Background Australia is known to have had one of the highest per-capita asbestos consumption rates, yet there are few contemporary reports on malignant mesothelioma trends.
Methods Data on 10 930 people with malignant pleural mesothelioma (MPM) and 640 people with malignant peritoneal mesothelioma diagnosed in Australia during 1982–2009 were analysed. Observed incidence rate trends were quantified. Incidence rates were projected up to 2030 using observed incident cases during 1982–2012. The relative per-decade change in excess mortality during 1999–2009 was estimated.
Results During 1982–2009, acceleration in MPM age-standardised incidence rates were highest for women and those aged 75 years and above, with average annual percentage changes of +4.9 (95% CI 3.6 to 6.2) and +7.2 (95% CI 5.4 to 9.0), respectively. Age-standardised incidence rates for men with MPM aged 0–64 years decelerated rapidly during 2003–2009, an average annual percentage change of −5.1% (95% CI −7.6% to −2.5%). Overall, male age-specific MPM incidence rates in the age group of 65–74 year during 2010–2030 are projected to decline with rates projected to increase for older men and women with MPM. There was a statistically significant 16% relative reduction in the excess mortality rate (EMR) up to 5 years postdiagnosis for people diagnosed with malignant pleural and peritoneal mesothelioma combined in 2009 compared with those diagnosed in 1999, an EMR ratio of 0.84 (95% CI 0.77 to 0.92).
Conclusions Australia's malignant mesothelioma incidence rates appear to have reached maximum levels but with differences over time by age, gender and tumour location. Improvements over time in survival provide a glimpse of hope for this almost invariably fatal disease.
Statistics from Altmetric.com
What this paper adds
There are few contemporary reports of population-based incidence and survival trends for malignant pleural mesothelioma (MPM) and malignant peritoneal mesothelioma (MPeM) in Australia, a country with one of the highest global age-standardised incidence rates.
We measured trends in age-standardised incidence rates during 1982–2009, a 28-year period, and projected age-specific incidence rates up to 2030 using observed incidence data for 1982–2012. We also modelled the per-decade change in population-based survival up to 5 years postdiagnosis during 1999–2009, with follow-up to 2011.
Since the early 2000s, age-standardised incidence rates have rapidly declined over time for men aged 0–64 years diagnosed with MPM, and at the same time an increase over time in the age-standardised incidence rates for older aged men occurred. MPeM remains rare in male and female groups.
There was a significant 16% reduction in excess mortality rates 5 years postdiagnosis for people diagnosed in 2009 compared with those diagnosed in 1999.
Overall, a peak in malignant mesothelioma incidence rates has occurred in Australia. However, the peak among men with MPM aged 75 years or more and women will not reach maximum levels until around 2020. Peak MPeM incidence in men will occur around 2015–2025, highlighting the need for evidence-based diagnosis and treatment guidelines to be developed for this group.
Improvements over time in survival provide optimism for people affected by malignant mesothelioma.
There is a large international variation in malignant mesothelioma incidence probably due to variation in per-capita levels of exposure to asbestos across countries.1 Australia is one of a number of high-income countries where there are wide-ranging negative consequences of asbestos use, including an epidemic of asbestos-related disease. The deleterious human health effects of Australia's very high asbestos consumption levels, which peaked at around 700 000 metric tonnes during 1970–1979,2 are well described using data from the Australian Mesothelioma Surveillance Program (1980–1985), the Australian Mesothelioma Register (1986–2007), and most recently the Australian Mesothelioma Registry (AMR; 2010–present).2–4 Since the early 2000s however, a comprehensive investigation of national malignant mesothelioma disease patterns and trends in Australia has not been published in peer-reviewed literature.
