Fax: +45 35257731
Trends in head and neck cancer incidence in Denmark, 1978–2007: Focus on human papillomavirus associated sites
Article first published online: 26 OCT 2010
Copyright © 2010 UICC
International Journal of Cancer
Volume 129, Issue 3, pages 733–741, 1 August 2011
How to Cite
Blomberg, M., Nielsen, A., Munk, C. and Kjaer, S. K. (2011), Trends in head and neck cancer incidence in Denmark, 1978–2007: Focus on human papillomavirus associated sites. Int. J. Cancer, 129: 733–741. doi: 10.1002/ijc.25699
- Issue published online: 26 MAY 2011
- Article first published online: 26 OCT 2010
- Accepted manuscript online: 28 SEP 2010 09:36AM EST
- Manuscript Accepted: 8 SEP 2010
- Manuscript Received: 18 MAY 2010
- Head-and-neck cancer;
- human papillomavirus;
- squamous cell carcinoma;
The aim of our study was to assess the overall trends in the incidence of head-and-neck cancer (HNC) among Danish men and women in 1978–2007, to describe the distribution and incidences of HNCs at different anatomical sites, and to determine whether the incidence of human papillomavirus (HPV)-associated cancers is increasing. Data were extracted from the nationwide Cancer Registry database. To assess the possible impact of HPV infection, the sites of squamous cell carcinomas were categorized as HPV-associated, potentially HPV-associated or HPV-unrelated. In total, 26,474 incident cases were identified and the overall incidence increased throughout the period. Significantly increasing incidence rates were notably seen for tumors in the oral cavity (2.2% per year), tonsils (4.8% per year), oropharynx (3.5% per year) and hypopharynx (4.4% per year). A significantly decreasing incidence of lip cancer was observed among men (–5.0% per year). Cancers at HPV-associated sites (n = 3650) showed strongly increasing incidence rates, primarily in individuals < 60 years. In contrast, HNCs at sites not related to HPV infection showed a significant decrease (in men) or virtually no change in incidence (in women). Our results suggest a marked impact of HPV infection on the epidemiology of HNCs in Denmark. HPV16 is the type most often found in HNCs; thus, the recent introduction of vaccination against HPV may in the future prevent HPV-associated cancers of the head and neck.
The term head-and-neck cancer (HNC) covers a broad spectrum of anatomical sites, most of which are considered to be relatively rare in comparison with other malignant tumors. Nevertheless, in 2008 HNC was the fifth commonest cancer in men and the eleventh commonest in women in Denmark.1 These cancers are found predominantly in men, but the male: female ratio varies worldwide and by anatomical site. There is no consensus about which sites to include under the term “HNC,” but the lip, oral cavity, salivary glands, tonsils, oropharynx, nasopharynx, hypopharynx, nasal cavity, middle ear, paranasal sinuses, larynx and thyroid can be included. As the thyroid differs considerably from the other sites, it is often excluded in both clinical and epidemiological reports of HNC. Morphologically, squamous cell carcinomas predominate.
HNC mortality depends on the location and stage of the tumor at diagnosis. Survivors of HNC often have long-term side-effects of surgery or radiation, which affect their quality of life. Long-term sequelae involves emotional, social and physical disabilities, including disfigurement, motor and sensory nerve damage, eating difficulties, caries, xerostomia, dysphagia, weight loss and cutaneous fibrosis. In many patients, the tumors are diagnosed when they are young. The impact on public health costs of young age at diagnosis and long-term effects after curative treatment is substantial.
The well-known risk factors for HNC are tobacco and alcohol use, which acts synergistically, but also poor oral hygiene is regarded a possible risk factor for oral cavity cancer.2 Oral and oropharyngeal cancers are therefore most commonly seen in countries where tobacco and alcohol consumption are high, including southern Africa, Australia, Brazil, France, India, The Netherlands, Papua-New Guinea and Switzerland.3 In line with this the overall HNC rates have decreased in the US where tobacco and alcohol use is declining.4 Epidemiological studies worldwide have shown, however, increasing trends in the incidence of HNC at specific sites, despite the declining use of tobacco.5–7 These findings suggest that other etiological factors are involved, and high-risk types of human papillomavirus (HPV) have been reported to be risk factors for oral and oropharyngeal cancer.8, 9
The aim of our study was to assess the overall incidence of HNC among Danish men and women in 1978–2007, to describe the distribution and trends in incidence rates of HNCs at specific anatomical sites and to assess whether the incidence of cancers possibly associated with HPV infection is increasing. Previous studies of HNC incidence addressed trends in the incidence of tumors at one site (often tonsils),7, 10–12 from only one region of a country,6, 10, 12, 13 or during a shorter period.5, 6, 12–14 We examined the nationwide trends for HNCs overall and for 11 specific sites over a 30-year period.
