Incidence trends of adult primary intracerebral tumors in four Nordic countries

Abstract

Brain tumors are some of the most lethal adult cancers and there is a concern that the incidence is increasing. It has been suggested that the reported increased incidence can be explained by improvements in diagnostic procedures, although this has not been totally resolved. The aim of our study was to describe the incidence trends of adult primary intracerebral tumors in four Nordic countries during a period with introduction of new diagnostic procedures and increasing prevalence of mobile phone users. Information about benign and malignant primary intracerebral tumor cases 20–79 years of age was obtained from the national cancer registries in Denmark, Finland, Norway and Sweden for the years 1969–98 and estimates of person-years at risk were calculated from the information obtained from national population registries. Annual age standardized incidence rates per 100,000 person-years were calculated and time trends analyses were carried out using Poisson regression. The overall incidence of all intracerebral tumors ranged from 8.4–11.8 for men and 5.8–9.3 for women, corresponding to an average annual increase of 0.6% for men (95% confidence interval [CI] = 0.4, 0.7) and 0.9% for women (95% CI = 0.7, 1.0). The increase in the incidence was confined to the late 1970s and early 1980s and coinciding with introduction of improved diagnostic methods. This increase was largely confined to the oldest age group. After 1983 and during the period with increasing prevalence of mobile phone users, the incidence has remained relatively stable for both men and women. © 2003 Wiley-Liss, Inc.

Adult primary brain tumors are a heterogeneous group of tumors that are relatively uncommon, but among the most lethal of all adult cancers.1, 2 Introduction of radiologic imaging technologies such as computed tomography (CT) during the 1970s and magnetic resonance imaging (MRI) during the 1980s, together with stereotactic biopsy have significantly enhanced the diagnosis of brain tumors. An increase in the incidence has been reported, especially among the elderly,3, 4, 5, 6, 7, 8 and it has been debated whether this increase is due to a true increase in incidence, caused by an unidentified risk factor, or reflects improved diagnostic techniques.9, 10, 11, 12

The etiology of primary brain tumors is likely to vary with the histologic type of the tumor and tissue origin, even though the possibility of overall etiologic factors exists. A wide range of possible etiologic factors have been evaluated, such as genetic factors and environmental determinants, encompassing physical, chemical, and biologic agents.13, 14 However, few studies have convincingly linked specific exposures to brain tumors, and the evidence remains inconsistent. Known risk factors such as exposure to high doses of ionizing radiation to the head during childhood15, 16, 17 and hereditary factors such as neurofibromatosis and tuberous sclerosis14 are uncommon and explain only a small proportion of all brain tumor cases. Moreover, the variation in age, gender and ethnic group-specific incidence of different subtypes of primary brain tumors supports the idea of differences in etiology.12, 13, 14

One of the most controversial risk factors discussed currently is the exposure to radiofrequency fields emitted by mobile phones. The rapid increase in mobile phone use during the last decade has led to a growing concern among the public about the possibility that mobile phone use may be associated with an increased risk of brain tumors. Epidemiological studies published so far have found little evidence for an increased risk,18, 19, 20, 21, 22, 23, 24 yet, all of them have low power to study long-term exposure. The population in the Nordic countries has used mobile phones since the 1980s and there has been an exponential increase of users during the 1990s (Fig. 1). In other words, during the last year a large part of the Nordic population has increased their exposure to radiofrequency fields directly to the head.

Figure 1. Annual proportion of mobile phone subscribers in four Nordic populations during 1987-98. Denmark is indicated as— —, Finland as– - - –, Norway as —— and Sweden as - - -.

The aim of our study was to explore the incidence trends of adult primary intracerebral tumors in Denmark, Finland, Norway, and Sweden during a period with introduction of new diagnostic procedures and increasing prevalence of mobile phone users in the population.

MATERIAL AND METHODS

Anatomic site of all tumors are coded according to a modified version of the 7th revision of the International Classification of Diseases (ICD7) in the national cancer registries in Denmark, Finland, Norway and Sweden. This version of ICD7 includes both malignant and benign brain tumors. In addition to ICD7, each country has updated the coding procedure over the years. Denmark has coded tumors according to the 9th revision (ICD9) from 1978, Norway according to the 10th revision (ICD10) from 1993, and Sweden according to ICD9 during 1987–92, and ICD10 from 1993 onward. Additionally to the ICD coding, the national cancer registries have coded the tumors according to histopathology. Denmark has coded according to the International Classification of Diseases for Oncology (ICD-O) from 1978, Finland according to a nationally developed system, Norway according to the Manual of Tumor Nomenclature and Coding (Motnac)25 and ICD-O from 1993, and Sweden according to WHO/HS/CANC/24.1 histology code26 and Systematized Nomenclature of Medicine (SNOMED)27 from 1993.

