Search

BACKGROUND As part of a nested case-control study of benign proliferative breast disease (BPBD) conducted within the cohort of women participating in the Alberta breast screening programme, an analysis of all women who had a benign breast biopsy between 1990 and 1995 was undertaken to identify the epidemiological risk factors for BPBD. METHODS The breast biopsies of all eligible women were re-reviewed by a panel of four pathologists using Page's classification for benign breast disease. Cases were 165 women whose biopsies, upon review, showed benign breast tissue changes ranging from sclerosing adenosis to atypical ductal hyperplasia. Controls were 217 women whose biopsies showed no evidence of any proliferative or neoplastic changes. In-person interviews were conducted with all study subjects. RESULTS Women with >/=25% fibroglandular breast tissue density, as compared to women with

The EUROCLUS study assembled incidence data for 13,551 cases of childhood leukaemia (CL) diagnosed between 1980 and 1989 in 17 countries (or regions of countries). These were referenced by location at diagnosis to small census areas of which there were 25,723 in the study area. Population counts, surface area and, hence, population density were available for all these small areas. Previous analyses have shown limited extra-Poisson variation (EPV) of case counts within small areas; this is most pronounced in areas of intermediate population density (150-499 persons/km2). In this study, the data set was examined in more detail for evidence that variations in incidence and EPV of CL are associated with population density. Incidence showed a curvilinear association with population density and was highest in areas which were somewhat more densely populated (500-750 persons/km2), where the incidence rate ratio relative to areas having > or = 1000 persons/km2 was 1.16 (95% confidence interval 1.07-1.26) and the P value for quadratic trend across eight strata of population density was 0.02. Incidence in these areas is uniformly elevated and showed no evidence of heterogeneity (i.e. EPV). Statistically significant evidence of EPV was evident amongst some of the areas previously classified as intermediate density areas (specifically, those with a density of 250-499 persons/km2, P < 0.001 for CL). These results were interpreted in terms of the current aetiological hypotheses for CL which propose that exposure to localised epidemics of one or more common infectious agent may contribute to the development of leukaemia. They suggest that such epidemics arise regularly in moderately densely populated areas and also sporadically in areas which are somewhat less densely populated. Although other interpretations are possible, these results may assist in the identification of characteristics which infectious agents must possess if direct or indirect causes of CL.

The epidemiology of Hodgkin’s disease suggests that it is a heterogeneous condition comprising more than one disease entity. The Epstein–Barr virus (EBV) is present in the Reed–Sternberg cells of a proportion of cases and is likely to play a role in the pathogenesis of these cases. In this study we show that EBV association rates vary with age at diagnosis. We suggest that Hodgkin’s disease can be divided into three disease entities on the basis of EBV association and age, thereby providing biological support for the multiple aetiology hypothesis proposed by MacMahon (Cancer Res 1966; 26: 1189–1290).

OBJECTIVE To improve understanding of the etiology of IDDM by analyzing spatial and space-time distribution of the incidence in children. RESEARCH DESIGN AND METHODS Statistical tests to detect clustering were applied to a population-based register of 1,490 children (aged 0–16 years) with IDDM in Yorkshire, northern England. The Knox test analyzed clustering in space and time, and the Potthoff-Whittinghill test quantified spatial differences in incidence between small-area census units (electoral wards). The Potthoff-Whittinghill test was conditioned for childhood population density and deprivation (Townsend index). RESULTS Both tests demonstrated clustering of IDDM in Yorkshire children. Space-time and spatial clustering is strongest in the younger children (0–4 and 5–9 years of age), even after conditioning for known associations. Clustering was more common in the county of Humberside during the years 1982–1985 and in wards of low population density (< 0.26 0- to 16 year-old subjects per hectare). CONCLUSIONS The study revealed a nonrandom space-time distribution of IDDM in children not accounted for by known covarying demographic factors. The Potthoff-Whittinghill test has not previously been applied to childhood IDDM. The new finding of strong clustering in young children is consistent with early exposure, possibly in utero, to infectious agents or localized environmental sources.

