Cancer Council Australia

Larger Text Smaller Text Print

Lifetime risk of lung cancer in Australian workers



 

Download Slides

Conference:

kNOw cancer risks at work, Cockle Bay Sydney, May 2015

 

Presenter:

Dr Renee Carey, School of Public Health Curtin University

 

Title:

Lifetime risk of lung cancer in Australian workers

 

Presentation outline:

Occupational exposure to substances known to cause lung cancer has been found to be common in Australia, with approximately 2.9 million workers estimated to be exposed in 2012. Dr Renee Carey outlines the results of a study showing exposures that contribute most to the overall future burden of occupational lung cancer. The results provide information on where to target preventive measures.

Read transcript

Today, I am going to talk about the lifetime risk of lung cancer in Australian workers. So, we know that lung cancer is a malignant tumour in the tissue of one or both the lungs and that it is the fifth most common cancer in Australia with a risk of diagnosis by age 85 being 1 in 16 or about 6%. Over 10,000 new cases of lung cancer were diagnosed in 2011, which represents about 9% of all cancer diagnoses and lung cancer is also the leading cause of cancer deaths so it accounts for about 19% of all deaths from cancer. Five-year survival rate is low of about 14% and we also know that lung cancer is much more common in men representing about 61% of cases.

So, a number of carcinogens have been associated with lung cancer, most well-known obviously being tobacco smoke including passive smoking, but other factors include air pollution, asbestos, diesel engine exhaust, ionizing radiation, metals like arsenic, chromium, nickel as well as silica. It is interesting to note most of these carcinogens can be encountered occupationally.

We conducted a recent study in the Australian Work Exposure Study or AWES. This looked at the current prevalence of exposure to 13 occupational lung carcinogens among Australian workers. These carcinogens included substances like acid mist, asbestos, metals, diesel engine exhaust, environmental tobacco smoke and silica. Overall, we found that about 31% of our sample were exposed to at least one lung carcinogen at work, exposure more likely in males with about 45% of males exposed compared to females at 13%. Prevalence estimates for the specific carcinogens ranged from 0.02% to 19% with the highest prevalence found for diesel engine exhaust at 19%, environmental tobacco smoke at 14%, silica at 7% and polycyclic aromatic hydrocarbons other than vehicle exhausts at 6%.

We extrapolated these numbers to the entire Australian working population and estimated that about 2.9 million workers are exposed to lung carcinogens at work. Many occupational groups were highly exposed to lung carcinogens, most commonly exposed groups were miners, so all of the miners in our sample were exposed, and heavy vehicle drivers - 99% exposed, followed by farmers, vehicle trades people like mechanics, automobile drivers - couriers for example and construction workers. Diesel exhaust was a common exposure in all of these groups, but other common exposures included tobacco smoke in construction workers and drivers, asbestos in vehicle workers and silica in the miners.

So AWES was able to provide us with data and information on what lung carcinogens workers were exposed to, who in particular is exposed and how prevalent those exposures are. However, what is not known is the contribution of current exposure to these carcinogens to the overall risk of lung cancer. So that is the proportion of future lung cancers, which can be attributed to current occupational exposure. This information is important, that calls attention to the issue, helps in the planning of policy responses and can help us to determine, which interventions and controls are most likely to be successful in preventing lung cancer.

The question we are asking in our current study, what proportion of future lung cancers will occur in those people who were occupationally exposed to lung carcinogens in 2012 as a result of their exposure?

We have from AWES, an estimate of the proportion of the working age population who were exposed in 2012 and we then want to look forward to see what proportion of lung cancers occurring in the future might occur in those people as a result of their exposure.

To illustrate this, we have a cohort of Australians who are of working age, aged 18 to 65 in 2012. We know that a proportion of these were exposed to at least 1 of 13 occupational lung carcinogens in that year. We can then estimate how many lung cancers would occur in this cohort in the future, what proportion would have occurred without any exposure - so the baseline, and then how many would occur in those who were exposed in 2012 as a result of their exposure, and it is this final group here that we are looking to estimate.

In order to do this, we have used the lifetime risk approach. This is a novel approach used to estimate the proportion of disease cases occurring in the future, which are due to exposure to agents in the proportion of the population exposed in a target year. So in our case, we are estimating the proportion of lung cancers occurring in the future, which will occur in those who are exposed in 2012. We look at all of the future cancers there in green and estimate how many of those might be expected to occur in the proportion of the population who were exposed in 2012.

This approach requires four inputs, firstly we need an estimate of the proportion of the population who are exposed currently, which we can obtain from AWES. We need a relative risk estimate for each exposure and so an estimate of the relationship between exposure and lung cancer risk. We have obtained a relative risk estimate for each of our 13 carcinogens from the literature using work conducted in the UK, which John referred to earlier, as a starting point. Third required input is an estimate of future lung cancer rates. We have used current and past incidence rates from the cancer registry and then projected these forward. Finally, we need an estimate of the future person-years at risk for our cohort.

