ysis of 56 episodes of transient cerebral hypoxia. Ann Neural 1994;36:233-237. 3. Rossen R, Kabat H, Anderson JP. Acute arrest of cerebral circulation in man. Arch Neurol Psychiat 1943;50:510-528. 4. Lin JTJ, Ziegler DK, Lai CW, Bayer W. Convulsive syncope in blood donors. Ann Neural 1982;11:525-528. 5. Stephenson JBP. Fits and faints. London: Mac Keith Press, 1990:41-58. 6. Duvoisin RC. The Valsalva maneuver in the study of syncope. Electroencephalogr Clin Neurophysiol 1961;13:622-626. 7. Halmagyi GM, Rudge P, Gresty MA, Sanders MD. Downbeating nystagmus: a review of 62 cases. Arch Neural 1983;40: 777-784. 8. Baloh RW, Spooner JW. Downbeat nystagmus: a type of central vestibular nystagmus. Neurology 1981;31:304-310. 9. DeJong JMBV, Cohen B, Matsuo V, Uemura T. Midsagittal pontomedullary brain stem section: effects on ocular adduction and nystagmus. Exp Neural 1980;68:420-442. 10. Zee DS, Yamazaki A, Butler PH, Giicer G. Effects of ablation of flocculus and paraflocculus on eye movements in primate. J Neurophysiol 1981;46:878-899. 11. Siesjo BK. Brain energy metabolism. Chichester: John Wiley & Sons 1978:133-138. 12. Cervos-Navarro J , Diemer NH. Selective vulnerability in brain hypoxia. Crit Rev Neurobiol 1991;6:149-182. 13. Thurston SE, Leigh R, Abel LA, Dell'Osso LF. Hyperactive vestibulo-ocular reflex in cerebellar degeneration: pathogenesis and treatment. Neurology 1987;37:53-57. 14. Leigh RJ, Zee DS. The neurology of eye movements. Philadelphia: F.A. Davis, 1991:457-458. Celiac disease with cerebral calcium and silica deposits: X-ray spectroscopic findings, an autopsy study P. Toti, MD; P. Balestri, MD; M. Cano; P. Galluzzi, MD; T. Megha, MD; M.A. Farnetani, MD; M.L.D. Palmeri, MD; M. Vascotto, MD; C. Venturi, MD; and A. Fois, MD Article abstract-An increased incidence of seizures and cerebral calcifications, usually bilateral and located in the occipital cortex, has been reported in celiac patients. The histology of cerebral lesions is not well defined, and their pathogenesis is only speculative. We report the autopsy results of a patient with celiac disease, seizures, and cerebral calcifications who died following a cerebral hemorrhage caused by Fisher-Evans syndrome. Calcifications were restricted to the cortical gray matter and composed of aggregates of small calcified spicules. Calcium deposition was present as psammoma-like bodies, along small vessels, and within neurons. X-ray spectroscopy of the calcified areas revealed that calcium (43%)and silica (57%) were present in the lesions. High silica content was also found in the cerebral hemorrhagic fluid. Silica toxicity has to be considered in regard to the pathogenesis of the cerebral lesions and of the seizures. NEUROLOGY 1996;46:1088-1092 The existence of a syndrome characterized by celiac disease (CD), partial seizures, and cerebral calcifications (CC) was first proposed in 1988,' although a single case had been previously described,2 as well as an increase in the prevalence of epilepsy among patients with CD.3 More recently, several series and case reports have been described, mainly by Italian author~.~-'O The combination of partial epilepsy and CC, in which CD had not been actively sought, was previously labeled atypical Sturge-Weber syndrome without cutaneous manifestations.'J* However, patients with epilepsy, CC, and CD differ from those with Sturge-Weber syndrome as follows: (a) absence of choroid plexus enlargement or cerebral atrophy detected via CT and MRI; (b) absence of abnormal deep cerebral veins detected via enhanced CT; ( c ) presence of bilateral rather than unilateral calcification in most cases; and (d) absence of neurologic A retrospective study on adult patients diagnosed with CD in childhood has shown that epilepsy and CC only occur in patients following a n unrestricted diet.I2 A gluten-free diet started immediately after the onset of epilepsy and early in childhood seems to be effective in controlling ~ e i z u r e s .However, ~ the mechanisms of brain damage are unclear. Folic acid deficiency has been suggested, but this is one of the most common biochemical abnormalities in uncomplicated CD and, therefore, seems unlikely." Since CD may be associated with different autoimmune d i s o r d e r ~ , ~the ~ -possibility '~ that CC depend on autoimmune or immuno-complex-related endothelial inflammation has been p r ~ p o s e dHowever, .