Information about incident benign and malignant brain tumor cases, coded as ICD7 193.0, in men and women aged 20–79 years was obtained from the national cancer registries in Denmark, Finland, Norway and Sweden. We used the information available for the period 1969–98. In the overall analysis, we defined primary brain tumors as intracerebral tumors and excluded neurinomas, meningiomas, lymphomas and pineal gland tumors. All intracerebral tumors combined, glioma, other intracerebral tumors (i.e., medulloblastoma and supratentorial primitive neuroectodermal tumor), and unspecified brain tumors were analyzed for the entire study period. The glioma incidence was also analyzed according to age at diagnosis, where age was categorized in three 20-year age groups (20–39, 40–59, 60–79 years). The glioma group was further analyzed by anatomic site. Finland was excluded from this analysis because the Finnish cancer registry lacks detailed information about anatomic site. In Denmark and Sweden information on anatomic site was analyzed from 1987 and in Norway from 1993.

In addition, more detailed analyses were made for the years 1993–98 using data from Denmark, Norway and Sweden. In these analyses intracerebral tumors were further defined as tumors in the neuroepithelial tissue with ICD-10 code C71 and the International Classification of Diseases for Oncology (ICD-O). Tumors were classified into nine different subgroups according to cellular origin and grading as shown in Table I.

Table I. Histologic Classification of Neuroepithelial Brain Tumors (ICD-10 C71) After ICD-O-2 Codes Given by the World Health Organization

Histologic type	ICD-O-2
Glioblastoma	9440–9442, 9460
Oligodendroglioma	9382, 9450–9451
Pilocytic astrocytoma	9421
High-grade astrocytoma	9380, 9401/3
Low-grade astocytoma	9400-9401/1, 9410–9421
Neuronal-glial	9490, 9505–9506
Other high-grade	9390/3, 9392, 9470–9473, 9490
Other low-grade	9383–9390/0, 9391, 9492
Unspecified	8000

Annual incidence rates per 100,000 person years were calculated for men and women separately, standardized for age in 5-year age groups, with the Nordic countries in 1985 as standard population (Table II). Estimates of person years at risk were calculated from the information obtained from national population registries. The incidence rates in the figures are presented as 2- or 3-year moving averages. Three-year moving averages were used only in the age specific analyses because of small numbers in each age category.

Table II. Population Weights from 1985 in Denmark, Finland, Norway, and Sweden for Ages 20–79 Years Used in the Aged Standardized Incidence Calculation

Weight	Age
0.107	20–24
0.102	25–29
0.105	30–34
0.116	35–39
0.100	40–44
0.081	45–49
0.073	50–54
0.074	55–59
0.074	60–64
0.065	65–69
0.058	70–74
0.045	75–79

Analyses of incidence time trends were carried out using Poisson regression to estimate the average annual changes in the incidence over time with corresponding 95% confidence intervals (CI) using data for each individual year. In these analyses, the study period was divided into three separate time periods to determine whether there was evidence of an upward or downward trend during the defined years. The earliest period was 1969–76, a time before CT was in widespread use. The second period 1977–83 was chosen as a time period when a new technology (CT) for diagnosis was under development, but before the introduction of handheld mobile phone technology. The third period 1984–98 was chosen as a time period when CT was well established and the use of handheld mobile phones began and increased rapidly.

Poisson regression was also used to calculate incidence rate ratios (IRR) for each year with corresponding 95% CI. We used 1983 as reference year in our IRR calculations and adjusted for age and country. The year 1983 was chosen as the reference for two reasons. First, CT was introduced during the 1970s in Denmark, Finland, Norway and Sweden and was well established in clinical use in the beginning of the 1980s and second, handheld mobile phones were not used before 1984.

