Trends in occupational disparities for exposure to carcinogenic, mutagenic and reprotoxic chemicals in France 2003–10

Abstract

Background: To explore trends in social and occupational inequalities in terms of exposure to carcinogenic, mutagenic and reprotoxic chemicals (CMR) for French employees. Methods: Our study assessed data from the French national cross-sectional survey of occupational hazards (SUMER) that was conducted in 2003 and 2010. We included all of the 27 CMR agents that were classified by the International Agency for Research on Cancer or European Union regulations as being known or presumed to have CMR potential in humans. Trends in prevalence and degree of exposure were examined using multilevel logistic regression analysis. Results: The number of employees exposed to CMR agents decreased by 17.5% between 2003 and 2010. The only CMR entities for which exposure rates increased are not considered to be proven CMRs according to the European Union regulations. With the exception of apprentices, there was an overall decrease in exposure prevalence for all employees. This decrease occurred, however, to different extents. The decrease in the risk of exposure to CMR agents was much greater for those on permanent contracts, managers, and in enterprises with more than 500 employees. Nonetheless, in situations where there was potential for exposure, companies with fewer than 10 employees were in fact able to decrease the degree of risk more than the others. Conclusions: Our results confirm the relevance of reinforcing regulatory restrictions for CMR products, while also indicating that monitoring of trends in disparities will allow public health policy makers to better evaluate progress made toward reducing disparities that affect vulnerable populations.

Introduction

A large number of agents classified by the International Agency for Research on Cancer (IARC) as being carcinogenic or probably carcinogenic to humans, can be found in occupational settings.¹ It has been estimated that 4–8.5% of all types of cancer are linked to occupational exposure. In France this amounts to a rate of 15 000–20 000 cancers annually.^2–4 The role of risk factors in the workplace is however often underestimated due to uncertainty regarding threshold effects in particular,² or due to the substantial delay between exposure and the occurrence of the disease. Consequently, it is estimated that only 15–30% of occupational cancers have been formally compensated.^2,5

In light of these findings, general public health policies and specific policies for the prevention of occupational cancers have been implemented in France over the past 15 years. For example, the Occupational Health Plan 2005–09 aimed to develop knowledge regarding the risks of occupational exposure, and to strengthen the efficiency of the regulations. More recently, the Occupational Health Plan 2010–14 focused specifically on prevention of the risks of exposure to materials that have a carcinogenic, mutagenic, or reprotoxic (CMR) effect. The national Health-Environment Plan 2009–13 had the same focus of reducing occupational exposure to CMRs. A clearly stated objective of the French Cancer Plan 2009–13 was to reduce the number of employees exposed to CMR products by 100 000 in the 2009–12 timeframe. This is why it struck us as pertinent to analyse changes in exposure to CMR agents between 2003 and 2010, as well as to probe for inequalities in exposure. We wished to determine whether there was a discernible decrease in exposure to CMR products in this 7-year period. Of particular interest, if such a decrease could be confirmed, was the question whether it occurred unequally and hence to the detriment of some employees for whom the exposure may have in fact increased.

Methods

Study population

The SUMER survey is a national cross-sectional survey^6–12 that is conducted periodically by the French Ministry of Labor and the French Directorate for Research, Studies and Statistics (DARES) to assess occupational risks among a representative sample of the French employee population. The survey is based on a two-level sampling involving voluntary physicians (1800 in 2003 and 2400 in 2010) who randomly selected, over a 3-month period, employees (56 345 in 2003 and 53 940 in 2010) among those who had been invited for the mandatory periodical medical check-ups. 49 984 workers in 2003 and 47 983 in 2010 consented to participate (response rate 89% for both surveys). The physicians assessed the employees’ exposure during a period of 1 week to different chemical, physical and biological agents. The occupational physician relied on both, the employee statements as well as the job description including individual and collective protections and the physician’s expertise. In case of doubt, the physician could perform a more in-depth workplace assessment. For each CMR identified, the physician assessed the exposure duration and intensity. The 2003 SUMER survey was repeated in 2010 using the same methodology to allow comparability over time. Minor differences between the two editions involved:

• The target population covered by the SUMER survey (17.5 and 21.7 million employees representing 80% and 92% of all French employees in 2003 and 2010, respectively). We focused our analysis on workers included in both surveys, i.e. workers in the private sector and in public hospitals in metropolitan France.