Background Various methods of analysis have been used to study age-period-cohort models. The main aim of this paper is to illustrate and compare three such methods. Those of Clayton and Schifflers, Robertson and Boyle, and De Carli and La Vecchia. The main differences between these methods lie in their approach to distinguish between linear-period and linear-cohort effects. Clayton and Schifflers do not attempt to solve this identification problem, whereas Robertson and Boyle, and De Carli and La Vecchia attempt to tackle this question. Methods In order to study the assumptions and problems of these methods, we analysed data from 2678 subjects aged 30-84 in Yorkshire, UK, who were diagnosed with non-Hodgkin's lymphoma (NHL) during the period 1978-1991. Loglinear Poisson models were used to examine the effects of age, period and cohort. Results All three methods of analysis agree that, after stratification for sex and county, the age-standardized rate has been increasing at about 5% per year. The Robertson-Boyle method differed from the Clayton-Schifflers method in showing a significant non-linear cohort effect, and a significant county-cohort interaction. The method of De Carli-La Vecchia agreed more closely with Clayton-Schifflers than with Robertson-Boyle. Conclusions The linear increase in incidence would lead to a doubling of the number of cases within 15 years. There is controversy over whether the identification problem can be solved and should be solved. Many authors would not rely on the results of the methods of Robertson and Boyle, or De Carli and La Vacchia.

This article presents additional follow-up analysis of women aged 45‐49 from the Edinburgh Randomized Trial of Breast Cancer Screening. The screening protocol included four mammographic examinations at two-year intervals and seven annual clinical examinations. Altogether, 21,774 women aged 45‐49 were recruited from 1978 to 1985 using cluster randomization. After 10‐14 years of follow-up, breast cancer mortality has been reduced by 12% to 18% (rate ratios, with and without adjustment for socio-economic status, are 0.88 and 0.82 respectively, with 95% confidence intervals [CIs] of 0.55‐1.41 and 0.51‐1.32). These benefits are smaller than that reported previously with shorter follow-up. This article also presents data from an observational study that compared survival beyond baseline (50‐52 years) of women first offered screening before and after age 50. Based on six-year data, the results suggest that earlier screening confers follow-up benefit (hazard ratio for later screening = 1.60; 95% CI: 0.96‐2.67), but these findings are not statistically significant. The trial is too small to yield statistically significant results by itself, but can make useful contributions to overview and meta-analyses. [Monogr Natl Cancer Inst 1997:22:31‐35]

Objective— To determine the pathological and biological characteristics of breast cancers diagnosed by screening and examined at the Edinburgh University pathology department. Methods— These cancers were classified by screening status: never screened (n=111), prevalence screen detected (n=105), and previously screened (n=74). The last category arose in women who had been regularly screened during the trial; the cancers were diagnosed as interval cases before the first invitation to service screening (n=33) or were incidence screen detected at that time (n=41). Results— Association (for operable invasive cancers, n=250) of cancer characteristics with screening status reflects influences of biology (aggressiveness) or chronology (time of diagnosis), or both. The prognostic indicators tumour grade, histological type, and oestrogen receptor status were found in a smaller percentage of the patients with poor prognosis among the prevalence screen detected cases (9%, 77%, 18%) than among those previously screened (29%, 84%, 35%). The chronological factors size and node status were found in a smaller percentage of patients with poor prognosis among women previously screened (31%, 24%) than among those never screened (62%, 39%)). Apart from these two, no other factors improved the diagnosis in the previously screened group compared with the never screened group. Conclusions— These results suggest that favourable characteristics of screen detected cases are often due to the effects of length bias on “biological factors” and fail to show that current local screening practice has succeeded in advancing the diagnosis of breast cancers to a less aggressive phase.

