Incidence trends of mesothelioma in Norway, 1965–1999

Authors


Abstract

Asbestos exposure is considered to be the only important risk factor for malignant mesothelioma. The importation of asbestos to Norway increased after World War II and peaked in 1970. Stringent regulations took effect in 1977, and importation and use of asbestos practically ended in Norway in the late 1970s, until importation was prohibited in 1982. Our study aimed to analyze the incidence of mesothelioma in Norway according to temporal variation, to study the consequences of the use of asbestos and the asbestos ban effectiveness. An age-period-cohort model was used to analyze time trends for pleural mesotheliomas. From 1965–1999, the annual number of pleural mesotheliomas rose gradually both in males and females, and the highest annual number of pleural mesotheliomas was recorded in 1999 with 73 new cases diagnosed. The age-adjusted log linear drift of malignant mesothelioma of the pleura during the observation period rose 31.1% per 5 years among men and 15.9% among women. In 1995–1999, the age-adjusted incidence rate for men was 16.6 per million person-years for men and 2.3 for women. Cohort-specific risks increased for men born up to around 1935. After this the risks seem to stabilize. The rates were determined by age and by birth cohort. The delayed period effect of the asbestos regulation by the late 1970s will probably have its greatest effects on the mesothelioma rates around 2010. © 2003 Wiley-Liss, Inc.

Mesothelioma is defined by its histology. It occurs almost exclusively in the pleura and peritoneum. Mesotheliomas were rarely noted until Wagner et al., in 1960, reported a large case series among South African crocidolite miners and residents living near asbestos mines.1 Since then, the incidence of mesothelioma has been increasing in many countries,2, 3, 4, 5, 6, 7 especially in industrialized areas, and the incidence among men is considered a good indicator of the epidemic of asbestos-related diseases. In addition to the elevated mesothelioma incidence in areas with shipbuilding industry2 and asbestos mines,1 an increased risk has also been documented among workers in asbestos mills and factories, insulation manufacture and installation, gas mask manufacture, railroad machinery and other occupations involving inhalation of asbestos dust.8 Among men with mesothelioma of the pleura, the attributable risk for exposure to asbestos has been estimated at 88%, and among men with mesothelioma of the peritoneum 58%.9 The situation among women is less clear. In the Nordic countries, the proportion of avoidable mesotheliomas among men in the year 2000, if occupational exposure to asbestos was eliminated, was estimated at 83%.10 There is, however, international variation in mesothelioma rates, partly related to differences in the completeness of case ascertainment11 and partly related to differences in industrialization and asbestos exposure. Although European cancer registries have described constant increases in the incidence rates of pleural mesothelioma during the last decades,4, 5, 7, 12, 13, 14 most developing countries in Africa, Central and South America and Asia still show low incidence rates.15 The Netherlands, Scotland and England have the highest mesothelioma incidence rates in Europe.15 Outside Europe, Australia, an important asbestos producer, has the highest mesothelioma incidence rates.15

A prominent feature of mesothelioma is the long duration between initial exposure to asbestos and onset of the disease. The latency period is described to be on average 30–40 years, but the range varies considerably.16, 17 Peto et al.18 showed that mesothelioma death rates rise in proportion to about the third power of duration since initial exposure to asbestos, with the disease rarely occurring within 20 years of first exposure.

Asbestos importation started in Norway during the 1920s. Norway's only asbestos-cement factory began production in 1942. The importation of asbestos increased after World War II and peaked in 1970. The main use of asbestos in Norway was for manufacturing asbestos-cement products and for insulation in ships and industry. The first report on malignant mesothelioma in Norway was published in 1960.19 A case series by Semb, published in 1963,20 referred to the investigation by Wagner et al. but reported that “the present series gave no correlation to occupation.” By the time of the First International Congress on the Biological Hazards of Exposure to Asbestos in 1965, the carcinogenic and fibrogenic properties of asbestos had been generally accepted.21 In the beginning of the 1970s, public interest and concern about the working environment increased in the Scandinavian countries. Construction and insulation workers demanded the use of asbestos for insulation purposes be prohibited. In Norway, pressure from trade unions resulted in the formation of the Asbestos Committee, which was established by the Board of the Directorate of Labor Inspection in 1975.22 New regulations took effect in 1977, and due to these stringent regulations, importation and use of asbestos practically ended in Norway in the late 1970s (Fig. 1). The import of asbestos has been prohibited since 1982.

Figure 1.

Import of raw asbestos to Norway after World War II (based on information from Statistics Norway, 2002).

