Hemolytic uremic syndrome (HUS) is the most common cause of childhood acute renal failure in developed countries. Early recognition of this condition leads to optimum management and improved outcome. Complications are not limited to the kidney, and neurological complications may be devastating. Understanding the etiology and pathogenesis of this condition is incomplete. There is currently no specific therapy that will alter the clinical course of HUS. The following case report describes a typical course of a patient with moderately severe HUS in which neurologic manifestations dominated the clinical picture.
An 11 -month-old male presented to local medical authorities with a history of 2 days of bloody diarrhea during an epidemic of gastroenteritis in a remote Northern community. He was thought to have a good urine output and was well hydrated. He was sent home. The following day he returned, looking slightly dehydrated, and he was irritable. He was admitted to a local hospital. That night, he had a generalized clonic seizure that lasted for 1 1 minutes. After 5 mg of diazepam was given intravenously, the seizure stopped. He continued to have periodic 15 to 20 second clonic episodes. After additional diazepam and intravenous phénobarbital, his respirations ceased and he required intubation. He was transferred to the Winnipeg Children's Hospital by air ambulance for further treatment. The patients 6-year-old sister had been transferred to Children's Hospital 3 days earlier with a diagnosis of HUS. She died after initiation of dialysis due to nonhemorrhagic pericardial tamponade. A 4-year-old sister of the patient also was admitted to hospital for treatment of HUS 2 days previously. She maintained an acceptable urine output with intravenous furosemide therapy, received packed cell transfusions, and did not require dialysis.
On admission, the patient was found to be a normal size child for age who was normotensive with good capillary filling and a heart rate of 181 beats/minute. He was not edematous. An abdominal examination was normal. Laboratory investigations revealed a white blood cell count of 30 600, of which 87% were polymorphonucleocytes, and a hemoglobin of 112 q/L A microangiopathic hemolytic anemia was detected on the blood smear. His platelet count was 63 000. Blood urea nitrogen level was 1 1 mmol/L with a creatinine level of 104 µ????/L. Serum electrolytes revealed a sodium level of 128 mmol/L, a potassium level of 4.3 mmol/L, and a carbon dioxide level of 14.1 mtnol. Liver function tests and a serum amylase were normal. His serum albumin was 20 g/L, and the prothrombin and partial thromboplastin times were normal.
Over the following days, the patients platelet count fell to 43 000. The hemoglobin fell to 62 q/L, and the patient was transfused with packed red blood cells. The initial urine output was monitored with a Fofey catheter and was 0.5 cc/kg/hour. A furosemide infusion was started at 0.5 mg/kg/hour and titrated according to the urine output up to a dose of 2 mg/kg/hour. The urine output increased to 2 cc/kg/ hour and to >2 cc/kg/hour by the fourth hospital day, following which the furosemide infusion was tapered and discontinued. Mechanical ventilation, which had been instituted on admission, was discontinued on the second hospital day. Hypertension (135/70 mm Hg) developed on the third hospital day and was treated initially with nifedipine and intravenous labetalol.
The patient steadily improved, and on the 19th day, he had a normal neurological examination and audiogram, and was discharged home. Two months after presentation, he was normotensive ( 102/48 mm Hg) with a creatinine of 55 µp???/L, a hemoglobin of 12.9, and 0.3 g/L of protein in the urine, with a specific gravity of 1.030. At follow-up visits, 1 and 2 years after his illness, he remained normotensive. His urinalysis is completely negative, and he appears to be a normal child.
Hemolytìc uremie syndrome was described by Gasser et al in 1955.1 Since then, thousands of cases of HUS have been repotted. Initial mortality rates were reported to be close to 100%, but with current management in developed countries, mortality rates are reported to be 5% or less. Early diagnosis due to a high index of suspicion, followed by meticulous management of the deranged physiology of these patients, has resulted in improved outcome.2
In North America, the vast majority of cases of HUS are classified as the "epidemic" variety. These are patients who have an infection by a verocytotoxinproducing Escherichia coli (VTEC). The most common serotype is E coii O157:H7, although other serotypes may produce verotoxin. The toxins produced by VTEC also are known as Shiga-like toxins (SLT) and are subdivided to SLT-I and SLT-2. Shiga-like toxins bind to glycolipid receptors on the surfaces of sensitive cells. The toxin is internalized and then inhibits protein synthesis, leading to cell damage and death. Shiga-like toxin has a very short half-life in serum as it is rapidly bound by sensitive cells. It causes endothelial damage and, in animal models, direct injury of the central nervous system and mucosal erosions of the gastrointestinal tract.