Summary. .To determine whether increased rates of childhood acute lymphoblastic leukaemia (ALL) which had been reported in isolated areas could be attributed to higher proportions of households owning cars, ecological analysis was performed with electoral wards as area units. Previous data were re-analysed using the proportion of households having no car, taken from the 1981 census of England and Wales, as an additional explanatory factor. A specialist registry of haematopoietic malignancies covering one-third of England and Wales (3270 electoral wards) recorded 438 cases of ALL in children diagnosed 1984-89, including 304 in the childhood peak (at ages 1–7 years). Relative risks were derived by Poisson regression of total childhood ALL and at ages 1–7 years for groups of electoral wards classified by isolation and car ownership. Multivariable analyses included adjustment for socio-economic status. No evidence of increased risk of childhood ALL in areas where more households own cars was found. ALL at ages 1–7 years was inversely associated with car ownership (risk in wards with least cars relative to those with most cars = 2.28,95% CI: 1.12-4.64). The associations with geographic isolation persisted after allowing for car ownership (risks for ALL in young children in isolated relative to built-up areas = 2.19,95% CI: 1.44-3.33). Levels of car ownership cannot explain the increased rates of childhood ALL which have been observed in isolated areas. No support has been found for a hypothesis relating these excesses to benzene exposure of children transported in cars. The previous explanation for the small area variation of childhood ALL in terms of geographical isolation and interpretation in terms of exposure to common infections continues to be justified.

The Leukaemia Research Fund Data Collection Study (DCS) is a specialist registry of leukaemias and lymphomas. The present study involves 494 cases of Hodgkin's disease (HD) registered with the DCS between 1985 and 1989. This entire data set has been tested for localised spatial clustering using an established nearest neighbour method with 18% of all cases in young people classified as clustered (P < 0.05). No clustering was found in older cases. Subsamples were selected from the registered cases for a pilot study in which case clustering, herpes virus antibody titres and Epstein-Barr virus (EBV) presence within the Reed-Sternberg (RS) cells (EBV-RS status) were investigated together. Firstly, a case-control study of HD in young people or nodular sclerosing (NS) subtype (39 HD cases and 26 healthy controls) found significant elevation of antibody titres to EBV-viral capsid antigen (VCA), EBV-early antigen (EA) and human herpes virus 6 (HHV-6) in HD cases compared with controls. EBV viral genome was present in 5 cases and 4 of these were in clusters of HD in young people. Elevation of antibody titres to the EBV antigens was not associated with case clustering or EBV-RS status. Antibody titres to HHV-6 differed significantly between EBV-RS+ and EBV-RS- cases (P = 0.04). Geometric mean titres for HHV-6 for EBV-RS+ and EBV-RS- cases were 11.5 and 73.7, respectively, with the former lower than the control value of 20.5. Secondly, a cluster study included all other cases (n = 14) in clusters containing known EBV-RS+ cases. 3 further cases were EBV-RS+ positive but no cluster consisted entirely of positive cases. Overall, 5/16 clustered, 2/12 peripheral and 1/25 random cases in these studies were EBV-RS+ (P = 0.017). The interpretation of these results in terms of shared aetiological exposures of cases within clusters and the roles of EBV and HHV-6 is discussed, and hypotheses for testing in future studies proposed.

The first section of this chapter provides an overview of the epidemiology of Hodgkin’s disease (HD), much of which is well known. Particular attention will be paid to the possible involvement of infectious agents in the aetiology and here attention becomes focused on young people, especially young adults. The second part considers clustering of HD in much greater depth. It is a basic belief of the author that clustering should not be considered in isolation, but forms an integral part of the investigation of infectious agents and HD. For this reason there is a single concluding paragraph covering the two sections.

