575
29. Scott GE, Tarlow D, McCrae MA. Detailed structural analysis of a genome rearrangement in bovine rotavirus. Virus Res 1989; 14: 119-28. 30. Allen AM, Desselberger U. Reassortment of human rotaviruses carrying rearranged genomes with bovine rotavirus. J Gen Virol 1985; 66: 2703-14. 31. Hundley F, Clark B, Chrystie I, Wood D, Beards G, Desselberger U. of rotaviruses from an Genome heterogeneity isolated immunodeficient, chronically infected child. J Virol 1987; 61: 3365-72. 32. Kalica AR, Flores J, Greenberg HB. Identification of the rotaviral gene that codes for hemagglutination and protease enhanced plaque formation. Virology 1983; 125: 194-205. 33. Hoshino Y, Sereno MM, Midthun K, Flores J, Kapikian AZ, Chanock RM. Independent segregation of two antigenic specificities (VP3 and VP7) involved in neutralization of rotavirus infectivity. Proc Natl Acad Sci USA 1985; 82: 8701-04. 34. Estes MK, Graham DY, Mason BB. Proteolytic enhancement of rotavirus infectivity: molecular mechanisms. J Virol 1981; 39: 879-88. 35. Offit PA, Blavat G, Greenberg HB, Clark HF. Molecular basis for rotavirus virulence: role of gene segment 4. J Virol 1986; 57: 46-49. 36. Flores J, Midthun K, Hoshino Y, et al. Conservation of the fourth gene among rotaviruses recovered from asymptomatic newborn infants and its possible role in attenuation. J Virol 1986; 60: 972-79. 37. Chen D, Burns JW, Estes MK, Ramig RF. The phenotypes of rotavirus reassortants depend upon the recipient genetic background. Proc Natl Acad Sci USA 1989; 86: 3743-47. 38. Flores J, Perez-Schael I, Kapikian AZ. Approaches to rotavirus vaccination. In: Farthing MJG, ed. Viruses and the gut. London: Swan, 1989: 109-19.
Anonymous HIV testing (p 516) we published the results of anonymous testing for human immunodeficiency virus (HIV) infection among pregnant women by use of Guthrie cards from neonatal screening. The study, conducted by Professor Peckham and her colleagues, covered three of the four Thames Regions in Britain over a one-year period to June, 1989. Approximately 115 000 samples were tested. The results show striking contrasts: in inner London 0-49 per thousand women were positive, outside London the rate was twelve times lower at 0 04 per thousand, while in outer London intermediate rates were observed. In all, 29 positives were identified. Owing to the general nervousness surrounding community-based studies of HIV infection and related behaviour, many important studies or programmes of surveillance have been slow to start, so this study, completed despite at times noisy and abusive criticism, is especially Last week
welcome. Control
of
the
AIDS epidemic depends overwhelmingly reducing the risk of transmission For this purpose behaviour. by modifying identification of where in the community transmission is taking place, and to what degree, is essential. The Cox report’ was the first systematic attempt in the UK to assess the state of the epidemic and to make short-term predictions. Its most noteworthy feature was the great range of uncertainty, which led the working group to recommend strongly that methods be developed for ascertaining seroprevalence rates more accurately, including the use of anonymised testing on a national scale. The update of the Cox report, published in January, also underlined these uncertainties, which result from an increasingly inexcusable lack of information on changing on
seroprevalence.
area where uncertainty has been reduced in the time that has elapsed considerably between the two reports is in prediction for the AIDS epidemic in the homosexual community. The authors of the 1990 report feel that less than a two-fold range covers the likely number of new AIDS cases in 1993, or the number of HIV-positives at the end of 1988. This compares with a 20 to 30 fold range for prediction of AIDS cases in 1993 due to intravenous drug use or heterosexual transmission. Part of this greater precision is arithmetic-larger numbers increase the proportionate precision. However, underlying the statistical projections were two sets of independent information about homosexual males: namely, anonymous testing without consent of homosexual attenders at a genitourinary medicine clinic in London over a 5-year periodand studies of sexual lifestyles4 among the London homosexual community. Both studies gave strong support to the statistical inference from the AIDS data that HIV transmission had reduced substantially among homosexuals in the period 1984 to 1987. Since the HIV testing was not population based, estimates of absolute numbers of infected individuals in the population could not be derived, but inferences could clearly be drawn about the temporal shape of the HIV epidemic, emphasising both the value and the limitations of testing sentinel high-risk groups.
