Article Text

PDF

Prevalence of family histories of breast cancer in the general population and the incidence of related seeking of health care
  1. C E Jacobi1,
  2. M A Jonker2,
  3. N J D Nagelkerke2,
  4. J C van Houwelingen2,
  5. G H de Bock1
  1. 1Department of Medical Decision Making, Leiden University Medcial Centre, Leiden, The Netherlands
  2. 2Department of Medical Statistics, Leiden University Medical Centre, Leiden, The Netherlands
  1. Correspondence to:
 C E Jacobi MSc, Leiden University Medical Centre, Department of Medical Decision Making (J10-S), PO Box 9600, 2300 RC Leiden, The Netherlands; 
 c.e.jacobi{at}lumc.nl

Statistics from Altmetric.com

Breast cancer is the most widespread cancer in women in western countries, currently accounting for one third of all cancers in women.1,2 The lifetime risk of developing breast cancer is 9% in the United Kingdom,3 10% in The Netherlands,1 and 12% in the United States.4 About 15%–20% of all cases of breast cancer are thought to be familial,5,6 and about 5%–10% of all breast cancers are attributed to the known breast cancer susceptibility genes BRCA1 and BRCA2.7–9

In most European countries nationwide breast cancer screening programmes are available for women of 50 years and older,10 whereas comprehensive screening for breast cancer in women under the age of 50 has been found cost ineffective. It is widely acknowledged that women who have relatives with breast cancer have an increased risk of developing breast cancer themselves.11,12 Consequently, screening for breast cancer might be advisable and cost effective in the subgroup of women under the age of 50 with a familial predisposition to breast cancer. We therefore started a study on the cost effectiveness of breast cancer screening in women under 50 with a familial predisposition to this cancer. For this study we needed information on the distribution of family histories of breast cancer among women under the age of 50 in the general population who might visit a healthcare doctor with concerns about their family history of breast cancer. Several studies have reported information on the percentage of women with a family history of breast cancer among cases of breast cancer and healthy controls.11 However, no such information is available for the general population. Other researchers have made attempts to gain information on the percentage of women in the general population who might be at an increased risk for breast cancer by using a family history questionnaire.13 The response rate to the questionnaire was, however, too low to draw useful conclusions. We have chosen to simulate a general population to gain this information.

The aim of this study was to estimate the prevalence of family histories of breast cancer among women aged 30–50 in the general population and to compute the incidence of related seeking of health care.

METHODS

Design

This study was performed as part of a study on the cost effectiveness of screening women under the age of 50 with a familial predisposition to breast cancer. We needed information on the prevalence and distribution of family histories of breast cancer among women aged 30–50 in the general population. As this information is not directly available, we simulated data with a model and combined them with data available from medical publications.

Simulation of a population including prevalent breast cancer

We simulated a population of 1 million women aged 30–50 and their first and second degree relatives, that is, sibs, parents, maternal and paternal aunts, and uncles and grandparents. Third degree relatives were not included, as previous research has shown that information on breast cancer in third degree relatives gives little additional information on the individual risk of breast cancer.14 The simulated pedigrees were independent and the structures of the pedigrees were based on actual pedigrees of 256 women under 50 years of age who had visited our hospital in the past seven years. The Dutch population statistics15 on age and survival were used as input variables in the simulation model to assure that our simulated population was comparable to the general Dutch population. We additionally compared the simulated population and the Dutch population for family structure, by using the mean number of sibs as found in the simulated population and the Dutch population.15 We found that both populations were comparable, indicating that the 256 actual pedigrees that we used for the simulation may be representative of general Dutch pedigrees.

Key points

  • No information is available on the distribution of family histories of breast cancer among women under the age of 50 in the general population. In this study we simulated a population of women aged 30–50 including their first and second degree relatives to gain this information.

  • One third of the women aged 30–50 years in the population had at least one relative with breast cancer.

  • Of the women with at least one relative with breast cancer, four-fifths had only second degree relatives with breast cancer, and one-fifth had at least one first degree relative with breast cancer.

  • Most women who seek health care because of a family history of breast cancer are those who have high risk pedigrees.

Using a genetic model,16 ancestors of the simulated pedigrees were assigned breast cancer polymorphisms, which were inherited by their offspring according to mendelian rules. The genetic model was created to describe correctly familial clustering of breast cancer by using the available information on the two known breast cancer susceptibility genes BRCA1 and BRCA2 from medical publications, that is, the mutated allele frequencies17,18 and penetrance functions.19,20 Because the BRCA1 and BRCA2 genes only explain a small proportion of the familial clustering of breast cancer,7 an additional gene was modelled accounting for the remaining familial clustered breast cancer using relative risks for breast cancer among women with relatives with breast cancer, which were derived from a large meta-analysis.11 This genetic model assigned the observed breast cancer incidence in The Netherlands1 to the simulated population (fig 1), and also provided us with information on the familial clustering of breast cancer in the general population.

Figure 1

Age specific penetrance of breast cancer among women in the Dutch population and the simulated population.

Prevalence of family histories of breast cancer

In the simulated population, the prevalence of family histories of breast cancer was assessed. For the classification of various risk groups for breast cancer, we determined five different family histories for the women with at least one relative with breast cancer (table 1). Within each of the five groups we determined the percentage of women who had a relative with breast cancer affected under the age of 50 and at the age of 50 or older, to define the family histories more precisely.

