Pituitary Hemorrhage in Hemolytic Uremic Syndrome Toshiro Nagai, MD*t, Osamu Hasegawa, MD**, Atsutoshi Tuji, MD**, Yasunori Kamiyama, MD**, Masataka Honda, MD**, and Hiroshi Ito, MD** A 5-year-old girl with hemolytic uremic syndrome is reported with pituitary hemorrhage in the acute phase. Seizures, hemiplegia, and transient blindness were observed. These symptoms cannot be attributed to the pituitary hemorrhage; however, pituitary hemorrhage should be added to the list of neurologic lesions that can occur in hemolytic uremic syndrome. Nagai T, Hasegawa O, Tuji A, Kamiyama Y, Honda M, Ito H. Pituitary hemorrhage in hemolytic uremic syndrome. Pediatr Neurol 1992;8:75-6. prodromal gastroenterocolitis manifested by watery-loose and bloody stools. On admission, she was alert with pale skin but without petechiae or ecchymosis. Blood pressure was 110/70 mm Hg. The platelet count was 71,000/mm 3 and fragmented erythrocytes were observed on peripheral smear. Hemoglobin was 10.7 gm/dl, hematocrit 30%, BUN 102 mg/dl, serum creatinine 4.5 mg/dl, serum sodium 139 mEq/L, potassium 4.7 mEq/L, chloride 103 mEq/L, and calcium 7.3 mg/dl. Blood gas analysis disclosed pH 7.32 and bicarbonate 15 mEq/L. IgM antibody assessed by the Widal method against somatic antigen for E. coli strain 0-157:H7 was increased on admission (x 320) and decreased 1 month later (x 80). Antiverotoxin antibody concentration determined by the ELISA method revealed high antibody levels of both antiverotoxim and antiverotoxin2. Stool culture confirmed the presence of E. coli 0-157:H7. Three hours after admission, she developed generalized tonic seizures with poor response to antiepileptic drugs. Funduscopic examination revealed no hemorrhage and no cotton-wool spots on the retina. The initial cranial CT on admission disclosed a high-density lesion in the pituitary gland with no evidence of large-vessel infarctions, no hemorrhage in the centrum semiovale, and no cerebral edema (Fig 1). There were no abnormalities of the geniculate bodies, calcarine cortex, or optic radiations. The child's hospital course was characterized by prolonged oliguria lasting about 1 month, blindness for longer than 2 months, and spastic hemiplegia for 2 months. Continuous, ambulatory, peritoneal dialysis therapy was successfully performed without hypertension or electrolyte imbalance. During the period of blindness, pupillary responses to light were elicited but were slow and incomplete. Three months after admission, she had no neurologic deficits and her renal function was normal. Serial CT scans obtained weekly in the acute phase revealed a persistent high-density lesion in the pituitary. Magnetic resonance imaging (MRI) performed 1 month after admission demonstrated high intensity in the Tl image and low intensity in the T2 image (Fig 2), suggesting hemosiderin deposit in the pituitary. Two months after admission, endocrinologic studies were performed. Luteinizing hormone and thyrotropin-releasing hormone loading tests Introduction Neurologic complications of hemolytic uremic syndrome (HUS) include generalized or focal seizures, alterations of consciousness, hemiparesis, and coma [1,2]. Reported computed tomographic (CT) findings range from normal to diffuse cerebral edema, large-vessel infarctions, diffuse, multiple small-vessel infarctions, or hemorrhage [3-6]. The causes of HUS are considered to be multifactorial but, recently, verotoxin-producing Escherichia coli infection was proposed as one of the causal factors [7]. We report a girl with HUS probably associated with E. coli 0-157:H7 infection. She demonstrated severe neurologic complications with pituitary hemorrhage on CT in the acute phase of HUS. Case Report An epidemic of HUS occurred at a kindergarten in the eastern Tokyo Metropolitan area in October, 1990. During this period we examined a 5-year-old girl who was admitted to our hospital with anuria after From the *Division of Pediatrics; Tokyo Metropolitan Kiyose Children's Hospital; Tokyo, Japan; tDivision of Pediatrics; Urawa City Hospital; Urawa, Japan; ~Division of Pediatrics; Keio University School of Medicine; Tokyo, Japan. Figure 1. On admission, unenhanced axial CT (5 mm thick section) demonstrates high-density lesion at the pituitary gland. This high-density lesion remained unchanged for more than 1 month. Communications should be addressed to: Dr. Nagai; Division of Pediatrics; Tokyo Metropolitan Kiyose Children's Hospital; 1-3-1 Kiyose Umezono; Tokyo, Japan 204. Received July 30, 1991; accepted September 16, 1991. Nagai et al: Pituitary Hemorrhage 75 press the optic chiasm and cause a visual disturbance. Wc could not demonstrate such compression ~f the optic chiasm on CT or MRI; however, we speculate that her transient blindness may be related to pituitary hemorrhage. Since the initial description by Karmali et ~tl. [71, the causal relationship between HUS and antiverotoxinproducing E. coli, especially E. coli 0-157:tf7, infection has been suspected [9,12-14]. We are uncertain whether there is any correlation between antiverotoxin-producing E. coli infection and pituitary hemorrhage. Thesc two processes