Tel.: 31-10-7038468, Fax: 31-10-7038474
Prevalence of multiple malignancies in the Netherlands in 2007
Version of Record online: 25 MAY 2010
Copyright © 2010 UICC
International Journal of Cancer
Volume 128, Issue 7, pages 1659–1667, 1 April 2011
How to Cite
Liu, L., de Vries, E., Louwman, M., Aben, K., Janssen-Heijnen, M., Brink, M., Coebergh, J. W. and Soerjomataram, I. (2011), Prevalence of multiple malignancies in the Netherlands in 2007. Int. J. Cancer, 128: 1659–1667. doi: 10.1002/ijc.25480
- Issue online: 28 JAN 2011
- Version of Record online: 25 MAY 2010
- Manuscript Accepted: 7 MAY 2010
- Manuscript Received: 29 JAN 2010
- Dutch Cancer Society. Grant Number: EMCR 2008-4132
- second primary;
As the number of cancer survivors increases in the Netherlands, there is a concomitant increase in patients with multiple malignancies (MMs), the prevalence of which needs to be assessed to estimate care needs. This study analyzed incidence data on all malignant cancers diagnosed between 1989 and 2006 retrieved from the population-based Netherlands Cancer Registry. The point prevalence of MMs was determined on January 1, 2007. Of all cancer survivors in 2007, 30,064 (7% of the total) were patients with MMs. Their median age was 74 (interquartile range 71–76) years. Ninety two percent (i.e., 27,660) of these patients had two cancer diagnoses. The most common subsequent cancers being squamous cell skin cancer (5,468), colorectal cancer (4,634), and breast cancer (3,959). High frequency of combinations included: (i) female breast and genital cancers (any order), (ii) urinary tract and prostate cancers (any order), (iii) Hodgkin's lymphoma and subsequent female breast cancer and (iv) non-Hodgkin's lymphoma and subsequent squamous cell skin cancer. As the number of cancer survivors continues to increase and their survival improves, MMs are becoming more important in the field of cancer surveillance.
The number of cancer survivors is increasing due to improvements in early detection programs and treatment. In 2005, it was estimated that 400,000 people in the Netherlands had a previous cancer diagnosis which corresponds with 2.5% of the total population.1 By 2015, the number of cancer survivors is predicted to increase by 38% to 692,000 individuals, representing 4% of the total Dutch population.2 An additional cancer diagnosis is one of the main concerns in cancer survivors and, therefore, merits attention. In 2001, in the USA 8% of the prevalent cancer patients had multiple malignancies (MMs),3 and one of six (16%) newly diagnosed cases in 2004 had MMs.4 The International Agency of Research on Cancer (IARC) has published a series of studies assessing the relative risk of developing second malignancies after a first cancer.5–12 Similar studies have been performed in the Netherlands.13–17 These sources show that cancer patients have an increased risk of developing a second cancer compared to the general population. Furthermore, compared to patients with a single malignancy, patients with MMs often exhibit a worse 5-year relative survival following diagnosis of one or more secondary cancers.18, 19 Moreover, the six aspects of quality of life (i.e., physical functioning, pain, general health perception, energy, social functioning and role limitations due to emotional problems) have been shown to decrease significantly after a second breast cancer diagnosis.20 Data on quality of life after the diagnosis of a second cancer at another site are scarce.
Because of the increasing number of cancer survivors, MMs will become an increasingly important topic in both cancer surveillance and epidemiology. A comprehensive description of patients with MMs would help to estimate care requirements and guide future research on MMs.21, 22
This study aimed to provide a comprehensive overview of the point-prevalence of MMs in the Netherlands as of January 1, 2007.
Material and Methods
The population-based Netherlands Cancer Registry (NCR) provided incidence data on all malignant cancers, including MMs, diagnosed between 1989 and 2006. Information on vital status is updated through an annual linkage with the Dutch Municipality Register. This follow-up information is complete up since October 1994. Therefore, all patients diagnosed with a cancer before October 1994 who did not develop a subsequent cancer were censored at their last date of follow-up. If subsequent malignancies are diagnosed later, the follow-up of the patient can be reconstructed. Detailed description of NCR data can be found elsewhere.23 The definition of multiple primaries in the Netherlands follows guidelines proposed by the international Association of Cancer Registries (IACR) and IARC.24 According to these guidelines, a primary cancer is one that originates in a primary site or tissue and is thus neither an extension nor a recurrence or a metastasis. The recognition of the existence of two or more primary cancers does not depend on time. Only one cancer can be recognized as rising in an organ or pair of organs or tissue (as defined by the three-character category of the ICD or the topography of the ICD-O) unless the histology is different.
