Increasing incidence and survival of head and neck cancers in Denmark: a nation-wide study from 1980 to 2014

Abstract

Background

The purpose of the study was to determine trends in age-adjusted incidence rates (AAIR) and survival probability in head and neck cancers (HNCs) in the Danish population from 1980 to 2014.

Material and methods

All patients registered with HNC in the nationwide Danish Cancer Registry from 1980 to 2014 were included. We evaluated the AAIR per 100,000 and the average annual percent change (AAPC). The relative survival probability at 5 years was calculated in relation to gender, anatomical location and histology, and we constructed age-period-cohort models of incidence.

Results

About 34,606 patients were included (64.7% men). The AAIR increased from 9.1 per 100,000 in 1980 to 17.4 per 100,000 in 2014 with an AAPC of 2.1%. The greatest incidence increase was observed in oropharyngeal cancer (AAPC: 5.4%) followed by hypopharyngeal cancer (AAPC: 4.2%). Adenocarcinomas had the highest AAPC (5.0%) followed by squamous cell carcinomas (AAPC: 2.0%). The AAPC was significantly higher in women (2.4%) compared with men (1.6%). For all HNC patients, the relative survival at 5 years rose significantly from 49.0% in 1980–1984 to 62.4% in 2010–2014. Women had a significantly higher survival than men with a relative survival of 61.7% compared to 50.0% in men. Laryngeal cancer had the best survival probability of cancers in the upper aerodigestive tract with hypopharyngeal cancer having the poorest survival.

Conclusion

This nation-wide study showed a significant rise in incidence of HNC for men and women along with a significant increase in relative survival. Oropharyngeal cancer had the highest increase in incidence followed by hypopharyngeal cancer which showed the poorest survival of HNCs.

Introduction

Head and neck cancer (HNC) encompasses a heterogeneous group of cancers and covers several anatomical sites in the head and neck region. Worldwide, HNC is the sixth most frequent malignancy accounting for more than 980,000 cases annually and is responsible for considerable morbidity and mortality [1]. The vast majority of cancers in the upper aerodigestive tract are squamous cell carcinoma (SCC) accounting for more than 90% of oral and pharyngeal cancers [2,3]. Alcohol and tobacco consumption remain the most common risk factors for the development of HNC although the number of HNC cases due to tobacco consumption varies worldwide [4]. Emerging oncogenic viruses such as human papillomavirus (HPV) and Epstein Barr virus (EBV) have also influenced the increase in HNC [5,6]. Notably, HPV has been shown to be the main cause of the recent and significant rising incidence of oropharyngeal cancer especially in Northern Europe [7–9]. Epidemiology studies of the general trend of HNC in a nation-wide, population-based setting are sparse.

This study of the trends of HNC is important since it provides an assessment of the overall development of HNC along with a comparison of locations and histology types. The aim of this study was to evaluate changes in incidence and survival rates for HNC cancers diagnosed between 1980 and 2014 in Denmark.

Material and methods

Data included in this study was derived from the Danish Cancer Registry (DCR) [10] and from the central population register (CPR) [11]. The CPR has since 1968 issued all Danish residents with a unique personal identifications number which enables accurate individual-linkage of data. The DCR contains data on the incidence of all cancers diagnosed in Denmark from the year 1943, and registration became mandatory in 1987.

Information on age at diagnosis, date of diagnosis, cancer location, and histology was derived from the DCR. Vital status (e.g., date of death and date of emigration) were obtained from the CPR. All patients registered with an HNC in the DCR between 1980 and 2014 were included. During this period, several classification schemes were used. Until 2004, patients were coded according to ICD-O (International Classification of Diseases for Oncology). These codes were converted to ICD-10 codes (10th revision of the International Classification of Diseases) [12]. Based on the ICD-10 codes, patients in this study were divided into nine groups: Oral cavity, oropharynx, nasopharynx, hypopharynx, sinonasal cavity, thyroid, major salivary glands, larynx and other (Supplementary Table S1).