~ this theory does not explain the prevalent location (cortical- From the Departments of Pathology (Drs. Toti, Megha, and Palmeri) and Pediatrics (Drs. Balestri, Farnetani, Vascotto, and Fois), University of Siena, Siena. Italy; the Department of Pathology (M. Cano), University of Nebraska, Omaha, NE; and the Division of Neuroradiology (Drs. Galluzzi and Venturi), Siena. Italy. Supported by M.U.R.S.T. 60%. Received April 10, 1995. Accepted in final form July 26, 1995. Address correspondence and reprint requests to Dr. Paolo Toti, Istituto di Anatomia Patologica, Via delle Scotte, 53100 Siena, Italy. 1088 C!opyright 0 1996 by the American Academy of Neurology Figure 1. A n emergency brain CT reveals irregular calcifications in both occipital lobes and a large intraparenchyma1 hemorrhage in the left occipital lobe. Another smaller calcification is present in the right frontal lobe. subcortical, and mainly occipital) of the calcifications. In this paper, we describe the autopsy results of an adolescent patient affected by the syndrome of CD, seizures, and CC. On the basis of the interesting and unexpected results we found, we discuss the pathogenesis of the CC and seizures. Case report. The patient, a female, presented two episodes of complex partial seizures and one generalized febrile seizure between the ages of 4 and 7 years. A CT revealed the presence of bilateral occipital and right frontal calcifications, which were unchanged in a subsequent examination performed 4 years later (figure 1). The patient was treated with phenobarbital and had no more seizures. When examined in the Department of Pediatrics at the age of 12, an EEG revealed spike and wave discharges in the occipitoparietal region. Height and weight were in the 10th percentile, hemoglobin was 7.2 g/dL, serum folic acid was reduced, and antigliadin and antiendomysial antibody levels were elevated. A jejunal biopsy confirmed the diagnosis of CD, and the patient was treated with a gluten-free diet. Hemoglobin and folic acid levels normalized, as well as antigliadin antibodies; antiendomysial antibodies remained elevated. When the patient was 15, she developed petechiae, bleeding, and hemolytic anemia. Fisher-Evans syndrome (immune hemolytic anemia with thrombocytopenia) was diagnosed. Despite treatment with intravenous gamma globulin, corticosteroids, and immunosuppressive drugs, the platelet count remained very low. The patient suddenly developed headache and vomiting and went into coma. A CT showed a large, inoperable cerebral hemorrhage (figure l),and the patient died 4 days later. Autopsy results. A complete autopsy was performed about 12 hours after death. The cause of death was severe cerebral edema, with herniation of the cerebellar tonsils. Blood was present in the meningeal space caused by hemorrhage arising from the left occipital lobe. Ancillary findings included hemorrhagic gastritis and pyelitis. The bone marrow was extremely hypercellular and the red pulp of Figure 2. Hematoxylin-eosin. Small and scattered cortical calcifications. The meningeal vessels appear completely normal and free of calcifications. Bar = 300 Fm. the spleen hyperplastic. Histologic examination of the kidneys showed the presence of free hemoglobin and casts of epithelial cells in the tubules. Moreover, calcium stain revealed the presence of small and sparse tubular calcifications. In the brain there were three distinct foci of dense calcification (right frontal lobe and left and right occipital lobes) composed of aggregates of small calcified spicules. These foci appeared to originate at the interface between the white and gray matter and spread to the surface of the cortex. No large vessels, either arteries or veins, were directly affected by the calcifying process. There were no other gross abnormalities identified in the brain, choroid plexus, or meninges; specifically, no foci of pial angiomatosis were identified. In each instance, the calcifications were restricted to the cortical gray matter. Microscopically, at the boundaries between the calcified areas and the normal brain, it was possible to