Optimal allocation of cancer detection, diagnostic, therapeutic and palliative care services can be guided by estimates of the future burden of disease.5 To predict the peak of Australia's malignant mesothelioma incidence curve, Leigh et al6 combined knowledge of Australia's asbestos consumption (where crocidolite and chrysotile exposure peaked around 1965 and 1975, respectively), with data from Western Australia where crocidolite was mined, and the national curve of malignant mesothelioma incident cases from 1945 to 1995. The investigators estimated that Australia's malignant mesothelioma epidemic would peak in 2010. The predicted incident case curve was estimated for the total mesothelioma population. It is timely to review these predictions, as female age-standardised incidence rates for malignant mesothelioma are increasing over time, and there also appears to be an increase over time in the proportion of older people being diagnosed with malignant mesothelioma.7 It is also likely that there are different disease trajectories for malignant pleural mesothelioma (MPM), the most common form of malignant mesothelioma comprising around 93% of all Australian cases diagnosed,3 ,6 and malignant peritoneal mesothelioma (MPeM).8
Malignant mesothelioma is an almost invariably fatal disease.9 Various approaches are available for measuring population-based changes over time in cancer survival with relative survival analysis the most common. Relative survival ratios are estimated using data on the observed number of deaths in the cancer population with data on the number of deaths from all other causes taken from population life tables. Regression modelling is a useful extension of the relative survival approach that allows for simultaneous consideration of prognostic factors10 ,11 such as age, gender and histological subtype relevant to malignant mesothelioma.12–14 However, these regression approaches have had limited use for estimating changes over time in surviving malignant mesothelioma.
In this study, we aimed to quantify population-based trends in incidence and survival for people diagnosed with MPM and MPeM in Australia. To understand patterns of disease observed in the past, segmented regression methods were used to estimate changes over time in age-standardised incidence rates during 1982–2009 separately by gender and age group. To test whether a peak in incidence has been reached, as predicted by Leigh et al,6 we used age–period–cohort (APC) models with natural cubic splines to project future incidence rates between 2010 and 2030 for combinations of gender and age groups. Relative survival was modelled to estimate the relative per-decade change in excess mortality rates (EMR) for people diagnosed in 2009 compared against 1999.
Materials and methods
Cancer registration data
The primary data source in this study was the Australian Cancer Database (ACD). The ACD is a data collection of all primary, malignant cancers diagnosed in Australia since 1982 where data are compiled at the Australian Institute of Health and Welfare (AIHW) from cancer data provided by the eight Australian statutory population-based state and territory cancer registries. Information about a cancer diagnosis in Australia is received from various sources including hospitals, pathology laboratories, radiotherapy centres, and birth, death and marriage registries. Anonymised individual unit records for all unique cases of malignant mesothelioma diagnosed in Australia were extracted from the ACD using the code C45 from the International Statistical Classification of Diseases and Related Health Problems, 10th revision (ICD-10), including mapping of diagnosis and cause of death codes to ICD-10 from 1982 onwards. The key characteristics of this case series are described in online supplementary table S1; cell counts less than five were suppressed to protect confidentiality. Recent cancer registration data from the AMR between 1 July 2010 and 31 December 2012 were also available in this study. A unique feature of this data collection is that pathologically confirmed cases of malignant mesothelioma recorded at the state or territory levels are fast-tracked to the AMR, in order to collect, if possible, information about a person's exposure to asbestos. The key characteristics of this case series are described in online supplementary table S2. AMR data, including a data update provided to the investigators on 7 September 2015, were used in the projection of incidence rates up to 2030.
The focus for this study was a description of trends in incidence and survival between 1982 and 2009 separately for MPM (ICD-10 C45.0) and MPeM (ICD-10 C45.1). A total of 99 people were diagnosed as having malignant pericardial mesothelioma and other or unknown tumour location (ICD-10 codes C45.2-C45.9). These records were excluded from analyses. Information about a person's tumour stage at diagnosis was not recorded on the ACD. However, data on histological subtype were available and were coded using the International Classification of Diseases for Oncology (ICD-O-3), 3rd edition. For the survival analyses conducted in this study, we grouped histological subtype into three categories: epithelioid subtype (ICD-O-3 code M9052/3); non-epithelioid subtype (ICD-O-3 code M9051/3 (sarcomatoid subtype) and M9051/3 (biphasic subtype)); and not otherwise specified histological subtype (ICD-O-3 code M9050/3). Complete (or partial) dates of death were suppressed to the investigators. Instead, the number of days survived since diagnosis was provided allowing for the calculation of an estimated date of death. For survival analyses, we used an approximated date of diagnosis where the day of diagnosis was set to the first day of the month. Fact and date of death were obtained through a probabilistic data linkage undertaken at the AIHW of the ACD records with records from the National Death Index (further information on the National Death Index can be found at http://meteor.aihw.gov.au/content/index.phtml/itemId/480010). In our study, if a person was not recorded as having died, they were assumed to be alive at the end of follow-up (31 December 2011).