Material and Methods
Cases of HNC were identified from the population-based Danish Cancer Registry database, which covers the entire Danish population of 5.5 million (2009).15 Since 1943, hospitals, general practitioners and specialists have reported all new cancer cases to the database. It contains information on virtually all newly diagnosed cases, including topography, morphology, date of diagnosis and birth date of each patient. To ensure that registration is as complete as possible, the Cancer Registry is continuously supplemented with information from the Causes of Death Registry and the National Patient Registry. Since 1978, HNCs have been classified by both topography and histological type, and we therefore studied the incidence trends from 1978 to 2007. The study was approved by the Danish Data Protection Agency.
Identification and classification of HNCs
To identify cases of HNC in 1978–2007 we extracted categories C00–C14 and C30–C32 of the International Classification of Diseases, 10th edition (ICD-10) (Table 1). Information about morphology was available from the International classification of diseases for oncology (ICD-O) morphology codes. Only invasive cancers were included (i.e., with three as the last digit). We excluded lymphomas, sarcomas (including Karposi sarcoma) and a few other tumors (primarily nerve tumors), with ICD-O3 codes 8800–8936, 8950–9110, 9120, 9124, 9130, 9140 and 9170–9948. The HNCs were then grouped into 11 anatomical sites: lip, oral cavity, salivary glands, tonsils, oropharynx, nasopharynx, hypopharynx, nasal cavity including middle ear, sinuses, larynx, etc. (lip, oral cavity, pharynx) (Table 1).
All HNC cases of squamous cell origin (morphology codes 8050–8084, 8120–8131) were classified into three groups: The first two groups, the HPV-associated sites and the potentially HPV-associated sites, were created based on the classification made by Ryerson et al.5 (Table 1). They defined the HPV-associated sites as those reported as being associated with HPV in previous studies. Cancers of the potentially HPV-associated sites, Ryerson et al. chose on the basis of their anatomical similarity to the cancers in the first group, even though an association with HPV infection has not yet been established. Finally we created a third group, the HPV-unrelated sites. This group consisted of the remaining sites, which we assumed were unrelated to HPV infection (Table 1).
Incidence rates were reported as numbers of new cases per 100,000 person–years, and they were age-standardized according to the 2000 standard world population.16 Incidence rates were calculated by 5-year periods according to gender and age (<60 and ≥60 years), and male: female ratios were calculated using the age-standardized incidence rates. Analyses were performed separately for all HNC and for specific sites. The average annual percentage change (APC) was estimated on the assumption of a log-linear Poisson distribution of the observed cancers; statistical significance was considered at p < 0.05.
Between 1 January 1978 and 31 December 2007, 27,772 new cases of HNC were identified. We excluded 1,016 lymphomas, 181 sarcomas and 101 other tumors, resulting in 26,474 cases (Table 1). The majority (88.6%) of these HNCs were squamous cell carcinomas; 4.2% were adenocarcinomas (primarily in the salivary glands), 2.9% were epithelial carcinomas not otherwise specified and 1.2% were mucoepidermoid carcinomas (primarily in the salivary glands) (data not shown).
An obvious male predominance was seen, with 72% men and 28% women. The age-standardized incidence rates increased from 13.32 per 100 000 in 1978–1982 to 18.84 per 100,000 in 2003–2007 among men (APC, 0.4%; 95% CI, 0.2–0.6) and from 4.00 to 6.75 per 100,000 in the same periods among women (APC, 1.3%; 95% CI, 1.0–1.6) (data not shown). The age standardized incidence increased significantly (p < 0.001) more for men aged < 60 years at diagnosis (APC, 1.0%) than those aged >60 years (APC, 0.0%), while no difference was noted for women (APC<60, 1.4% vs. APC≥60, 1.3%, p = 0.676).