distinguish three main types of calcium deposition (figures 2, 3, and 4). The first was characterized by sharply demarcated, round, 30-40 pm psammoma-like bodies, without any identifiable relationship with cells, Figure 3. Hematoxylin-eosin. From left to right along the gyrus, one can appreciate the progressive increase in the number of calcified cortical foci and corresponding decrease in the number of neurons. Bar = 240 pm. April 1996 NEUROLOGY 46 1089 Figure 4. (a and 6). Hematoxylineosin. The three main types of calcium deposits are easily identifiable i n the cortical gray matter: calcium deposits inside a neuron (a, 7 i ; sharply demarcated, round, 3040-pm psammoma-like bodies (b, 7 I ; small, granular structures in the perivascular space along venules and capillaries (b, t 1. Bars = 50 pm. vessels, or other brain structures. The second featured small, granular calcium deposition along small vessels, mainly venules and capillaries. Endothelial cells were spared by the calcification process, which clearly starts from the outer part of the vessels (basal lamina or perivascular space, or both). Sometimes it was possible to identify one or two layers of deposition within such sites. Thirdly, focal, scanty areas wherein calcium appeared to be deposited in neurons could be identified. Toward the central part of all the calcified foci, the calcifications were much larger and assumed an irregular shape, probably due to fusion of smaller areas of calcification (figure 5). These areas were devoid of neurons, with a mild degree of reactive gliosis. In all foci of calcification, histochemical stains confirmed the presence of calcium and also established that lesser amounts of iron were admixed with the calcium. Careful histologic examination of numerous random sections of the remaining brain (cerebral cortex, cerebral white matter, thalamus, brainstem, cerebellum, choroid plexus) did not identify other calcifications or any other abnormalities; specifically, nowhere were inflammatory cells identified. Calcium, silica, iron, aluminum, phosphorus, and nickel contents in the brain, kidneys, and blood were evaluated by energy-dispersive x-ray spectroscopy (Kevex-7000) with a Scanning Electron Microscopy Philips 515 at 20 kV. Re- Figure 5. Hematoxylin-eosin. Towards the central part, calcified foci fuse and assume an irregular shape. I n this picture, a small artery ( and a venule ( t t I are also visible; calcium deposition clearly starts from the outer part of the vessel wall (perivascular spaces andlor basal lamina) and spares endothelial cells. Bar = 100 pm. 1090 NEUROLOGY 46 April 1996 sults were expressed in relative atomic percentage. The brain deposits were composed of calcium (43% ? 3%)and silica (57% 2 3%). These substances were undetectable in the brain regions devoid of histologic lesions. Iron was present in both lesional and normal areas, in comparable amounts. Phosphorus, nickel, and aluminum were undetectable, in both calcified and normal areas. X-ray analysis of the kidney disclosed the presence of high levels of silica and aluminum, and lesser amounts of calcium. High levels of silica and aluminum were also found in the blood collected in the subarachnoid space during autopsy. Discussion. Conditions in which calcified foci may be found diffusely in the brain, often affecting the basal ganglia, are considerably frequentI7 (encephalitis caused by necrotizing agents or HIV, Leigh syndrome and other encephalopathies related to mitochondrial diseases, disorders of calcium metabolism ). However, in our case, these conditions were ruled out by the patient’s history, clinical and laboratory evaluation, a n d pathologic examination. On t h e other hand, the presence of cortical brain calcification is a much rarer pathologic event. In SturgeWeber syndrome, intracranial calcifications affect hamartomatous vessels in the hypervascularized meninges, the outer cortical layers, and the media and adventitia of non-hamartomatous cerebral mediumsized arteries and veins.1RMoreover, the deposition of lime salts seems to be preceded by the appearance of a layer of hyalinized connective tissue in the wall of malformed vessels and by diffuse atrophy, neuronal cell loss, and gliosis in the underlying gyri.’