Changes over time in age-standardised incidence rates
First, individual unit records for MPM and MPeM were analysed separately to estimate age-standardised incidence rates per 100 000 person-years for men, women, and men and women combined, using the Australian 2001 standard population and, to enhance interpretation of findings in the international context, the Segi world standard population. We also estimated age-standardised incidence rates by three truncated age groups (0–64, 65–74 and 75 years or more). Rates were calculated using the Stata ‘dstdize’ command. These rate data, along with the number of cases, and midyear population estimates (1982–2009) drawn from cancer incidence and mortality data from the AIHW were used to estimate average annual percentage changes (AAPC) during 1982–2009, and annual percentage changes (APC) for each model-selected regression segment between 1982 and 2009. To estimate these values, we used Joinpoint Trend Analysis Software, V.220.127.116.11 (June 2015), produced by the Surveillance, Epidemiology and End Results programme at the US National Cancer Institute. Joinpoint regression is a type of non-linear regression (see online supplementary appendix 1) made up of a number of continuous linear segments, and is often used in quantifying trends in cancer incidence and mortality data. Sparse data were observed for MPeM. Specifically, values of 0.0 were often observed when estimating age-standardised incidence rates for each year of diagnosis when stratified by gender and age group resulting in the inability to calculate segmented regression estimates for some combinations of gender or age. AAPC and APCs were estimated using age-standardised incidence rate data using both Australian and international age-standardisation weights (tables 1 and 2 and online supplementary tables S3 and S4, respectively).
Projections of age-specific malignant mesothelioma incidence rates
Whereas Leigh et al6 predicted the future curve of incident cases by using data on exposure to asbestos (see online supplementary figure S1), we have focused on projecting age-specific incidence rates from historical trends allowing for simultaneous assessment of age at diagnosis, diagnosis period, and birth year effects. Individual unit records for MPM and MPeM were aggregated into two separate files containing information for each combination of age at diagnosis grouped into 18 5-year age groups, single calendar year of diagnosis (1982–2012), single cohort year calculated as the difference between the calendar year of diagnosis and the age at diagnosis (using the midpoint of the 5-year age group), gender, number of cases, midyear population estimates (1982–2009) drawn from cancer incidence and mortality data from the AIHW,7 and population projections until 2030 obtained from the Australian Bureau of Statistics.15 A categorical age variable was created for people aged 0–64, 65–74 and 75 years or more. APC models were then applied with natural cubic splines (see online supplementary appendix 1) using the Stata ‘apcspline’ command developed by Sasieni.16 The apcspline command allows for two link functions to be used, with the first being a logarithmic link function and the second a power 0.2 link function consistent with other cancer incidence projection methods such as NordPred.17 ,18 Using the default values, the number of cubic spline knots varied depending on the type of effect. Three knots were used for cohort effects, five knots used for the period effect, and, to allow for greater flexibility, six knots for the age-effect. A methodological challenge in APC models is to account for non-linear cohort and period effects, which requires an assumption about the degree of change over time in the projected rates, referred to as drift. We used the default drift value of 8 years. Once values from the cubic spline age–period model were estimated, the ‘grmean’ Stata command was then used to plot observed rates (1982–2012), and fitted rates (1982–2030) for the combination of truncated age group and gender. These plots provide fitted rates for both the logarithmic link function and the power 0.2 link function. We also applied the grmean command to project MPM incidence rates for those aged 65 years or more by 5-year age group for men and women separately.
