of patients with CNS symptoms allowed early detection of subtle vision dysfunction, visual perception deficit, and mild cognitive disabilities.
It was previously reported that HUS patients with partial seizures tend to present epilepsy or abnormal movements after HUS recovery (Dhuna et al., 1992;Hue et al., 1992;Koehl et al., 2010).However, none of our patients developed abnormal movements during mediumand long-term follow-up, and seizures or EEG abnormalities at the acute phase did not determine a poor outcome (only the patient with previously diagnosed cortical dysplasia presented focal seizures).SHU mortality has decreased in recent years, from 25% in the 1980s to 2% in more recent publications (Rivero et al., 2004).Despite this low mortality rate, a small percentage of patients with neurological symptoms at the acute stage subsequently present neurological sequelae.In our series of 25 children with neurological symptoms, one patient died and 5 had medium-term neurological complications (hemiparesia, cognitive delay or visual perception deficit).The rates of medium-term neurologic morbidity (20%) and mortality (4%) were similar to those of other authors (Hahn et al., 1989;Erikson et al., 2001).Only in one patient after 3 years of follow-up were there persistent minor neurological sequelae (slight cognitive, visual perception and visual construction impairments), with gradual improvement despite the absence of significant changes on MRI and visual evoked potentials monitoring.Although neurocognitive impairment is not frequently reported in HUS (Roche et al., 2008), neuropsychological evaluation and follow-up of these children, especially when basal ganglia (mainly putamen) and cortico-subcortical regions are damaged at the initial brain MRI, helps to identify neurocognitive disabilities.Even if they are not severe, a good neurofunctional diagnosis and rehabilitation can help patients with their school performance and day-to-day life.

Conclusions
In summary, HUS is not yet completely understood from a physiological point of view.The most common neurological manifestations in the acute phase are drowsiness, stupor, irritability and convulsions.Neurological morbidity is important: it affects 20% of children with acute neurological presentation (8% of all patients with HUS).Seizures at presentation were not a risk factor for poor outcome in our series.Electrophysiological abnormalities at the acute phase tend to normalize; when they persist, clinical expression is very subtle.Importantly, brain lesions may persist during follow-up despite clinical recovery.No clear correlation can be established between MRI findings and long-term clinical outcome.Neurocognitive evaluation of children with neurological impairment in the context of SHU should be part of the medium-and long-term follow-up in these patients.

Table 2 .
Results of the pathologic tests in patients with HUS and neurologic symptoms at the acute phase and MRI control.

Table 3 .
Diagnostic tests, medium-and long-term outcome of patients with neurologic symptoms at the acute phase; (*) abnormal test.