The
one
Attention must now focus on the epidemic as it relates to intravenous drug use and heterosexual transmission. How many people are infected? Has behaviour changed among at-risk groups? Where is the epidemic heading? A national programme of anonymous testing has been announced in the UK and its rationale describedand non-Government money has been forthcoming for a national study of sexual lifestyles. Results of these investigations will form part of the mosaic of information that needs to be assembled to understand the epidemic fully.
Anonymous testing needs large numbers of individuals from whom blood samples are taken for another purpose. With prevalence rates among women averaging about 1 in 10 000, large means 105 or more. There are few sources on such a scale, and even fewer that are readily accessible, but one is formed by pregnant women. Both the dried blood spots taken from newborn babies and the sample taken for rubella testing are obvious possibilities. Professor Peckham and her colleagues took note of the fact that coverage in the UK by Guthrie card screening is almost complete-1 of 493 cases of congenital hypothyroidism was missed in a nationwide survey.66 The total reagent costs amounted to CO. 12 per sample. Centralisation of testing at regional laboratories greatly facilitates the collection of the material. The system is therefore practicable, cheap, and comprehensive. Approval was obtained from the ethical committees responsible for the neonatal screening programme, but in future it appears that the
576
ethical committee of each district (ie, about seventy in the Thames Regions) will have to be approached. Such a policy would seriously hinder continuation of the study. There is a strong case for establishing regional or even national ethical committees for studies of regional or national scope if there are no issues peculiar to particular districts. The
the basis of HIV the prevalence figures among general population of active women has been criticised. No sexually information on risk factors will be available; many sexually active women intend never to become pregnant; pregnancy terminations will be missed, and these women may well be at higher risk of HIV infection; and with the right to opt out of anonymous testing no scheme can claim to be comprehensive. Emphasis on the right to opt out may well come to be regretted. In the USA, no individual rights are recognised over an anonymised residuum of a blood sample taken for other purposes. With possibilities of testing pools of 10 or even 50 samples simultaneously, and of randomly selecting only some samples for testing (so that no individual would ever know whether or not her sample had been tested), it is difficult to see who benefits from the right to opt out. The degree to which the right is exercised will need careful monitoring. Terminations clearly pose a problem and parallel studies will be required to evaluate the distortion that they induce. That pregnant women do not represent all sexually active women has to be treated in this context as an unfortunate fact of life. There is no other source of material on that scale, and unless evidence of serious bias is forthcoming, the reduction in statistical uncertainty due to the numbers would seem far to outweigh a minor bias due to self-selection. The lack of risk factor information becomes a drawback once prevalence becomes appreciable. Thus, in the results of the latest study, one would like to know where and how the 0-49 per thousand prevalence in inner London arises. In some areas of London it would therefore seem appropriate to conduct named, with consent, testing in parallel with anonymous testing. The former provides risk-factor information, the latter provides lack of bias. The 4 per 100 000 prevalence outside London is a less urgent target for further work. use
of pregnant
women as
Even with all these caveats, the value of Professor Peckham’s results can be clearly seen, and it is not difficult to envisage the increasing value that will accumulate as a time-series of such results becomes available covering the UK. Greatly protracted discussion on the rights and wrongs of anonymous testing has delayed the introduction of a national programme in Britain by several years, to the serious detriment of public health control of the disease. This latest study provides baseline data two years earlier than would otherwise have been the case.
1. Short term prediction of HIV infection and AIDS in England and Wales: report of a working group (the Cox report). London: HM Stationery 2.
3.
4.
Office, 1988. Working group report to Director of Public Health Laboratory Service (the Day report). Acquired immune deficiency syndrome in England and Wales to end 1993—projections using data to end September 1989. Commun Dis Rep Jan 1990 (suppl). Gill ON, Day NE, Adler MW. Monitoring the prevalence of HIV. Br Med J 1989; 299: 1295-98. Carne CA, Weller IVD, Johnson AM, et al. Prevalence of antibodies to human immunodeficiency virus, gonorrhoea rates, and changed sexual
behaviour in homosexual men in London. Lancet 1989; i: 656-58. 5. Evans BA, McLean KA, Dawson SG, et al. Trends in sexual behaviour and risk factors for HIV infection among homosexual men, 1984-7. Br Med J 1989, 298: 215-18. 6. Grant DB, Smith I. Survey of neonatal screening for primary hypothyroidism in England, Wales, and Northern Ireland 1982-4. Br Med J 1988; 296: 1355-58.
