of looking for associations between H pylori seropositivity and several variables including indices of childhood overcrowding. A stepwise logistic regression analysis of H pylori seropositivity was done on sex, social class, age, number of years living in Gwynedd, place of birth, sharing a bed as a child, and the number of people sharing a bedroom during childhood. This analysis indicated that increasing seropositivity is associated with manual social class (odds ratio 1-63, 95% confidence intervals [CI] 1-25-2-11) and sharing a bedroom as a child (odds ratio 1 58, 95% CI 1-19-2-12):
Sharing a bed as a Social class Non manual Manual Total
analysis of single channel, subdurally recorded, peri-ictal EEG signal.
(including sleep, performance of cognitive tasks, meditation,
anaesthesia, and dementia); and fractal analysis has been reported to enhance the computerised detection, prediction, and localisation of
epileptic activity. 4,5 These are early days and methodological problems remain to be resolved.6 Of greatest clinical significance are the limitations
imposed by the data requirements and by the complexity of most
estimate fractal dimension of the EEG.
Nevertheless, the effort needed to overcome the operational problems is worth making. By comparison with other neuropsychiatric investigations, EEG recording is remarkably safe,
acceptable, and available. The EEG signal as traditionally analysed and imaged may justifiably be regarded by many clinicians as little more than an inscrutable squiggle of ink, only vaguely or occasionally helpful-but this does not necessarily mean that it is intractably meaningless. If fractal analysis lives up to its promise this very accessible source of poorly comprehended data may yield exciting benefits for neuropsychological medicine
Maudsley Hospital, London SE5 8AZ, UK 1.
EDWARD BULLIMORE MICHAEL BRAMMER COLIN BINNIE GONZALO ALARCON
Pijn JP, Neerven JV, Noest A, Lopes da Silver FH. Chaos or noise m EEG signals; dependence on state and brain site. Electroenceph Clin Neurophysiol 1991; 79: 371-81.
Goldberger AL, West BJ. Chaos m physiology: health or disease? In: Chaos in biological systems. New York: Plenum, 1987: 1-4. 3. Jansen BH. Quantitative analysis of electroencephalograms: is there chaos m the future? Int J Biomed Comput 1991; 27: 95-123. 4. Arle JE, Simon RH. An application of fractal dimension to the detection of transients in the electroencephalogram. Electroenceph Clin Neurophysiol 1990; 75: 296-305. 5. Iasemedis LD, Sackellares JC, Zaveri HP, Williams WJ. Phase space topography and the Lyapunov exponent of electrocorticograms in partial seizures. Brain Topog
1990; 2: 187-201. 6. Mayer-Kress G, Layne SP. Dimensionality of the human electroencephalogram. Ann N Y Acad Sci 1987; 504: 62-87.
Helicobacter pylori infection and overcrowding in childhood SIR,-Helicobacter pylori infection is accepted as a cause of both gastritis and chronic type B gastritis.* It may be implicated in the aetiology of gastric cancer, infection by the organism being especially common in underdeveloped countries, where gastric
is frequently encountered.2 H pylori is strongly associated intestinal-type gastric cancerbut, according to Correa and
Ruiz,’ it promotes rather than initiates carcinogenesis. Barker et aP have shown an association between overcrowding in childhood and the later occurrence of gastric cancer, which they argue suggests the involvement of a transmissible agent such as H pylori. There is evidence of intrafamilial spread of the infection,6 which provides further support for this hypothesis. We have lately completed a study of nearly 1000 blood donors in Gwynedd, UK, with the aim
32% 45% 38%
(128/400) (138/310) (266/710)
45% 54% 49%
(57/128) (68/127) (125/255)
(185/528) (206/437) 965
Our findings provide support for the hypothesis that infection with H pylori may be one link between an overcrowded childhood environment in which close physical contact is more likely, and the later development of gastric cancer. Other aspects of the relation between overcrowding and H pylori seropositivity are being considered. Centre for Applied Statistics, School of Mathematics, University College of North Wales, Bangor, Gwynedd LL57 1UT, UK
O. P. GALPIN C. J. WHITAKER A. J. DUBIEL
Dooley CP, Cohen H. The clinical significance of Campylobacter pylori. Ann Intern Med 1988; 108: 70-79. 2. Editorial. Campylobacter pylori becomes Helicobacter pylori. Lancet 1989; ii: 1019-20. 3. Parsonnet J, Vandersteen D, Coates J, Sibley RK. Helicobacter pylori in intestinal and 1.
diffuse type gastnc cancer. Abstracts of the 1990 Interscience Conference on Antimicrobial Agents and Chem., Atlanta, USA. 4. Correa P, Ruiz B. Campylobacter pylori and gastric cancer. In: Rathbone BJ, Heatley RV, eds. Campylobacter pylori and gastroduodenal disease Oxford: Blackwell, 1989. 5. Barker DJP, Coggon D, Osmond C, Wickham C. Poor housing in childhood and high rates of stomach cancer in England and Wales. Br J Cancer 1990; 61: 575-78. 6. Drumm B, Perez-Perez GI, Blaser MJ, Sherman PM. Intrafamilial clustering of Helicobacter pylori infection. N Engl JMed 1990; 322: 359-63.
Prostate-specific-antigen-con A binding ratio in benign prostate hyperplasia and prostate cancer SIR,-Prostate-specific antigen (PSA) is a valuable biochemical adjunct in screening and follow-up for prostate cancer.1,2 However, since in 21-68% of patients with benign prostatic hyperplasia serum PSA is also raised, measurement of this marker does not allow differentiation between the two forms of prostatic disease. On the assumption that PSA in benign disease and PSA produced by prostatic cancer cells are differently glycosylated, the degree of binding to concanavalin A (con A) has been used to distinguish between the two forms of disease. Barak et aP reported significantly different PSA-con A non-binding ratios for these two disease types. However, in a comparable study Chan and Gao4 found considerable overlap between the ratios, and therefore concluded that the PSA-con A ratio had no clinical application. To analyse these reported discrepancies we explored the binding of PSA to varying of con A. We analysed serum samples from 20 patients with benign prostate hyperplasia (group 1) and 20 with active prostate cancer (group 2). All patients were selected on the basis of a positive serum PSA value (group 1, mean 61 ng/ml, range 5-487; group 2, mean 183 ng/ml, range 5-853). Con A ’Sepharose’ (Pharmacia) was washed on a glass filter with con A buffer (0-5 mol/1 NaCl,1 mmol/1 MnClz, 1 mmol/1 MgClz, 1 mmol/1 CaClz, and 01 mol/1 sodium acetate buffer; pH 6-5). Con A was subsequently dried in air. Con A buffer was added to 2 g of dry con A sepharose up to a volume of 10 ml. 1 ml of this suspension was poured in a small tube and 1 ml con A buffer was added. After centrifugation for 10 min at 1000 g, the supernatant was decanted; filter paper was used to remove the last drops. The tube with this con A precipitate is defined as the con A tube. Samples with PSA concentration above 100 ng/ml were diluted with normal human serum to a concentration of 25-75 ng/ml. In the first step of the assay 500 pl serum was pipetted into a amounts
A tube and incubated for 2 h at