691
Only in HBsAg-negative cases with C.A.H. and high inflammatory activity a linear deposition of IgG at the membranes of isolated rabbit hepatocytes could be observed. The antibody could be absorbed by human liver-specific lipoprotein (H.L.P. i) as isolated and characterised by Meyer zum Buschenfelde and Miescher.6 This fraction is an organ-specific but not species-specific membrane antigen. The detection of IgG on rabbit hepatocytes did not correlate with the antibody activity measured by passive hasmagglutination technique using H.L.P. I as antigen. The results show that in serum of patients with HB,,Agnegative C.A.H. and high inflammatory activity a cytophilic antibody occurs which is directed against native liverspecific membrane antigen. The corresponding antigen of the membrane-fixed IgG in HBsAg-positive cases is still unknown. Second Department of Internal Medicine, University of Mainz, 65 Mainz, Langenbeckstrasse 1, Germany.
U. HOPF K. H. MEYER W. ARNOLD.
ZUM
BÜSCHENFELDE
PARVOVIRUS-LIKE PARTICLES IN HUMAN FÆCES
SIR,-With reference should like
offer
to
your leader of Feb. 1
(p. 257)
additional information regarding the 22 nm. particles in human fseces. In our recent paper 8 we showed that these particles were identical in size, morphology, and buoyant density to known animal parvoviruses. As a result of this work we feel justified in stating that the 22 nm. fsecal particles are parvoviruses. Furthermore, we have shown that known isometric bacteriophages of the same order of size as these human parvoviruses are morphologically distinct. In work shortly to be published we have shown that the frequency of antibody to these human parvoviruses in an adolescent population is of the order of 30% to 40%, which is far higher than has been reported for known isometric bacteriophages. For example, the frequency of humoral antibody in the general population to 0 X 174, which is a member of a group of common small isometric bacteriophages9 and is also a potent immunological antigen, 10 , 11 is less than 5%." We realise that final classification of this human parvovirus awaits isolation of the virus in cell-culture. we
to
some
W. K. PAVER E. O. CAUL S. K. R. CLARKE.
Public Health Laboratory,
Myrtle Road, Kingsdown, Bristol BS2 8EL.
BILIRUBIN LEVELS IN NEONATES
SIR,-Dr Calder and his colleagues (Dec. 7, p. 1339) puzzling that bilirubin levels in infants delivered by
find it
caesarean
section
are
lower than those in infants delivered
vaginally. I suggest that this is due
to
the fact that the infant born
by cassarean section is always above the level of the placenta, and therefore receives less blood than the vaginally delivered infant who is kept below placental level. Several years ago 6. 7.
8.
Meyer zum Büschenfelde, K. H., Miescher, P. A. ibid. 1972, 10, 89. Hopf, U., Meyer zum Büschenfelde, K. H., Freudenberg, J. ibid. 1974, 16, 117. Paver, W. K., Caul, E. O., Clarke, S. K. R. J. gen. Virol. 1974, 22, 447.
Bradley, D. E. Nature, 1962, 195, 622. Ching, Y. C., Davis, S. D., Wedgewood, R. J. J. clin. Invest. 1966, 45, 1593. 11. Peacock, D. B., Jones, J. V., Gough, M. Clin. exp. Immun. 1973, 13,
Dr Mavis Gunter demonstrated that the infant weighed several ounces less when held above placental level than when he was lowered and weighed beneath this level. The delivery of less blood to the caesarean infant means fewer red cells and therefore less haemolysis and lower bilirubin. 71 Collins
Street, Melbourne, Victoria 3000, Australia.
HIGH-DENSITY LIPOPROTEIN AND ATHEROSCLEROSIS SIR,-Dr G. J. Miller and Dr N. E. Miller (Jan. 4, p. 16) hypothesise that a reduction in plasma-high-density-lipoprotein (H.D.L.) concentration may accelerate the development of atherosclerosis by impairing the clearance of cholesterol from the arterial wall. Apart from their own work, showing a negative correlation between plasmaH.D.L.-cholesterol concentration and cholesterol-pool sizes, the bulk of their argument is based upon an inverse relationship between H.D.L.-cholesterol concentration and several predictive factors for the development of coronary heart-disease (C.H.D.) revealed by an exhaustive search of the epidemiological literature. Using reported data from 13 studies, Miller and Miller have shown a statistically significant negative correlation between L.D.L.-cholesterol and H.D.L.-cholesterol concentration by plotting the means of these variables obtained from each individual study, but without reference to the variance of these means or to the numbers and age-structures of the populations from which the means are derived. We consider that The negative correlation they found may be spurious and due to a combination of population differences, age effects, and the use of different analytical techniques in the studies cited (which varied in publication date from 1953 to 1974). A more valid approach would be to seek correlations between H.D.L.-cholesterol and L.D.L.-cholesterol within the same population group. However, in our study of 1005, randomly selected, London local-government employees1 we have shown a statistically significant positive correlation between H.D.L.-cholesteroland total-cholesterol concentrations for both men (n=502, r=0-33, and women (n=503, r=0-47, p<0001). This p<0-001) significant positive relationship persists when multiple regression analysis is used to correct for the possible independent effects on H.D.L.-cholesterol concentration of age, degree of obesity, fasting-plasma-triglyceride, and
blood-sugar. Again using multiple regression analysis we found no statistically significant, independent correlation between H.D.L.-cholesterol concentration, on the one hand, and either degree of obesity, fasting blood-sugar, or log. triglyceride concentration (three other possible risk factors for c.H.D.), on the other. The hypothesis of an inverse relationship between H.D.L.cholesterol concentration and other risk factors for C.H.D. is thus not supported when analyses are made within a population group and due allowance is made by multivariate methods for interrelationships between risk factors
dependence on age. In fact our own study showed positive relationship between H.D.L.-cholesterol concentra-
and their a
tion and total-cholesterol concentration and rather contradicts the hypothesis that a lowered plasma-H.D.L. concentration may hasten the development of atherosclerosis. Department of Medicine, Guy’s Hospital Medical School, London SE1 9RT.
9. 10.
497.
KATE CAMPBELL.
JOHN H. FULLER R. J. JARRETT H. KEEN S. L. PINNEY.
Fuller, J. H., Jarrett, R. J., Keen, H., Pinney, S. Proceedings of 1st International Congress on Obesity, London, 1974 (in the press). 2. Burstein, M., Samaille, J. Clinica chim. Acta, 1958, 3, 320. 1.