Changes over time in survival 5 years following a malignant mesothelioma diagnosis
First, patterns over time in survival for adults aged 15–99 years diagnosed with MPM and MPeM during 1999–2009, with vital status follow-up to the end of 2011, were assessed by estimating 5-year relative survival ratios separately by collapsed age group (15–54, 55–74 and 75+ years), gender, histological subtype, and calendar period (1999–2002, 2003–2006, 2007–2009) (see online supplementary figures S2 and S3). People whose basis of diagnosis was recorded as ‘death certificate only’ (1% of our case-series) were excluded. We estimated relative survival for people diagnosed up to the end of 2006 using the cohort approach; all remaining cases were estimated using the period approach commonly used for recently diagnosed cases where full 5-year follow-up data are not available.19 The Ederer II method20 was used to calculate expected survival. Relative survival was calculated using the ‘strs’ Stata command.21 ,10
Next, data for all cases diagnosed between 1999 and 2009, inclusive, were used to run EMR models (see online supplementary appendix 1) separately for MPM, MPeM and for MPM and MPeM combined. This model also specifies a log-link function for the Poisson error distribution. The EMR model included a covariate for the per-decade calendar period specified as a continuous variable with 1999 as the reference year, gender, age group (15–54, 55–64, 65–74, 75–84 and 85+ years), histological subtype, annual follow-up time since cancer diagnosis, and an interaction term for age group and follow-up time (75–85 and 85+ years age groups with first and second years of follow-up only, otherwise age and follow-up treated as separate main effects). EMR modelling was undertaken using the ‘rs’ Stata command,21 and adapting the regression modelling commands for the parameter estimates required for this study.
Number of people with malignant pleural and peritoneal mesothelioma
During 1982–2009, 10 930 people were diagnosed with MPM and 640 people were diagnosed with MPeM (see online supplementary table S1). The majority of these cases occurred in men but with differences by tumour location, with men making up 85% and 73% of all MPM and MPeM cases, respectively. People with MPeM tended to be younger—the median age at diagnosis for MPeM was 64 years compared with 70 years for MPM. Around 60% of MPM and MPeM cases were recorded as having not otherwise specified histological subtype. Between 1 January 2011 and 31 December 2012, 1251 people were newly diagnosed with MPM, and 77 incident cases of MPeM were recorded using data from the AMR (see online supplementary table S2).
Changes over time in age-standardised incidence rates
During the 1982–2009 period, the AAPC in age-standardised incidence rates among people with MPM was highest for women, an AAPC of +4.9 (95% CI 3.6 to 6.2), and among people aged 75 years and over, an AAPC of +7.2 (95% CI 5.4 to 9.0) (table 1, online supplementary table S3). AAPC estimates for people with MPeM were less stable; nevertheless, the AAPC during 1982–2009 for men and women combined with MPeM was +2.4 (95% CI 0.9 to 3.9).
We also assessed the results for each regression segment for people with MPM (table 2, online supplementary table S4). These segmented regression results provide information about differences in the rate of change over time during 1982–2009. Focusing on changes over time in the age-standardised incidence rate for women with MPM, the APC accelerated over time during 1982–2002, with an APC of +6.7 (95% CI 5.3 to 8.1), with evidence of deceleration in the most recent period, an APC of −0.2 (95% CI −3.5 to 3.3) during 2002–2009. This acceleration and subsequent deceleration was observed also in men, and men and women combined with MPM. The best regression model fit for people with MPM aged 0–64 years was with two joinpoints (ie, three segmented regression estimates). The estimates show a steady increase in the ASIR between 1982 and 1994 (APC estimate: +4.6 (95% CI 3.3 to 7.3)), levelling off of the incidence rate between 1994 and 2003 (APC estimates of +0.9 (95% CI −0.9 to 2.7)), and then a sharp decline between 2003 and 2009 (APC estimate: −5.1 (95% CI −7.6 to −2.5)). These patterns for people with MPM aged 0–64 years differed in older age groups. For example, the APC for people with MPM aged 75+ years was +11.0 (95% CI 7.7 to 14.3) during 1982–1996 and continued to increase, but to a lesser extent, during 1996–2009 (APC of +3.2 (95% CI 1.5 to 4.9).