So, we have used mortality rates obtained from the Australian Bureau of Statistics as well as lung cancer incidence rates to calculate future person-years at risk. We then used these four inputs to calculate the lifetime risk of lung cancer in our cohort and then in our exposed population. So, the first step is to calculate the lifetime risk of lung cancer in our cohort in the working age population of Australia. This risk is regardless of exposure and is calculated as the number of cases (D) divided by the number of people in the population (N). These calculations are done separately for men and women. We estimated the lifetime risk of lung cancer in our cohort as 5.6% for men and 4.4% for women. This translates to over 408,000 future lung cancers in men and just under 320,000 in women.

The next step is to calculate the excess lifetime risk for each exposure. This is the excess risk of lung cancer in the exposed, which results from exposure and again is calculated separately for each exposure. The formula looks something like this. It takes into account the relative risk, the number of people exposed, number people in the cohort and lifetime risk in the cohort.

We can then use the excess lifetime risk to calculate the future excess number and future excess fraction for each exposure. So, the future excess number or FEN is the number of excess cancers occurring in the exposed population, basically the cases attributable to that exposure. This is calculated as the product of the excess lifetime risk and the number of people exposed. The future excess fraction or the FEF is the proportion of future cancers in the exposed population which are due to exposure to that carcinogen. So, again each calculated separately for each carcinogen. We then need to combine the future excess fractions for all exposures to get a final answer to our question. We combined these using the complement of the product of complements, which are just for the possibility of simultaneous exposure or cross carcinogens. We can then get an overall future excess fraction and future excess number.

This table will show the overall results. As I mentioned before, we can see that our cohort would be expected to get around the 728,000 cancers over their lifetime, 408,000 in men and around 320,000 in women. Around 26,000 of those will be in the exposed population and due to exposure, which translates to an overall future excess fraction of 3.6%. You can see that these numbers are much higher in males than in females as we would expect based on our exposure results where males were more likely to be exposed than females.

The following two tables show the results by individual exposure for those exposures which contribute more than 50 cancers to the overall burden. First of all the future excess number or the number of cancer registrations attributable to each carcinogen, so as we can see here silica contributes the most to the overall burden with just under 5,800 cancers followed by nickel and diesel engine exhaust. It is also worth noting here that environmental tobacco smoke is still contributing over 3800 registrations despite legislation which bans smoking in workplaces. Again, as we can see in all cases, the number is higher in males. So, this table shows us a future excess fraction and we can see here the same pattern, silica is the largest contributor followed by nickel and diesel engine exhaust. As well as answering our question regarding the proportion of future lung cancers which will occur in the exposed population as a result of exposure, we are also able to see where the greatest burden is coming from or which exposures are leading to the most cancers.

From the table we have just seen looking at the future excess numbers, we can see that the majority of cancer registrations overall are occurring among those exposed to silica and nickel, each with over 5000 exposures. So, if we wanted to focus on a single carcinogen to reduce future risk and reducing exposure to either of those is likely to have the most impact on future lung cancer cases. However, if we were interested in reducing cases among females in particular, we would be better off focusing our efforts on environmental tobacco smoke, which is by far the largest contributor among women.

We can also potentially do similar analyses by occupation and industry looking at which occupational groups or industries are more likely to contribute the most cancer registrations.

We can also use this model to look at the potential effect of various interventions to reduce exposure to see where our preventive efforts would be best targeted. We can look at banning particular carcinogens, reducing exposure to carcinogens and closing high-risk industries, all of which would reduce our overall number of exposed people whether to one or to multiple carcinogens. We can also look at the effect of changing levels of exposure, say for example by changing tasks, which might lead to high levels of exposure to those that might lead to lower levels or increasing the use of protective equipment and we are aiming to look at those interventions in our future work.

We do need to note that the model is based on some assumptions. Firstly, we have used current prevalence of exposure to denote the proportion of people who are exposed. This means we assume a normal distribution around this prevalence. Some people may have been exposed for a long time in the past, some will continue to be exposed for a long time in the future and some will only be exposed for a short time, and this aligns with the definition of exposure I used to calculate the relative risks, which we used in our calculations, so they look at those who were ever exposed. Our second assumption which is somewhat related is that we do not need to take into account the later period and so as it has been mentioned a few times today generally we assume a period of somewhere between 10 and 50 years between exposure and cancer, but because in this model we assume that some people have already been exposed for a long time in the past those exposures may actually be contributing to cancer cases now and so therefore we do not take that latent period into account and finally we have used a modeled future incidence rates rather than present rates because we know that cancer incidence rates are changing, so for example reducing smoking rates and we need to take that into account when we are looking at future cancers.

In conclusion, we have used the lifetime risk approach to estimate the contribution of current exposure to occupational carcinogens to the overall risk of lung cancer. We have estimated that around 26,000 future lung cancers will occur in those who were exposed in 2012 as a result of their exposure, so about 3.6% of future cases. Our results can also be used to show us where the greatest burden is coming from, which exposures are contributing the most future registrations as well as occupations and industries. We can also look at the potential effect of future interventions, which is what we are starting to look at now.

So to conclude, I would just like to acknowledge the AWES cancer team some of whom are here today, as well as our funders. Thank you.


This page was last updated on: Thursday, January 28, 2016