MAGNETOENCEPHALOGRAPHY The electrical
activity of the brain produces changes in potential (the electroencephalogram, EEG); in current, detectable by the resulting magnetic fields (the magnetoencephalogram, MEG); and in electrical impedance. The EEG is a long-established method of investigating cerebral function; cerebral impedance imaging is in its infancy, but MEG is undergoing rapid development in many centres worldwide. The brain’s magnetic signals are extremely weak-about 10-13 tesla by comparison with the geomagnetic field of 5 x 10-5 tesla and the typical urban magnetic noise level of 5 x 10-7 tesla / JHz. MEG became feasible only with the development of sensitive superconducting quantum interference devices (SQUIDS) and gradiometers (systems of detector coils that discriminate against distant noise sources). MEG systems have evolved swiftly from cumbersome single-channel probes incapable of sampling the field at more than one location on the head at a time, to arrays of over twenty-five channels, now commercially available at prices of 1-2 million. The scalp EEG is only indirectly related to neuronal generators, since it arises from the volume currents that flow
throughout the brain and its coverings. The poor conductivity of the skull and inhomogeneous impedance of the brain attenuate and distort the spatial pattern of the it is difficult to locate the source. The theoretical advantage of the MEG is that the magnetic fields of the brain are not greatly affected by the medium through which they pass. In an idealised homogeneous spherical head, an elemental current generator produces a simple magnetic field which enters and leaves the head at two points or extrema. The underlying generator lies at the bisector of the line joining the extrema; its direction is perpendicular to this line and its depth is linearly related to the distance between the extrema.1 Departures from the idealised spherical head and variations in internal conductivity may lead to a volume current contribution that could distort the magnetic field. However, many experimental isofield maps obtained from evoked response and epilepsy studies2,3 show that the idealised model usually holds true, so the volume current contamination is small. The difficulty of deducing the underlying current distribution from either magnetic field (MEG) or electrical potential measurements (EEG) exemplifies the wider class of "inverse problems".4 Such problems seldom have a unique mathematical solution, and become insoluble when more than one spatially restricted generator is responsible for the observed patterns. However, given the constraints of anatomy and existing knowledge, it is often possible to decide between mathematically
EEG,
so
29. Scott GE, Tarlow D, McCrae MA. Detailed structural analysis of a genome rearrangement in bovine rotavirus. Virus Res 1989; 14: 119-28. 30. Allen AM, Desselberger U. Reassortment of human rotaviruses carrying rearranged genomes with bovine rotavirus. J Gen Virol 1985; 66: 2703-14. 31. Hundley F, Clark B, Chrystie I, Wood D, Beards G, Desselberger U. of rotaviruses from an Genome heterogeneity isolated immunodeficient, chronically infected child. J Virol 1987; 61: 3365-72. 32. Kalica AR, Flores J, Greenberg HB. Identification of the rotaviral gene that codes for hemagglutination and protease enhanced plaque formation. Virology 1983; 125: 194-205. 33. Hoshino Y, Sereno MM, Midthun K, Flores J, Kapikian AZ, Chanock RM. Independent segregation of two antigenic specificities (VP3 and VP7) involved in neutralization of rotavirus infectivity. Proc Natl Acad Sci USA 1985; 82: 8701-04. 34. Estes MK, Graham DY, Mason BB. Proteolytic enhancement of rotavirus infectivity: molecular mechanisms. J Virol 1981; 39: 879-88. 35. Offit PA, Blavat G, Greenberg HB, Clark HF. Molecular basis for rotavirus virulence: role of gene segment 4. J Virol 1986; 57: 46-49. 36. Flores J, Midthun K, Hoshino Y, et al. Conservation of the fourth gene among rotaviruses recovered from asymptomatic newborn infants and its possible role in attenuation. J Virol 1986; 60: 972-79. 37. Chen D, Burns JW, Estes MK, Ramig RF. The phenotypes of rotavirus reassortants depend upon the recipient genetic background. Proc Natl Acad Sci USA 1989; 86: 3743-47. 38. Flores J, Perez-Schael I, Kapikian AZ. Approaches to rotavirus vaccination. In: Farthing MJG, ed. Viruses and the gut. London: Swan, 1989: 109-19.