Projected age-specific incidence rates
We estimated observed and projected age-specific incidence rates per 100 000, during 1982–2012 and 2010–2030, respectively, by combination of tumour location, gender and truncated age group by fitting an APC model using natural cubic splines. Overall, we can conclude that a maximum level of the incidence of malignant mesothelioma has been reached but with differences by age group and gender in the incidence peaks (figure 1). Focusing on men with MPM, we found differences in projection estimates by age group. There was little change over time in the age-specific incidence rate for men with MPM aged 0–64 years. However, age-specific incidence rates for men with MPM aged 65–74 years appears to have peaked during 2005–2010. Incidence rates for men with MPM aged 75 years or more are expected to increase up to 2020 and then decline. For women with MPM, age-specific incidence rates for those aged 65 years and over will increase slightly up to 2020. Projection estimates show little change over time for age-specific rates in women with MPeM, but have projected changes over time for men (note the different y scale (0–5 per 100 000)). For men with MPeM, age-specific incidence rates in those aged 65 years or more are projected to peak during 2015–2025, somewhat later than for MPM.
We also estimated observed and projected age-specific incidence rates per 100 000 for people aged 65 years and above with MPM by 5-year age group and gender (figure 2). For men, a decrease over time in the age-specific incidence rate was projected for each 5-year age group, with the shape of the decline varying by age group. A slow increase over time in the age-specific incidence rate was projected for women aged 80–84 and 85 years or more. For all fitted rates, there was little difference between the estimates for the APC model using the logarithmic link and estimates where the power function was used.
Changes over time in survival up to 5 years after diagnosis
First, we assessed patterns in relative survival up to 5 years postdiagnosis separately by age group, gender, histological subtype and calendar period for MPM and MPeM (see online supplementary figures S1 and 2). From these data, differences in survival were present by age, gender and histological subtype with younger age groups, women and people with epithelioid histology having greater survival. There was little change over time in 5-year relative survival regardless of subgroup characteristics.
Second, to provide the most stable estimates, we measured relative differences in EMRs (the EMR ratio (EMRR)) 5 years postdiagnosis simultaneously by calendar period, age group, gender, follow-up time since diagnosis, and histological subtype for MPM and MPeM combined (table 3). Focusing on changes in survival over time, there was a reduction in EMRs for people with malignant mesothelioma during 1999–2009 with an EMRR per-decade of 0.84 (95% CI 0.77 to 0.92) or, put another way, a statistically significant 16% reduction in the EMR for people diagnosed in 2009 compared to 1999. Given the model parameterisation, this relative reduction per decade was assumed to apply uniformly across all age groups, gender, histological subtypes and each follow-up period. Focusing on relative differences in excess mortality by other covariates, statistically significant differences by age were apparent. Excess mortality was greater for older age groups when compared with people aged 15–54 years. For example, people aged 75–84 years had 40% greater excess mortality compared with the youngest age group, an EMRR of 1.40 (95% CI 1.14 to 1.71). However, this association was not observed for people aged 85 years or more where the point estimate was below the null. Regarding gender, women had 9% less excess mortality compared to men, an EMRR of 0.91 (95% CI 0.85 to 0.98). Compared to people with epithelioid subtype, an almost twofold statistically significant difference in excess mortality was found for people with non-epithelioid disease—an EMRR of 1.94 (95% CI 1.79 to 2.10).
Third, we also separately ran the EMRs for MPM and MPeM (see online supplementary tables S5 and S6, respectively). Given the composition of our case-series being weighted towards people with MPM, EMRR estimates for all mesothelioma cases presented in table 2 and MPM, online supplementary table S5 were similar, with the exception of the association between gender and survival that became non-significant. There were some differences in the EMRR estimates for people with MPeM (see online supplementary table S6). There was a 30% reduction in excess mortality comparing people diagnosed in 2009 to 1999, an EMRR of 0.70 (95% CI 0.48 to 1.03). Thus, a faster reduction in EMRs for people with MPeM than people with MPM is possible, but it is not established with any certainty.
