Information fusion for diagnosis coding support.

All patient-related medical information during a hospital stay in France, has to be collected and coded in the compilation of medical units discharge documents, according to a standardized approach. The process of describing a patient disease in terms of appropriate diagnostic codes is nevertheless, a non-intuitive operation for the physician. As a consequence, coding errors, inaccuracies and missing data are frequent, leading to potentially severe economical upshots. A coding support system developed to improve medical coding results, integrates three information processing methodologies, using the outputs from various Hospital Information System applications. Each methodology generates partial heterogeneous information, with considerable semantic variety. In order to properly synthesize these outputs, information fusion is required to produce enriched contextualized information, presented to the physician as an ordered list of suggested codes. This paper explores two information fusion approaches: voting system and possibilistic. Both methods are tested on a database of 1,000 discharge summaries, to show the interest of information fusion in this context. Results show that fusion methods perform better in most of the cases than partial information extraction methods.

ANTEROCOD: actuarial survival curves applied to medical coding support for chronic diseases.

For the practitioner, choosing diagnosis codes is a non-intuitive operation. Mistakes are frequent, causing severe consequences on healthcare performance evaluation and funding. French physicians have to assign a code to all their activities and are frequently prone to these errors. Given that most of the time and particularly for chronic diseases indexed information is already available, we propose a tool named AnterOcod, in order to support the medical coding task. It suggests the list of most relevant plausible codes, predicted from the patient's earlier hospital stays, according to a set of previously utilized diagnosis codes. Our method applies the estimation of code reappearance rates, based on an equivalent approach to actuarial survival curves. Around 33% of the expected correct diagnosis codes were retrieved in this manner, after evaluating 998 discharge abstracts, significantly improving the coding task.

Expression of interleukin-1 genes and interleukin-1 receptors in the mouse brain after hippocampal injury.

In order to evaluate the role of IL-1 production in post-traumatic brain, transcripts for IL-1 (alpha, beta, RA) have been quantified following RT-PCR, in hippocampus and cortex after injury of either hippocampus (Hip) or striatum (Stri). Moreover, 125I IL-1alpha binding sites have been directly quantified using binding experiments on brain sections and quantitative autoradiography. Under basal conditions, levels of PCR products were very low. On day 1, IL-1RA transcripts only were strongly increased in the hippocampus after Hip-lesions and in cortex after Stri lesion. Transcripts were back to control values on day 7 post-lesion. IL-1 receptor densities in the hippocampus (dentate gyrus) were decreased at day 1 around the site of the lesion (but not on the contralateral side) and were back to controls on day 7 indicating a transient and local IL-1 production in the surroundings of the lesion. No changes were found following Stri lesion. This study provides further evidence of the role of the IL-1 molecules family, notably IL-1RA, in the brain reaction to trauma.

Specific pattern of nitric oxide synthase expression in glial cells after hippocampal injury.

In the central nervous system (CNS), nitric oxide (NO) is thought to be involved in a variety of functions including synaptic plasticity, long term potentiation, and neurotoxicity. The aim of the present study was to investigate the expression of nitric oxide synthase (NOS) in the mouse CNS, following surgical injury to the hippocampus. NOS expression was assessed by histochemical detection of nicotinamide adenine dinucleotide phosphate diaphorase (NADPH-diaphorase) activity and immunohistochemistry of the inducible NOS (iNOS). Two days after injury to the CA1 hippocampal field, NADPH-diaphorase activity was detected in pyramidal and granular neurons and also in glial cells in the hippocampus, in contrast to the non-injured one where NADPH-diaphorase staining was observed only in a few interneurons. NADPH-diaphorase histochemistry combined with immunolabelling for GFAP and F4/80 demonstrated that these glial cells were astrocytes and microglia. This pattern of NOS expression is induced specifically after a hippocampal injury since lesion to the prefrontal or cerebellar cortex leads to NOS activity only in monocytes/macrophages like cells. Despite the large expression of NOS detected by NADPH-diaphorase histochemistry after lesioning the hippocampus, immunostaining for iNOS was confined to microglia. The fact that induction of high levels of NOS activity are detected in glial cells after a lesion to the hippocampus could be accounted for by the sensitivity of this structure to a high release of glutamate.