RESULTS

During 1969–98, a total of 43,120 (55% men) adult primary intracerebral tumor cases were identified. Incidence rates and trends in the countries evaluated were similar, with a slightly higher incidence in Denmark and a lower incidence among men in Finland. The gender distribution among cases was similar in all countries, with a slightly lower proportion of men in Finland (50%). The proportion of microscopically confirmed diagnoses (biopsy, operation material, autopsy or cytology) was stable (around 80%) during the time period of the study in all countries with the exception that during the last 10 years confirmation increased in Sweden (to ∼90%) and decreased in Denmark (to ∼70%).

All intracerebral tumors combined showed a small increase in the age standardized incidence during the study period with incidence rates ranging from 8.4 to 11.8 per 100,000 for men and 5.8 to 9.3 per 100,000 for women, corresponding to an average annual increase of 0.6% for men (95% CI = 0.4, 0.7) and 0.9% for women (95% CI = 0.7, 1.0). Glioma increased annually with an average of 0.7% for men (95% CI = 0.5, 0.9) and 0.6% for women (95% CI = 0.4, 0.8) (Fig. 2). The incidence in the group “other intracerebral tumors” has been fairly constant over the years, with only a minor increase. The incidence of unspecified brain tumors has increased slightly during the study period, except for the last 10 years when the incidence has been stable.

Figure 2. Age standardized incidence of adult primary intracerebral tumors by histological type shown as 2-years moving average in Denmark, Finland, Norway and Sweden, 1969–98. Gliomas are indicated as □ = men and ▪ = women. Unspecified tumors are indicated as ▵ = men and ▴ = women. Other intracerebral tumors are indicated as ○ = men and • = women.

The overall and glioma incidences for men during 1969–76 remained virtually constant with an annually change of −0.03% (95% CI = −1.2, 1.2) and −0.003% (95% CI = −1.4, 1.3) respectively. An upward trend in the incidence was observed during the period 1977–83, when the incidence of all intracerebral tumors combined increased annually with 1.7% (95% CI = 0.3, 3.1). For glioma a non significant annual increase of 1.4% (95% CI = −0.2, 3.0) was observed. For women, the incidence of all intracerebral tumors combined changed only slightly during the period 1969–76; on average 0.2% (95% CI = −1.0, 1.5) per year. During 1977–83 the incidence increased annually with 1.8% (95% CI = 0.4, 3.2). The corresponding numbers for glioma were 2.8% (95% CI = 1.2, 4.3) and 1.2% (95% CI = −0.6, 2.9), respectively. After 1983, the annual incidence of all intracerebral tumors combined decreased slightly for men with an annual average of −0.6% (95% CI = −1.0, −0.2) and for women with −0.4% (95% CI = −0.8, 0.0). During the same period the glioma incidence changed on average −0.1% (95% CI = −0.6, 0.3) for men and 0.2% (95% CI = −0.3, 0.7) for women.

Results from the time trend analyses of all intracerebral tumors combined and glioma showed that the rate ratios were higher after 1983 compared to the time period before 1983 for both men and women. As illustrated in Figure 3, most IRRs of all intracerebral tumors combined and glioma remained unchanged after 1983 in each separate analysis and only a few IRRs were significantly different from the reference year.

Figure 3. Incidence rate ratios (IRR) with 1983 as reference year with 95% CI of all adult primary intracerebral tumors ▵ = men and ▴ = women, and of glioma □ = men and ▪ = women in Denmark, Finland, Norway and Sweden, 1969–98. The distances between the lines are the same as one unit (1.0) IRR.

The most common anatomical locations of glioma were in the frontal and temporal lobes. The incidence of glioma in different anatomical locations varied over the studied period without a specific pattern, except for a decrease in the “unspecified location”, and a slight increase in “other locations” (other than frontal, temporal or parietal lobe) (Fig. 4).

Figure 4. Age standardized incidence as 2-years moving average of different anatomic sites of glioma for men (a) and for women (b) in Denmark, Norway and Sweden, 1987–98. Incidence before 1993 does not include Norway. ○ = frontal lobe, □ = temporal lobe, ▵ = parietal lobe, ▪ = other sites and ▴ = unspecified. The total number of cases for each anatomic site are presented in the figure.