Based on the ICD-O-3 classification via MORPHO3 registrations from the DCR [12], the histological information was derived, and 335 histological subtypes were identified. These were categorized into seven histology groups: Squamous cell carcinomas, adenocarcinomas, sarcomas, salivary gland malignant tumors, carcinomas unspecified, malignant melanomas and others. We excluded all hematolymphoid tumors due to altered registration procedure in 2003. Patients with missing follow-up status were excluded from the relative survival calculations (n = 26).

Studies on the incidence and survival of respectively oral cancer and oropharyngeal cancer have previously been investigated separately [13,14].

Statistical analysis

Statistical analysis was performed in R statistics version 3.3.3 [15]. Incidence rates per 100,000 were calculated using the direct method with the EpiTools package using the WHO world standard population as reference [16]. By calculating the age-adjusted incidence rate (AAIR), the fact that most cancer rates are age dependent, is circumvent. The average annual percent change (AAPC) was calculated using Joinpoint Trend Analysis Software v. 4.2.0.2, with growth assumed to be logarithmic with the formula ln(y) = xb. The Joinpoint regression analysis estimates possible significant changes in trends (trend breaks).

The relative survival rate was calculated as the ratio of the observed survival rate to the expected survival rate in Denmark matched by age, sex and calendar year with the R package relsurv [17]. Relative survival is defined as the all-cause observed survival in the cancer population under study divided by the expected survival of a comparable group in the general population. Relative survival is therefore a measure of the excess mortality associated with a diagnosis of the cancer. Expected survival was estimated using the Ederer II method [18]. Patients that were alive at the last date of follow-up were censored at this date (31 December 2016).

We evaluated the effect of age, calendar period, and birth cohort on incidence using the age-period-cohort (APC) model. The APC was calculated with the Epi package [19]. The reference year was 1900 for the cohort effect and 1982.5 for the period effects. Ages and periods where arranged in 5-year intervals. To avoid statistical instability the analysis was restricted to persons at ages 30–84 years as the number of cases before the age of 30 and after the age of 84 is low.

We considered a p value < .05 as statistically significant.

Results

A total of 34,606 patients were included of which 64.7% (n = 22,397) were men. The median age at diagnosis was 62.6 years (95% CI 62.3; 62.8) with no significant difference between men and women. The most frequent diagnosis bases on ICD-10 registration was laryngeal cancer accounting for 23.9% (n = 8279) of all HNCs, closely followed by oral cancer accounting for 23.2% (n = 8028) (Table 1). The most common histology was squamous cell carcinoma (69%; n = 23,894) followed by adenocarcinomas (12.9%; n = 4444).

Table 1. Age-adjusted incidence rates for head and neck cancers in Denmark 1980–2014.

