Articoli scientifici
19/12/2025

Accuracy of mesothelioma occurrence forecasts, characteristics of cases, and incidence rates in Lombardy Region (Northern Italy)

, , , , , ,

Introduction

Italy, with its about 60 million inhabitants, has been a great producer and consumer of all types of asbestos (4.5 million tons) from 1933 until 1992 – the year of the asbestos ban – and is among the countries with the highest mesothelioma burden.1-3 A National Mesothelioma Registry (Registro Nazionale Mesoteliomi, ReNaM) has been established at the National Institute for Insurance against Accidents at Work (Inail, Rome) according to the Italian law 308/2002: ReNaM task is to monitor mesothelioma occurrence in Italy. It is noteworthy that some Regions started registering mesothelioma cases earlier.4 In Lombardy Region (North-Western Italy), a mesothelioma registry was established in 2000 (Registro Mesoteliomi Lombardia, RML). Lombardy has been – and is nowadays – a highly industrialised Region and counts about 10 million people: for these reasons, in the period 2015-2019, 27% of mesothelioma cases in Italy were recorded in this Region.4
Given the long latency between asbestos use and mesothelioma occurrence, there is great interest in estimating the future mesothelioma burden for public health purposes. Forecasts of mesothelioma occurrence has been made in several countries using different modelling methods, mostly Poisson age-cohort or age-period-cohort (APC) models applied to incidence or mortality data. In Italy, projections of pleural cancers have been provided for the whole country,5-7 for Veneto Region (North-Eastern Italy),8 and for Sicily Region (Southern Italy)9. In Lombardy, the number of cases were estimated for the period 2013-2029,10 based on incidence rates of mesothelioma (all anatomic sites) in the period 2000-2012, using gender-specific age-cohort models.11
Notwithstanding the many projections, to the Authors’ knowledge, there are no studies that assessed the accuracy of predictions over time. The main objective of this work is to evaluate the accuracy of forecasts for Lombardy by comparing observed and predicted number of cases for the period 2013-2024. Moreover, clinical and occupational characteristics of mesothelioma cases were described and incidence rates were reported for the Lombardy Provinces.

Methods

Ascertainment of mesothelioma cases

The RML is a population-based registry. Since 2000, it collects all mesothelioma cases (all sites) reported among Lombardy residents at the time of diagnosis.12 The primary source is mesothelioma reporting (compulsory by law 277/1991 and law 81/2008) from every hospital. Completeness of ascertainment is periodically verified using several sources, notably pathology, hospital admission, and mortality databases. These sources become available to RML from the responsible institutions (Lombardy Region, pathology departments and others) with months of delay. Once this is information is acquired, it is necessary to contact many hospitals to request and obtain clinical/pathology/mortality records of potential cases. Finally, evaluation of clinical information is performed (including imaging, cytology, and histology) to verify diagnosis. 
Following ReNaM guidelines,4 final diagnosis is established on a case-by-case basis considering all available clinical information, classifying mesothelioma cases as “certain” (histological diagnosis, possibly with immuno-histochemical confirmation and imaging), “probable” (usually, cytology suggesting mesothelioma plus imaging), “possible” (positive imaging), or “non-mesothelioma”. Morphology is defined and coded according to the WHO histological classification and the International Classification of Diseases for Oncology (ICD-O), Third Edition.
These activities are completed as of 2024. Therefore, for the period 2000-2024, the available information includes counts of cases, gender, and mesothelioma site/morphology. 

Assessment of asbestos exposure

Mesothelioma patients (or their next-of-kin) are interviewed (most often inteviewes were performed face-to-face) by trained personnel using a standardised questionnaire to collect information on lifetime occupational history (industry, occupation, work environment characteristics).4 Information on lifetime extra-occupational asbestos exposure is also collected, including:

  • brushing or washing of dirty work clothes of an exposed next-of-kin (para-occupational or familial exposure);
  • ironing on asbestos-coated ironing-boards;
  • small repair works;
  • thermal insulation;
  • use of asbestos gloves or asbestos-containing objects (home-related exposure);
  • presence of industries in the vicinity, e.g., asbestos cement production plants (environmental exposure).

