Cancer and congenital abnormalities in Asian children: a population-based study from the West Midlands

British Journal of Cancer (1995) 72, 1563-1569 © 1995 Stockton Press All rghts reserved 0007-0920/95 $12.00 Cancer and congenital abnormalities in Asian children: a population-based study from the West Midlands JE Powell', AM Kelly', SE Parkes', TRP Cole2 and JR Mann' 'West Midlands Regional Children 's Tumour Research Group, The Children 's Hospital, Ladywood Middleway, Ladywood, Birmingham B16 8ET, UK; 2Department of Clinical Genetics, Birmingham Maternity Hospital, Queen Elizabeth Medical Centre, Birmingham, B15 2TH, UK. Summary Cancer and associated congenital abnormalities were investigated in Muslim and non-Muslim Asian children from the West Midlands. Cancer incidence rates were calculated for Indian (non-Muslim), Pakistani/Bangladeshi (Muslim) and white children diagnosed from 1978 to 1992. Incidence was significantly higher in the Pakistanis, with an age-standardised rate (ASR) of 163 cases per million per year, compared with 115 for Indian and 125 for white children. Among Asian cancer patients, congenital malformations were significantly more common in Muslim (21%) compared with non-Muslim (7%). In Muslims the malformation excess was caused by autosomal recessive and dominant disorders (in 8% and 5% of cases respectively). Cancer malformation/predisposition syndromes were found in 10% of Muslims, compared with 2% of non-Muslims. In 33% of the Muslims with malformations, childhood cancer and a malformation were also present in a close relative. None of the non-Muslims with malformations had a relative with childhood cancer. The cancer excess in Muslims may be partly related to inherited genes causing both malformations and cancer. The prevalence of autosomal recessive disorders may be related to consanguinity, which is common in the Pakistani Muslim population. The high incidence of autosomal dominant disorders may be related to older paternal age at conception, giving rise to spontaneous mutations. Keywords: childhood cancer; Asians; malformation; consanguinity; Muslim Worldwide variation in the pattern and incidence of childhood cancer in different ethnic populations is a wellestablished phenomenon (Parkin et al., 1988) that offers insight into the causes of the disease. In particular, the study of immigrant populations (Goodman et al., 1989a, b) may uncover environmental factors in the aetiology of diseases in which the disease pattern changes over time towards that seen in the indigenous population. In the UK, ethnic minorities now constitute up to 27% of the population in some metropolitan boroughs (Haskey, 1991). However, research into childhood cancer in these groups has been limited (Stiller et al., 1991), partly because of the difficulties of obtaining ethnic population data and partly because ethnic group is not routinely recorded in cancer registrations. Incidence rates have recently been calculated (Powell et al., 1994) for cancer in white and Asian children living in the West Midlands Health Authority Region, where nearly 10% of all children belong to Asian ethnic minorities. These revealed a different pattern of cancers in the two groups and suggested that Asian children were at significantly greater risk of cancer (relative risk 1.43, 95% CI 1.05-1.93) than white children. This excess may reflect differences in social conditions and/or genetic influences. Senior and Bhopal (1994) have observed that the word 'Asian' is too broad a term to classify migrants from the Indian subcontinent, who are culturally as well as geographically diverse. Furthermore, some groups have been settled in this country for a longer period and have become more westernised. In the West Midlands, Pakistanis and Bangladeshis are the most recent Asian immigrants, and compared with the Indians are of lower social class, often unemployed and have a poor command of English (Bundey et al., 1990). Religion is a factor that may be particularly relevant to malignant disease. Consanguineous marriages are common within some religious groups, particularly Pakistani Muslims (Bundey et al., 1990). High rates of congenital abnormalities are seen in the children of consanguineous Pakistani Muslims (Bundey et al., 1991) and consanguinity may perhaps contribute to the higher cancer incidence in Asian children Correspondence: JE Powell Received 10 January 1995; revised 4 July 1995; accepted 27 July 1995 (Powell et al., 1994). We therefore investigated cancer incidence and congenital abnormalities in Muslim and nonMuslim Asian children living in the West Midlands. Methods The West Midlands Regional Children's Tumour Research Group (WMRCTRG) maintains a population-based register of all cases of childhood cancer diagnosed in residents of the West Midlands Region and also reviews the histopathology of the cases (Muir et al., 1992). The WMRCTRG codes four categories of ethnic origin; these are 'White Caucasian', 'Asian', 'Black' and 'Other', the last category also including children of mixed race. From this register, cancers arising in white and Asian residents of the West Midlands Region between 1978 and 1992 were identified. For the Asian children, more detailed ethnic information (country of origin, religion and consanguinity) was obtained by review of their clinical records. The clinical records of white children were not examined. When hospital notes held insufficient information, and the patient was still living, details were requested from the patient's general practitioner. When religion was not stated, it was determined from an analysis of the names (Henley, 1979). Since details of parental consanguinity were available for only 57% of cases, religion was employed as an indirect measure of consanguinity. First cousin marriages are permitted for Muslims but prohibited for Sikhs and Hindus. A recent study in Birmingham revealed that 69% of Pakistani Muslim women and 30% of Muslims of other races are related to their husbands, while consanguinity in the mainly Sikh and Hindu Indians is as rare as it is in white people (0.7% and 0.4% respectively) (Bundey et al., 1991). By dividing the Asians into Muslims and non-Muslims two populations with respectively a high and low level of consanguinity were produced. Population figures for the West Midlands Region relate to country of origin rather than religion. However, the Asians could still be divided into two groups: the Indian and EastAfrican Asians who are predominantly non-Muslim (Sikh and Hindu); and the Pakistanis and Bangladeshis who are almost entirely Muslim. Ethnic population figures were $ M 1564 Cancer and congenital abnormalities in Asian children JE Powell et al obtained from the 1981 Labour Force Survey (OPCS, 1982), the 1986-88 Labour Force Surveys (Haskey, 1991) and the 1991 census (OPCS, 1992). Population figures for the remaining years were estimated by linear interpolation. Using these figures, age-standardised incidence rates (ASRs), standardised to the world population by the direct method (Boyle and Parkin, 1991), w (...truncated)