		Trend 1		Trend 2		Trend 3
Group	Cases (%)	Period	APC (95% CI)	Period	APC (95% CI)	Period	APC (95% CI)	AAPC (95% CI)	p
Total	34,606 (100)	1980–1984	3.4 (0.2;6.7)	1984–2004	1.5 (1.3;1.8)	2004–2014	2.6 (1.8;3.4)	2.1 (1.6;2.5)	<.05
Male	22,397 (64.7)	1980–2014	1.6 (1.4;1.7)	NA	NA	NA	NA	1.6 (1.4;1.7)	<.05
Female	12,209 (35.3)	1980–1994	1.3 (0.7;2.0)	1994–2007	2.2 (1.4;3.1)	2007–2014	4.6 (2.6;6.6)	2.3 (1.8;2.9)	<.05
Histology
Squamous cell carcinomas	23,910 (69.0)	1980–1986	3.7 (1.5;5.8)	1986–2004	1.4 (1.0;1.9)	2004–2014	2.1 (1.1;3.0)	2.0 (1.5;2.5)	<.05
Adenocar-cinomas	4444 (12.9)	1980–2014	5.0 (4.4;5.6)	NA	NA	NA	NA	5.0 (4.4;5.6)	<.05
Sarcomas	179 (0.5)	1980–2014	1.3 (−0.1;2.6)	NA	NA	NA	NA	1.3 (−0.1;2.6)	.1
Salivary gland malignant tumors	1238 (3.6)	1980–2014	1.7 (1.2;2.3)	NA	NA	NA	NA	1.7 (1.2;2.3)	<.05
Carcinomas unspecified	4076 (11.8)	1980–2014	−0.8 (−1.3;−0.3)	NA	NA	NA	NA	−0.8 (−1.3;−0.3)	<.05
Malignant melanomas	287 (0.8)	1980–2014	0.9 (−1.1;2.8)	NA	NA	NA	NA	0.9 (−1.1;2.8)	.4
Other-histology	472 (1.4)	1980–2014	−1.2 (−3.2;−1.1)	NA	NA	NA	NA	−1.2 (-3.2;−1.1)	.3
Location
Oral cavity	8028 (23.2)	1980–1996	3.2 (2.4;4.0)	1996–2014	0,9 (0.2;1.6)	NA	NA	2.0 (1.5;2.5)	<.05
Oropharynx	6267 (18.1)	1980–1986	9.2 (4.7;13.8)	1986–2014	4,6 (4.2;5.0)	NA	NA	5.4 (4.6;6.2)	<.05
Nasopharynx	911 (2.6)	1980–2014	−0.4 (−1.1;0.2)	NA	NA	NA	NA	−0.4 (−1.1;0.2)	.2
Hypopharynx	2032 (5.9)	1980–2014	4.2 (3.7;4.7)	NA	NA	NA	NA	4.2 (3.7;4.7)	<.05
Sinonasal	1932 (5.6)	1980–2014	0.0 (−0.5;0.5)	NA	NA	NA	NA	0.0 (−0.5;0.5)	.9
Thyroid gland	5177 (15.0)	1980–2005	1.77 (1.2;2.4)	2005–2014	8,1(5.1;11.2)	NA	NA	3.4 (2.5;4.3)	<.05
Major salivary gland	1667 (4.8)	1980–2014	0.8 (0.2;1.3)	NA	NA	NA	NA	0.8 (0.2;1.3)	<.05
Larynx	8279 (23.9)	1980–2014	−0.7 (−1.0;−0.4)	NA	NA	NA	NA	−0.7 (−1.0;−0.4)	<.05
Other location	314 (0.9)	1980–2014	0.4 (−0.7;1.5)	NA	NA	NA	NA	0.4 (−0.7;1.5)	.4

APC: annual percent change; AAPC: average annual percent change; 95% CI: 95% confidence interval; NA: Not applicable. Significant at p < .05.

Incidence

An increase for the total population was observed with two significant trend breaks; from 1980 to 1984 the APC was 3.4%. In the period 1984–2004, the APC declined to 1.5%. From 2004 and in the remaining period, the APC rose to 2.6%. In the period 1980–2014, the AAPC was 2.1% (95% CI 1.6; 2.5). The incidence of HNC increased from 9.1 per 100,000 in 1980 to 17.4 per 100,000 in 2014 (Figure 1(A)), corresponding to a total number of 1519 patients in 2014.

Figure 1. Age-adjusted incidence rates in the period 1980–2014 for (A) all head and neck cancers, (B) men and (C) women.

We observed an increase in incidence for both men and women; however, the rise in women was significantly higher with an AAPC of 2.3% (95% CI 1.8; 2.9) compared with men with an AAPC of 1.6% (95% CI 1.4; 1.7). A trend break was observed in women in 1994 where the APC increased from 1.3% to 2.2%, and increased additionally in 2007 to an APC of 4.6%. We observed that men had a higher incidence of HNC with an AAIR per 100,000 of 21.2 in 2014 (Figure 1(B)) compared with 13.9 for women (Figure 1(C)). Based on the anatomic location, a significant increase in incidence was observed for oral cavity, oropharyngeal, hypopharyngeal, thyroid and major salivary gland cancers but not for nasopharyngeal, sinonasal and the category composing other cancers (Figure 2). The only cancer with a significant decrease in incidence was laryngeal cancer with an AAPC of −0.7 (95% CI −1.0; −0.4). Oropharyngeal cancer had the highest increase in incidence with an AAPC of 5.4% (95% CI 4.6; 6.2).