Based on this information, lifetime occupational history is evaluated during weekly meetings of the RML personnel. Asbestos exposure is classified as occupational (certain, probable, or possible), para-occupational, home-related, or environmental. Cases ever exposed to asbestos at workplace are classified as occupationally exposed irrespective of other exposure sources. For patients never occupationally exposed to asbestos, the following exposure hierarchy is considered: para-occupational > environmental > home-related. However, the final decision is taken on a case-by-case basis, considering also information on time since first exposure and length of exposure to each source. Medico-legal procedures (e.g., reporting to health authorities and Inail to claim compensation) are undertaken when needed.

Data

Once all clinical and occupational information regarding mesothelioma cases is collected and evaluated, data is archived in a standardised ReNaM database. Moreover, coding of each work period is also performed directly in this database: industries and occupations are coded following respectively the Italian ATECO (Classificazione delle Attività Economiche) classification 1991 and CIP (Classificazione delle Professioni) classification 1991. Complete data are finally transmitted to ReNaM and are used to produce official reports for the Lombardy Region. Given the high number of cases, all these activities are time-consuming and are still ongoing for years 2022-2024. Therefore, a complete description of cases can only be made for the period 2000-2021. 
Population data by year, gender, age, and municipality of residence (used to calculate rates) is periodically downloaded from the Italian National Institute of Statistics (Istat) website (http://demo.istat.it/index_e.html).

Statistical analysis

In a previous paper,10 Poisson age-cohort models were fitted using 5-year categories for age at diagnosis (reference: 70-74 years) and birth cohort (reference: cohort 1920-1924) to make projections.11 The gender-specific age and cohort regression coefficients were then applied to population data (actual population data was used for the years 2000-2015, while estimated population data was used for the years 2016-2019) to calculate projections of the numbers of mesothelioma cases and their 90% confidence intervals in the years 2013 to 2029. 
In this work, observed and predicted numbers of cases were compared in the period 2013-2024, years in which ascertainment of mesothelioma cases is complete (see above). Moreover, clinical and occupational characteristics of cases were described in the period 2000-2021, years in which all patients or their next-of-kin have been interviewed and asbestos exposure has been evaluated (see above).
Also, for the period 2000-2021 (complete data), crude and age-standardised rates (ASR, age 0-99 years) were calculated by year and Province using various standard populations.
To evaluate incidence trends over time, the Lombardy 2001 standard population (in which 17.9% of residents was aged 65 or more years) was used.
For comparison with other Italian regions and countries, three populations were used as a standard: Italy 2001, Europe 2013, and World (Segi’s population). These standard populations have different age structure: Italy 2001 and Europe 2013 are older (18.7% and 19.5% aged 65 years or more, respectively), while World population is younger (7% aged 65 years or more). As a consequence, ASR using as standard the World population are expected to be lower, because this population gives more weight to younger ages, when mesothelioma rates are low. In Lombardy, the average proportion of people aged 65 years or more in the 2000-2021 period was 20.5%: therefore, crude rates are expected to be not very different from ASR calculated using Italy and Europe as standards. Data management and statistical analyses were performed with Stata 19 (StataCorp. 2025). Standardised rates were calculated using the Stata command distrate.13

Results

Observed versus predicted numbers of cases, 2013-2024

In 2013-2024, in both genders the observed cases fell within the 90% confidence bounds around forecasts, except for men in 2013 (Figure 1). In this period, the yearly count of cases was constant in men and slightly declining in women. The highest count of cases (both genders) was observed in 2013 (see online supplementary material, Figure S1). In men, 3,588 cases were observed, 435 more (36 per year on average) than the 3,153 cases predicted in the previous paper10; in women, 1,671 cases were recorded, 90 less (7 per year on average) than the 1,761 predicted10 (Table 1). Overall, in the period 2000-2024, 9,716 cases were recorded: 6,449 (66.4%) in men and 3,267 (33.6%) in women.

 

Characteristics of cases, 2000-2021

Interview and evaluation activities were completed as of 2021 for 5,568 men and 2,869 women (Table 2). The more represented age classes are 70-79 years in men and 75-84 years in women. Most patients were 65+ years old at diagnosis (77.9% men, 82.3% women). Pleural mesothelioma accounts for more than 90% of cases. Diagnosis was classified as certain in more than three quarters of cases. Epithelioid mesothelioma was by far (>60%) the more frequent histological type. Interviews were obtained for more than 90% of cases, in men mostly from the patient and in women in equal proportions from patients and their next-of-kin.