Hodgkin’s disease (HD) may be the rare outcome of a common infection which is influenced by host genetic susceptibility. We have analysed this hypothesis by determining the frequency of HLA-DPB1 alleles in two series of HD patients using molecular typing methods. One series consisted of a retrospective/prospective group of 118 patients over the age of 15 years, and the other a prospective group of 45 patients between the ages of 16 and 24 years. In both series, the proportion of patients typing for HLA-DPB1 *0301 was greater than that for the controls, suggesting that this may be an HD-susceptibility allele. Analysis of DPB1 alleles in relation to HD subtype also showed that the increase in *0301 was present in nodular sclerosing HD (HDNS) patients, as well as in mixed cellularity HD (HDMC) and lymphocyte predominant HD (HDLP) patients. Preliminary evidence was obtained suggesting an increase in *0401, and possibly *0501, in HDMC and HDLP. Analysis of *0301-like hypervariable region (HVR) associations with HD subtypes indicated an increase in an *0301-like HVR-C motif in HDNS, but not in non-HDNS. The frequency of *0301- and *0401 -like HVR-C and HVR-F amino acid residues also differed in HDNS and non-HDNS patients. The frequency of the HVR-C amino acid residues Asp55, Glu56 (*0301) was increased in HDNS, but not in non-HDNS patients, whereas the frequency of Ala55, Ala56 (*0401) was increased in non-HDNS, but not in HDNS patients. In addition, the frequency of the HVR-F amino acid residues Asp84, Glu85 and Ala86 (*0301) was greater in the HDNS than non-HDNS patients, whereas there was no increase in Gly84, Gly85, Pro86 (*0401). Although these are preliminary findings, they suggest that genetic susceptibility to an infectious aetiology in HD may reside at the level of DPB peptide-binding residues rather than with a specific allele.

A study of childhood malignancies in North Humberside for 1974-1986 has revealed significant spatial aggregations. Though no specific reason was found, certain geographical areas, tumour groups and putative risks were identified for further study. One possible risk was a tin smelter to the west of Hull. A new method of analysis (Stone's conditional Poisson maximum) was used and showed some evidence of increased risk close to the smelter. This was accounted for by the solid tumours (primarily central nervous system malignancies) and was not apparent for the leukaemias. The Poisson maximum method is a special case of isotonic regression; a more general isotonic regression method, also due to Stone, Maximum Likelihood Ratio (MLR), is also available, and for reasons described elsewhere we prefer the more general method. The present study begins by applying this to the childhood data and then to data for corresponding adult disease (leukemias and central nervous system tumours) in the same area and time period. In both children and adults the results are negative for the leukaemias. For the remaining tumour categories, there is a significant increase in risk close to the smelter, but it is not possible to identify any aspect of the smelter or feature in its vicinity as causative.

OBJECTIVE--To determine whether parental occupations and chemical and other specific exposures are risk factors for childhood leukaemia. DESIGN--Case-control study. Information on parents was obtained by home interview. SETTING--Three areas in north England: Copeland and South Lakeland (west Cumbria); Kingston upon Hull, Beverley, East Yorkshire, and Holderness (north Humberside), and Gateshead. SUBJECTS--109 children aged 0-14 born and diagnosed as having leukaemia or non-Hodgkin's lymphoma in study areas during 1974-88. Two controls matched for sex and date and district of birth were obtained for each child. MAIN OUTCOME MEASURES--Occupations of parents and specific exposure of parents before the children's conception, during gestation, and after birth. Other adults living with the children were included in the postnatal analysis. RESULTS--Few risk factors were identified for mothers, although preconceptional association with the food industry was significantly increased in case mothers (odds ratio 2.56; 95% confidence interval 1.32 to 5.00). Significant associations were found between childhood leukaemia and reported preconceptional exposure of fathers to wood dust (2.73, 1.44 to 5.16), radiation (3.23, 1.36 to 7.72), and benzene (5.81, 1.67 to 26.44); ionising radiation alone gave an odds ratio of 2.35 (0.92 to 6.22). Raised odds ratios were found for paternal exposure during gestation, but no independent postnatal effect was evident. CONCLUSION--These results should be interpreted cautiously because of the small numbers, overlap with another study, and multiple exposure of some parents. It is important to distinguish periods of parental exposures; identified risk factors were almost exclusively restricted to the time before the child's birth.