• The list of chemical agents has been extended between 2003 and 2010 based on policy advancements and substitutions undertaken by companies, as well as a result of increased knowledge on toxicities.¹³ Nevertheless, for the CMR agents, the difference between the two surveys is limited to the family of glycol ethers, not assessed in 2003. Overall, 27 CMR agents were identified as being common to both surveys.

Post-data collection adjustment accounted for the characteristics of occupational physicians, the periodicity of medical visits and the characteristics of non-respondents. The data were further weighted by industry sector, on gender, age, nationality, working time, occupation, company size and economic activity to ensure the representativeness of the survey samples and exposure frequencies of the target populations in 2003 and in 2010.

Ethics approval was granted by the French National Commission for Data Protection and Privacy (CNIL n° 762430 V1, 2009) and National Council on Statistical Information.

Exposure variables

The characteristics of the 27 CMR agents classified as being carcinogenic or probably carcinogenic or mutagenic to humans by the IARC (Groups 1 and 2 A) or classified as known, presumed, or suspected to have CMR potential for humans by the European Union regulations (categories 1 and 2 of the classification in effect in 2010) are shown in table 1.^7,8 Our analyses examined the changes (i) in the prevalence of exposure to at least one CMR agent in the workplace (proportion of employees exposed within 1 week); (ii) in exposure duration (reported as a categorical variable: <2 h, 2–10 h, 10–20 h, ≥20 h within 1 week) and (iii) in intensity (very low, i.e. slightly higher than the general population or at the limit of detection; low, i.e. less than 50% of the short-term exposure limit (STEL) value; high, i.e. around 50% of the STEL; very high, exceeding the STEL or equal to the exposure level of the population known to be the most exposed).

Additionally, a summary indicator of the degree of exposure, which combined the intensity and duration of exposure, was created. It was a dichotomous variable equal to 1 if the exposure was considered to be strong (a strong or very strong intensity or a slow intensity with duration > 2 h, or a very low intensity with duration > 10 h within the previous week) and 0 otherwise.⁹

Statistical methods

Descriptive statistics were used to examine trends in the prevalence, duration, intensity and degree of exposure to CMR agents. Moreover, the exposure prevalence was studied using a logistic regression with random effects^14–16 against a reference group of workers who were not exposed to CMR. The economic activity of the company was modeled with a random intercept to account for heterogeneity in the exposure probability at a sufficiently disaggregated level. The degree of exposure was studied by a logistic regression with random effects. Heterogeneity in exposure degree between the 27 identified CMR agents was taken into account by a random intercept.

The covariates included were four variables describing employee characteristics (age, gender, occupation and seniority), four variables related to job characteristics (employment contract, work hours, work schedules and main occupational duties) and four company characteristics (company size; geographical location; presence of a committee for health, safety, and work conditions (CHSCT); and economic activity of the company). CMR agents in the SUMER database were further classified using the following criteria (table 1): (1) whether a CMR agent is classified Category 1 or Category 2 by the European Union (2) whether a Binding Occupational Exposure Limit Value (BOELV) was in place for a CMR agent before 2010 and ‘0’ otherwise and (3) whether CMR agents were substitutable, based on their principal use, information provided by the French National Institute for Research, Occupational Safety and Health (www.inrs.fr); and by the French Agency for Food, Environmental and Occupational Health & Safety (www.substitution-cmr.fr). A dichotomous variable identifying the edition of the SUMER survey and interactions terms between explanatory variables and this dichotomous variable were introduced in the two models. Statistical analyses were conducted using STATA version 13.1 (StataCorp, College Station, TX).

Results

A decrease in exposure between 2003 and 2010

Table 2 presents the overall changes in exposure prevalence, duration, and intensity to CMR agents between 2003 and 2010. The proportion of employees exposed to at least one chemical CMR decreased from 13.7% in 2003–10.6% in 2010. The decrease was more pronounced for products classified as being category 1 by the IARC and for products classified as being category 1 or 2 by the European regulations. However, the duration of exposure and the intensity for CMRs were only slightly reduced.