In 2003–2007 the age-specific incidence rates in men increased sharply from the forties, peaked at 60–69 years and decreased steadily hereafter. In women age-specific incidence increased moderately from the forties onwards (data not shown).
The distribution of HNCs by anatomical site is shown in Table 2. The three commonest sites in the entire 30-year period were the oral cavity (27.8%), larynx (27.7%) and lip (12.4%). For all sites there was a male predominance, which was especially pronounced for cancers of the lip, larynx and hypopharynx. The trends over time (Fig. 1) were similar for men and women for cancers at almost all sites (except the lip).
The commonest site of HNC in 2007 was the oral cavity, and the age-standardized incidence rate for this cancer increased steadily throughout the study period, from 2.10 in 1978–1982 to 5.46 in 2003–2007 for men (APC, 2.6%; 95% CI, 2.2–3.0) and from 1.39 to 2.61 in women (APC, 1.5%; 95% CI, 1.1–1.9) (Figs. 1 and 2). As the ICD-10 topography code most often specified for oral cavity cancers was “oral cavity not otherwise specified,” it was not possible to establish the exact anatomical location.
The age-standardized incidence rates for tonsillar cancer increased significantly, from 0.63 in 1978–1982 to 2.61 in 2003–2007 among men (APC, 4.9%; 95% CI, 4.3–5.5) and from 0.25 to 0.88 among women (APC, 4.3%; 95% CI, 3.4–5.2), representing the largest percentage incidence increase of any of the 11 HNC sites (Figs. 1 and 2).
The incidences of cancers of both the oro- and hypopharynx showed increasing trends. From 1978–1982 to 2003–2007 the age-standardized incidence rate of hypopharyngeal cancer increased from 0.41 to 1.78 (APC, 4.4%; 95% CI, 3.6–5.2) for men and from 0.09 to 0.39 (APC, 3.8%; 95% CI, 2.6–4.9) for women (Figs. 1 and 2). The incidence rate of oropharyngeal cancer increased from 0.30 to 1.00 (APC, 3.3%; 95% CI, 2.5–4.1) in men and from 0.13 to 0.37 (APC, 3.7%; 95% CI, 2.5–4.9) in women (Figs. 1 and 2).
An opposite trend was observed for lip cancer, for which a significant decrease in age-standardized incidence, from 3.36 to 0.83 (APC, –5.0%; 95% CI, –5.4 to –4.6) was seen among men, dropping from the second to the sixth commonest HNC site (Figs. 1 and 2). The trend was due primarily to a decreasing incidence among men <60 years (APC<60, –7.3%, APC≥60, –4.3%; p < 0.001). Among women, the incidence was stable (APC, 0.1%; 95% CI, –0.7–0.9). Thus, the male: female ratio dropped from 10:1 to 3:1 during the study period. For the remaining HNC sites, the incidence remained relatively stable over time (Fig. 1).
Squamous cell HNCs and relation with HPV infection
HNCs of squamous cell origin constituted 88.6% (n = 23,466) of all HNCs (Table 1). When these carcinomas were classified into three groups on the basis of their anticipated association with HPV, they were distributed as follows: for men, 15.3% (n = 2645) HPV-associated HNCs, 56.5% (n = 9802) potentially HPV-associated and 28.2% (n = 4896) HPV-unrelated; for women, 16.4% (n = 1005) HPV-associated HNCs, 59.4% (n = 3638) potentially HPV-associated and 24.2% (n = 1480) HPV-unrelated (data not shown). Figure 3 shows the incidences over time of the three groups of HNCs according to gender. For men, increasing incidences of both HPV-associated and potentially HPV-associated sites were found, while the incidence of HNCs at HPV-unrelated sites decreased during the study period. For women, the pattern was similar but less pronounced (Fig. 3).
The group of HPV-associated cancers consisted of 58.2% cancers of the tonsils including the Waldeyer ring, 14.4% cancers of the base of the tongue and lingual tonsils and 27.4% cancers at other oropharyngeal sites. The estimated APC for the HPV-associated sites was 4.4% for men (95% CI, 4.0–4.9) and 4.1% for women (95% CI, 3.4–4.9). The increase was seen primarily for individuals < 60 years (men: APC<60, 5.1; APC≥60, 3.7; p = 0.005; women: APC<60, 5.0; APC≥60, 3.5; p = 0.040).