Mothers of 294 children aged 3-9 years were interviewed about their child's previous history of vaccinations and infections. The study group comprised healthy children selected at GP surgeries (n = 136) and those hospitalized for both minor and serious conditions (n = 168). The results evaluate the entire study group for the pattern of reporting vaccinations and infectious episodes under three years of age, comparing two different data sources-mothers' reports and GP records. Maternal recall accuracy is formally tested (Kappa statistic) for agreement with GP records. For vaccination history the results indicate that mothers and GPs are inadequate data sources, and suggest use of computerized health district records to improve data accuracy. Major differences on infectious history were evident for two types of questions-closed for specific infections and open-ended for 'other infections'. Specific infections (e.g. measles) were systematically over-reported by mothers compared with GP records, but mothers failed to 'time' the event. Concordance for 'other infections' was also poor, but, by contrast, GPs systematically over-recorded this range of infections, indicating substantial inaccuracy in mothers' recall. The authors suggest that GP records should be the preferred source of data for these 'other' infectious episodes, especially under one year of age. Improved questionnaire design would elicit more accurate maternal reports on specified infections for which a GP may well not be consulted and corroboration would be impossible.

Data from a specialist registry of haematopoietic malignancies in England and Wales (1984-1988) have been analysed to investigate variations of incidence by age and diagnostic subtype of lymphoid malignancies in young people (aged 0-24 years). Attention has been focussed on the role of community lifestyle indicators for small areas, derived from routine sources, in an ecological analysis. The predominant conditions were acute lymphoblastic leukaemia (ALL)--42.4%, and Hodgkin's disease (HD)--37.5%. Lowest overall incidence at approximately 8 years of age corresponded to the termination of the childhood peak for ALL. Opposite trends of incidence rates with distance from urban centres (urban distance) were observed for the two age groups: odds ratios (OR) for areas greater than 20 km from towns and cities were 1.46 (95% confidence interval 1.01-2.12) for ages 0-7 and 0.75 (95% confidence interval 0.56-0.99) for ages 8-24. For the younger group this was entirely attributable to ALL. HD, which was dominant in the older group, had highest incidence in connurbations but the gradient of risk for older onset ALL followed the overall pattern for this age group. A positive relationship with socioeconomic status was evident for both age groups but this was considerably stronger for the older cases (OR = 1.16, 95% confidence interval 1.01-1.33) than for the younger for whom it was not independent of urban distance. These results display an association between expression of lymphoid malignancies in young people and urban distance which is not attributable to socioeconomic status; for urban distances the distribution is shifted towards ALL and towards younger age at onset.

EDITOR,—Ciaran B J Woodman and colleagues' paper reporting the high rate of interval cancers in the North Western Regional Health Authority1 and the accompanying editorial2 will help to increase awareness that successful breast screening has several facets. Two view mammography at the first screen is a simple and effective way of improving the yield at prevalence screens, thus reducing the potential for “false negative” interval cancers. One of the principles behind mammographic breast screening is that the earlier in the detectable preclinical phase that this (or any) modality can achieve detection the longer the delay before …

Throughout we consider operators on a reflexive Banach space X. We consider certain algebraic properties of F(X), K(X) and B(X) with the general aim of examining their dependence on the possession by X of the approximation property. B(X) (resp. K(X)) denotes the algebra of all bounded (resp. compact) operators on X and F(X) denotes the closure in B(X) of its finite rank operators. The two questions we consider are:(1) Is K(X) equal to the set of all operators in B(X) whose right and left multiplication operators on F(X) (or on B(X)) are weakly compact?(2) Is F(X) a dual algebra?

It is well known that every B ∗ {B^\ast } -algebra is Arens’ regular and that its bidual is a B ∗ {B^\ast } -algebra. Wong has asked whether a dual A ∗ {A^\ast } -algebra of the first kind is Arens’ regular. It is shown that this is true in the topologically simple case; in the course of the proof it is shown that in this case the bidual is, modulo its radical, an A ∗ {A^\ast } -algebra of the first kind.