 

Evidence of occupational exposure was found for more than two thirds in men and about one third in women (Table 3). In 20.2% of men and 41.8% of women there was no evidence of past asbestos exposure. Median time since first occupational exposure (latency) was over 50 years. Most men were exposed to asbestos in metalworking and metallurgy industries (35.4%) and in construction (33.7%). Most women had been exposed in non-asbestos textile (53.3%) and clothing (13.4%) industries.

 

In both genders, the proportion of cases with occupational exposure decreased over time, while the proportion of cases with no evidence of exposure increased (Figure S2). The proportion of female cases exposed to asbestos in textile industries decreased in the last years, while exposure of men in other industries appears rather constant (Figure S3).

Incidence rates, 2000-2021

In men, ASR (Standard: Lombardy 2001) increased in the period 2000-2013 and then slightly declined, while in women rates were constant in the period 2000-2018 and the slightly decreased (Figure S4). 
In men, age-specific rates showed declines of 1%-6% per year for ages below 70 years, were stable for age 70-74 years, and increased by 3%-5% per year for ages 75+ years (Figure S5, upper panel). In women, age-specific rates showed a decrease of 3%-5% per year for ages below 70 years, were stable for age 70-74 years, and increased by 2%-3% per year for ages 75-84 years, and were stable for ages 85+ years (Figure S5, lower panel).
In the whole period, average crude rates and ASR rates for the whole Lombardy Region using Italy or Europe as standard population were between 5 and 6 per 100,000 person-years in men and between 2 and 3 in women; using World population as standard, rates were much lower: 2.3 (men) and 0.9 (women) (Table S1). In men, ASRs (standard: Italy 2001) higher than the Lombardy average were observed in the Provinces of Pavia, Bergamo, Lecco, and Milan (Table S1 and Figure 2). In women, the highest rates were observed in the Provinces of Pavia, Bergamo, Lodi, Monza and Brianza, and Sondrio.

Discussion

Observed versus predicted number of cases, 2013-2024

In this study, a fair agreement was found between numbers of observed cases and the 90% confidence bounds around forecasts that were estimated using age-cohort Poisson models for the period 2013-2024. Looking at average counts and predictions, cases in men were moderately underestimated and predicted cases in women were slightly overestimated. No indication of decrease of cases was found in men in the last years, while a slight decline was observed in women.
In comparing observed and predicted cases, it should be considered that forecasts had been made using population projections (actual for the period 2000-2015, estimated for the years 2016-2029).10 Actual population numbers in the period 2016-2024 were slightly lower (3 to 6%) than the estimates. Therefore, forecasts for the years 2016-2024 had been slightly overestimated.
Forecasts of mesothelioma occurrence were made in several countries using different models applied to incidence or mortality data (see lists of references in previous papers)6,7,10.
In Italy, the first projections of pleural cancer mortality (2000-2034) were made in men on data from 1970 to 1999.5 Corrections for cause of death misclassification and misdiagnosis were applied and several models were fitted: Poisson age-cohort and APC models predicted numbers of deaths (around 800 per year in 2015-2024) comparable with a model which took into account patterns of asbestos consumption over time. The peak was estimated to occur in 2020-2024, followed by a slow decline. 
A second Italian study used pleural cancer mortality 1970-2014 to make predictions (in men and women separately) for the period 2015-2039.6 Poisson age-period-cohort and age-cohort models produced almost identical predictions, with the peak of cases expected in the period 2020-2024 (according to the APC mode: 1,429 deaths per year, 1,066 men and 366 women), followed by a slow decline until 2035-2039 (1,145 deaths per year, 858 men and 288 women; numbers comparable to counts observed in 2010-2014).
A third Italian study used distributed lag non-linear  models (DLNM) applied to mortality data of the period 1970-2014 (men only) to make predictions until 2040, based on past asbestos consumption figures.7 According to that model, the peak of deaths was predicted to occur in 2021 (1,122 deaths), with a subsequent rapid decline (344 deaths predicted in 2039; a number comparable to the observed deaths in the 1980s).
A study conducted in Veneto Region used incidence data 1987-2010 to make predictions (both genders) for the years 2011-2026 by fitting Bayesian APC models.8 The peak occurred in 2010 and a slow decline was predicted thereafter.
A study in Sicily Region, based on 1998-2016 data on mesothelioma incidence, mortality, and survival, using an APC model, estimated that incidence and mortality will increase until 2026.9
In summary, national and regional Italian studies (excluding the one in Sicily), using age-cohort or APC models, indicated a slow decline after the peak of mortality or incidence of pleural cancer or mesothelioma: conversely, the DLNM model predicted a more rapid decline.
In the last report, ReNaM reported a peak of mesothelioma incidence in 2013 (1,854 cases, both genders); in 2014-2019, cases were 1,682, 1,740, 1,658, 1,622, 1,631, and 1,516, respectively; although in a few Regions registry activities have been stopped or greatly reduced in the last years (especially Campania since 2016), data suggests that the mesothelioma peak in Italy has already occurred and that cases are slowly decreasing over time.