may have occurred in this patient coincidentally; however, because verotoxin causes vascular toxicity in the kidney and neurotoxicity in the CNS at the same time [ 15 ], we speculate that there may be some correlation between antiverotoxin-producing E. c o t i infection and pituitary hemorrhage pathologically. Figure 2. Sagittal MR1 obtained 1 month after admission discloses a hypointense lesion at the pituitary, gland by T2 image. disclosed a normal response to luteinizinghormone,follicle-stimulating hormone, and thyroid-stimulatinghormone. Arginine (0.5 gm/kg) and insulin (0.1 IU/kg) loading tests revealed an inadequate response of growth hormone (3.0 ~ 9.3" 5.1 ~ 6.8 ng/ml, respectively). The adrenocorticotropic hormone (ACTH) and cortisol response to insulin loading also documented an unsatisfactory response: ACTH (31 ~ 42 pg/ml) and cortisol (7.96 -+ 17.27 ~ag/100ml). Discussion The pathogenesis of neurologic complications of HUS has not been elucidated. Many metabolic disturbances encountered in this s y n d r o m e - hypertension, volume overload, hyponatremia, acid-base changes, hypocalcemia, intravascular coagulation, and microthrombi in the central nervous system ( C N S ) - may be considered to be causal factors. Intravascular coagulation and microthrombi in the kidney are the hallmark of this syndrome; however, these changes have not been consistently present in the CNS in autopsied cases [1,8]. Our patient had a renal disorder; however, because her metabolic disturbance was relatively mild (normal sodium and calcium concentration and normal blood pressure), her neurologic complications cannot be attributed to the metabolic disturbances. Radiologic examination revealed no infarction, but pituitary hemorrhage was observed. Cranial CT changes reported in this syndrome are various [3-6]; pituitary hemorrhage, however, has not been reported. Clinically, she demonstrated inadequate response of ACTH and growth hormone to anterior pituitary gland stimulation tests. She now has no apparent sign of ACTH or growth hormone deficiency, but is being carefully monitored. Ocular involvement, especially cortical blindness, has been described occasionally in HUS [2,9,10] and retinal thrombi and opacity also have been reported [ 11 ]. The blindness usually is transient and no permanent blindness has been described in the literature. Our patient had no retinal lesions or detectable lesions of the optic tract or cortex on CT and MRI. Pituitary hemorrhage can corn- 76 PEDIATRIC NEUROLOGY Vol.8 No. 1 We are indebted to Tae Takeda (National Children's Research Center) for performingthe verotoxin study and would like to thank Dr. Mitsuru Osano, Keio University School of Medicine, for reviewing the manuscript. References [1] RooneyJC, AndersonMR, Hopkins IJ, Clinical and pathological aspects of central nervous system involvement in the hemolyticuremic syndrome.Aust Paediatr J 1971;7:28-33. [2] Sheth KL, Swick HM, Haworth N. Neurologic involvementin hemolytic-uremicsyndrome.Ann Neurol I986;19:90-3. [3] Mendelsohn DB, Hertzanu Y, Chaitowitz B, Cartwright JD. Cranial CT in the hemolytic uremic syndrome. J Neurol Neurosurg Psychiatry 1984~47:876-8. [4] DiMario FJ, Bronte-Stewart H, Sherbotie J, Turner ME. Lacunar infarctionof the basal ganglia as a complicationof hemolyticuremic syndrome. Clin Pediatr 1987;26:586-90. [5] Trevathan E, Dooling EC. Large thrombotic strokes in hemolytic-uremicsyndrome. J Pediatr 1987;I 11:863-6. [6] Hahn JS, Havens PL, Higgins JL et al. Neurologicalcomplications of hemolytic-uremicsyndrome. J Child Neurol 1989;4:108-i 3. [7] Karmali MA, Steele BT, Petric M, Lira C. Sporadic cases of hemolytic-uremic syndrome associate with faecal cytotoxin and cytotoxin-producingEscherichia coli in stools. Lancet 1983;19:619-20. [8] Gianantonio CA, Vitacco M, MendilaharzuF. Gallo GE, Sojo ET. The hemolytic-uremicsyndrome.Nephron 1973;11:174-92. [9] Martin DL, Kristine MPH, MacDonald L, White K, Soler JT, Osterholm MT. The epidemiologyand clinical aspects of the hemolytic uremic syndromein Minnesota.N Engl J Med 1990;323:1161-7. [10l Kahn SI, Tolkan SR, Kothari O, Garella S. Spontaneous recovery of the hemolytic uremic syndrome with prolonged renal and neurologicalmanifestations.Nephron 1982;32:188-91. [11] Siegler RL, Brewer ED, Swartz M. Ocular involvement in hemolytic-uremicsyndrome. J Pediatr t988;112:594-7. [12] Gransden WR, Damm MA, AndersonJD, Carter JE, Lior H. Further evidence associatinghemolyticuremic syndromewith infection by verotoxin-producingEscherichia coil 0157:H7. J Infect Dis 1986; 154:522-4. [13] Carter AO, BorczykAA, CarlsonJAK, KrishnanC, Kom DA, Lior H. A severe outbreak of Escherichia coil 0157:H7-associatedhemorrhagic colitis in a nursing home. N Engl J Med 1987;317:1496-500. [14] Milford DV, Taylor CM, Guttridge B, Hall SM, Rowe B, Kleanthous H. Hemolyticuremic syndromein the British Isles 1985-8: Association with verotoxinproducingEscherichia eoli, Part 1: Clinical and epidemiologicalaspects. Arch Dis Child 1990;65:716-2I. [15] KeuschGT, Grady GF, Mata LJ, McClever J. The pathogenesis of Shigella toxin. J Clin Invest 1972;51:1212-8.