The group of potentially HPV-associated sites consisted of 53.1% laryngeal, 30.2% oral cavity, 13.5% tongue and 3.2% oropharyngeal cancers. The increase in APC was moderate (APCmen, 0.7%; 95% CI, 0.4–0.9; APCwomen, 0.9%; 95% CI, 0.5–1.3) during the 30-year period and was less pronounced than that for HNC at HPV-associated sites. Patients <60 years tended to have a larger annual percentage increase than those ≥60 years; however the difference did not reach statistical significance (men: APC<60, 0.9; APC≥60, 0.5; p = 0.079; women: APC<60, 1.0; APC≥60, 0.8; p = 0.661).
The incidence of HNCs at HPV-unrelated sites in men decreased significantly (APC, –2.0%; 95% CI, –2.3 to –1.7); in women, however, the APC in HNCs at HPV-unrelated sites (APC, 1.1%; 95% CI, 0.5–1.7) was similar to that for the potentially HPV-associated sites, showing only a slight increase. For men, the incidence decreased significantly in both age groups (APC<60, −1.8; 95% CI, −2.4 to −1.2; APC≥60, −2.1; 95% CI, −2.5 to −1.7), the difference between the age groups being nonsignificant (p = 0.442). In contrast, no change was observed for women <60 years (APC, −0.0; 95% CI, −1.1 to 1.1), whereas the incidence increased significantly during the study period for women ≥60 years (APC, 1.5; 95% CI, 0.9–2.3). APC was significantly different between the two age groups in women (p = 0.017).
More than half the cancers at HPV-associated sites were tonsillar cancers including the Waldeyer ring. Figure 4 shows the incidence of tonsillar cancer by age. The incidence increased nearly six times during 1978–2007 in men younger than 60 years (APC<60, 6.3%; 95% CI, 5.4–7.2). For men ≥60 years a less pronounced, but still significant, increase in the incidence was observed (APC≥60, 3.6%; 95% CI, 2.7–4.5). The difference between the increase in incidence in the younger and older men was statistically significant (p < 0.001); this pattern was stronger for tonsillar cancer including the Waldeyer ring than for any of the other HPV-associated cancer sites (data not shown). We found a similar but nonsignificant age difference among women (APC<60, 5.6%; APC≥60, 3.8%; p = 0.056).
This population-based study of trends in HNC incidence has one of the longest follow-up times (30 years) to date. Our results are in line with previous studies reporting HNC to be most common in men, and showing increasing incidence rates for several specific sites. Importantly, we found that incidences of HNCs at HPV-related and potentially HPV-related sites have increased significantly during the past 30 years, especially in men, whereas those of cancers at sites probably not related to HPV decreased (in men) or showed virtually no change (in women). The results are even more pronounced than those reported by Ryerson et al.5 and suggests that HPV plays an important role in HNC.
The largest increase in incidence was observed for cancers of the tonsils, with an approximately fourfold increase in both men and women over the past 30 years. Similar increasing trends have been reported in other countries. In Sweden, the average APC in the incidence of tonsillar cancer was 2.6% in men and 1.1% in women between 1960 and 2003.7 In Finland, the incidence doubled between 1956 and 2000.17 Significant increases in the incidence of tonsillar cancer were also found in south-east England between 1995 and 2004, in Dutch men (APC, 1.7%) between 1989 and 2006 and in men aged <65 years in the United States between 1973 and 2001.6, 14, 18, 19 A study in Sweden, in which HPV DNA was detected by polymerase chain reaction, showed an increasing proportion of HPV-positive tonsillar cancers over time, from 23% in the 1970s to 93% in 2006–2007.10 In the head-and-neck area, HPV is most strongly associated with tonsillar cancer and may be the main explanation for the widespread increase in incidence seen for this site.20–22
Our observation of significantly increased rates of oral cavity and oropharyngeal cancers also parallels observations from Europe and the United States.6, 13, 14, 22, 23 The increase in oral cavity cancer is especially noticeable, as the number of cases far exceeds the numbers at other sites for which increases have been observed, including the tonsils. As the topography of most of the oral cancers in our study was not specified, we could not assess whether the increase was due to tongue cancers, as has been seen in several other countries.6, 14, 19, 24
Another pronounced APC in our material was seen in cancers of the hypopharynx, for both men (4.4%) and women (3.8%). Although the initial number of cases was low, a trend is evident. The results of other studies of hypopharyngeal cancer are equivocal. Significant increases were found in Germany between 1996 and 2005 (men, 2.40–4.44 per 100,000; women, 0.16–0.76 per 100,000) and in Dutch women between 1989 and 2006 (2.5% per year; stable for men).13, 14 In south-east England, however, Doobaree et al.6 found a significant decrease between 1995–1999 and 2000–2004 (age-standardized incidence rate, 0.19–0.14 per 100,000), and in the United States a decrease of 35% was seen between 1975 and 2001.19 It is not clear whether HPV plays a role in hypopharyngeal cancer, but the epithelial structure and the increase in incidence seen in our study are similar to those of the oral cavity, tonsils and oropharynx, which could point to a common etiological factor.