The purpose of this paper is twofold. In [6] Tomiuk gives a representation theorem for a topologically simple right complemented algebra that is also an annihilator algebra. We strengthen this and then give a converse, so as to characterise right complemented algebras among respectively primitive Banach algebras and primitive annihilator Banach algebras. Our second aim is to investigate the relationship between the different annihilator conditions—left annihilator, right annihilator, annihilator, and dual—when imposed on a complemented algebra. Tomiuk [6] has already shown that a right complemented semisimple algebra that is a left annihilator algebra is an annihilator algebra; further, a topologically simple bi-complemented algebra that is also an annihilator algebra is dual. We show that for a topologically simple right complemented algebra all four annihilator conditions are equivalent. Further, for a semi-simple Banach algebra the first three are equivalent provided it is right complemented, and if it is also left complemented, then they are equivalent to duality.

In an effort to reduce the cost of breast cancer screening several studies have explored the possibility of using risk factors to select a high-risk group of women and then restrict screening to that group. The results of these studies have been almost entirely negative and so it is not possible at present to classify any woman as at such low risk that she need not be screened. Nevertheless it is well known that some groups of women can be identified as being at higher risk than the general population. In this study it is assumed that each woman will be offered one screen at which risk factor information will be collected. The usual screening policy is then one of uniform intervention in which the interval to the next screen is the same for all women: the interval that is currently recommended in the UK is three years. An alternative is a risk strategy in which the time to the next screen depends on the woman's risk status; thus the total number of screens available to the population are distributed according to risk status. Using data from the Edinburgh randomised trial of breast cancer screening these policies have been compared. It is estimated that the proportion of cases detected by screening in the three years following the completion of the initial screening round could be raised from 60% to 67% by adopting a risk strategy. Lead time benefits are also quantified as are the comparisons for an established screening programme.

Women aged 45-64 in 78 general practices in the city of Edinburgh were followed up for five to seven years and all cause mortality noted. Standardised mortality ratios were calculated for the individual practices. Postcodes were available for a 20% sample of these women and were used to retrieve relevant measures of social class and deprivation from the 1981 census for the smallest division, the enumeration district. Weighted averages gave socioeconomic variables at the level of the general practice. High positive correlations were found between standardised mortality ratios and the socioeconomic variables, with the highest being for percentage overcrowding. This study established that the relation between deprivation and excess mortality can be shown in general practices in one large city and gave a direct relation for women without reference to their husbands' occupations, thus obviating problems of assigning social class. The data also partially refute the "social drift" hypothesis as an explanation of the association between mortality and social class.

For randomised population trials the unit of randomisation is normally the individual person. In some situations, however, investigators take other groups as basic unit and one such design is cluster randomisation. Considerable attention has been given to this design recently in statistical and epidemiological literature. The Edinburgh randomised trial of breast cancer screening is an example which takes general practices as clusters of patients. The experience of this trial is reported here. Mortality from all causes, cardiovascular disease and lung cancer over the first 5 year period of follow up are examined. We found that spurious mortality differences were present in the analyses, which do not allow for socio-economic status. From examination of methods of adjusting for this, we conclude that allowance can be made at the analysis stage, and it is intended that this approach will be adopted when breast cancer mortality is analysed in the Edinburgh trial. Nevertheless, we recommend that for future studies with outcome related to socio-economic status, randomisations which use this design be stratified by socio-economic criteria where this is feasible.

Day and Walter derived methods of joint maximum likelihood estimation for the sojourn time distribution and the false negative rate for a screening programme. Their methods are not directly applicable to a programme which uses alternate screening by two modalities whose sojourn times and false negative rates will differ. A modification is proposed and the results applied to data from the Edinburgh Randomised Trial of Breast Cancer Screening. This enables the effects of mammography and clinical examination to be separated. It is estimated that in a programme using both modalities 79 per cent of tumours arising in regularly screened women would be detected by screening and if the clinical examination were omitted this figure would be reduced by 5 per cent. The confidence intervals are, however, quite wide.