Characteristics of cases, 2000-2021

In the period in which registry activities have been completed (2000-2021), almost 8,500 cases were recorded, with a male to female ratio of 2, mostly among individuals aged 65+ years. Evidence of occupational exposure was found for the majority of men, but for only a minority of women. The low proportion of occupationally exposed women (in previous studies) led some authors to hold that that most cases in women are “idiopathic”, i.e., due to unknown determinants other than asbestos; this view is not supported by the scientific literature and possibly biased by undisclosed conflicts of interests.14
A couple of factors may explain the low fraction of occupationally exposed women. First, unrecognised occupational exposures: historically, most studies had been made in industrial sectors in which most workers were men. Thus, asbestos exposure might have been overlooked in sectors where the workplace was predominantly composed of women. As an example, thanks to RML investigations, only at the beginning of this century the widespread use of asbestos has been demonstrated in the textile industry.15 Second, unrecognised environmental exposures. Situations of heavy asbestos pollution (e.g., residence near asbestos cement production plants) are easily identifiable, but are quite rare. Conversely, it is very difficult or nearly impossible to investigate widespread low/moderate exposures, because individuals are unaware of nearby asbestos sources (e.g., deteriorated asbestos roofs in the neighbourhood). To address this problem, an ecological study was performed in Lombardy showing that the spatial distribution of cases without identified asbestos exposure (in majority women) matches the distribution of occupationally exposed cases.16 This finding suggests that a large part of mesothelioma cases may have occurred due to unrecognised occupational or environmental asbestos exposures.
The median latency was very high for occupational cases (more than 50 years in both genders). Latency figures were provided for descriptive purposes only: inference on age at diagnosis or latency based on cases only is incorrect.17-19 
In men, most cases had been exposed to asbestos in metal working and metallurgy and in the construction sector. The construction industry is responsible for a very high cancer burden (due to exposure to asbestos and other carcinogens), either for mesothelioma or lung cancer.20,21 In women, a very large proportion of patients has been exposed in non-asbestos textile and clothing industries, as previously observed.22 
The decrease of proportion of cases with occupational exposure over time is expected to grow in the next years as we move away from the year of asbestos ban in 1992.

Incidence rates, 2000-2021

Although age-adjusted incidence rates in the 2000-2021 were slightly increasing in men and stable in women, the number of cases showed a strong increase, especially in men, from about 200 per year in the early years to around 300 per year in the recent years. At first sight, this may seem contradictory, but can be in part explained by the aging of the population over time. Yearly rates were increasing only for older individuals. Crude incidence rates (and hence the number of cases) are driven by the high rates at older ages, and this effect gets stronger as the population gets older over time: the population aged 65+ years increased from 14.1% in 2000 to 20.2% in 2021 in men and from 20.7% to 25.5% in women. Again, these values witness the efficacy of the asbestos ban, because older people had more occasions of occupational exposure than youngers.
ASR using different standards produced different results (similar and higher ASR when using Italy or Europe as standards, lower ASR using World). It must be remebered that ASR differences are due to different age structures. The standard World (Segi’s) population is younger and thus gives more weight to rates at younger ages, when mesothelioma rates are low, thus producing lower ASR.
When analysing rates by Province, the same pattern observed on data from 2000 to 2016 was confirmed: in Provinces with rates higher than the regional average, there were many factories of various industrial sectors, including metalworking and metallurgy, construction, non-asbestos textile and clothing production, chemical and plastics, motor-vehicle, and rubber production.22
By far, the highest incidence rates in both genders were observed in the Province of Pavia. This is due to exceptionally high rates in the municipality of Broni and the nearby municipality of Stradella. Previously, it was shown that both suffered a large mesothelioma burden due to asbestos exposure (occupational and environmental) associated with an asbestos cement plant (Fibronit) which has been operating in Broni in the period 1932-1993.23,24