Figure 2. Age-adjusted incidence rates stratified on anatomical locations in the period 1980–2014.

Regarding histological type an increase was observed in squamous cell carcinomas, adenocarcinomas, salivary gland malignant tumors and the category other. Adenocarcinomas had the greatest rise in incidence with an AAPC of 5.0% (95% CI 4.4; 5.6). A decrease was observed in carcinoma unspecified with an AAPC of −0.8 (95% CI −1.3; −0.3).

Reporting to the DCR became mandatory in 1987 thus to investigate if this affected the incidence, we analyzed the AAPC for the period 1980–1986, 1987–2014 and 1980–2014. No significant difference between the AAPC for the three periods existed (Supplementary Table S2).

Relative survival

Patients were followed for a median period of 3.5 years (95% CI 3.4; 3.6) (range 0–37 years). The relative survival at 5 years rose significantly from 49.0% (95% CI 47.0; 51.0) in 1980–1984 to 62.4% (95% CI 61.0; 63.8) in 2010–2014 (Table 2 and Figure 3(A)). Women had a significantly better survival than men with a relative survival after 5 years of 61.7% (95% CI 60.7; 62.7) compared to men with a survival probability of 50.0% (95% CI 49.3; 50.8) (Figure 3(B)). Based on location, we observed that thyroid cancer had the best survival probability at 5 years of 77.5% (95% CI 76.2; 78.8) (Figure 3(C)) for all HNCs. Laryngeal cancer had the best survival probability for cancers in the upper aerodigestive tract with a survival probability at 5 years of 58.9% (95% CI 57.6; 60.2). Hypopharyngeal cancer had the worst survival probability with a relative survival at 5 years of 22.8% (95% CI 20.9; 24.9) (Figure 3(C)).

Figure 3. Relative survival based on (A) year of diagnosis, (B) gender, (C) anatomical location and (D) histology. CO: Oral cavity; OP: Oropharynx; NP: Nasopharynx; HP: Hypopharynx; SN: Sinonasal; TG: Thyroid gland; SG: Major salivary gland; La: Larynx; other: other location; SCC: Squamous cell carcinomas; Adeno: Adenocarcinomas; Sarc: Sarcomas; Car: Carcinomas; Mela: Melanomas; SGC: Salivary gland malignant tumors; Oth: Other-histology.

Table 2. Univariate and multivariate Cox regression analyzes of overall survival for patients with head and neck cancer in Denmark 1980–2014.