There have been many studies of individual risk factors for breast cancer; most of the factors concerned may be broadly grouped into demographic and dietary, reproductive history, endocrine related, family history of breast cancer, and previous history of breast disease. Some of these studies have examined the combined effect of these factors. The present case-control study does this in the context of a randomised controlled trial of breast cancer screening. The relative risks that we have obtained are, in general, of similar magnitude to those in other reports. The relevance of the results to a screening programme is discussed.

It is usually accepted that an older age of menopause is associated with an increased risk of breast cancer. This is often interpreted as a statement about postmenopausal women; however, some authors, eg ref. 1, explicitly and others, eg ref. 2, implicitly take it as embracing both an increased risk for postmenopausal women who have had a late menopause and an increased risk for older premenopausal women who are still menstruating. We have recently examined the role of menstrual status and age at menopause in a study of risk factors in a population of screened women.3 Our results there and the difficulties we encountered in relating them to the literature have motivated this study. We begin with a survey of the literature: comparison of reported studies is difficult because of considerable variations in the definitions of menstrual status (as well as absence and vagueness of definitions), the frequent absence of any "menopausal" category, and a lack of age-specific figures. The only clear consensus is that there is a higher risk among women aged 50-54 who are still menstruating than among those who are postmenopausal. Later we propose a modification of the "two linear component" model for age incidence: this includes increased incidence at the time of the menopause and a subsequent deficit. The intention is to test this model using data from the Edinburgh Breast Screening Trial, and the suitability for this purpose of the data which are being collected is discussed.

In view of the possible introduction of screening programmes, this study compares oestrogen receptor (ER) levels in a series of women whose primary tumour was detected by screening and an age-matched consecutive series of women whose tumours were diagnosed after symptomatic presentation. Because of missing data and other statistical considerations, the comparison was made using T1 and T2 categories of tumour only. Some differences were found: the distribution of ER levels was significantly different in the two groups, with more extreme values in the symptomatic series; the screening series, however, had more moderate/rich ER levels than the symptomatic group. Tumours of special pathological type (for example, tubular, cribriform, lobular, medullary, and mucoid) were more likely to be ER-moderate or -rich, and there were more of these tumours in the screening series. The relationship of these findings to tumour growth rate is discussed. The study highlights the difficulty of obtaining sufficient tissue for conventional DCC biochemical assays from the small non-invasive tumours found by screening, and suggests that newer alternative methods employing monoclonal antibodies may be required for such types of tumour.

The EUROCLUS project included information on residence at diagnosis for 13351 cases of childhood leukaemia diagnosed in the period 1980-89 in defined geographical regions in 17 countries. A formal algorithm permits identification of small census areas as containing case excesses. The present analysis examines spatial-temporal patterns of the cases (n = 970) within these clustered areas. The objectives were, first, to compare these results with those from an analysis conducted for UK data for the period 1966-83, and, second, to extend them to consider infant leukaemias. A modification of the Knox test investigates, within the small areas, temporal overlap between cases in a subgroup of interest at a putative critical time and all other cases at any time between birth and diagnosis. Critical times were specified in advance as follows: for cases of acute lymphoblastic leukaemia aged 2-4 years, the 18-month period preceding diagnosis; for cases of total leukaemia aged 5-14 years, 1 year before to 1 year after birth; and for infant cases (diagnosed < 1 year), 1 year before to 6 months after birth. Each of the analyses found evidence of excess space-time overlap compared with that expected; these were 10% (P = 0.005), 15% (P= 0.0002) and 26% (P= 0.03) respectively. The results are interpreted in terms of an infectious origin of childhood leukaemia.