Definition of prevalence of MMs
The point prevalence of cancer is defined as the proportion of people alive with cancer at a certain point in time, disregarding the moment of disease onset. We determined the point prevalence of patients with one cancer and patients with MMs on January 1, 2007. Patients with MMs were defined as those with a primary invasive cancer followed by one or more additional cancers (invasive/in situ) between 1989 and 2006. These patients were classified as “ever diagnosed patients with MMs” and were categorized into three groups according to the vital status: alive, deceased and lost to follow-up (LFU). In this study, those alive on January 1, 2007 represent the prevalent cases of MMs. First primary invasive cancers were defined as first malignancies. Multiple primary cancers were defined according to international guidelines for multiple primary cancers.24 We chose the presented first cancers based on the SEER data in the period from 1973 till 2002.25 After a first cancer, if the 25-year cumulative incidence of the second cancers is larger than 10%, the first cancers were selected to be presented and were displayed in the following 13 anatomical sites/systems: mouth and pharynx, colorectal and anus, soft tissue, malignant melanoma of the skin, skin (squamous cell carcinoma), breast, female genital, male genital, prostate, kidney, urinary tract (renal pelvis, ureter, bladder, urethra), Hodgkin's and non-Hodgkin's lymphoma. All other first malignancies were grouped into a category “other.” Ovarian carcinomas of borderline malignancy were excluded, as were basal cell carcinomas of the skin, which were only recorded in one regional registry.26
Patients with MMs were described by gender, number, status on the last date of follow-up (alive, deceased and LFU), age on January 1, 2007 (grouped into six age categories: 0–39, 40–49, 50–59, 60–69, 70–79, 80+) and cancer types. Also calculated were: age at diagnosis of the first and subsequent cancers, and time intervals between cancer diagnoses. Age at each cancer diagnosis and time interval between diagnoses were presented as median and interquartile range (IQR). In addition, we calculated the proportion of survivors according to the number of cancer diagnoses (1, 2, 3 and 4 or more) and the period from the last diagnosis until the end of the study (<1 year, 1–2 years, 3–4 years and ≥5 years). Patients with MMs were grouped by site of first cancer. Statistical analyses were performed using SAS 9.0.
Between 1989 and 2006, a total of 1,351,621 individuals in the Netherlands were diagnosed with cancer (data not shown); 85,676 (6%) of whom were patients with MMs. On January 1 2007, 424,340 of these patients were still alive, 7% (30,064) included patients with MMs. Ninety two percent of these patients (n = 27,660) had two cancers. Of all patients with multiple cancers diagnosed between 1989 and 2006, 58% (49,335) had died and 7% (6,277 patients) were LFU. Median age at first cancer diagnosis was 65 years (IQR: 62–67 years), and the median age on January 1, 2007 was 74 years (IQR: 71–76 years). The period between two consecutive diagnoses ranged from 0.5 to 3 years (median, 1 year). As the number of cancers diagnosed per patient increased, the time interval between the diagnoses decreased (Table 1).
On January 1, 2007 there were no surviving patients with more than five cancer diagnoses. Men were more often diagnosed with MMs than women (8% vs. 6%, respectively). Most of the prevalent patients with MMs (42%, increasing from 25 to 30% for the 2nd cancer and 51–65% for the fourth or following cancers) were diagnosed within 1 year of their last cancer diagnosis (Table 2).
MMs were most commonly observed among elderly patients (median 74 years), but MMs prevalence decreased among people aged 80 years or older. Less than 1% (274/30,064) of the MMs patients were younger than age 39 years (data not shown).
The youngest patients with MMs were survivors of Hodgkin's lymphoma (median age on January 1, 2007 55 years) and those with a first male genital cancer (97% testicular cancer; median age on January 1, 2007 62 years) (Table 3).
The frequency ranks of the first, second, third, fourth and higher-order cancers are shown in Figures 1a–1c. The most commonly occurring second cancers were squamous cell cancer of the skin (n = 5,468), colorectal cancers (n = 4,634) and breast cancers (n = 3,959). Higher order cancers (≥3) were most frequently found for cancers of the skin (squamous cell skin cancer, n = 579), colon and rectum (n = 333) and urinary tract (n = 308).
Table 3 gives an overview of the distribution of subsequent cancers according to the type of the first and subsequent cancers. The shortest time interval between the first and second cancer was observed among survivors with a first urinary tract cancer (1 year); the longest time interval was found for those with Hodgkin's disease survivors (6 years). Overall, on January 1, 2007 the patients with MMs had lived with cancer (from the first diagnosis onwards) for 8 years (median), ranging from a median of 6 years for prostate cancer patients to a median of 11 years for survivors of Hodgkin's disease. The lowest proportion of subsequent cancers was found among first male genital cancers, being only 3%. The highest proportions were observed for urinary tract cancer (15%), mouth and pharynx cancer (12%), squamous cell carcinoma of the skin (10%) and colorectal cancer (9%).