In addition to SLT, pathogenic E coli release lipopolysaccharide (LPS). Lipopolysaccharide released by dying E coli may be preferentially absorbed into the circulation through mucosal erosions induced by SLT. Once LPS is absorbed into the circulation, it may activate white cells, which in turn may aggravate endothelial injury and endothelial cell detachment, increase procoagulant activity, and liberate free radicals capable of causing oxidative damage to cell membranes. Thus, LPS and SLT may have complementary synergistic effects, leading to endothelial cell damage and the process of HUS.
In addition to damage to endothelial cells and activation of white blood cells, the coagulation system also is believed to be involved in the pathogenesis of HUS. A number of abnormalities of prostaglandin I2 have been detected in patients with HUS. Prostaglandtn I2 synthesis by the endothelial cell is believed necessary to maintain the platelet in its inactive state. Intravascular platelet activation in HUS therefore may be induced by decreased prostaglandin I2 synthesis. In addition, thromboxane A2 levels (an indicator of platelet activation) are increased during HUS. Von Willebrand's factor multimers are unusually large and cause increased in vitro aggregation of platelets from normal donors. A degrading factor of von Willebrand's factor multimers may be deficient in HUS. Renal biopsies from patients with active HUS demonstrate large amounts of fibrin deposited within the glomeruli, which is thought to be due to reduced fibrinolytic activity. Red cells are injured by passage through the fibrin thrombi forming on the damaged endothelial surface of renal arterioles. Increased lipid peroxidation and abnormal vitamin E metabolism, which also has been observed in patients with HUS, may contribute to hemolysis. Thrombocytopenia is thought to be due to platelet destruction and aggregation in the kidneys and other body organs. Platelets that remain in circulation are degranulated and have impaired aggregation capabilities.
The intravascular pathology may occur in any organ in the body, but the sequelae of this process are observed most profoundly in the kidney and the gastrointestinal and central nervous systems. The changes in the glomeruli depend on when during the clinical course a biopsy is taken and whether the HUS is epidemic versus nonepidemic in form. During the initial active stage, glomeruli show thickening of capillary walls and endothelial swelling with a double contour appearance of the basement membrane on light microscopy. There are deposits of fibrillar material beneath the endothelial cells. Localized areas of fibrin may be deposited in the glomerular tufts, and the mesangium may have a fibrillar appearance. In more severe forms of HUS and in the nonepidemic variety the arteries are more severely involved and may cause ischemic changes in the downstream glomeruli. Thrombi and aneurysmal dilatations of small vessels also may be seen. The interstitium of the kidney may show edema, deposits of fibrous tissue, or scattered inflammatory cell infiltration. In severe cases, necrosis of tissue is present with areas of hemorrhage and infarction.
O'Regan and Rousseau3 have suggested that the pathology in the kidney includes a superimposed uric acid nephropathy induced by precipitation of uric acid in the renal tubules. This leads to a potential therapeutic strategy of aggressive use of diuretics to maintain urinary output as was used in the patient described above. The pathological process in other organs is similar to that which occurs in the kidney Small thrombi may be seen in widespread vessels. Small infarcts and fibrin thrombi have been reported in the brain. The colon may have areas of extensive necrosis.