Abstract 269 Introduction: Markers in the human leukocyte antigen (HLA) class I locus and infectious mononucleosis (IM) have both been associated with Epstein-Barr (EBV)-positive Hodgkin lymphoma (HL). IM is caused by late infection with EBV. To further explore associations between HLA class I genes and EBV-positive HL and to investigate potential interactions between genetic (HLA) and environmental risk factors in the development of HL, we performed a case-only study of 934 HL cases. Methods: HLA-A and B typing were performed by PCR sequence specific oligonucleotide assay. Logistic regression was used to estimate odds ratios (ORs) of EBV-positive HL relative to EBV-negative HL for HLA-A and B genotypes and other known risk factors. Results: Increasing age, male sex, HLA-A*01 and a previous history of IM were all associated with a statistically significant increased odds of EBV-positive HL while HLA-A*02 and HLA-B*07 were associated with a protective effect. Significant interactions were detected between HLA-A*01 and HLA-B*07 and between prior IM and HLA-A*02, such that HLA-B*07 abolished the increased odds associated with HLA-A*01 and HLA-A*02 abrogated the increased odds associated with prior IM. In our final model of EBV-positive HL, HLA-B*07 was associated with a dominant protective effect (OR=0.57, 95% confidence interval (CI): 0.34–0.94); HLA-A*02 was associated with a protective effect which was additive (for individuals with no prior IM, ORper allele =0.69, 95% CI: 0.50–0.97); and HLA-A*01 was associated with an increased odds in patients who were HLA-B*07-negative (ORper allele=2.15, 95% CI: 1.54–3.01). Prior IM was independently associated with EBV-positive HL (OR=3.35, 95% CI: 1.67–6.73). This resulted in a >32-fold higher odds of EBV-positive HL in HLA-A*01 homozygotes with IM compared to HLA-A*02 homozygotes without IM. Conclusions: We have identified strong associations between both genetic and environmental risk factors and EBV-positive HL. Our findings suggest that HLA class I-restricted, cell-mediated responses to EBV and a prior history of IM play critical roles in the pathogenesis of EBV-positive HL. Disclosures: No relevant conflicts of interest to declare.

Background: The Collaborative Group on Hormonal Factors in Breast Cancer has brought together and reanalysed the worldwide epidemiological evidence on the relation between breast cancer risk and use of hormonal contraceptives. Methods: Individual data on 53 297 women with breast cancer and 100 239 women without breast cancer from 54 studies conducted in 25 countries were collected, checked, and analysed centrally. Estimates of the relative risk for breast cancer were obtained by a modification of the Mantel-Haenszel method. All analyses were stratified by study, age at diagnosis, parity, and, where appropriate, the age a woman was when her first child was born, and the age she was when her risk of conception ceased. Findings: The results provide strong evidence for two main conclusions. First, while women are taking combined oral contraceptives and in the 10 years after stopping there is a small increase in the relative risk of having breast cancer diagnosed (relative risk [95% Cl] in current users 1·24 [1·15-1·33], 2p

The interpretation of reports of clusters of childhood leukaemia is difficult, first because little is known about the causes of the disease, and second because there is insufficient information on whether cases show a generalized tendency to cluster geographically. The EUROCLUS project is a European collaborative study whose primary objective is to determine whether the residence locations of cases at diagnosis show a general tendency towards spatial clustering. The second objective is to interpret any patterns observed and, in particular, to see if clustering can be explained in terms of either infectious agents or environmental hazards as aetiological agents. The spatial distribution of 13351 cases of childhood leukaemia diagnosed in 17 countries between 1980 and 1989 has been analysed using the Potthoff-Whittinghill method. The overall results show statistically significant evidence of clustering of total childhood leukaemia within small census areas (P=0.03) but the magnitude of the clustering is small (extra-Poisson component of variance (%) = 1.7 with 90% confidence interval 0.2-3.1). The clustering is most marked in areas that have intermediate population density (150-499 persons km[-2]). It cannot be attributed to any specific age group at diagnosis or cell type and involves spatial aggregation of cases of different ages and cell types. The results indicate that intense clusters are a rare phenomenon that merit careful investigation, although aetiological insights are more likely to come from investigation of large numbers of cases. We present a method for detecting clustering that is simple and readily available to cancer registries and similar groups.