 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
| |
|
||||||||||||||||
| |
|
||||||||||||||||
|
|
 Cancer incidence in young adults in Canada: preliminary results of a cancer surveillance project
Loraine D Marrett, Jennifer Frood, Diane Nishri, Anne-Marie Ugnat and the Cancer in Young Adults in Canada (CYAC) Working Group Abstract Surveillance of cancer in young adults has been neglected, despite Sir Richard Doll's having emphasized its importance a decade ago. This report describes the patterns, time trends and regional variation in cancer incidence in Canada's young adults. In 1987-1996, 97,469 cancers were diagnosed in Canadians aged 20-44, with almost two-thirds in females. Ten types of cancer accounted for 83% of diagnoses in women and 74% in men. The most common cancers in young women were breast, cervix, melanoma, thyroid and ovary, and in young men were testis, non-Hodgkin's lymphoma, melanoma, colorectal and lung. Although incidence rose only slightly for total cancer between 1969 and 1996, it increased dramatically for several specific types of cancer: lung (women), melanoma, testis, thyroid and non-Hodgkin's lymphoma. Incidence declined for a few cancers (colorectal, lung (men), cervix and ovary). Lung cancer incidence was significantly lower than the Canadian average in Prairie women and non-significantly high in Quebec (both sexes), while the rate of melanoma was significantly low in Quebec (both sexes) and high in women in the Pacific region. Key words: incidence; neoplasms; surveillance; time trends
Introduction Although a diagnosis of cancer in young adulthood (ages 20-44 years) is a relatively rare event, the consequences of such a diagnosis are great: at the time they are diagnosed with cancer, these individuals have most of their potential years of life ahead of them, and so may either spend decades living with the effects (physical, reproductive, social, emotional and spiritual) of cancer diagnosis and treatment or have tragically shortened lives, with major repercussions on their families and on society in general. Surveillance of cancer patterns and trends in young adults has been neglected, despite considerable surveillance activity in both children and the population as a whole. Sir Richard Doll, in his plenary address at the meeting of the Society for Epidemiologic Research in Buffalo, New York in 1991, emphasized the importance of surveillance in this age group: "The trends in young adults are, I suggest, by far the most important for assessing our progress against cancer for two reasons. First, because the trends can reflect only relatively recent changes in the prevalence of carcinogenic agents and are not confused by the effects of changes in the distant past and, second, because young people tend to adopt new habits before the old."1 This report presents preliminary data from a project being undertaken by the Cancer in Young Adults in Canada (CYAC) Working Group (members are listed in the Appendix). The main purpose of the project is to describe cancer patterns in those aged 20-44 in Canada, including identification of the most common forms of cancer and description of the trends over time and regional variation in the incidence of these common cancers. Materials and methods Materials Cancer incidence data for the period 1969-1996 were obtained from Health Canada. These data originate with the provincial and territorial cancer registries, but have been provided to Statistics Canada to form the National Cancer Incidence Reporting System (1969-1991) and the Canadian Cancer Registry (1992-1996).2 Data in the Canadian Cancer Registry are internally linked at Statistics Canada so that patients registered with the same diagnosis in more than one province are counted only once. Health Canada receives a copy of this file without nominal information. Health Canada provided frequencies by year of diagnosis, sex, five-year age group at diagnosis, region and type of cancer for those aged 20-44 at diagnosis. Cancer type is coded according to the 9th Revision of the International Classification of Diseases (ICD-9).3 (Note that non-melanoma skin cancers, ICD-9 code 173, are not included.) Canada has been divided into six regions: Pacific (British Columbia); Prairie (Alberta, Saskatchewan, Manitoba); Ontario; Quebec; Atlantic (Nova Scotia, New Brunswick, Prince Edward Island, Newfoundland); and North (Northwest Territories and Nunavut, Yukon). Health Canada also provided corresponding population data.4 Methods Cancer types for each sex were ranked on the basis of the number of cases diagnosed during the most recent 10-year period, 1987-1996. Those occurring most frequently were included in more detailed analyses. Incidence rates were age-standardized in five-year age groups (20-24, 25-29, 30-34, 35-39, and 40-44) to the age distribution of the 1991 Canadian population.5 Trends in incidence rates over the 28-year period 1969-1996 were examined for all of Canada by sex for total cancer, all types of cancer affecting both sexes (total cancer minus reproductive and male and female breast cancers), and each of the most common cancer types. In order to estimate the average annual percent change (AAPC), the logarithm of the annual age-standardized rate was modelled as a function of the year of diagnosis using linear regression and SAS PROCs REG and GLM.6 Both linear and quadratic terms were included in the initial model. If the quadratic term was non-significant (p>0.05), the linear model was assumed to be adequate and the AAPC was estimated by the formula AAPC = (exp(beta)-1) * 100; 95% confidence limits were likewise calculated by transforming the confidence limits for the estimated slope. When a significant quadratic term suggested that the trend over time was not linear, the AAPC was not estimated. Trends