Overall, the incidence of glioma was higher among men than women, but the gender difference was smaller in the youngest age group (Fig. 5). The incidence remained relatively stable over time in the 20–39-years and 40–59-year age groups. During 1984–98, the incidence changed on average −0.2% (95% CI = −1.2, 0.9) per year for men and 0.3% (95% CI = −1.0, 1.5) for women in the 20–39-year age group. The corresponding results for the 40–59-year age group was −0.7% (95% CI = −1.4, 0.0) for men and 0.03% (95% CI = −0.9, 0.9) for women. In the oldest age group (60–79 years) an increase was observed in both men and women. Overall, the average annual increase was 1.0% (95% CI = 0.7, 1.3) for both men and women. The main increase, although not statistically significant, was seen 1977–83 with an average of 1.6% (95% CI = −0.8, 4.1) per year for men and 2.3% (95% CI = −0.4, 4.9) for women. During 1984–98, the incidence changed on average −0.5% (95% CI = −0.3, 1.2) per year for men and 0.4% (95% CI = −0.4, 1.1) for women.

Figure 5. Incidence as 3-years moving average of age specific analyses of glioma in Denmark, Finland, Norway and Sweden, 1969–98. Ages 60–79 years are indicated as □ = men and ▪ = women, ages 40–59 years are indicated as ▵ = men and ▴ = women, ages 20–39 years are indicated as ○ = men and • = women.

Glioblastoma was the most common histological type with incidence rates ranging from 3.3 to 5.1 per 100,000 for men and from 2.1 to 3.5 per 100,000 for women in 1993–98 (Fig. 6). Incidence for the more rare histological types not presented in Figure 6 were all <0.3 per 100,000. During this period, the incidence remained virtually unchanged for all diagnoses except for glioblastoma where an average annual increase could be observed in both men (9.2% [95% CI = 6.2, 12.1]) and women (8.8% [95% CI = 5.3, 12.4]), and for unspecified brain tumors where a decrease could be observed during the same period. Figure 7 shows the incidence of glioblastoma and unspecified tumors for the groups 20–59-year age group and 60–79-year age group. The increase of glioblastoma was confined to the oldest age group where an average annual increase could be observed in both men (12.4% [95% CI = 8.2, 16.6]) and women (10.5% [95% CI = 5.9, 15.0]). The incidence of unspecified brain tumors in this age group decreased annually for men with −5.3% (95% CI = −10.0, −0.5) and changed on average −0.8% (95% CI = −5.7, 4.0) per year for women.

Figure 6. Age standardized incidence in Denmark, Norway and Sweden as 2-years moving average in the period 1993–98 for men (a) and women (b). □ = glioblastoma, ▵ = high grade astrocytoma, ○ = low grade astrocytoma, ▪ = oligodendroglioma and ▴ = unspecified.

Figure 7. Age standardized incidence in Denmark, Norway, and Sweden as 2-years moving average in the period 1993–98 for glioblastoma and unspecified tumors for men (a) and women (b). Ages 20–59 years ▵ = glioblastoma, ▴ = unspecified. Ages 60–79 years □ = glioblastoma, ▪ = unspecified.

DISCUSSION

In this population based study of adult intracerebral tumors we found an increase in the overall incidence in the Nordic countries during 1969–98 that was confined to the late 1970s and the early 1980s. The increase was most apparent in the 60–79-year age group, whereas the younger age groups had more stable incidence rates. Since 1984, the incidence has been stable, or even shown a minor decreasing trend. A similar pattern was seen for glioma, with the exception of the 60–79-year age group in which an increasing incidence during the last part of the 1990s was observed. The latter increase is partly explained by a decrease in the incidence of unspecified brain tumors during the same time period.

Our study has a number of advantages. The Nordic countries share many important features, including similar health care systems, nationwide cancer registries of high quality, and population registries based on a unique personal identification number assigned to all residents in the Nordic countries. All physicians in hospitals and other establishments for medical treatment under public or private administration must report all cases of cancer to the national cancer registry. Furthermore, reports are also submitted by pathologists and the two independent notifications systems ensure a high completeness of coverage.28, 29, 30, 31 The registries also cover cases without microscopic or radiological confirmation and tumors detected at autopsy. The registries include both malignant and benign brain tumors, and therefore a detected change in the brain tumor incidence cannot be explained by a shift of the classification of benign tumors as malignant or vice versa. Changes in the classification systems are unlikely to affect the trends in the overall analyses, where broad diagnostic categories were analyzed. Furthermore, localization and histopathology were coded according to the older versions in parallel with the new versions of the ICD coding system in all countries except Norway. We analyzed specific histological types from 1993 when a uniform coding system was used in all included countries. The registration deficit of information on brain tumors in the national cancer registers is regarded to be <2%.28, 29, 30, 31

