Incidence of prostate and urological cancers in England by ethnic group, 2001-2007: a descriptive study

Urological cancers account for about 14 % of cancers diagnosed globally and more than a fifth of all cancers in Europe [1]. There is also significant international variation in incidence and the aetiology of urological cancers remains poorly understood. Identifying the extent of ethnic variation can contribute to our understanding of aetiology and assist in planning care for different ethnic groups. Unfortunately international comparisons are of limited value as registration systems vary in their quality; there are systematic variations between health systems and systematic biases exist in the way different populations access care [2].

The UK is a multi-ethnic society, with ‘non-White’ ethnic groups making up around 14 % of England's population in 2011. British (South) Asians - Indians, Pakistanis and Bangladeshis—form the largest group of about 6 %, and British Blacks - Black Africans (mainly from Nigeria, South Africa, Ghana and Somalia) and Black Caribbeans (predominantly from Jamaica)—are second at about 3 %, with Chinese (mainly from Hong Kong) about 1 % [3]. Studies have shown South Asians in the U.S. to have lower rates of kidney, bladder, prostate and testicular cancer than Whites and Blacks [4]. South Asians, however, are a heterogeneous group with varied socio-cultural practices and the risk of urological cancers within each individual ethnic group is unknown. Further, these data do not consider socioeconomic status, and are therefore subject to under-reporting in ethnic groups with decreased access to care [5].

Although it has long been known that there are differences in the incidence of many cancers by ethnic group [6] and in access to healthcare (including screening) due to socioeconomic disadvantages [7], studies of cancer incidence in ethnic groups in the UK have been of limited accuracy in the past due to the incomplete ethnicity data held by cancer registries. Various techniques have been used to try and overcome this problem, including using country of birth, the calculation of proportional incidence ratios and assigning ethnicity on the basis of name [8–10]. However, all these methods have significant limitations and the most accurate method is to use self–assigned ethnicity (as has been done in the census since 1991) which allows us to use the same method of assigning ethnicity in the numerator and denominator.

From 1995, self-assigned ethnicity has been recorded in the National Health System’s Hospital Episodes Statistics (HES) database, and HES records can now be linked to cancer registrations, providing more reliable information on ethnicity [11]. Although the recording of routine ethnicity data in primary care is still limited [12], hospital data is much better and has improved markedly in the last 20 years, with the percentage of missing ethnicity values falling from 35 % in 1998 to less than 10 % by 2009 [13]. In England, consistency of diagnostic methods, reporting and registration procedures across the entire health system removes significant biases intrinsic to databases in many other countries.

Our objective was to compare the incidence of kidney, bladder, prostate and testicular cancer amongst ethnic groups (Indian, Pakistani, Bangladeshi, Black African, Black Caribbean and Chinese) in England, to each other and to Whites.

The methods used in this study were broadly the same as those described in our previous studies [14–16] and are summarized below.

We compared incidence rates for urological cancers in the main ‘non-White’ ethnic groups in England – South Asian (Indian, Pakistani and Bangladeshi), Black (African and Caribbean) and Chinese to Whites and to each other. There was considerable variation by ethnic group, even when age and socioeconomic factors were taken into account. Overall, urological cancers were diagnosed less often in all the ‘non-White’ ethnic groups, except prostate cancer in Blacks, which demonstrated a higher incidence than in Whites. Amongst South Asians, we demonstrated that the incidence of urological cancers substantially differed between Pakistanis, Bangladeshis and Indians, supporting the notion that South Asians should not be viewed as a single ethnic group. Similarly, large differences existed between Black African and Caribbean populations, highlighting the need to differentiate between these two groups.

The different patterns of cancer risk between ethnic groups suggest that our findings are unlikely to be due to systematic reporting biases in any of the ethnic groups compared to Whites. Our previous work using the same dataset showed increased risks of some gastrointestinal and haematological cancers in ethnic minority groups further supporting the absence of an under-reporting bias [14, 15]. Using self-assigned ethnicity is also more reliable than other measures of ethnicity (e.g. name analysis) as it uses the same measure of ethnicity in the numerator and denominator. We also adjusted for socioeconomic status, a potential confounder in studies of health and ethnicity, particularly when comparing Pakistanis, Bangladeshis and Blacks due to their higher levels of deprivation [5].

Our findings are consistent with current literature although we are not aware of any previous studies which present incidence by individual ethnic group for kidney, bladder and testes cancers. (which grouped all ‘Blacks’ and all ‘South Asians’ together.)

For renal cancer, a previous study showed that South Asians had a lower incidence of renal cell carcinoma compared with Whites, consistent with our results [24]. Smoking is a known risk factor for renal cell carcinoma [25] and our results are consistent with smoking prevalence by ethnic group [26]. Some reports have demonstrated higher incidence of renal cell carcinoma in African Americans than Whites [27], contrary to our findings. This discrepancy may be attributable to the fact that these studies were done in the USA where black populations may have different countries of origin.

For bladder cancer, previous reports indicate a lower incidence amongst south Asians [28]. Again, smoking is a known risk factor for bladder cancer [29] and our results are consistent with smoking prevalence by ethnic group with the exception of Chinese [26] who have the lowest smoking prevalence but the highest bladder cancer incidence. This may be due to genetic factors or other ethnicity-specific risk factors (e.g. dietary soy) which is frequently used in Chinese cuisine, has been associated with increased bladder cancer risk [30].