were displayed graphically using three-year moving average rates. Regional variation in incidence was examined by calculating age-standardized rates by sex, cancer type and region for the period 1987-1996. A regional rate was considered significantly different from that for all provinces and territories combined if its 95% confidence interval (calculated using a binomial approximation) excluded the all-Canada rate. Except for total cancer, the North was excluded from the regional analysis because of small numbers for many of the cancer types; the North reported only 270 cancers diagnosed in young adults in this decade. Results Common cancers Between 1987 and 1996, 97,469 cancers were diagnosed in young adults. Cancer occurred about twice as often in women as in men aged 20-44: there were 60,803 cancers diagnosed in young adult women and 36,666 in young adult men (Figure 1). Restricted to those cancers that can occur in both sexes (i.e., excluding cancers of the reproductive system and of the female breast), the numbers were closer, with a slight excess of men (28,426 women; 31,165 men). The most common cancers for each sex are also shown in Figure 1. Ten types accounted for 83% of the cancers in young women and 74% of those in young men. The female breast was by far the most common site of cancer (n = 20,680), representing 21% of the cancers diagnosed in both sexes combined and 34% of cancers in women. Several of the other top-ranked cancers arose in the reproductive system:
cervix and ovary (ranks 2 and 5 respectively) in women and testis (rank
1 in men). Although the ranks differed between sexes, some cancers were
common in both men and women: non-Hodgkin's lymphoma, melanoma, Hodgkin's
disease, and cancers of the colorectum, lung and brain. Thyroid cancer
was much more common in women (n=4,562) than in men (n=1,206, not shown). |
||||||||||||||||||||||||||||||||||||||||||||
|
FIGURE 1
a Non-melanoma skin cancers not included. Surface areas of circles are proportional to the numbers of cancers. b Bone and connective tissue. |
|||||||||||||||||||||||||||||||||||||||||||||
|
Trends over time Cancer incidence increased slightly over the period 1969-1996 in both young men (0.66% per year) and young women (non-linear trend) (Figure 2 and Table 1). The consistent and substantial female excess for total cancer is evident in Figure 2. However, when limited to cancers that can occur in both sexes, rates were similar for men and women although there was a slight excess in men. Rates have been increasing significantly for this subgroup of cancer types (0.56% per annum in males and 1.23% in women). Figure 3 presents trends for the most common cancers, while Table 1 shows the results of the regression analysis. Perhaps the most striking feature of these trends is the strong increase evident for a number of cancers. The increase was linear and more than 2% per year for testicular cancer (2.73% per year); thyroid cancer in men (2.83% per annum); and non-Hodgkin's lymphoma in both sexes (3.69% in men and 2.68% in women, per annum). For some additional types of cancer, substantial increases occurred over the time period (Figure 3: lung, women; melanoma, both sexes; thyroid, women) but because the rate of change was not consistent over time, a single AAPC could not adequately summarize the trend. Of these, it is encouraging to note that recent melanoma trends are either flat or downward and that the increase in lung cancer incidence in women has slowed in recent years. Significant linear decreases occurred for colorectal cancer (both sexes), lung cancer (men only) and cancer of the ovary, although a downward trend is also evident for cervical cancer (non-linear because of a recent slowing in the rate of decline). Although the trend for breast cancer is also non-linear, incidence has been stable over the past decade. Figure 3 also permits description of the male to female rate ratios. While incidence was consistently higher in males for non-Hodgkin's lymphoma, females had higher rates of melanoma and much higher rates of thyroid cancer. Over time, male and female incidence rates have converged for lung and colorectal cancers and diverged somewhat for non-Hodgkin's lymphoma and thyroid cancer. FIGURE 2
a Standardized to the 1991 Canadian population age distribution. b Excludes non-melanoma skin cancers TABLE 1
|
|||||||||||||||||||||||||||||||||||||||||||||
|
FIGURE 3
a Standardized to the 1991 Canadian population age distribution. Note that the scale is different for the breast, cervix and ovary cancer graph.
|
|||||||||||||||||||||||||||||||||||||||||||||
|
Regional variation There was virtually no variation in total cancer incidence across the country, although rates were slightly lower in the North and higher in Ontario (Figure 4). There were a few striking regional differences in site-specific incidence. Quebec had higher rates of lung cancer (non-significant) and lower incidence of melanoma (significant) for both men and women. The Prairies had low rates of lung cancer, significantly so in women, while the Pacific region had high rates of melanoma (significant in women). Ontario's females had considerably higher rates of thyroid cancer than those in other regions (non-significant); a slight excess was evident in Ontario men as well. There was an apparent west- to-east gradient in colorectal and ovarian
cancer incidence while the reverse was true for testicular cancer and
for melanoma (with the exception of the very low rates in Quebec). |
|||||||||||||||||||||||||||||||||||||||||||||
|
FIGURE 4
a Standardized to the 1991 Canadian population age distribution. b Non-melanoma skin cancers not included. * Significantly different from the Canadian rate (p < .05). Note that scales are different for total cancer and breast cancer.