Anonymous HIV testing (p 516) we published the results of anonymous testing for human immunodeficiency virus (HIV) infection among pregnant women by use of Guthrie cards from neonatal screening. The study, conducted by Professor Peckham and her colleagues, covered three of the four Thames Regions in Britain over a one-year period to June, 1989. Approximately 115 000 samples were tested. The results show striking contrasts: in inner London 0-49 per thousand women were positive, outside London the rate was twelve times lower at 0 04 per thousand, while in outer London intermediate rates were observed. In all, 29 positives were identified. Owing to the general nervousness surrounding community-based studies of HIV infection and related behaviour, many important studies or programmes of surveillance have been slow to start, so this study, completed despite at times noisy and abusive criticism, is especially Last week
welcome. Control
of
the
AIDS epidemic depends overwhelmingly reducing the risk of transmission For this purpose behaviour. by modifying identification of where in the community transmission is taking place, and to what degree, is essential. The Cox report’ was the first systematic attempt in the UK to assess the state of the epidemic and to make short-term predictions. Its most noteworthy feature was the great range of uncertainty, which led the working group to recommend strongly that methods be developed for ascertaining seroprevalence rates more accurately, including the use of anonymised testing on a national scale. The update of the Cox report, published in January, also underlined these uncertainties, which result from an increasingly inexcusable lack of information on changing on
seroprevalence.
area where uncertainty has been reduced in the time that has elapsed considerably between the two reports is in prediction for the AIDS epidemic in the homosexual community. The authors of the 1990 report feel that less than a two-fold range covers the likely number of new AIDS cases in 1993, or the number of HIV-positives at the end of 1988. This compares with a 20 to 30 fold range for prediction of AIDS cases in 1993 due to intravenous drug use or heterosexual transmission. Part of this greater precision is arithmetic-larger numbers increase the proportionate precision. However, underlying the statistical projections were two sets of independent information about homosexual males: namely, anonymous testing without consent of homosexual attenders at a genitourinary medicine clinic in London over a 5-year periodand studies of sexual lifestyles4 among the London homosexual community. Both studies gave strong support to the statistical inference from the AIDS data that HIV transmission had reduced substantially among homosexuals in the period 1984 to 1987. Since the HIV testing was not population based, estimates of absolute numbers of infected individuals in the population could not be derived, but inferences could clearly be drawn about the temporal shape of the HIV epidemic, emphasising both the value and the limitations of testing sentinel high-risk groups.
The
one
Attention must now focus on the epidemic as it relates to intravenous drug use and heterosexual transmission. How many people are infected? Has behaviour changed among at-risk groups? Where is the epidemic heading? A national programme of anonymous testing has been announced in the UK and its rationale describedand non-Government money has been forthcoming for a national study of sexual lifestyles. Results of these investigations will form part of the mosaic of information that needs to be assembled to understand the epidemic fully.
Anonymous testing needs large numbers of individuals from whom blood samples are taken for another purpose. With prevalence rates among women averaging about 1 in 10 000, large means 105 or more. There are few sources on such a scale, and even fewer that are readily accessible, but one is formed by pregnant women. Both the dried blood spots taken from newborn babies and the sample taken for rubella testing are obvious possibilities. Professor Peckham and her colleagues took note of the fact that coverage in the UK by Guthrie card screening is almost complete-1 of 493 cases of congenital hypothyroidism was missed in a nationwide survey.66 The total reagent costs amounted to CO. 12 per sample. Centralisation of testing at regional laboratories greatly facilitates the collection of the material. The system is therefore practicable, cheap, and comprehensive. Approval was obtained from the ethical committees responsible for the neonatal screening programme, but in future it appears that the
576
ethical committee of each district (ie, about seventy in the Thames Regions) will have to be approached. Such a policy would seriously hinder continuation of the study. There is a strong case for establishing regional or even national ethical committees for studies of regional or national scope if there are no issues peculiar to particular districts. The
the basis of HIV the prevalence figures among general population of active women has been criticised. No sexually information on risk factors will be available; many sexually active women intend never to become pregnant; pregnancy terminations will be missed, and these women may well be at higher risk of HIV infection; and with the right to opt out of anonymous testing no scheme can claim to be comprehensive. Emphasis on the right to opt out may well come to be regretted. In the USA, no individual rights are recognised over an anonymised residuum of a blood sample taken for other purposes. With possibilities of testing pools of 10 or even 50 samples simultaneously, and of randomly selecting only some samples for testing (so that no individual would ever know whether or not her sample had been tested), it is difficult to see who benefits from the right to opt out. The degree to which the right is exercised will need careful monitoring. Terminations clearly pose a problem and parallel studies will be required to evaluate the distortion that they induce. That pregnant women do not represent all sexually active women has to be treated in this context as an unfortunate fact of life. There is no other source of material on that scale, and unless evidence of serious bias is forthcoming, the reduction in statistical uncertainty due to the numbers would seem far to outweigh a minor bias due to self-selection. The lack of risk factor information becomes a drawback once prevalence becomes appreciable. Thus, in the results of the latest study, one would like to know where and how the 0-49 per thousand prevalence in inner London arises. In some areas of London it would therefore seem appropriate to conduct named, with consent, testing in parallel with anonymous testing. The former provides risk-factor information, the latter provides lack of bias. The 4 per 100 000 prevalence outside London is a less urgent target for further work. use
of pregnant
women as
Even with all these caveats, the value of Professor Peckham’s results can be clearly seen, and it is not difficult to envisage the increasing value that will accumulate as a time-series of such results becomes available covering the UK. Greatly protracted discussion on the rights and wrongs of anonymous testing has delayed the introduction of a national programme in Britain by several years, to the serious detriment of public health control of the disease. This latest study provides baseline data two years earlier than would otherwise have been the case.