Strength and limitations of this study

The present study has two major strengths. First, the high data quality of the Registry: completeness of case identification is evaluated and updated regularly by using different case sources; moreover, clinical records are reviewed for each affected individual to verify diagnosis. Second, interviews could be obtained for more than 90% of patients.
The main limitation of this study is that complete data were available only for the period 2000-2021 (the reasons are explained in the “Methods” section). A further limitation is that, in a sizable percentage (especially in women), interviews could only be obtained by next-of-kin, who may not recall important details of patients’ occupational history and extra-occupational asbestos exposure. As explained above (see “Methods” section), there are important delays in case identification (mostly independent of RML); as a result, given the poor prognosis of mesothelioma, a large number of cases, when contacted to request consent to interview, are too ill to sustain interview or even dead.

Conclusions

In this study in Lombardy, a fair agreement was found between observed mesothelioma cases and forecasts published in 2016. The peak occurred in 2013, but, in the following years, there was no indication of decrease of yearly number of mesothelioma cases in men and only a slight decrease in women. The highest burden concerns individuals aged 75 or more years. These data indicates that in 2025-2029 there still will be between 350 and 400 cases per year. The legacy of asbestos regarding mesothelioma is expected to continue for many years, especially in elderly people. The decrease of cases with occupational exposure over time and the low mesothelioma rates in youngers witness the usefulness of asbestos ban in 1992.


Conflicts of interest: Dario Consonni and Carolina Mensi served as consultants in litigations concerning asbestos-related diseases.

Funding: the present work was partially supported by the Istituto Nazionale per l’Assicurazione contro gli Infortuni sul Lavoro (Inail), Rome, Italy: BRiC ID66/2022.

Acknowledgements: the authors wish to thank the patients affected by mesothelioma and their next-of-kin for granting interviews; the personnel of various institutions (hospital departments, ATS-SPSAL, UOOML) for collaboration in reporting and collecting information for cases. 