Hemolytic uremic syndrome has been recognized throughout the world, but there are regions such as Argentina, California, and Central North America where it is thought to be more common. Evidence from a number of sources suggest an increase in the rate of HUS. For example, the incidence in children under the age of 5 in the United States increased fiom 1.6 per 100000 in 1979 to 5.8 per 100 000 in 1988. This is believed to be due to an increase in infections with VTEC. The incidence also may increase transiently with epidemics occurring in isolated communities, day-care centers, and institutions in which children reside.4
Typically, most cases of epidemic HUS occur between the age of 6 months and 4 years, with an approximately equal sex distribution. In the northern hemisphere, most patients with epidemic HUS present between late spring and early lñll. In some studies, there is a slight preponderance of patients from higher socioeconomic strata.5
Patients presenting with HUS can be divided on clinical grounds into two broad groups: epidemic and nonepidemic. The epidemic form occurs more frequently. It is less likely to have severe sequelae and is rarely recurrent. Outbreaks usually are associated with summer diarrhea caused most frequently by E coli 0157:H7, although other serotypes have been reported. Other bacteria reported in association with HUS include Shigella dysenteriae, Salmonella typhimurium, Campylobacter, Yersinia, and others. Verocytotoxinproducing E coli may be transmitted by food and from person to person. Verocytotoxin-producing E coIl is a commensal in the intestine of beef cattle, and bacteria may contaminate the surthce of beef during the slaughtering process. Ground beef may be contaminated throughout with VTEC. If the hamburger is not thoroughly cooked, VTEC may survive and cause infection. Epidemics have been reported in list foodrestaurant patrons and in individuals consuming unpasteurized milk and cheese. Person-to-person spread is an important method of transmission according to recent studies conducted in Canada. In addition, it has been responsible for outbreaks in day-care settings and nursing homes.5
A prodrome of diarrhea may last up to 12 days, but typically is only 3 to 4 days long. The diarrhea may be watery or bloody, and is then associated with abdominal discomfort. As the diarrhea is resolving, the child is pale secondary to the development of anemia and irritable with or without seizures due to neurological involvement. Edema and other signs of fluid overload may be present depending on the balance of fluid intake and output. A history of oliguria may be absent despite oligo or anuria because the presence of watery stools may mimic urine output in the diaper.
Nonepidemic (atypical) HUS comprises less than 10% of the total number of cases in unselected series of HUS. There is no seasonal distribution, and it may occur at any age, but it is more likely to occur in children older than those seen in typical epidemic HUS. A diarrheal prodrome is rare, and respiratory infections frequently have been reported prior to the onset of other symptoms. Some patients have a slow and progressive course, with an almost "insidious" development of HUS. Hemolysis may occur well before renal symptoms. Hypertension and neurological symptoms are seen more commonly.
Figure. Computed tomography scan of a 4-year-old girl who presented with a seizure after 3 days of epidemic summer diarrhea secondary to verocytotoxin-producing Escherichia coli. The lesion is a hemorrhagic region in the anterior capsule associated with marked cerebral edema, loss of gray-white differentiation, and collapse of the ventricles. This patient died shortly after this scan was taken.
Nonepidemic HUS may be inherited or associated with drugs such as oral contraceptives, cyclosporine, mitomycin-C, cisplatin, and vincristine, and infections such as those caused by varicella and enteroviruses. Kawasaki disease, pregnancy, and rhabdomyolysis have been associated with nonepidemic HUS. Infections with Streptococcus pneumonioe may lead to nonepidemic HUS caused by release of streptococcal neuraminidase. Circulating neuraminidase damages endothelial cells and exposes the ThomsenFriedenreich antigen. This antigen is present on red cells, platelets, and glomeruli. When it is exposed, it may react with circulating anti-Thomsen-Friedenreich IgM antibody, leading to HUS.
Familial HUS has been divided into varieties that are epidemic or nonepidemic. In epidemics of HUS, it is common to find families with more than one person manifesting HUS, usually within 2 to 3 weeks of each other. In these situations, the prognosis for each child is the same as with typical epidemic HUS. This may be due simply to person-to-person transmission of the infectious agent in a family with a susceptible genetic background. The other form of familial HUS is diagnosed when multiple members of a family develop HUS separated by periods of months or years. Typically, the children develop HUS at approximately the same age. Autosomal dominant and autosomal recessive inheritance patterns have been reported. Relapses occur frequently, and the prognosis is similar to atypical nonepidemic HUS.6
These children are often acutely ill, and investigation and management occur simultaneously. Attention is directed first to maintenance of the airway, breathing, and circulation. A quick neurological examination will reveal evidence of irritability, decreased level of consciousness, and lateralizing neurological signs. The cardiovascular examination focuses on the fluid status, presence of congestive heart failure, pulmonary edema, elevated jugular venous pressure, ascites, and peripheral edema. The abdominal examination may demonstrate distention with increased or decreased bowel sounds, peritonitis, perforation, and rectal prolapse. Examination of the skin may reveal petechiae at sites of restraint or under and distal to blood pressure cuff application.