Asbestos exposure is considered to be the only important risk factor for malignant mesothelioma. In recent years, however, infection with simian virus 40 (SV40), which was a possible contaminant of the polio vaccines used in Norway from 1956–1962 and offered to small children, school children, pregnant women and occupationally exposed persons, is also considered to be a risk factor for mesothelioma.23 The risk to specific birth cohorts, especially from the period 1949–1962, may thus have been modified by this factor.

The aim of our study was to analyze the incidence of mesothelioma in Norway according to temporal variation, to study the consequences of the use of asbestos and to study the asbestos ban effectiveness.

MATERIAL AND METHODS

The investigation is based on data from the Cancer Registry of Norway, which has complete records of cancer cases from 1953. All hospitals are obliged to notify the Cancer Registry about all new cancer cases in the country. Moreover, the registry receives information on cancer histology and morphology from pathology laboratories and information on death certificates from Statistics Norway whenever cancer is mentioned. In Norway, each inhabitant has a unique personal identification number, which makes identification simple and reliable. Until 1970 all cancer cases were coded by a pathologist at the registry, later by specially trained personnel supervised by a pathologist. All cases coded as malignant neoplasm of the pleura, peritoneum or other sites specified as primary (International Classification of Diseases 7th edition [ICD-7] code 158, 163 and 199), and histologically coded as mesothelioma (code 9053 according to the Manual of Tumor Nomenclature and Coding, 1968 edition) were included in the investigation. In Norway histologic examination of tumors was not systematically carried out before 1965, and we therefore chose to start our trend study from this time.

An age-period-cohort (APC) model24 was used to analyze time trends for pleural mesotheliomas. The data were organized in 10 five-year age groups (40–44, …, 80–84, 85+), 7 five-year diagnostic periods (1965–1969, …, 1995–1999) and 16 ten-year synthetic (overlapping) birth cohorts (<1884, 1880–1889, …, 1950–1959). The model can be written as Dap = exp(Aa + Pp + Cc + log Nap), where Dap is the observed number of cases in age group a, period p, Nap, is the corresponding number of person-years, Aa, Ppand Ccare the effects of age, calendar period and birth cohort, respectively, and c = p − a, and the linear drift is defined through Cc.24Dapis assumed to be Poisson distributed. Pointwise confidence intervals for cohort-specific risks (exp(CcCc)) were based on standard normal approximation theory. The first birth cohort of men that was observed in all 7 diagnostic periods was born around 1910 and was used as the reference cohort (c').

RESULTS

Table I shows pleural mesothelioma cases grouped by calendar periods and age groups for each gender. From 1965–1999, the annual number of pleural mesotheliomas rose gradually both in males and females, and the highest annual number of pleural mesotheliomas was recorded in 1999 with 73 new cases diagnosed (not shown in table). The annual number of pleural mesothelioma cases in men rose from 8 in 1965 to 62 in 1999 and in women from 3 in 1965 to 11 in 1999. The age-adjusted log linear trend (drift) of malignant mesothelioma of the pleura during the observation period rose 31.1% per 5 years among men and 15.9% among women (not shown in table), and the age-adjusted incidence rates are still increasing both for men and women (Fig. 2).

Table I. NUMBERS OF MESOTHELIOMAS OF THE PLEURA REGISTERED IN THE CANCER REGISTRY OF NORWAY (OBS) AND INCIDENCE RATES PER MILLION (RATE) IN NORWAY BY AGE, GENDER AND PERIOD
Period Age
25–3435–4445–5455–6465–7475–8485+Total
Men         
 1965–1969Obs12411175141
 Rate0.91.73.210.824.717.016.64.3
 1970–1974Obs04325238066
 Rate0.03.92.523.131.424.10.06.7
 1975–1979Obs02132337234102
 Rate0.01.911.823.446.762.755.810.2
 1980–1984Obs0193345272117
 Rate0.00.89.029.853.368.323.511.5
 1985–1989Obs08215166475198
 Rate0.05.120.450.974.7108.452.619.1
 1990–1994Obs03154886625219
 Rate0.01.911.652.798.3132.948.520.6
 1995–1999Obs1222631116818285
 Rate0.61.214.565.9137.3133.7155.826.1
 TotalObs22287254385240351028
 Rate0.22.410.435.368.585.858.914.4
Women         
 1965–1969Obs003251011
 Rate0.00.02.41.86.02.50.01.4
 1970–1974Obs202411111
 Rate1.60.01.63.51.12.19.41.1
 1975–1979Obs010774019
 Rate0.01.00.05.87.37.20.01.9
 1980–1984Obs010854220
 Rate0.00.80.06.84.96.411.81.9
 1985–1989Obs0213104222
 Rate0.01.41.02.99.35.89.52.1
 1990–1994Obs03261112337
 Rate0.02.01.66.310.516.312.23.4
 1995–1999Obs00571518248
 Rate0.00.03.57.215.922.77.14.3
 TotalObs271337544410170
 Rate0.20.81.64.98.010.38.12.3
Figure 2.