|
|||||||||||||||||||||||||||||||||||||||||||||
|
Discussion Overall, cancer occurs relatively infrequently in young adults: diagnoses in this group accounted for only 8.7% (10,331/118,631) of all newly diagnosed cancers in Canada in 1996.7 In contrast, about 1% of cancer diagnoses occur in 0-19 year olds and 90% in those aged 45 and over. The types of cancer that occur most often in young adults and their relative frequencies are different from those in both older adults and children. They represent a mix of cancers common in children and adolescents (brain cancer, Hodgkin's disease, non-Hodgkin's lymphoma), those very common in older adults (breast, colorectal and lung), and some which are not particularly common in either (melanoma, testis, thyroid). The sex ratio is also very different from that for cancers at other ages. While there is a male excess of cancer both in childhood/adolescence and after age 60,4 females have a striking excess of cancer in young adulthood, with a female-to-male incidence rate ratio of about two. The female excess is explained almost entirely by the much greater frequency of breast and reproductive system cancers in women compared with men. Several of the cancers that occur commonly in this age group have been increasing in incidence over the nearly 30-year time span of these data, some dramatically (melanoma, thyroid cancer and non-Hodgkin's lymphoma in both sexes, lung cancer in women and testis cancer in men). The reasons for some of these increases have been established (e.g., increasing sun exposure for melanoma and increased smoking for female lung cancer) but for some (e.g., testis cancer), the reasons are unknown). Although a few strong regional patterns are evident, caution is required in interpreting regional differences because Canada's cancer registry is really an amalgam of separate province-level registries, each with its own operating procedures, supportive legislation and registration rules. As Sir Richard Doll has made clear,1 cancer surveillance in young adults may be particularly informative. Cancers in this age group may represent "sentinel events", providing warning of the effects of new or changing exposures/behaviours, including early adoption of protective behaviours. Thus, examination of cancer trends and patterns in this age group may alert the researcher to the need to seek information on possible explanatory exposures/behaviours, and may even provide clues to what these might be. This will be the focus of future study by the CYAC Working Group. Further, for cancers that are also common in older age groups, trends in young adults may forecast future trends in older adults. For example, one possible interpretation of the decline/stabilization of melanoma incidence in this age group over the most recent decade is that the young have been adopting "sun smart" behaviour that has translated into an end to the increase in the incidence of melanoma. One can speculate that melanoma rates may stabilize and eventually decline in future decades in the older age groups, where to date incidence has continued to increase as the current generation of young adults ages. Conclusions and future directions Despite Sir Richard Doll's recommendation a decade ago, there has been virtually no systematic surveillance of cancer patterns and trends in young adults. The CYAC Working Group was established to rectify this situation, at least in the Canadian context. At a workshop in Toronto in October 2000, the Working Group adopted a protocol for a study to describe cancer in young adults in Canada over the past three decades. The primary objectives of the project are to describe the most important forms of cancer in young men and young women in Canada at the present time; to document incidence and mortality trends from these cancers and for important histologic subtypes thereof, where appropriate; and to interpret the trends in terms of likely responsible risk factors/exposures. This paper presents preliminary data only for the first and part of the second of these objectives. In the future, the CYAC Working Group will focus on the remaining objectives. In particular, its members will be conducting literature reviews and searching for sources of data to help address hypotheses suggested by incidence patterns and the literature, with the ultimate goal of making recommendations for public health actions and priority research. The Working Group will also be refining and updating the results presented here using data provided directly by the provincial cancer registries. This will result in more current and more comparable data than presented herein, and will permit investigation and understanding of data anomalies or artefacts that might affect interpretation of trends and regional variation. For example, rules for what constitutes a second primary cancer differ between registries; unless standardized, variations in rules could result in artefactual regional differences for some cancers, particularly those of the breast, skin (melanoma) and colon where multiple cancers of the same organ are relatively common. In addition, quality and completeness of cancer data vary across registries and across time. The individual registries are in the best position to help the Working Group understand the changes and limitations of their data, as these relate to interpreting regional patterns and trends. Acknowledgements Health Canada's Centre for Chronic Disease Prevention and Control provided funds for the workshop held in Toronto in October 2000 at which the CYAC Working Group developed the protocol for this project. Thanks to Mrs. Gini Hunter and Ms. Sheila Wing for their help in organizing a successful workshop. References
Author References Loraine D Marrett, Jennifer Frood, Diane Nishri, Division of Preventive Oncology, Cancer Care Ontario, Toronto, Ontario Anne-Marie Ugnat, Surveillance and Risk Assessment Division, Centre for Chronic Disease Prevention and Control, Health Canada, Ottawa, Ontario The Cancer in Young Adults in Canada (CYAC) Working Group: see Appendix Correspondence: Loraine D Marrett, Division of Preventive Oncology, Cancer Care Ontario, 620 University Ave., Toronto, Ontario, M5G 2L7; Fax: (416) 971-6888; E-mail: loraine.marrett@cancercare.on.ca
[Previous] [Table of Contents] [Next] |
|||||||||||||||||||||||||||||||||||||||||||||
 |
|||
|
Last Updated: 2002-06-20 |  |
|