Laboratory investigations include culture of the stool for E coii and verotoxin-producing organisms. Some laboratories may have the capability to look for the verotoxin gene using polymerase chain reaction in stool samples. Acute and convalescent sera for serologie evidence of verotoxin and lipopolysaccharide antibodies may be diagnostic. A urinalysis typically demonstrates dysmorphic red cells, some protein, and relatively few casts.
The complete blood cell count will demonstrate evidence of anemia with a microangiopathic hemolytic process. A platelet count may be low and the reticulocyte count high. Electrolytes frequently show evidence of hyponatremia secondary to hypo-osmolar fluid intake. Urea and creatinine levels may be elevated. Liver function tests are usually normal. Serum amylase and Upase levels may be elevated, indicating pancreatitis. Coagulation studies typically show elevated fibrinogen degradation products with normal prothrombin and partial thromboplastin times. A computed tomography or magnetic resonance imaging study of the brain may be necessary if there are localizing neurological signs or a deeply depressed level of consciousness (Figure). Diffuse or focal cerebral edema may be seen in the absence of identifiable thrombosis or hemorrhage and has excellent potential for recovery. Flat and upright abdominal radiographs may be taken to look for evidence of perforation.
The differential diagnosis includes other causes of giomerulonephritis or vasculitidies. However, these are unlikely to present with a microangiopathic hemolytic anemia unless associated with severe and malignant hypertension. Sepsis with or without a disseminated intravascular coagulation may present with acute renal failure and a microangiopathic process. The history and evidence of fever and other signs of infection may help to identify sepsis. In addition, septic patients tend to be hypotensive with poor perfusion and should have positive blood cultures.
A coagulopathy with abnormal prothrombin and partial thromboplastin times also mitigate against the diagnosis of HUS. Renal vein thrombosis may present associated with acute renal failure, abdominal pain, hematuria, and thrombocytopenia with or without a microangiopathic picture. Usually, however, the abdominal pain is significant, there is renal enlargement clinically, and a history of giomerulonephritis before the presentation. Doppler renal ultrasound with or without venography are helpful in differentiating HUS from renal vein thrombosis.
Hepatorenal syndrome may present with oliguria and a microangiopathy; however, symptoms and signs of liver dysfunction accompanied by evidence of a hepatic biosynthetic failure and elevation of serum ammonia would point to primary liver disease and not HUS.
Fluid management is usually prescribed to provide replacement of ongoing losses. These include insensible water loss (400 cc/m2), the urine output, and ongoing watery stool loss. Applying the usual "maintenance" fluid calculations to these children may be dangerous due to oliguria. In the child with a large volume of watery stools, it is helpful to insert a Foley catheter to separate urine output from stool output. If the patient is in a situation of fluid overload, fluid intake should be reduced appropriately. Asymptomatic hyponatremia is treated by using isotonic fluids and fluid restriction. If hyponatremia is symptomatic, small doses of hypertonic saline calculated to raise the serum sodium by 2 to 3 mmol/L are given, and further correction is carried out slowly.
Anemia is treated with transfusions of packed red blood cells if the hemoglobin is <70 or if symptoms and signs of cardiac dysfunction are present. Platelet transfusions usually are not prescribed unless active bleeding is occurring or to transiently raise the platelet count in preparation for a surgical procedure. Antibiotics are not prescribed unless strong clinical indications are present. The routine use of antibiotics is controversial and is believed by some to increase the release of verotoxin and LPS by intestinal bacteria. Anti-motility agents have been associated with more severe disease perhaps due to retention of infectious organisms. If the urine output is less than 1 mL/kg/ hour, a trial dose of furosemide ( 1 mg/kg given as a bolus increasing to 3 to 4 mg/kg given over 20 minutes) may increase the urine output. If this occurs, a continuous infusion of furosemide (0.5 to 2 mg/kg/ hour) may be used with the goal to wean the infusion as rapidly as possible and to maintain a urine output of 2 mL/kg/hour or better. Other protocols use frequent (every 2 to 4 hours) bolus doses of 2 to 3 mg/kg/dose.