Age-adjusted incidence rates of pleural mesothelioma according to the World Standard Population (1966) per million person-years among men and women in Norway, 1965–1969 to 1995–1999.

The annual number of pleural mesotheliomas among men younger than 55 years of age increased until 1985–1989 with the highest annual number in 1987, when 9 cases were recorded, but has remained relatively stable since this period (Fig. 3). Among men older than 54 years of age, the annual number and rates clearly continued to increase until 1999 (Fig. 3). In 1995–1999, the age-adjusted incidence rate for men was 16.6 per million person-years for men and 2.3 for women.

Figure 3.

Incidence rates of malignant mesothelioma of the pleura by age group per million person-years among men in Norway, 1965–1969 to 1995–1999.

Table II shows the number of mesotheliomas of the peritoneum and incidence rates per million in Norway by age and gender in the 3 ten-year periods from 1970 to 1999. The incidence rates among men remained stable during the period. Among women, the incidence rates rose from 0.1 in the 1970s to 0.9 in the 1990s. More cases of peritoneal mesothelioma were recorded among women than among men in the 1990s (19 vs. 14). Mesothelioma at other sites (pericardium, mediastinum, tunica vaginalis and unspecified) occurred in 17 men and 10 women and have not been included in the analyses.

Table II. NUMBERS OF MESOTHELIOMAS OF THE PERITONEUM REGISTERED IN THE CANCER REGISTRY OF NORWAY (OBS) AND INCIDENCE RATES PER MILLION (RATE) IN NORWAY BY AGE AND GENDER, 1970–1999
Period Age
25–3435–4445–5455–6465–7475+Total
Men        
 1970–1979Obs11042210
 Rate0.30.50.01.81.32.90.5
 1980–1989Obs02137215
 Rate0.00.70.51.44.12.40.7
 1990–1999Obs11343214
 Rate0.30.31.12.11.82.10.6
Women        
 1970–1979Obs0000112
 Rate0.00.00.00.00.51.00.1
 1980–1989Obs03122210
 Rate0.01.10.50.91.01.50.5
 1990–1999Obs11249219
 Rate0.30.30.72.14.51.30.9

Table III shows the goodness of fit for the different models for pleural mesothelioma. An age-cohort model fit the data well (degrees of freedom [df] of 45 vs. deviance of 43.5). An age-drift model fit the data reasonably (p-value for goodness of fit 0.12), but adding nonlinear cohort effects improved the model (p = 0.01). Adding nonlinear period effects to the model did not significantly improve the model fit (change in deviance: 5.8 with 5 df, p-value = 0.34). The cohort-specific risks relative to men born around 1910 are shown in Figure 4. Cohort-specific risks increased for men born up to around 1935. After this, the risks seemed to stabilize. The width of the confidence intervals indicates large uncertainties regarding whether the risk has started to decline for the youngest cohort.

Table III. AGE-PERIOD-COHORT ANALYSIS OF PLEURAL MESOTHELIOMA IN NORWEGIAN MEN IN 1965–1999: DEVIANCE AND DEGREES OF FREEDOM (DF) FOR VARIOUS MODELS
ModelDevianceDf.p-value goodness of fitModels to compareChange in deviance (Δdf)p-value for change in deviance
(I) age334.160<0.001   
(II) age + drift72.1590.12(II) vs. (I)262 (1)<0.001
(III) age + period63.6540.17(III) vs. (II)8.5 (5)0.13
(IV) age + cohort43.5450.54(IV) vs. (II)28.6 (14)0.01
(V) age + period + cohort37.7400.57(V) vs. (IV)5.8 (5)0.33
Figure 4.

Risks for different birth cohorts relative to men born around 1910, with 95% pointwise confidence intervals. The cohort effects are estimated in the age-cohort model.

DISCUSSION

Our present study shows that the risk of pleural mesothelioma is still increasing among men in Norway. There also is an increasing trend among women but at a much lower level. The rates among men are determined by age and by birth cohort with the highest incidence in the cohorts born 1935–1945. Norway's asbestos consumption peaked in the early 1970s, probably during the most active working period of these cohorts. The only well-established risk factor for mesothelioma is exposure to asbestos. In Norway, mainly men have been occupationally exposed to asbestos in asbestos-cement production, insulation work, shipyards and construction industry, and the differences in incidence among men and woman strongly support an occupational origin to the increased risk.