For testicular cancer, previous studies have also revealed a lower incidence amongst Asians and Blacks [31] consistent with our data. This may be due to inter-ethnic variations in environmental factors acting prenatally or early in childhood [32]. Cryptorchidism, a known risk factor of testicular cancer, may also vary between ethnicities, with reports of reduced incidence amongst black babies [33] but data is not available for South Asians.

For prostate cancer, studies in the USA have shown increased incidence in men with African ancestry, even after migration to areas of lower prevalence [34, 35]. Our finding that Black African and Black Caribbean’s demonstrated a higher incidence of prostate cancer than Whites is also consistent with previous UK studies [28, 31].

The specific cause of increased prostate cancer risk amongst Blacks is not known. A recent review of known risk factors for prostate cancer found limited environmental explanation for the racial differences in incidence [36]. Nonetheless, there have been a number of dietary factors which have been independently implicated, including intake of animal fats and products [37]. Increased risk amongst Blacks has also been attributed to genetic factors including: variants of the genes of the enzymes involved in androgen biosynthesis and metabolism, such as SRD5A2, CYP17, and CYP3A4; the C-genotypes CD14; NAT2 and NER genetic variants; and polymorphisms at 17q21 and 8q24 [38–42]. Although the much higher rates seen in Black Africans and Caribbean in the UK compared to their regions of origin is consistent with a change of environment causing the increase, the very low rates seen in sub-saharan Africa almost certainly underestimate the true incidence due to under-diagnosis and under-reporting. We also note that rates are generally higher in black populations everywhere—sub-Saharan Africa and the Caribbean, (where they are almost double compared to other similarly developed regions) [1], Blacks in the US and Blacks in the UK. This suggests that the increased risk seen in Blacks in the UK is most likely due to the change in environment in genetically susceptible populations.

The role of genetics in determining variation in risk is reinforced by the observation that both black Africans and Caribbeans display increased prostate cancer incidence, despite different countries of origin; lifestyles and environments.

Consistent with our findings, previous reports have also demonstrated reduced prostate cancer incidence in Asians compared with Whites [28], and within the subset of Asians: Indians, Pakistanis and Bangladeshis [43–47]. The reduced incidence amongst South Asians has been associated with religion, differences in diet, reduced vitamin D levels and early-life sun exposure [48, 49]. Other risk factors associated with prostate cancer are lack of exercise, antioxidants and saturated fat, smoking, alcohol consumption, socioeconomic status and carcinogen exposure (e.g. radiation and arsenic) [50–53] and it is possible that differences in these explain some of the reduced incidence in South Asians. Further, it has been suggested that South Asians meet with more obstacles when accessing health care resources, receiving less diagnostic and screening (PSA) tests [54]. Within South Asian ethnic groups, Bangladeshis displayed a substantially lower incidence of prostate cancer; although the cause of this is likely multifactorial, obesity has been linked to prostate cancer [55], and Bangladeshis show markedly lower obesity prevalence compared to Indians and Pakistanis [26]. Prostate cancer incidence in Chinese was higher than in South Asians and this may in part be contributed to by increased soy, genistein and daidzein intake [56].

The comparison of rates between ethnic groups in England and their countries of origin is problematic for a number of reasons. Firstly, population-based cancer registries simply do not exist in many of these countries, particularly in the areas from where the majority of migrants originate e.g. Punjab & Gujarat in India, Kashmir & Punjab in Pakistan, Sylhet in Bangladesh, Jamaica in the Caribbean and Somalia, Nigeria & Ghana in Africa. and Even where registries exist the quality is very variable and there are differences in cancer registration practices [2]. Rates in these developing countries are also likely to be underestimated due to under-diagnosis and under-ascertainment, and in access to screening and early detection, particularly for prostate cancer [2]. Migrants are also a selective group and may not be representative of the population from which they arose and they may be more or less healthy than the population in their native country [6].

Whilst acknowledging these limitations, the increased incidence of prostate, kidney and testicular cancers in all ethnic groups in England compared to their countries of origin is consistent with changes in environmental risk factors and the differences in screening, diagnosis and registration. For bladder cancer the picture was more mixed with higher incidence seen in the countries of origin of Pakistanis, Black Caribbean and Chinese—the reasons for this are unclear.

The main limitation of our study was that we do not have individual level information on most exposures. Ethnicity information was missing for 25 % of cancer registrations—however, this is lower than previous studies [28, 44] and the sensitivity analysis produced similar results. Recording of ethnicity in HES has improved markedly in the last 20 years, with the percentage of missing ethnicity values falling from 35 % in 1998 to less than 10 % by 2009. The quality of the ethnic coding in HES has also been assessed and no ethnic group is widely misrepresented in HES data for England [13, 14].

Also, the group ‘British White’ inevitably included some ‘Other (non-British) Whites’ as the ethnic category ‘Whites’ included both British Whites and ‘Other Whites’ prior to 2003 although this was less than 5 % of the ‘total White’ category based on the data post-2003 so would not materially affect the results for British Whites. We were therefore also unable to compare British Whites to ‘Other Whites’, and given that the ‘Other White’ population has increased rapidly since 2004 (due to migration from the European Union), future studies should look at incidence in this group as well.

This is the first and largest study to investigate the differences in incidence of urological cancers by ethnic group in England. The very high incidence of prostate cancer in both black Africans and Caribbeans, despite differing lifestyles and environments and countries of origin is most likely due to the change in environment in genetically susceptible populations. The large differences in incidence we found with other cancers by ethnic group are not readily explained by known risk factors, which suggest that important, potentially modifiable, risk factors are yet to be discovered.