Dialysis is required in one third to one half of patients. Some studies have suggested that improvement in morbidity and mortality is associated with early dialysis. Typical indications for dialysis include oligo or anuria unresponsive to or unsuitable for furosemide use, marked fluid overload, acidosis unresponsive to bicarbonate, hyperkalemia unresponsive to treatment with potassium-binding resins, and severe hypertension unresponsive to pharmacologie management.
In most patients, peritoneal dialysis is the preferred dialysis modality as long as there is no obvious evidence of peritonitis or bowel perforation. The peritoneal dialysis catheter may be inserted quickly at the bedside or in the operating room. The presence of gastrointestinal dysfunction does not prevent adequate dialysis from occurring, and dialysis itself does not predispose the intestines to perforation. In situations where peritoneal dialysis is contraindicated or unavailable, hemodialysis is also satisfactory. During the recovery phase of HUS, it is common practice to provide supplementation with vitamin E and folie acid for a few weeks to months to replace losses incurred by the hyperactive bone marrow response.
In summary, meticulous and aggressive supportive care is all that is required in most patients with HUS. Many other therapies have been tried and have lost favor or have proved to be unhelpful. These include treatments with vitamin E, fresh frozen plasma, intravenous immunogtobulin, heparin, antibiotics, dipyridamole streptokinase, and aspirin. Plasmapheresis is helpfiil in patients with nonepidemic HUS, but has not been shown to be helpful in patients with epidemic HUS.
The end of the acute phase typically is heralded by a rise in platelet count and a reduction of packed cell transfusions. Anuria rarely lasts more than 2 weeks. When the child is oligo/anuric and dialysis is initiated, the Foley catheter should be removed as it serves as an avenue for ascending infection. Once urine output returns, it may be difficult to separate urine from watery stools, and it is common for urea and creatinine levels to fell before it is recognized that urine output is increasing.
Nutritional management is important in these patients. While their intestine is nonfunctioning, aggressive total parenteral nutrition is necessary. They may be intolerant of lipid preparations due to their high triglycéride levels. Glucose intolerance may be seen in patients with pancreatitis. Once bowel sounds, flatus, and stool return, initiation of oral feeding should be carried out slowly. The presence of a peritoneal dialysis catheter is not a contraindication to feeding. It is common to see hypertension develop or increase in severity during the recovery phase of HUS.
Most children with HUS are irritable and have an impaired level of consciousness out of keeping with their degree of uremia or illness. This is thought to be a manifestation of underlying processes occurring in the brain. Occasionally, children may present with a thrombotic or hemorrhagic stroke, which may be manifested by lateralizing neurological signs or evidence of a neurological catastrophe. Seizures are reported to occur in 20% to 40% of patients. The risk of seizures is increased in patients who have hyponatremia. Long-term sequelae reported in one study included cortical blindness, persistent seizures, and mental retardation in a total of 3% of their population.7 A recent study by the Canadian Pediatrie Kidney Disease Research Center indicates preliminary findings of a more subtle cognitive defect in a significant percentage of survivors of HUS. This remains uncorroborated. Central nervous system disease is the leading cause of death in HUS in most studies.
Cardiorespiratory complications of HUS include hypertension, which is typically seen in association with fluid overload and during the recovery phase. Hypertension may persist for months to years and is associated with chronic renal disease. Cardiac complications include sudden death from coronary artery thrombosis and the rare complication of pericardia! effusion and tamponade (Birk PE, Ogborn MR1 Lacson AG. 1994. Unpublished data).
The gastrointestinal system may experience widespread microangiopathy. Colitis may be severe with transmurai infarction and a perforative peritonitis during the acute phase. These severe lesions may heal by scarring and lead to secondary stenosis of the small and large bowel that may present as a subacute or acute bowel obstruction. Intussusception, gangrene, and rectal prolapse are reported. Pancreatic involvement may be more common than previously appreciated with presentations of typical pancreatitis or insulin-dependent diabetes. Hepatic involvement may manifest with hepatomegaly, increased transaminases, and rarely cholestasis.