Expression of TNF alpha in central neurons of Lewis rat spinal cord after EAE induction.

Experimental allergic encephalomyelitis (EAE), an animal model for multiple sclerosis (MS), is a demyelinating autoimmune disease of the central nervous system (CNS). The proinflammatory cytokine TNF alpha, as an endogenous mediator of inflammation, plays an important role in the pathogenesis of EAE disease. In this study, we demonstrate the presence of TNF alpha in spinal cord of Lewis rats, during the critical phase of EAE. The expression of TNF alpha is observed mainly in the gray matter of thoracic and lumbar levels of the spinal cord, in the motoneurons and interneurons of the ventral horn. Surprisingly, one month after recovery, we still found an intense TNF alpha-neuronal expression, including in the cervical region, and this positivity lasted up to 40 days after recovery, with, however, a decrease in its intensity. These results suggest that central neurons respond directly to massive infiltration of lymphocytes and macrophages after the breakdown of the blood-brain barrier (BBB), by producing TNF alpha cytokine. In addition, neuronal-TNF alpha detection in the recovery stage of EAE may suggest a role other than its classical action in promoting inflammatory processes.

Widespread neuronal expression of c-Fos throughout the brain and local expression in glia following a hippocampal injury.

Fos oncoprotein is an immediate early gene product and a marker of cell activation following a variety of insults. We have previously shown that a mechanical lesion to the hippocampus of adult mice induces a neuronal expression of the cytokines interleukin-1alpha (IL-1alpha) and tumor necrosis factor-alpha (TNF alpha) whereas a lesion to the striatum does not. The role of these inflammatory cytokines in the pathophysiology of central neurons is still unclear. The present work was undertaken to study a possible correlation between the central expression patterns of c-Fos on the one hand and IL-1alpha and TNF alpha on the other hand. We show that Fos is expressed in a majority of brain neurons after a unilateral lesion to the hippocampus whereas it is confined to the site of injury when applied to the striatum, as previously described for the expression of the cytokines.

Identification and topography of neuronal cell populations expressing TNF alpha and IL-1 alpha in response to hippocampal lesion.

In previous studies, we have shown that a traumatic lesion to the hippocampus of adult mice induces the transitory expression of TNF alpha and IL-1 alpha by neurons of different brain areas and also by glial cells at the site of injury. The aim of the present study was to establish whether the expression of TNF alpha and IL-1 alpha is restricted to defined subpopulations, or else is common to most of the central neuronal populations. Using polyclonal anti-GAD 67, anti-TH and monoclonal anti-ChAT, and anti-5-HT antibodies in a double-labeling immunohistochemical procedure in combination with murine anti-TNF alpha and anti-IL-1 alpha polyclonal antibodies, we show that most GABAergic, catecholaminergic, and serotoninergic neurons, and a subgroup of the cholinergic neurons, express these cytokines. Although not immunohistochemically characterized, neurons in some glutamatergic structures such as the hippocampus and the prefrontal cortex also express these cytokines. Thus, we conclude that the capacity of central neurons to express cytokines like TNF alpha and IL-1 alpha in reaction to a brain injury is not restricted to peculiar neuronal subtypes, but could include most of the neuronal populations of the brain.

Differential oligodendroglial expression of the tumor necrosis factor receptors in vivo and in vitro.