Table 1 details the product by product changes in exposures between 2003 and 2010. The number of exposed employees decreased for all CMR agents, with the exception of phthalates, cobalt, cadmium in its various forms, metallic carbides and crystalline silica. Agents with the most pronounced decreases in exposure were rubber fumes (–67%); the three solvents benzene (–33%), tetrachloroethylene (–40%) and trichloroethylene (–62%) even though their exposure durations in the previous week and exposure intensities remained stable; and to asbestos (–33%), which has been banned since 1997 in France.

Changes in exposure inequalities

Table 3 provides descriptive statistics of the prevalence, duration, intensity and degree of exposure to CMR agents according to various employee and job characteristics. An overall decrease in exposure to CMR agents occurred for all employee categories, with the exception of apprentices. Already in 2003, apprentices were by far the most exposed category of employees. It can also be seen that the decrease in exposure to CMR agents did not occur in a homogeneous manner for all employees. For example, the decrease was particularly pronounced in companies with more than 500 employees, while it was more limited in those with fewer than 10 employees. The decrease in the degree of exposure did not apply to non-skilled workers; temporary workers; employees engaged in handling, storage, and logistics; and employees between 46 and 60 years of age. Contrary to the other categories, these employees underwent a substantial increase in the proportion of employees exposed with a strong exposure between 2003 and 2010. Yet unskilled workers represented the socio-professional category with the most pronounced decrease in the exposure prevalence. Similarly, the proportion of exposed employees importantly decreased for the 56–60 year old age categories.

Multiple regressions (table 4) show that the influence of specific characteristics on the prevalence and the degree of exposure increased. For example, compared to managers, blue-collar workers and services workers already had a higher probability in 2003 of being exposed to CMR agents and to have been subjected to a substantial degree of potential harm when exposure did occur and the situation has been worsened in 2010. The situation in regard to administrative staff also deteriorated in this same seven-year period. Thus while administrative staff had a probability of exposure and of pronounced exposure that was statistically similar to that for managers in 2003, they were exposed more often and to a greater extent in 2010. The sector of production, manufacturing and construction; and the sector of installation, repair and maintenance were the most affected by workplace exposures. In 2010, there was no significant improvement, except in the sector for handling, logistics and warehousing; and in the sector of administrative secretaries, typists and receptionists. For the sector of engineering, research and development, the probability of exposure increased in 2010.

Yet some effects were reversed between 2003 and 2010. For example, workers with permanent contracts were the most exposed employees in 2003, while they were the least exposed in 2010. Thus, the decrease in exposures occurred to the detriment of employees with fixed-term contracts and particularly for the most precarious, such as apprentices. At the company level, micro-enterprises had the highest exposure probabilities in 2003 and 2010, compared to companies with 500 or more employees or with 10–49 employees; and this gap widened over the 7 year period. Nonetheless, companies with fewer than 10 employees were better able to manage their degree of exposure relative to the others. In particular, they were able to catch up with companies with more than 500 employees, such that in 2010 these two categories of enterprises exhibited similar degrees of exposure with equivalent sectors and jobs. Lastly, the probability of having been subjected to substantial exposure was not significantly associated with the product category in the European Union regulations and with BOELVs. It was, however, significantly reduced by substitutability of a given CMR agent, regardless of the time period.

Discussion

The decrease in occupational exposures to CMR products between 2003 and 2010 in France occurred in the context of tighter regulations. For example, the 23rd of December 2003 Decree on the prevention of chemical risks fundamentally changed the rules by imposing the requirement to perform prior risk assessments, regular measurements of concentrations and for occupational physicians to inform employees of the risks of exposure to CMR agents. Similarly, the 9th of February 2006 Decree established BOELVs for certain CMR agents (e.g. wood dust, benzene, diethylamide, lead). Thus, the safe limit values are an important part of broader risk prevention strategies, and they have the advantage of providing a benchmark for the minimal level in regard to health protection. They also help in providing a better understanding of the complex notion of chemical hazards, as well as their control and prevention. Awareness of the risks of occupational exposure to CMR agents has also been heightened by the implementation on the 1st of January 2007 of the REACH regulations for streamlining restrictions and for improvement of the regulatory framework of the European Union in regard to chemicals. Moreover, this decrease can also be explained by changes in production processes, or by substitution with safer products. For example, the substantial decreases in exposure to the three solvents benzene, perchlorethylene, and trichloroethylene have occurred through use of substitute products such as detergents or non-halogenated solvents used for degreasing in the mechanical and metallurgical industries and as solvents in dry cleaning facilities. Moreover, since the 1st of March 2013, no new dry cleaning devices using perchlorethylene may be used in France, thus resulting in a gradual replacement of conventional dry cleaning machines with alternative solutions such as, for example, the aqua-cleaning process.