Table 1

Family histories of breast cancer in the general population of women aged 30–50 (n=1 000 000)

Incidence of related healthcare visits

To determine the incidence of seeking health care among women with a certain family history of breast cancer, the following variables were abstracted from the published reports: (1) the distribution of various family histories in general practice and (2) the number of attendees in general practice who newly sought health care because of a family history of breast cancer. Two previous studies reported the variety in family histories of breast cancer in general practice, both from a large primary care centre which corresponded to five general practitioner (GP) practices21 and from a sample of 200 GP practices.22 These family histories of breast cancer were registered if a woman visited the GP with concerns about a personal risk of breast cancer. The distribution of the various family histories of breast cancer in general practice, as reported by these studies, is shown in table 2. The number of new attendees with a family history of breast cancer in general practice was obtained through the first study: in the five year period 1997–2001 the number of new attendees in general practice who sought care because of their family history of breast cancer was on average 3.1/1000 women/year (SD 1.8, range 1.4–5.3) (unpublished update to De Bock et al21). In the Dutch healthcare system, which is similar to that of the United Kingdom, practically every person is registered with a general practice, regardless of his or her medical condition. People who seek care usually start at a general practice.

Table 2

Incidence of seeking health care because of a personal risk for breast cancer among unaffected women aged 30–50 with at least one first or second degree relative with breast cancer

We computed the average, the minimum, and the maximum incidence of those seeking health care for the various groups of women with a family history of breast cancer by using the distribution of family histories in general practice and the average (3.1), the minimum (1.4), and the maximum (5.3) number of new attendees combined with the distribution of family histories as found in our simulated population. For each group, the incidence of those seeking health care was proportional to the ratio of the distribution of family histories in general practice and the distribution of family histories in the population, by which the sum of the incidence of the various family history groups resulted in the observed number of new attendees.

RESULTS

Prevalence of a family history of breast cancer

The prevalence of family histories of breast cancer in the population of women aged 30–50 (n=1 million) is presented in table 1. Of all the women in this population, 1.0% (n=0 166) had breast cancer, 64.3% (n=643 082) were unaffected and had no relatives with a history of breast cancer, and 34.7% (n=346 752) were unaffected and had at least one relative with breast cancer.

In the population, 12.7% of the women (n=127 118) had at least one first degree relative or at least two second degree relatives affected with breast cancer, of whom 55.9% (n=71 100) had at least one first degree relative with breast cancer.

Incidence of related healthcare visits

In table 2, we calculated the incidence of those seeking health care among women with various family histories of breast cancer. Women with one second degree relative with breast cancer and no affected first degree relatives comprised 21.96% of the general population and 0.8% in general practice and their incidence of seeking health care was 0.00011 (range 0.00005–0.00018). By contrast, women with one first and two or more second degree relatives with breast cancer or at least two first degree relatives with breast cancer comprised 1.09% of the general population and 16.7% in general practice and their incidence of seeking health care was 0.04888 (range 0.02208–0.08357). Thus, the incidence of seeking health care among women with a strong affected family history of breast cancer is higher than of those with a less affected family history.

DISCUSSION

The present study has shown that almost 35% of the women aged 30–50 have at least one first or one second degree relative with breast cancer, of whom most (63.3%) only have one affected second degree relative. In the other groups of women with relatives with breast cancer (12.7% of all women aged 30–50) the family history of breast cancer might represent an increased risk for the disease. The incidence of seeking health care varies from 0.00011 among women with only one second degree relative with breast cancer to 0.04888 among women with one affected first degree relative and two or more second degree relatives and women with two or more affected first degree relatives, resulting in 3.1 new people seeking health care/year/1000 women aged 30–50 in the population.

Some remarks on our results should be made. Firstly, we have based our estimates on the age specific incidence of breast cancer in The Netherlands. As this varies between countries, the estimates could be slightly different for other populations. Secondly, we have not included ovarian cancer or other related cancers in the family history. This may have resulted in an underestimation of the number of women at increased risk, but is in line with the current guidelines for general practice, which only include breast cancer among family relatives.23,24 Thirdly, the number of new attendees with a family history of breast cancer in general practice was retrieved over a five year period from one large primary care centre, which corresponds to five GP practices. By using a relatively small sample of GP practices, some underestimation of the consultation rate might be introduced. We, however, think that these figures are representative of the general population as other researchers found a slightly higher rate in a sample that included women at all ages over 1625 instead of ages 30–50 as in our population.

Our results indicate that about 13% of women aged 30–50 have at least one first degree relative or at least two second degree relatives with breast cancer, which may relate to an increased risk for the disease. Future studies should assess whether screening these women or subgroups is worthwhile and cost effective. The estimates from the current study could be used to plan management strategies in health care for these women and to compute the cost effectiveness of various treatment strategies in this subgroup of women. In addition, future studies should also be directed at gaining insight into the sensitivity and specificity of preventive strategies for women under the age of 50 to decide if the currently available preventive strategies are effective in these women.

In conclusion, about 13% of the women aged 30–50 have a family history of breast cancer that may relate to an increased risk for breast cancer and most of the women who seek health care because of a family history of breast cancer are those women who have higher risk pedigrees. The estimates that we provided may be used in future studies that investigate management strategies and compute the cost effectiveness of monitoring women aged 30–50 who might be at an increased risk for breast cancer.

Acknowledgments

We gratefully acknowledge Job Kievit and Perla Marang-van de Mheen for their valuable contribution to an earlier draft of this paper, and Wilbert van den Hout and Margot de Waal for their contribution to the estimation of data. This study was financially supported by the National Health Insurance Board of The Netherlands (College voor Zorgverzekeringen).

REFERENCES

View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.