The cytokine tumor necrosis factor-alpha (TNF alpha) has been proposed to play a key role in the degenerative processes observed in demyelinating diseases such as multiple sclerosis (MS). In the immune system the cellular responses to TNF are mediated by two different receptors: TNF-RI, which is involved in cell death, and TNF-RII, which has been shown to mediate cell proliferation. We investigated the oligodendroglial expression of TNF-RI and -RII. In vivo, in normal adult rodent brain, oligodendrocytes express TNF-RII but not TNF-RI. However, after 3 days in culture, both types of receptors were expressed by mature oligodendrocytes, purified from 4-week-old rats, suggesting that expression of TNF-RI was induced by either the isolation process or the culture conditions. This inducibility of TNF-RI may explain the differences in oligodendrocyte cell death reported in various experimental conditions and in the pathology of MS lesions.

Oligoclonal beta-galactoside-specific gamma immunoglobulins allow the immunocytochemical detection of cellular antigenic determinants expressed in mouse brain after surgical injury.

Histological brain sections were probed with human oligoclonal lectin-like IgGs (L-IgG) purified from normal serum. In intact brain, antigenic determinants for these IgG were restricted to some blood vessel endothelial cells. By contrast, during the inflammatory reaction following a surgical injury, these determinants were detected at the cell surface of different cell types, within and near the lesion site. The cells reacting with L-IgG consisted of endothelial cell, mature astrocytes, activated microglial and ependymal cells.

Distribution of [3H]clonidine binding sites in the brain of the convulsive mutant quaking mouse: a radioautographic analysis.

The radioautographic analysis of [3H]clonidine binding was performed on brain slices from the convulsive mutant mice quaking and their controls of the same strain. In the quaking mice significant increases were observed mostly in the brainstem and the cerebellum, but also in a few regions of the forebrain, such as the lateral and medial thalamic nuclei, the medial geniculate nucleus, the amygdala and the hypothalamus. Other regions, such as the cerebral cortex and the hippocampus, which are classically involved in various models of epilepsy, but not in the quaking mice, did not show any modification of [3H]clonidine binding. A high degree of correlation was found between the structures with an increased density of alpha 2-adrenoceptor binding sites and the distribution of regions from which seizures can be elicited by intracerebral electrical stimulation in head-restrained quaking mice. This comparison emphasizes the role of noradrenaline acting at the level of alpha 2-adrenoceptors in the epileptic syndrome of the quaking mutants.

Postnatal evolution of the gamma-aminobutyric acid/benzodiazepine receptor complex in a model of inherited epilepsy: the quaking mouse.

Binding assays of [3H]muscimol and [3H]-flunitrazepam have been performed on brain homogenates of brainstem, cerebellum, and forebrain of genetically epileptic quaking (qk) mutant mice 20, 40, 70, and 90 days old and their corresponding controls of the same strain (C57BL/6J). The endogenous gamma-aminobutyric acid (GABA) content has been determined in various brain regions of 70-day-old qk and control mice. Finally, the behavioral effects of diazepam, of the mixed GABAA/GABAB receptor agonist progabide, and of the selective GABAB receptor agonist baclofen have been assessed in adult qk mutants. Our results strongly suggest a lack of involvement of GABAergic neurotransmission in the inherited epilepsy of the qk mutant mouse.

Supernumerary locus coeruleus neurons as a determinant of inherited epilepsy in the convulsive mutant mouse quaking.

In quaking mice (a genetic model of epilepsy with an increased number of noradrenergic neurons) bilateral electrolytic coagulation of locus coeruleus (LC) in adult mice inhibited the convulsions elicited by somatic stimulations while neonatal 6-hydroxydopamine (6-OHDA) treatment remained ineffective upon the convulsions. Biochemical effects of the two treatments differed only in the brainstem where electrolytic lesion decreased while 6-OHDA treatment increased noradrenaline (NA) and 3-methoxy 4-hydroxyphenylethyleneglycol (MHPG) levels. Our results suggest that supernumerary LC neurons mediate the convulsions of the mutants through an action presumably restricted to the brainstem.

Modifications of sphingomyelin and phosphatidylcholine metabolism by tricyclic antidepressants and phenothiazines.