Acute renal failure is the classic renal manifestation. Some patients, however, have extremely mild renal disease with more severe extrarenai manifestations. Patients have been reported where hemolysis was the only significant clinical problem, and other patients have been reported who presented with pseudomembranous colitis with minimal changes in the urinary sediment. There is a correlation between the severity of acute renal failure and the long-term outcome. In patients with the epidemic variety of HUS, anuria of >8 days is a poor prognostic feature. Robson et al8 found that the decline in glomerular filtration rate was related directly to the square root of the duration of anuria.
More than 90% of the children who survive an episode of HUS have a return of life-sustaining renal function. Persistence of proteinuria past 1 year after the illness is a poor prognostic sign and is strongly associated with long-term renal disease.9 Studies of patients followed for an average of 10 years show between 7% and 28% of the patients with a decreased glomerular filtration rate. Other studies suggest that the most common abnormality is a decrease in the functional renal reserve. Clearly, studies with a longer follow-up are necessary. After 1 year of follow-up, if the child has normal protein excretion, normal glomerular filtration rate, and normal blood pressure, the prognosis is believed to be optimistic.
Children who do sustain permanent renal failure are good candidates for kidney transplantation. Most studies have reported satisfactory outcomes, although one study showed an inordinately high rate of recurrent HUS in the transplanted kidneys.10 This may be secondary to the use of cyclosporin-based immunosuppression.
Although great strides have been made in determining the cause and pathogenesis of HUS and in management of the acute illness, much work remains to be done. Although we may be able to maintain the life of most patients through the acute phase of the illness until they can heal, we are no closer to a therapy that either prevents HUS or mitigates the acute phase of the disease process. Death and disability due to central nervous system involvement and multisystem disease is still common and resistant to our current management strategies. Current strategies undergoing investigation include use of antitoxins to neutralize the effect of VTEC and a study of an oral verotoxin-neutralizing agent that may prevent HUS. Public health education currently is the most important modality with which to decrease the incidence of HUS in North America. Families must be taught to ensure that hamburger is thoroughly cooked. Plates and utensils that have touched uncooked hamburger should not be used to serve cooked hamburger for fear of reinfection. In families or institutions where epidemic summer diarrhea is occurring, standards of cleanliness and handwashing must be enforced. General pediatricians must participate in these efforts by educating their patients about the risks of summer diarrhea and undercooked meat, and by keeping a high index of suspicion to allow for early recognition and management of those patients with HUS.
The authors thank Janice Martin for typing the manuscript and Dr Alice Patton for allowing them to use the Figure.
1. Gasser VC, Gautier E, Steek A. ei al. Hamolytisch-urämische syndromei bilaterale nierenrindennekrosen bei akuten erworbenen hämolytischen anämien. Sfhurö Med Wochensdn. 1955;38:905-909.
2. Scewatt CL, Tina LU. Hemolytic uremie syndrome. MKW Rn. 1993:14:218-225.
3. O'Regan S, Rousseau E. Hemolytic uremie syndrome: urate nephropathy superimposed un an acute gbmerulopathy! An hypothesis. CIm Nephro/. 1988; 3Or 207-21 0.
4. Martin DL. MacDonald KL, While KE. Soler JT, Osterholrn MT. The epidemiology and clinical aspects of the hemolytic uremie syndrome in Minnesota. N EngiJMed. 1990;323:1 161-1 16?.
5. Rowe PC, Orbine E, Wells GA. el al. Epidemiology of hemolyric-uremic syndrome in Canadian children from 1906 to 1988. J Pediaa. 1991;! 19t218-224.
6. Kaplan BS, Chesney RW1 Drummond KN. Hemolytic uremie syndrome in families. NEngUMid. 1975;292:1090-1093.
7. Loirat C. Baudouin V, Sonsinu, E. Marian i-Kurd] ian E Eìion ]. Hemolytic -uremie syndrome in the child. Adwmces in NejAroIcg}. 1993;22:I41-I68.
8. Robson LM, Leung AKC, Kaplan BS. Hemolytic uremie syndrome. CUIT Probi Pediatr. 1993:23:16-33.
9. Milfbrd DV. White RHR. Taylor CM. Prognostic significance of proteinuria 1 year after onset of diarrhea-associated hemolytic uremie syndrome J Pediaa. 1991;! 18:191194.
10. Hebert D, Sibley RK, Mauei SM. Recurrence of hemolytic uremie syndrome in renal transplant recipients. Kidney Ira. 1986;30:S51-S58.