In this study, we assessed whether the incidence of malignant mesothelioma has increased or decreased over time in Australia. We found evidence for significant increases over time in MPM age-standardised incidence rates during 1982–2009 for both genders and all age groups. However, the speed and direction of these have varied over time by age and gender. For example, the age-standardised incidence rate for people with MPM aged 0–64 years appears to have three distinct change periods with the most recent period showing a −5.1% annual decrease during 2003–2009. This differed from those aged 75 years and above, where the incidence rate increased by +11% between 1982 and 1996, and continued to increase, but at a substantially lower rate of +3.2%, during 1996–2009. Overall, age-specific incidence rate projections up to 2030 suggest that a maximum level of malignant mesothelioma incidence in Australia has been reached but with differences by age and gender. Importantly, our results show that male MPM age-specific incidence rates in those aged 65–74 years have declined during 2005–2010, and with evidence of an increase in age-specific incidence rates for men with MPM aged 75 years, and women with MPM aged 65 years and over such that age-specific incidence will peak for these two groups during 2015–2020. Incidence for men with MPeM aged 65 years and over is projected to peak during 2015–2025. Focusing on survival following a diagnosis of mesothelioma, we found a 16% relative reduction in the EMR for people diagnosed in 2009 compared with those diagnosed in 1999.
Since the late 1970s, Australia's use of asbestos has been regulated starting with the introduction of occupational asbestos exposure limits that varied by asbestos type.22 ,23 The majority of asbestos consumption was phased out during the 1980s, but with use of chrysotile, was not completely prohibited at that time. A formal prohibition was introduced in 2003 for the use, importation, transportation, sale, storage and manufacturing of all forms of asbestos and asbestos-containing products. In this study, we were not able to directly test whether with the prohibition of asbestos in its various forms over time in Australia, has led to a decrease in incidence of malignant mesothelioma. However, our findings are consistent with those of Leigh et al6 where incidence of mesothelioma was predicted to peak in 2010 by taking into account the peaks of chrysotile and crocidolite consumption in Australia in the mid-1970s and 1960s, respectively. Notably, the 2003 prohibition did not require removal of asbestos products in situ. When disturbed either through occupational or non-occupational activity in situ asbestos poses a potential public health risk. Non-occupational asbestos exposures, including home renovation, are causing increasing concern in Australia.3 ,24 ,25
The findings in this study, and those shown by Leigh et al,6 are generally consistent with the evidence regarding the distribution of observed latency periods between exposure to asbestos and risk of malignant mesothelioma, commonly reported to occur between 15 and 50 years of age.26–29 Notably, a recent study by Reid et al30 showed that the rate of increase in MPM incidence reduces at about 45 years since first exposure to asbestos. For MPeM though, Reid et al found that the rate of increase in MPeM incidence continued to rise over and above 45 years since the first exposure to asbestos. These findings provide a plausible explanation for the differences in the projected male age-standardised incidence rate changes over time for MPM (incidence rates projected to decrease) and MPeM (incidence rates projected to increase) shown in our study. The projected changes over time in age-specific MPeM incidence rates also have therapeutic implications. Evidence-based guidelines for the diagnosis and treatment for MPM in the Australian context have recently been developed by van Zandwijk,31 yet there is no equivalent Australian clinical standard for MPeM. Given that diagnostic procedures and treatment options differ between MPM and MPeM,32–34 it will be important to develop similar MPeM guidelines in Australia as for MPM to ensure that all malignant mesothelioma patients receive optimal care.