Phenothiazines and tricyclic antidepressants, when added to culture medium, gave rise in several types of cells (C6 rat glioma cells and human fibroblasts), to a decrease in lysosomal sphingomyelinase activity. The effect of chlorpromazine and desipramine was dose dependent, and was observed after 3 hours of incubation with the drugs at concentrations ranging between 1 and 10 microM. In C6 glioma cell cultures, the decrease in sphingomyelinase activity was related to the clinical effectiveness of phenothiazines, tricyclic antidepressants and derivatives. Incorporation of (choline-14C) sphingomyelin showed that the metabolic pathway implying the synthesis of phosphatidylcholine from the hydrolysis of sphingomyelin and/or transfer of phosphorylcholine to phosphatidylcholine was also partially reduced.

Altered postnatal ontogeny of alpha 1- and alpha 2-adrenoceptor binding sites in the brain of a convulsive mutant mouse (quaking).

Binding assays of [3H]dihydroalprenolol ([3H]DHA), [3H]prazosin and [3H]clonidine have been performed on whole brain (minus cerebellum) homogenates of the convulsive mutant mice quaking (qk) and the controls of the same strain (C57BL/6J:B6). In 70-day-old mutants (which fully exhibit the qk convulsive phenotype), the binding of [3H]DHA to beta-adrenoceptor binding sites was not different from the controls, whereas the binding capacities of [3H]prazosin and [3H]clonidine to alpha 1-and alpha 2-adrenoceptor sites, respectively, were greatly enhanced. The biphasic ontogenic pattern of alpha 2-adrenoceptors had a greater amplitude in the brain of 30- to 90-day-old mutants than in the corresponding B6 controls. In mutants younger than 30 days or older than 90 days, the number of alpha 2-adrenoceptor sites was not modified. The number of alpha 1-adrenoceptor binding sites was increased in the brain of the mutants, only in animals older than 70 days. In younger mice, the postnatal modulation of alpha 1-adrenoceptor sites was identical to the controls. Regional studies were performed in 70-day-old mice. [3H]clonidine binding was increased in the brainstem of the mutants, and to a lesser extent in the cerebral cortex, while it was slightly diminished in the hypothalamic area. [3H]prazosin binding was also increased in the brainstem of the mutants, and decreased in the olfactory bulbs. Our results suggest that the convulsions of the qk mutants are selectively associated with modifications of alpha- and not beta-adrenoceptor binding sites.(ABSTRACT TRUNCATED AT 250 WORDS)

Increased number of locus ceruleus noradrenergic neurons in the convulsive mutant quaking mouse.

Noradrenergic cell bodies in the locus ceruleus of the convulsive mutant quaking mouse and the control of the same strain were visualized using histofluorescence and tyrosine hydroxylase-like immunoreactivity. Cell counts performed with the two techniques gave closely similar results within each strain, indicating a 50% increase in the number of noradrenergic neurons in the midportion of the mutants' locus ceruleus when compared to the controls. This result gives histological support to the increased noradrenergic neurotransmission previously described in the brain of this mutant. Thus, the abnormally high activity of the noradrenergic system appears to be a primary effect of the mutation, associated with the convulsions of this animal model of epilepsy.

Chronic cobalt-induced epilepsy: noradrenaline ionophoresis and adrenoceptor binding studies in the rat cerebral cortex.

Several studies indicate that brain noradrenaline (NA) depletion facilitates the occurrence of epileptogenic syndromes in various animal models. In cobalt-induced epilepsy in the rat activity is associated with a cortical NA denervation. In order to search for cortical adrenoceptor modifications, inonophoretic studies and adrenoceptor binding assays were performed. At the period of maximal seizure activity, there was a significant supersensitivity of cortical neurons to the ionophoretic application of NA. An increase in the density of beta-adrenoceptor binding sites was observed. No modification in alpha 1- and alpha 2-adrenoceptor binding sites was found. This suggests that in cobalt-induced epilepsy there is a denervation supersensitivity which rests on a selective involvement of beta-adrenoceptors.

Desipramine elicits the expression of opiate receptors and sulfogalactosylceramide synthesis in rat C6 glioma cells.