In absolute terms, multiple cancers were most frequent among survivors of female breast cancer (n = 5,774), colorectal (n = 5,169) and prostate cancer (n = 3,862).The most frequent combinations between first and subsequent cancers were as follows (i) breast and female genital cancers (any order), (ii) urinary tract and prostate cancers (any order), (iii) Hodgkin's lymphoma and subsequent breast cancer, (iv) non-Hodgkin's lymphoma and subsequent squamous cell skin cancer.
This study assessed the burden of MMs in terms of prevalence within the Dutch population of which the average incidence of cancer was 404 (per 100,000) from 1989 to 200623 and exhibited a 5-year crude survival of 60% in the period 2003–2007.27 The prevalence of MMs included current survivors with at least one subsequent cancer diagnosis after a first invasive cancer within the observation period from 1989 to 2006. With a median follow-up time of 8 years, 30,064 patients were diagnosed with MMs, representing 7% of all cancer survivors on January 1, 2007 in the Netherlands. A previous study in the USA reported 8% of cancer survivors had MMs in 2002.3
The most common subsequent cancers were cancers of the skin (squamous cell skin cancer), colorectal cancer and female breast cancer. The prevalence of subsequent cancers is influenced by the age at first diagnosis, incidence and survival of the first and subsequent cancers and the maximum and median length of follow-up time (e.g., history of cancer registry). The three most common subsequent cancers mentioned above are those with high incidence and a relatively good prognosis (5-year relative survival >50%) in the Netherlands.28 Although lung cancer is one of the most common cancers and shares its main risk factor (smoking) with many other cancers, it contributes little to the prevalence of MMs because of its poor prognosis (5-year survival ∼10%).29
The current population with MMs was composed mainly of elderly subjects (median age,74 years). The average age at diagnosis of first cancer was 65 years. The treatment of second or higher order cancers is complex due to old age and co-morbidity and might need to be tailored individually.30, 31 The youngest patients with MMs had Hodgkin's lymphoma (55 years) and male genital cancer (62 years) as a first primary. Their young age at first cancer diagnosis (40–50 years) combined with the good prognosis (10-year relative survival is 75–95%) led to a long life expectancy after first diagnosis.28 However, the risk of developing a subsequent cancer was 1.4 to 3.0 times higher compared to the general population.32–34 It is plausible that prevention by means of lifestyle changes (such as quitting smoking) might decrease this group's risk of a second cancer and other chronic diseases.35 However, the benefit and harm of intervention programs (for either primary or secondary prevention) remain to be investigated.17
We observed four pairs of cancers that frequently occur together. It should be noted that this observed association on prevalence should not be simply interpreted as a result of shared risk factors, which is normally interpreted in the incidence data. Cancer prevalence is an interaction of various factors such as risk of getting cancer (incidence), risk of dying from cancer (mortality) and duration of time lived with cancer (survival) and many more. Therefore, the etiological link between the first and subsequent merits attention, but should be considered only one of the multitude of factors which determine the prevalence of MMs and, more particularly, the prevalence of specific combinations of multiple cancers. There are etiological theories concerning the development of subsequent cancers: (i) specific treatment effects6, 9; (ii) shared risk factors between first and subsequent cancers (lifestyle or environmental factors12, 36); (iii) shared genetic predisposition and (iv) combinations of the above three.5, 10, 37 Adverse effects of treatment (i.e., radiotherapy) may explain the high prevalence of breast cancer among Hodgkin's lymphoma survivors.38–40 Although second lung cancer is also associated with radiotherapy,35 the poor prognosis of lung cancer contributed to its low proportion of multiple cancers in this cancer type. Shared risk factors, e.g., hormonal risk factors such as nulliparity, obesity, and hormone-replacement therapy, probably relates breast cancer to female genital cancers.41, 42 Moreover, carriers of BRCA1/2 may contribute 2–5% in developing breast and ovarian cancers, especially among younger patients.43 As to the last theory, the combination of the previous three factors has hardly been studied and is difficult to explain with our data. In addition to the above general theories, disease or treatment-induced immunosuppression may explain the high frequency of squamous cell skin cancer after non-Hodgkin's lymphoma.44
Finally and importantly, enhanced early detection programs (e.g., clinical follow-up and screening programs) may explain some of the high frequency cancer pairs, especially when the time interval between the first and the second cancer diagnosis is short; for instance, in case of prostate and urinary tract (usually bladder) cancer, the short interval between the first and the second cancer diagnosis (≈1 year) may reflect the common diagnostic process of these two cancers.45
Recent Dutch studies of long-term survivors of breast, Hodgkin's and non-Hodgkin's lymphoma, prostate and endometrial cancer showed that disease progression (recurrence, metastasis and detection of other primary cancers) negatively affects health status and quality of life.20, 46 As data on quality of life among MMs patients are scarce, more studies are needed before designing interventions to improve the quality of life in this population.