References

  1. Delgermaa V, Takahashi K, Park EK, et al. Global mesothelioma deaths reported to the World Health Organization between 1994 and 2008. Bull World Health Organ 2011;89(10):716-24, 24A-24C. doi: 10.2471/BLT.11.086678
  2. Park EK, Takahashi K, Hoshuyama T, et al. Global magnitude of reported and unreported mesothelioma. Environ Health Perspect 2011;119(4):514-18. doi: 10.1289/ehp.1002845
  3. Gariazzo C, Gasparrini A, Marinaccio A. Asbestos consumption and malignant mesothelioma mortality trends in the major user countries. Ann Glob Health 2023;89(1):11. doi: 10.5334/aogh.4012
  4. Marinaccio A, Binazzi A, Di Marzio D, ReNaM GdL. Il Registro Nazionale dei Mesoteliomi. Ottavo Rapporto. Rome: Inail; 2024.
  5. Marinaccio A, Montanaro F, Mastrantonio M, et al. Predictions of mortality from pleural mesothelioma in Italy: a model based on asbestos consumption figures supports results from age-period-cohort models. Int J Cancer 2005;115(1):142-47. doi: 10.1002/ijc.20820
  6. Oddone E, Bollon J, Nava CR, et al. Predictions of mortality from pleural mesothelioma in Italy after the ban of asbestos use. Int J Environ Res Public Health 2020;17(2):607. doi: 10.3390/ijerph17020607
  7. Oddone E, Bollon J, Nava CR, et al. Effect of asbestos consumption on malignant pleural mesothelioma in Italy: Forecasts of mortality up to 2040. Cancers (Basel) 2021;13(13):3338. doi: 10.3390/cancers13133338
  8. Girardi P, Bressan V, Merler E. Past trends and future prediction of mesothelioma incidence in an industrialized area of Italy, the Veneto Region. Cancer Epidemiol 2014;38(5):496-503. doi: 10.1016/j.canep.2014.08.007
  9. Usticano A, Tumino R, Dardanoni G, et al. Trends and projections of pleural mesothelioma incidence and mortality in the national priority contaminated sites of Sicily (Southern Italy). Epidemiol Prev 2020;44(4):218-27. doi: 10.19191/EP20.4.046
  10. Mensi C, De Matteis S, Dallari B, et al. Incidence of mesothelioma in Lombardy, Italy: exposure to asbestos, time patterns and future projections. Occup Environ Med 2016;73(9):607-13. doi: 10.1136/oemed-2016-103652
  11. Clayton D, Schifflers E. Models for temporal variation in cancer rates. I: Age-period and age-cohort models. Stat Med 1987;6(4):449-67. doi: 10.1002/sim.4780060405
  12. Mensi C, Termine L, Canti Z, et al. The Lombardy Mesothelioma Register, Regional Operating Centre (ROC) of National Mesothelioma Register: organizational aspects (Italian). Epidemiol Prev 2007;31(5):283-89.
  13. Consonni D, Coviello V, Buzzoni C, Mensi C. A command to calculate age-standardized rates with efficient interval estimation. Stata J 2012;12(4):688-701.
  14. Baur X, Frank AL, Soskolne CL, Oliver LC, Magnani C. Malignant mesothelioma: Ongoing controversies about its etiology in females. Am J Ind Med 2021;64(7):543-50. doi: 10.1002/ajim.23257
  15. Chiappino G, Mensi C, Riboldi L, Rivolta G. Asbestos risk in the textile industry: final confirmation of data from the Lombardy Mesothelioma Registry. Med Lav 2003;94(6):521-30.
  16. Catelan D, Consonni D, Biggeri A, et al. Estimate of environmental and occupational components in the spatial distribution of malignant mesothelioma incidence in Lombardy (Italy). Environ Res 2020;188:109691. doi: 10.1016/j.envres.2020.109691
  17. Consonni D, Barone-Adesi F, Mensi C. Comment on ‘The latency period of mesothelioma among a cohort of British asbestos workers (1978-2005)’: methodological problems with case-only survival analysis. Br J Cancer 2014;111(8):1674. doi: 10.1038/bjc.2013.823
  18. Oddone E, Terracini B, Mirabelli D, et al. Comment on: Malignant mesothelioma diagnosed at a younger age is associated with heavier asbestos exposure. Carcinogenesis 2019;40(3):488-89. doi: 10.1093/carcin/bgy179
  19. Consonni D. Something is missing: what’s wrong with using age at diagnosis/death or latency among cases. Epidemiol Prev 2013;37(1):85-88.
  20. Stella S, Consonni D, Migliore E, et al. Pleural mesothelioma risk in the construction industry: a case-control study in Italy, 2000-2018. BMJ Open 2023;13(8):e073480. doi: 10.1136/bmjopen-2023-073480
  21. De Matteis S, Consonni D, Lubin JH, et al. Impact of occupational carcinogens on lung cancer risk in a general population. Int J Epidemiol 2012;41(3):711-21. doi: 10.1093/ije/dys042
  22. Mensi C, De Matteis S, Catelan D, et al. Geographical patterns of mesothelioma incidence and asbestos exposure in Lombardy, Italy. Med Lav 2016;107(5):340-55.
  23. Mensi C, Riboldi L, De Matteis S, Bertazzi PA, Consonni D. Impact of an asbestos cement factory on mesothelioma incidence: global assessment of effects of occupational, familial, and environmental exposure. Environ Int 2015;74:191-99. doi: 10.1016/j.envint.2014.10.016
  24. Consonni D, De Matteis S, Dallari B, et al. Impact of an asbestos cement factory on mesothelioma incidence in a community in Italy. Environ Res 2020;183:108968. doi: 10.1016/j.envres.2019.108968
Vai all'articolo su epiprev.it Versione Google AMP