Possible explanations for the increasing incidence rates may be changes in risk factors or tumor biology. During the past few decades, the level of exposure in the Danish population to the traditional risk factors for HNC has not increased. The prevalence of smoking has decreased since 1970, and oral hygiene has improved.25 The Danish population has a higher alcohol intake than other Scandinavian and European countries, but alcohol consumption has been stable since the 1970s.25 Consumption of fresh fruit and vegetables has been reported to be protective against oral cancer.26 Although only 12% of the Danish population follow the recommendation to eat a minimum of 400 g/day of fruit and vegetables, the mean intake of fruit and vegetables has increased.25 More Danish men than women are exposed to the traditional risk factors and fewer men are exposed to the protective factors.25 The differences in incidence between men and women in our study may partly be attributable to these risk factors, but it is unlikely that the significantly increasing trends in the incidences of cancers at several sites in both genders can be sufficiently explained in this way.
In previous studies, the prevalence of HPV has been found to be 4–80% in oral cancers, 15–85% in tonsillar cancers and 14–57% in oropharyngeal cancers,10, 20, 27–30 and the increased incidences of these cancers are thought to be if not solely then at least partly due to HPV infection. The wide variation in HPV prevalence may be due to differences in HPV detection methods and also to geographical variation.8 Information on the prevalence of HPV in the normal oral mucosa is poorly documented.31 As HPV16 is the predominant HPV type in oral, tonsillar and oropharyngeal carcinomas, HPV vaccination might stabilize or even decrease the incidence of these cancers in the future.8, 17, 20, 27
If HPV is an etiological agent in HNCs, the HPV-associated cancer incidence could be anticipated to be affected by sexual behavior. Risk factors for HPV infection in the oral cavity include lifetime number of sexual partners, lifetime number of oral sexual partners, young age at first intercourse, oral contraceptive use, a history of genital warts or Chlamydia infection, HIV infection and increasing age.8, 32–35 HPV is the commonest sexually transmitted agent, with a cervical prevalence of 26.4% in the Danish female population and a lifetime risk for infection of up to 80%.36, 37 D'Souza et al.38 investigated the prevalence of HPV in the oral cavity in healthy individuals and compared it with their sexual behavior. They found that not only oral sex but also open-mouthed kissing were associated with acquisition of an oral HPV infection. Hemminki et al.39 showed excess incidences of tonsillar and tongue cancer in the husbands of cervical cancer patients. Prophylactic HPV vaccines for women may therefore also reduce the incidence of oral HPV infection among men.
Generally, the cancer sites with the highest annual percentage increases (oral cavity, tonsils, oro- and hypo-pharynx) are also the ones dominated by multilayer of stratified squamous epithelium. This epithelium is also predominant in lip cancer, but here it is cornified, which could explain the lack of association to HPV and thus no increase in incidence. Lip cancer decreased significantly in the Danish male population, in particular among men aged <60 years. Finland, Scotland and the United States have seen similar trends in the male population.12, 19, 40 The risk factors for lip cancer include exposure to ultraviolet radiation, tobacco use and chronic irritation. The decreasing trends may be due to changes in the prevalence of smoking, and a decrease in the use of unfiltered cigarettes. However, the population in general may also be increasingly less exposed to sunlight, due to a reduction in the number of people living in rural areas.41 A reduction of outdoor occupancy would have a similar effect. High rates of survival and a decreasing incidence forecast a future reduction in the number of lip cancer cases.