To our knowledge, this is the first population-based, nation-wide report on the prevalence of MMs that includes all cancer types (excluding basal cell carcinomas of the skin). This ensures the representativeness and validity of the prevalence estimates.
Concerning the accuracy and completeness of cancer diagnoses in the NCR, most cases were histologically confirmed cancers retrieved from the nation-wide pathology network (PALGA), and therefore, a high accuracy of cancer diagnosis may be expected. Furthermore, the National Registry of Hospital Discharge Diagnosis system collected data on patients who were only diagnosed clinically, which increases the completeness of the NCR data. Finally, the application of IARC/IACR rules for multiple primary cancers24 facilitates external comparison of the study results, as long as the same coding rules are applied.
The main limitation of the current prevalence data is the probable underestimation of MMs cases. First, 18 years of follow-up does not yield the lifetime prevalence but represents about 90% of the full estimate.47 Second, 7% of ever diagnosed multiple cancer patients were LFU, which may have led to an underestimate of prevalence. Third, we were unable to include the most common cancer: basal cell carcinoma. Although this is the case in most registries, if this cancer had been taken into account, higher MMs prevalence values would have emerged. Finally, the coding rules that were used to record multiple primaries should be taken into consideration when making cross-countries comparisons. For example, in USA, the Surveillance, Epidemiology, and End Results (SEER) use different rules than those defined by IACR/IARC.48 Differences between the coding guidelines include: the existence of two or more primary cancers do not depend on time in the IACR/IARC guidelines, however, only metachronous cancers (occurring 2 or more months after initial diagnosis) are recorded as separate primaries in the SEER guidelines; furthermore, regarding cancers originating from paired organs (e.g., ovary, Wilm's cancer, Retinoblasma), SEER guidelines treat them as independent, whereas, the IACR/IARC guidelines regard them as one single cancer unless a histological difference exists. In general, the IACR/IARC guidelines are more restrictive than those applied by SEER. Moreover, there are unresolved debates regarding coding of multiple primaries.49 However, it is yet to be investigated how such coding differences affect differences in recorded prevalence.
In conclusion, on January 1 2007, 30,064 patients were alive with MMs, representing 7% of all cancer survivors and 0.2% of the total Dutch population. The estimate is subject to a possible 10–15% of underestimation. As the number of cancer survivors continues to increase by 3–5% annually and prognosis improves with each year survived,50 MMs are becoming more important in the field of cancer surveillance.
The authors would like to thank Dr. J.A. (Jan Anne) Roukema for his valuable comments and advice.
- 1Dutch Comprehensive Cancer Centres. Available at: http://www.ikcnet.nl/page.php?id=1525&nav_id=97. Accessed on July 14, 2009.
- 2Kanker in Netherland. Trends, prognoses, en implicatie voor zorgvraag. 2004. 2004.
- 4Ries LAGMD, KrapchoM, StinchcombDG, HowladerN, HornerMJ, MariottoA, MillerBA, FeuerEJ, AltekruseSF, LewisDR, CleggL, EisnerMP, et al., eds. SEER cancer statistics review, 1975–2005. Bethesda, MD: National Cancer Institute, 2007.
- 19Multiple tumours in survival estimates. Eur J Cancer 2009; 2: 2., , , , , , , .
- 23Dutch Comprehensive Cancer Centres, Available at: http://www.ikcnet.nl/page.php?id=96, Accessed on July 14, 2009
- 24International Association of Cancer Registries (IACR). International rules for multiple primary cancers. Asian Pac J Cancer Prev 2005; 6: 104–6.
- 25New malignancies among cancer survivors: SEER cancer registries, 1973–2000. Bethesda, MD: National Cancer Institute, 2006., , , , , , .
- 27Available at: http://www.ikcnet.nl/system/image_viewer/index.php?recID=1756&image=1, Accessed on July 14th 2009.
- 28Available at: http://www.ikcnet.nl/page.php?id=96, Accessed on July 14th 2009.
- 41Epidemiology of breast cancer In: DoneganWL SJe, ed. Cancer of the breast, 5th edn. Philadelphia: Saunders, 2002. 111–32., .
- 42Epidemiology of gynecologic cancers. In: HoskinsWJ PCYRe, ed. Principles and practice of gynecologic oncology, 4th edn. Philadelphia: Lippincott Williams & Wiklkins, 2005. 3–32., .
- 48NCI. 2006b. SEER Program Coding and Staging Available at: Manual:http://seer.cancer.gov/tools/codingmanuals/, Assessed on July 14th, 2009.
- 49Revising the multiple primary and histology coding rules. J Regist Manag 2007; 34: 81–6., , , , , .