The increase in intracerebral tumor incidence coincides with the introduction and more widespread use of CT and rates has since then been stable, which is in accordance with previous studies.32, 33, 34, 35 CT was introduced during the 1970s in the studied countries, and was established in clinical practice in the early 1980s. The introduction of CT has greatly enhanced the diagnostic procedure and increased the detection rate of brain tumors, and it is reasonable to assume that changes in clinical routines has led to an increased number of detected cases, especially among older patients.32, 33

In accordance with previous studies,8, 14, 36 we found the lowest intracerebral tumor incidences in the younger age groups and an increasing incidence with age. Our results also confirm previous observations of a higher incidence of glioma in men than women,14, 37, 38 although in the 20–39-year age group the difference between genders was small. The most likely explanation is that the gender difference is most pronounced for high grade gliomas, which are more common among older persons, whereas low grade gliomas dominate among the young.37

Gliomas can develop anywhere in the brain, but the most frequent anatomic locations are temporal, frontal and parietal lobes.37 We found no indication of any consistent changes in these locations over time. However, tumors with unspecified sites were decreasing during the observation period, whereas tumors in other anatomic sites in total were increasing. Adding Norway to the analysis from 1993 has led to a slight overestimation of the changes in the trends for other sites and unspecified sites because the incidence in other sites is higher in Norway compared to Sweden and Denmark, and the incidence in unspecified sites is lower. However, when Norway is excluded completely from these analyses, a decrease in unspecified sites and an increase in other sites are still evident. One explanation might be that MRI facilitates detection of tumors in particular regions of the brain, such as the brain stem and posterior fossa that are less easily visualized by other methods.39 A more detailed analysis of the different anatomic sites combined in the group “other locations” showed that the sites of brain stem and posterior fossa seem to be responsible for the increased incidence in this group (data not shown).

Brain tumors vary widely by site of origin and morphologic features. It is obvious that the validity of the data depends largely on the accuracy of the specific diagnoses and the proportion of histologically confirmed cases. Therefore, the incidence of specific brain tumor types may not be comparable over time or between hospitals or countries. A true time trend increase in the incidence of a specific histologic type may be difficult to distinguish from an apparent increase due to improved diagnostic methods. A shift between diagnostic categories may not reflect a real change in the incidence, but rather a shift in diagnostic procedures. In the analyses of specific histologic types during the period 1993–98, we found an increase in the incidence of glioblastoma, and a decrease in the incidence of unspecified tumors. This pattern was confined to the older age group, and can probably be explained by application of more rigorous diagnostic procedures among older patients.

Hand-held analogue mobile phones were introduced in late 1980s, but the more widespread use of mobile phones in the general Nordic population began with the introduction of digital phones (GSM) in the early 1990s. The early users were predominantly young men with corporate subscriptions, whereas during the later part of the 1990s usage has become more common also among teenagers and older persons. There is a growing concern among the public that exposure to radiofrequency fields from mobile phones might increase the risk of brain tumors. Our results do not indicate any short term effect of mobile phone use on the intracerebral tumors incidence at the population level. If exposure to radiofrequency fields from mobile phones has a short-term promotional effect on intracerebral tumor development, we would have expected to see an increase in the incidence among men in the 40–59-year age group during the later part of the study period and an increase in the topographic location considered to be primarily exposed (temporal and parietal lobes),40, 41 assuming that no other risk factors have decreased since the late 1980s. Mobile phones have not been in use long enough in the general population, however, to allow evaluation of long term effects.

Incidence rates of brain tumors tend to be higher in the Nordic countries compared to the rest of the world, and other European countries and North America have a higher incidence than Africa, Asia and South America.42 The reason for this differences in incidence between continents remains unclear, although the availability of health care technology is likely to explain some of the differences.

In summary, the incidence of adult intracerebral tumors in the Denmark, Finland, Norway and Sweden has been stable or even declined over the last 15 years. Increases in the incidence during the late 70s and early 80s coincided with improved diagnostic methods, and were largely confined to the oldest age groups. The overall incidence has remained stable during the period after the introduction and wide spread use of mobile phones.

Acknowledgements

The authors acknowledge the cancer registries in Denmark, Finland, Norway and Sweden for providing cancer data.