In the course of our studies on lipidoses induced by amphiphilic drugs, we have investigated the ef- of desipramine, a tricyclic antidepressant, on glial cells in culture. We noted that the addition of desipramine to the culture medium of C6 glioma cells resulted in the modification of the lipid profile of the cell membranes. Of particular interest was the presence, in the desipramine-treated cells, of an additional lipid comigrating on thin layer chromatography with sulfogalactosylceramide (S-GalCer). Addition of radiolabelled sulfuric acid in the culture medium of the desipramine-treated cells resulted in the incorporation of [35S]sulfate in the newly synthesized lipid. Furthermore, this lipid was localized selectively by indirect immunofluorescence using a specific rabbit anti-S-GalCer antibody on the cell surface of desipramine-treated, but not control, C6 cells. Desipramine also increased the activity of 3'-phosphoadenosine-5'-phosphosulfate sulfotransferase (the enzyme responsible for the synthesis of S-GalCer). Since it has been suggested that S-GalCer may be involved in opiate receptors, we looked for opiate binding sites on C6 glioma cells after exposure to desipramine. We found that dihydromorphine was able to bind to the desipramine-treated C6 cell membrane. The binding of [3H]dihydromorphine (180 fmol/mg protein) was stereospecific and had a KD of 30-60 nM. Furthermore, morphine reduced both the basal and isoproterenol-stimulated cyclic AMP levels of the desipramine-treated C6 cells. This effect was blocked by naloxone. In these respects, the opiate binding sites induced after treatment of C6 glioma cells with desipramine fulfill the requirements of a true opiate receptor.

Effect of tricyclic antidepressants on sphingomyelinase and other sphingolipid hydrolases in C6 cultured glioma cells.

Cationic amphiphilic drugs, which include tricyclic antidepressants, have been shown to give rise to lipidoses under experimental conditions, with a general increase of lipids especially phospholipids. We report here an early and important decrease in sphingomyelinase activity in C6 glioma cells cultured in the presence of imipramine or desipramine at final concentrations of 0.01 and 0.05 mM. The effect was both dose-dependent and time-dependent and was observed before any lipid accumulation. Cerebroside beta-glucosidase and cerebroside beta-galactosidase had normal activities under the same experimental conditions and thus there was no general effect on membrane-bound sphingolipid hydrolases. A decrease of sphingomyelinase activity has been previously reported for two amphiphilic compounds, perhexiline maleate and AY 9944. These results suggest a potential function of sphingomyelinase in the mode of action of these drugs.

Tricyclic antidepressant desipramine induces stereospecific opiate binding and lipid modifications in rat glioma C6 cells.

Incubation for 48 hours of C6 glioma cell cultures with 10(-4)M tricyclic antidepressant desipramine gave rise to a quantitative increase of total lipids and to qualitative modifications of glycosphinegolipids involving detection by thin-layer chromatography of spots migrating according to cerebroside and sulfatide and presence of an abnormal ganglioside pattern. These lipid modifications were associated with the appearance of stereospecific binding of opiates (dihydromorphine) with a dissociation constant of 30-60 nM. These results favor an important role of lipids in opioid receptor function.

Lipid analysis in nerve biopsy specimens of hypertrophic neuropathy.

We performed a lipid analysis on nerve biopsy specimens in two cases of degenerative hypertrophic neuropathy. Quantitative analysis of the major lipid classes, ie, cholesterol, cerebrosides, sulfatides, ethanolamine phospholipids, phosphatidyl-choline, phosphatidyl-serine, phosphatidyl-inositol, sphingomyelin, and gangliosides, were performed. The two cases exhibited extreme decreases in levels of lipids that could be related to the very low myelin content of these nerves. Cholesterol and phospholipid levels were especially reduced. Cerebrosides and sulfatides were not modified in the same proportion, as could have been predicted from the degree of demyelination. This relative glycolipid increase could be due to the very high Schwann cell proliferation.