			Univariate		Multivariate
	Cases	Death	HR (95%CI)	p	HR (95%CI)	p
Total	34,580	24,639
Male	22,378	16,965	Ref		Ref
Female	12,202	7674	0.89 (0.87;0.92)	.00	0.99 (0.96;1.02)	.61
Histology
Squamous cell carcinomas	23,894	18,521	Ref		Ref
Adenocarcinomas	4438	1980	0.59 (0.57;0.62)	.00	0.71 (0.67;0.75)	.00
Sarcomas	179	126	1.13 (0.95;1.35)	.17	1.16 (0.98;1.39)	.09
Salivary gland malignant tumors	1238	732	0.61 (0.57;0.66)	.00	0.63 (0.58;0.68)	.00
Carcinomas	4073	2678	0.79 (0.76;0.83)	.00	0.86 (0.82;0.9)	.00
Malignant melanomas	286	220	0.87 (0.76;0.99)	.04	1.01 (0.88;1.16)	.91
Other	472	382	0.90 (0.81;0.99)	.04	0.99 (0.89;1.1)	.85
Locations
Oral cavity	8025	6290	Ref		Ref
Oropharynx	6259	4313	1.08 (1.04;1.13)	.00	1.03 (0.99;1.07)	.14
Nasopharynx	911	712	1.34 (1.24;1.45)	.00	1.25 (1.15;1.35)	.00
Hypopharynx	2032	1778	1.88 (1.79;1.99)	.00	1.85 (1.75;1.95)	.00
Sinonasal	1929	1552	0.78 (0.73;0.82)	.00	0.82 (0.78;0.87)	.00
Thyroid gland	5172	2049	0.58 (0.55;0.61)	.00	0.63 (0.59;0.68)	.00
Major salivary gland	1665	1123	0.64 (0.6;0.69)	.00	0.76 (0.71;0.81)	.00
Larynx	8273	6567	0.74 (0.72;0.77)	.00	0.71 (0.69;0.74)	.00
Other	314	255	1.31 (1.16;1.49)	.00	1.36 (1.2;1.54)	.00
1980–1984	3374	3168	Ref		Ref
1985–1989	3758	3447	1.03 (0.98;1.08)	.24	1.02 (0.98;1.07)	.34
1990–1994	4101	3621	1 (1;1)	.03	1.06 (1.01;1.12)	.01
1995–1999	4717	3883	1.07 (1.02;1.12)	.01	1.06 (1.01;1.11)	.02
2000–2004	5156	3835	1.1 (1.05;1.15)	.00	1.09 (1.04;1.14)	.00
2005–2009	6121	3718	1.02 (0.97;1.07)	.38	1 (0.95;1.05)	.94
2010–2014	7353	2967	0.91 (0.86;0.96)	.00	0.92 (0.87;0.97)	.00

HR: hazard ratio; CI: Confidence interval.

Regarding histology, adenocarcinomas had the best relative survival at 5 years [77.6% (95% CI 76.1; 79.1)]. Malignant melanomas had the worst relative survival at 5 years of 45.2% (95% CI 38.8; 52.6) followed by squamous cell carcinomas with a relative survival of 47.8% (95% CI 47.1; 48.5) (Figure 3(D)).

Age-period-cohort model

For all HNCs, the highest incidence was observed in the age group 65–70 years (Figure 4). The same trend was observed in men (Supplementary Figure 4(B)) whereas women had the highest incidence of cancer in the age group 70–75 years (Supplementary Figure 4(C)). When considering the cohort effect, the rate ratio increased until 1925. A continued decrease was observed until after 1940 were the trend shifted and a rise began, continuing for the remaining cohorts. When stratifying based on gender the pattern was similar.

Figure 4. Age-period-cohort model for all head and neck cancers. Abbreviation: LCI: lower confidence interval; UCI: upper confidence interval; RR: rate ratio.

In the period, effect an increase was observed until approximately the year of 1995. A decrease was observed from 1995 until 2005 where an increase was observed again. The same trend was identified in men. Contrarily, in women, a rise was observed until approximately the year of 1995. After 1995, a continued decrease in the remaining period was observed.

Discussion

This nation-wide study included a total number of 34,606 patients diagnosed with HNC in Denmark in the period 1980–2014. We investigated AAIR and the survival probability stratified on location, histology and gender.

We observed that the AAIR for all HNCs increased significantly, but with great heterogeneity in trends depending on location, histology and gender. Our study showed that men have a higher incidence of HNC than women. One of the major risk factors of HNC is alcohol consumption. A decrease in alcohol consumption has been seen in Denmark the last 10 years [20]. However in 2013, 21% of the Danish adult population consumed more than the recommended alcohol amount, and more men than women drank above the recommended amount [20]. This could, in part, explain the excess in incidence in men compared to women.