1. Short term prediction of HIV infection and AIDS in England and Wales: report of a working group (the Cox report). London: HM Stationery 2.
3.
4.
Office, 1988. Working group report to Director of Public Health Laboratory Service (the Day report). Acquired immune deficiency syndrome in England and Wales to end 1993—projections using data to end September 1989. Commun Dis Rep Jan 1990 (suppl). Gill ON, Day NE, Adler MW. Monitoring the prevalence of HIV. Br Med J 1989; 299: 1295-98. Carne CA, Weller IVD, Johnson AM, et al. Prevalence of antibodies to human immunodeficiency virus, gonorrhoea rates, and changed sexual
behaviour in homosexual men in London. Lancet 1989; i: 656-58. 5. Evans BA, McLean KA, Dawson SG, et al. Trends in sexual behaviour and risk factors for HIV infection among homosexual men, 1984-7. Br Med J 1989, 298: 215-18. 6. Grant DB, Smith I. Survey of neonatal screening for primary hypothyroidism in England, Wales, and Northern Ireland 1982-4. Br Med J 1988; 296: 1355-58.
MAGNETOENCEPHALOGRAPHY The electrical
activity of the brain produces changes in potential (the electroencephalogram, EEG); in current, detectable by the resulting magnetic fields (the magnetoencephalogram, MEG); and in electrical impedance. The EEG is a long-established method of investigating cerebral function; cerebral impedance imaging is in its infancy, but MEG is undergoing rapid development in many centres worldwide. The brain’s magnetic signals are extremely weak-about 10-13 tesla by comparison with the geomagnetic field of 5 x 10-5 tesla and the typical urban magnetic noise level of 5 x 10-7 tesla / JHz. MEG became feasible only with the development of sensitive superconducting quantum interference devices (SQUIDS) and gradiometers (systems of detector coils that discriminate against distant noise sources). MEG systems have evolved swiftly from cumbersome single-channel probes incapable of sampling the field at more than one location on the head at a time, to arrays of over twenty-five channels, now commercially available at prices of 1-2 million. The scalp EEG is only indirectly related to neuronal generators, since it arises from the volume currents that flow
throughout the brain and its coverings. The poor conductivity of the skull and inhomogeneous impedance of the brain attenuate and distort the spatial pattern of the it is difficult to locate the source. The theoretical advantage of the MEG is that the magnetic fields of the brain are not greatly affected by the medium through which they pass. In an idealised homogeneous spherical head, an elemental current generator produces a simple magnetic field which enters and leaves the head at two points or extrema. The underlying generator lies at the bisector of the line joining the extrema; its direction is perpendicular to this line and its depth is linearly related to the distance between the extrema.1 Departures from the idealised spherical head and variations in internal conductivity may lead to a volume current contribution that could distort the magnetic field. However, many experimental isofield maps obtained from evoked response and epilepsy studies2,3 show that the idealised model usually holds true, so the volume current contamination is small. The difficulty of deducing the underlying current distribution from either magnetic field (MEG) or electrical potential measurements (EEG) exemplifies the wider class of "inverse problems".4 Such problems seldom have a unique mathematical solution, and become insoluble when more than one spatially restricted generator is responsible for the observed patterns. However, given the constraints of anatomy and existing knowledge, it is often possible to decide between mathematically
EEG,
so