Similar trends, so far without any indications of effects of preventive measures, have been observed in Sweden,7 Denmark4 and France.25 In Finland the annual number of cases is still increasing for men older than 65 years of age but has not increased for younger men or women since 1990.5 Also mortality data from England, Wales and Scotland indicate that mesothelioma deaths are increasing and that the rates are determined by age and birth cohort.3 The asbestos use in Sweden, Denmark and Finland has been quite parallel to the use in Norway.5, 7, 12 Also in many other European countries, the use of asbestos began to decrease after the mid-1970s. This may explain the similar, still increasing trends of mesothelioma incidence observed in many European countries.

In an age-period-cohort model, the effects of etiologic factors are studied indirectly, and calendar period and birth cohorts are used as proxy variables. The effect of asbestos is naturally modeled through cohort effects, since men born around the same period of time would be exposed to similar levels of asbestos. This is reflected in our model fit, where significant nonlinear cohort effects were observed. The strict regulation of asbestos use by the late 1970s and the total ban of asbestos import in 1982 would most likely affect all cohorts and be seen as a period effect in the model. Due to the latency time, this effect would not materialize immediately but gradually over the following years. This should appear as nonlinear period effects in observation periods following the time of the ban. We did not, however, find significant period-specific departures from the linear drift. One reason for this might be that we have only followed the population for about 20 years after the ban. If the median latency time is 30–40 years, as is commonly described,16, 17 the delayed period effect of the asbestos regulation in 1977 would have its greatest effects on the rates around 2010.

We find it problematic to make forecasts for the number of cases of mesotheliomas in the future, based on projecting current cohort-specific risks. It is reasonable that the strict asbestos regulations introduced in 1977 will produce period effects in the next 5–10 years. Since these are future effects, it is not possible to estimate them from past observations, and making predictions without correcting for these period effects would most likely produce overestimates.

Changes in diagnostic procedures may bias trend studies. First by the 1960s, growing numbers of primary pleural and peritoneal tumors were recognized, and malignant mesothelioma became generally accepted as a distinct clinicopathologic entity.26 Including very old people in the analyses may, however, induce bias such as underestimation due to poor elucidation and competing diagnoses among old patients.

It is believed that different types of asbestos have different potency to cause mesotheliomas. It is not known which fiber types were imported to Norway. Information from Statistics Norway, however, shows that more than 70% of the asbestos import was from Canada or the USSR, countries that are known to export mainly chrysotile. This indicates that at least this proportion was made up of chrysotile, which is supposed to be less carcinogenic than colored asbestos like amosite and crocidolite.27 But asbestos was also imported from, among others, South Africa, Rhodesia and Australia, all countries with export of amphibole asbestos. Mesothelioma has mainly been associated with exposure to amphiboles,28 and exposure to amphiboles (crocidolite, amosite) may also explain the still increasing mesothelioma risk among men in Norway.

The incidence rates of mesothelioma of the peritoneum were almost similar in men and women. However, misclassification may occur and complicate the interpretation of the results. The diagnosis of peritoneal mesothelioma is difficult, both clinically and histopathologically, and incomplete biopsy material may lead to uncertainty. In a study of peritoneal mesotheliomas in Danish women, about 30% of the peritoneal mesothelioma cases were found to be misclassified ovarian or gastrointestinal tumors.29 Misclassification may probably also occur the other way around, although the extent of this has not been described. The attributable risk for exposure to asbestos has been estimated at 20% among women with peritoneal mesothelioma.9 The uncertainty according to asbestos exposure and peritoneal mesothelioma risk made us choose only pleural mesotheliomas for the age-period-cohort analyses.

Although asbestos is considered to be the principal cause of mesothelioma, factors such as infections and radiation also have been suggested to contribute to the etiology.30, 31 Infection by SV40 may independently contribute to mesothelioma induction.23 Whether infection with the SV40 has modified the mesothelioma trend is difficult to evaluate due to small population and few cases.

In conclusion, the risk of pleural mesothelioma is still increasing among men in Norway. There also is an increasing trend among women but at a much lower level. The rates are determined by age and by birth cohort with the highest incidence in the cohorts born up to around 1935. After this the risk seems to stabilize. The delayed period effect of the asbestos regulation by the late 1970s will probably have its greatest effects on the mesothelioma rates around 2010, but due to the long latency period for developing mesothelioma, the final conclusion about ban effectiveness will first be drawn in 10–20 years.

Ancillary