In most European countries and the United States, the incidence of laryngeal cancer is decreasing.6, 14, 19, 42 In our study, the incidence was stable. Studies of the prevalence of HPV in laryngeal cancers are smaller and less convincing than those of the oral cavity and oropharynx, with HPV detected in 0–59% of laryngeal cancers.5, 43–46 The mucosa of the larynx consists of squamous epithelia, as in the cancer sites for which increasing incidence rates have been observed, at mechanically exposed areas. It is possible that the incidence of a minor subset of laryngeal cancers associated with HPV might have increased at the same time as a decrease in a larger subset of tobacco-associated laryngeal cancers, resulting in a more or less unchanged picture overall. Evidence for an association between laryngeal cancer and HPV is therefore not available.
It has been suggested that HPV-related HNCs may constitute a different type of HNCs than those related to alcohol and tobacco. Several studies have reported a more favorable prognosis of HPV-positive tumors,23, 43, 47, 48 and a high viral load has been suggested to be associated with a better prognosis, although the mechanism is unknown.49 Llewellyn et al.26 showed that a considerable proportion (26%) of young patients (<45 years) with oral cavity tumors had had little or no exposure to tobacco and alcohol. It is still not clear, however, whether tobacco and alcohol are co-, independent or synergistic factors in the HPV-associated cancers. In many studies, new cases of HPV-associated HNCs appear to arise in younger people;11, 23, 24, 29, 32 In line with this, we observed the largest increase in incidence among men aged <60 years. This was particularly pronounced for tonsillar cancer including the Waldeyer ring, supporting the assumption that the association with HPV is strongest for this cancer site.20, 21 Changes in sexual behavior, with oral sex being more common than before, may account for some of the increase in incidence of HPV-associated cancers in the past few decades in younger generations.
Our study has several important strengths. It was nationwide, covered a long period (30 years) and addressed trends in the incidences of HNCs at all 11 sites individually. The potential limitations are due to the fact that the head and neck is a complex anatomical area and precise distinction of some sites is not always straightforward, resulting in a risk for misclassification in the registration of these cancers by clinicians. Furthermore, the classification as HPV-associated sites was based on the currently available knowledge about the relationship between HNC and HPV and not on actual detection of HPV DNA.
In conclusion, our data show increasing incidence rates for a subset of cancers in the head and neck area. Especially the HPV-associated sites have increased in incidence, suggesting a marked impact of HPV infection on the epidemiology of HNCs in Denmark. Particularly noteworthy is the dramatic increase in tonsillar cancer including the Waldeyer ring in men <60 years, where the incidence increased nearly six times from 1978 to 2007. Clinicians and researchers should be aware of this and closely monitor these cancers. Population-based reports of HNC are important in recognizing new trends and planning future research. Thus, our study may form the basis for further studies of the potential effects and outcome of current treatment and prophylactic interventions, such as HPV vaccination.
The authors thank Pernille Clausen for programming support.
- 1National Board of Health. The Danish cancer registry 2008, 13rd edn., vol. 5. Copenhagen: National Board of Health, 2009.
- 3Chapter 2: The burden of HPV-related cancers. Vaccine 2006; 24 ( Suppl 3): S/11–S25., .
- 13Head and neck cancer in Germany: a site-specific analysis of survival of the Thuringian cancer registration database. J Cancer Res Clin Oncol 2009 [Epub ahead of print]., , , , , , , .
- 15Statistics Denmark. Population and election. Key figures on population [In Danish]. Copenhagen. Statistics Denmark, 2009. Available at http://www.statistikbanken.dk. Accessed September 10, 2010.
- 16Cancer in five continents, vol. 8. IARC Scientific Publications No. 155. Lyon: IARC, 2002., , , , .
- 20IARC. Human papillomaviruses. IARC monographs on the evaluation of carcinogenic risks to humans, vol. 90. Lyon: IARC, 2007. Available at http://monographs.iarc.fr. Accessed September 10, 2010.
- 25http://www.si-folkesundhed.dk. Accessed September 10, 2010., , . Public health report, Denmark 2007 [In Danish]. Copenhagen: National Institute of Public Health, 2007;185-194, 209-234, 247-260. Available at
- 31Prevalence of human papillomavirus (HPV) in oral cavity and oropharynx. Braz J Otorhinolaryngol 2006; 72: 272–82., .
- 41Statistics Denmark. Population and election. City inventory 1. January [In Danish]. Copenhagen: Statistics Denmark. Available at http://www.statistikbanken.dk. Accessed September 13, 2010.