The higher AAPC among women compared to men could be due to change in smoking habits, a major risk for the development of HNC. Today the proportion of daily smokers (>15 cigarettes daily) is equal across genders minimizing the differences in tobacco-related cancers between genders. This is in contrast to smoking habits in 1970 where 68% of men were daily smokers compared to only 47% of women [21].

In regards to location a rise in incidence rate was observed for oral, oropharyngeal, hypopharyngeal and thyroid cancer which parallels trends observed in other European countries [13,22–24]. Oropharyngeal cancer had the highest increase in incidence, which in other studies has been attributed to an increase in HPV-driven carcinogenesis [7,25]. Notably, the HPV-driven oropharyngeal cancer is associated with significantly better oncological outcomes when compared to the HPV-negative, e.g., tobacco/alcohol, induced cancer [26]. Laryngeal cancer was the only cancer to significantly decrease in incidence. Similar trends in laryngeal cancers have been observed in other studies [22,27], and may be attributed to the general decrease in tobacco smoking [28]. However, as smoking is a risk factor for all HNC, and as the incidences of other HNCs rose from 1980 to 2014 despite the decreasing tobacco use, it is possible that tobacco use might impact laryngeal cancer more than other HNCs.

We also investigated the survival probability of patients diagnosed with HNC. Hypopharyngeal cancer had the lowest survival probability with a survival at 5 years at only 22.8% (95% CI 20.9; 24.9). Similar trends are seen in other studies and the reason for the lower survival rate has been argued to be due to advanced stage at diagnosis [29]. This contrast to laryngeal cancer which has the highest survival probability of the upper aerodigestive tract cancers at 5 years at 58.9% (95% CI 57.6; 60.2). This trend is also observed in other studies [29]. The explanation to this is probably that glottic cancer, the most abundant form of laryngeal cancer, often are diagnosed in early stages due to an early presentation of symptoms [28], together with a minimal risk of regional metastasis due to the lack of lymphoid drainage from the vocal cords.

Thyroid cancer had the highest survival probability of all HNCs. These cancers are predominantly adenocarcinomas and only few cases are squamous cell carcinomas. Adenocarcinomas have a significant better survival than squamous cell carcinomas. Further, thyroid cancer is often diagnosed at an early stage. These factors aid the beneficial survival of thyroid cancer patients [30].

The data included in this study was derived from DCR, which covers the entire Danish population, and gives an excellent opportunity to investigate national changes in HNC trends. Further, Denmark provides universal, tax-financed healthcare to all citizens with uniform standards of treatment diminishing referral- and selection-bias. A great strength in this study is thus the non-selected group of patients in a definite geographical area.

In regard to survival of any cancer, the TNM classification is an important prognostic factor. Our study does not include the TNM classification. The diagnostic tools have improved significantly during our follow-up period, and in 2007 a fast-track program was introduced in Denmark to reduce waiting time for diagnosis and treatment [31]. This could be a contributing factor to the increased survival but might also contribute to the bias of stage migration following faster diagnostic and treatment procedures in 2007. In the 1990s, immunohistochemistry was implemented as a diagnostic tool for malignant melanomas [32]. This is worth noticing as it could affect the incidence of malignant melanomas [33]. Since the introduction of the fast-tract program in 2007, treatment of head and neck cancer is initiated within one month. All included patients were treated according to local protocol and national guidelines. The management of HNC varies depending on location and stage of the disease. Management has focused on single modality treatment to ensure the least possible amount of side effects, however severe cases and widespread disease is treated with combined modality treatment.

Our study comprises a non-selected nation-wide setting, and provides a comprehensive presentation of the trends in incidence and survival of HNC in Denmark in 1980–2014. We showed a significant increase in incidence and survival in HNCs. The rising incidence underlines HNC as an emerging public health problem.

Disclosure statement

The authors declare no potential conflicts of interest.

Additional information

Funding

Kathrine Kronberg Jakobsen was funded by The Danish Cancer Society (grant number: R165-A10483-16-S7) and The University of Copenhagen (grant number: A5090). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.