None of the five CMR agents which exhibited an increase in their exposure prevalence during the study period are classified as known CMRs by the European Union, and the restrictions regarding their use are therefore less stringent. However, cadmium and crystalline silica had been classified for years as category 1 by IARC. Use of products or processes resulting in exposure to crystalline silica, cadmium, cobalt, or to phthalates is, however, often unavoidable since they are essential from a technical perspective in sectors such as aerospace, construction, the metallurgical industry and the rubber/paint/plastics industries. Moreover, it is not surprising that exposure to cobalt, cadmium and metal carbides have changed in the same direction, since these products are associated with each other in some manufacturing processes. Part of the increase in the number of employees exposed to cobalt and cadmium can probably be attributed to the production of batteries, for which demand continues to expand rapidly.^17,18 The elevated increase in exposure to phthalates is due to the widespread use of this family of compounds (adhesives, detergents, pharmaceuticals and cosmetic products…) and due to the consequent increase in production levels. By contrast, exposure to asbestos is decreasing gradually over time since asbestos has been banned in France since the 1st of January 1997 and removal of asbestos from buildings and equipment is less and less performed.

The changes in exposures between 2003 and 2010 differed not only in terms of the CMR products but also according to the employee and the company categories. The decrease in overall exposure to CMR agents occurred to the detriment of employees on fixed-term contracts and apprentices in particular. Since the occupational regulations in France require abolition or substitution of CMR agents whenever this becomes technically possible, it seems likely that companies preferentially aimed their efforts toward jobs and employees with an elevated risk of being exposed for longer periods due to their long-term association with the company. Along these lines, apprentices may also tend to temporarily take health risks with the aim of being awarded a contract to undertake vocational training. This has become invaluable for gaining long-term employment in a job market where having a degree and experience have now become entrenched as clear necessities. The inequalities in exposure have in essence become a key reflection of the differences between very precarious employment and the more stable forms of employment. We found that the decrease in the risks of exposure to CMR agents was much more pronounced for permanent employees and middle management. At the company level, this decrease was much more noticeable in establishments with more than 500 employees. This could be explained by an increased ability of larger companies to commit financial resources to substitution and prevention policies, while also outsourcing certain processes associated with increased CMR exposure to third parties.^19–23 Lastly, the requirement to implement a CHSCT in companies with more than 50 employees²⁴ can, strictly speaking, reinforce the effect of the company size, and we indeed noted a more pronounced decrease in exposures to chemical CMR agents in establishments where a CHSCT was in place. Nonetheless, companies with less than 10 employees were able to achieve a better reduction in the degree of exposure relative to the others, with a more pronounced decrease in the probability of being subjected to substantial exposure. This is most likely a result of the implementation of the national Occupational Health Plan, for which the top priority target was prevention and protection in small companies. Consequently, our results indicate that the measures stipulated by these regulations aimed at small companies have preferentially impacted on being able to control occupational exposures to CMR agents, rather than in terms of their elimination as such. More effort should hence be made in this regard, and this is why small companies have again been earmarked as priority targets in the Occupational Health Plan for 2009–14.

The most encouraging result from this study was the finding that tightening of regulatory conditions for CMR products appears to have exerted a positive effect on risk reduction. The number of employees exposed to CMR agents decreased by 17.5% between 2003 and 2010, which amounted to a much more pronounced decrease than for chemical entities in general.⁸ Furthermore, monitoring trends in exposure disparities will allow public health policy makers to better evaluate progress made toward reducing disparities that affect vulnerable populations, and to adjust and improve national prevention and protection strategies.

Acknowledgements

This study was funded by the French Directorate of Research, Studies and Statistics Coordination (Direction de l'animation de la recherche, des études et des statistiques–DARES), of the French Ministry of Labour, in the framework of the call for ‘Risques du travail: autour de SUMER 2010’ project.

Funding

This study was funded by the French Directorate of Research, Studies and Statistics Coordination of the French Ministry of Labour.

Conflicts of interest: None declared.

References