Several country-specific studies have been published estimating the future incidence of malignant mesothelioma. However, it is difficult to obtain a comprehensive global overview of future malignant mesothelioma projections due to the fact that there are differences between the studies on the data collection, statistical approaches used, and heterogeneous exposure levels to asbestos within and across countries. Nevertheless, several recent studies are informative. Krupoves et al35 projected MPM incidence rates during 2008–2032 in Canada, specifically in Quebec, an area that accounts for 95% of Canada's asbestos production, and for Canada excluding Quebec. Using an age-cohort analysis, Krupoves et al showed that MPM incidence rates will peak in Quebec and Canada excluding Quebec between 2003 and 2012 and will decline after that point. A study from the USA published by Moolgavkar et al8 projected MPM and MPeM incidence. Consistent with our findings, a substantial decrease over time was projected for MPM incidence among men with a concomitant increase in women. However, contrary to our findings, they also reported an increase in the annual number of female MPeM cases and a decrease in the number of male MPeM cases, but with little trend over time for male or female MPeM incidence rates. They, along with Price and Ware,36 suggest that MPeM incidence rate trends have not been influenced by occupational asbestos exposure in the USA. In our study though, we found a significant increase over time in MPeM incidence rates for men and women combined during 1982–2009 with these rates projected to increase during 2010–2020 for men aged 75 years or more. Although the incidence rate of MPeM is lower than for MPM, our findings show that a similar pattern has emerged with time for the male incidence of MPM and MPeM suggesting that there is a no threshold causal relationship with asbestos exposure. A number of studies have also reported projected future malignant mesothelioma mortality rates in Italy and Great Britain where rates are projected to peak during 2011–2015 and 2012–2024, respectively.37 ,38
There is a paucity of population-based studies investigating changes over time in malignant mesothelioma survival, particularly those using relative survival approaches. However, there is growing interest in whether advances in chemotherapy have led to population-based improvements in malignant mesothelioma survival. Damhuis et al39 used data on 4731 people with malignant mesothelioma from the Netherlands Cancer Registry to investigate the impact of chemotherapy with cisplatin-antifolate combinations on population-based survival, measured using Cox regression methods. They reported an increase in median survival between 1995–1998 and 2005–2006 of 2 months (10.1 and 13.1 months, respectively) and concluded that chemotherapy usage had increased and coincided with an improvement in survival at the population level. We are only able to interpret that the per-decade improvements in survival reported in our case series occurred during the time in which pemetrexed became widely available in Australia particularly from 2008 onwards. Pemetrexed is regarded as the optimal chemotherapy treatment for people with MPM.40 Further research is warranted to more accurately address this research question.
Our study shows that, overall, the incidence rate of malignant mesothelioma appears to have reached maximum levels, but that in the next few years there will be further increases in older women diagnosed with pleural mesothelioma, and older men with peritoneal mesothelioma, before their peak is inevitably reached. Evidence of population-based survival improvements 5 years following a malignant mesothelioma diagnosis should provide optimism among those affected by malignant mesothelioma.
The authors thank the staff at the Australian Mesothelioma Registry (AMR) (the AMR is a collaborative arrangement between Safe Work Australia, Comcare and the Cancer Institute NSW), the Australian Cancer Database based at the Australian Institute of Health and Welfare, and state-cancer and territory-cancer registries that provided access to the de-identified unit record data for people diagnosed with malignant mesothelioma analysed in this study. This publication includes data made available by the AMR, owned by the Commonwealth as represented by Safe Work Australia. The authors also thank the Australian Institute of Health and Welfare for linkage of AMR records to the National Death Index.
Contributors MJS led the study, undertook the epidemiological analyses and was the primary author of the manuscript. JL, TD and BA provided input regarding Australian asbestos exposure and subsequent disease trends. NvZ contributed content on clinical and treatment aspects. JMY provided input on the overall descriptive epidemiological study and the interpretation of national cancer registration data. All authors revised the manuscript for intellectual content.
Funding This research was funded by a Translation Program Grant (11/TPG/3-06) awarded to the Asbestos Diseases Research Institute by the Cancer Institute NSW. JY and MS were funded through an Academic Leader in Cancer Epidemiology award (08-EPC-1-01) from the Cancer Institute NSW.
Disclaimer The views expressed in this publication are those of the authors and do not necessarily represent the views of Safe Work Australia or the AMR.
Competing interests None declared.
Ethics approval New South Wales Population Health and Services Research Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Owing to the conditions of ethics approval for this study, the authors cannot provide individual unit record data. However, the authors are able to provide the point estimates and 95% confidence limits to other investigators.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.