Introduction

Cigarette smoking is the most important preventable cause of increased morbidity and mortality due to a number of diseases. It is widely accepted that cigarette smoking causes lung cancer and other diseases of the lung, such as chronic obstructive pulmonary disease, emphysema and pneumonia.1–6 In addition, smoking is a risk factor for malignant tumours of the pharynx, oral cavity and the urogenital bladder.7 8 Smoking of cigarettes also increases the risk of cardiovascular diseases such as myocardial infarction, stroke and arteriosclerosis.9–12

Although smoking causes a variety of diseases and a high smoking prevalence and permanent occupational disability are both great burdens on the population level, data about the impact of smoking habits on occupational disability are sparse. Only a few studies have evaluated the effect of smoking on occupational disability in detail, so far.13–17 In a cohort study of construction workers from Germany we have previously found a significantly increased risk (SIR) of occupational disability compared with the general work force (SIR 1.47, 95% CI 1.41 to 1.53).18 In particular, risk of occupational disability, which is an important end point that causes major social and economic problems,19 was found to rise significantly with number of cigarettes smoked.17

However, in our former analysis,17 the size and follow-up time of the cohort were too small to assess dose-response patterns or the impact of smoking on disability due to different causes in detail. Therefore, we enlarged the cohort and extended follow-up in order to assess dose-response patterns with respect to overall and cause-specific disability.

Materials and methods

Study population

The study cohort at baseline comprised 19 421 male employees from the German construction industry, aged 25 to 59 years, who were working as bricklayers (n=6204), painters (n=2947), labourers (n=2874), plumbers (n=2804), carpenters (n=2594) or plasterers (n=1998). They participated in a routine occupational health examination by the Institution for Statutory Accident Insurance and Prevention in the Building Trade in Württemberg (a region with about 5.4 million people in the South of Germany) between 1 August 1986 and 31 December 1992. This occupational health surveillance is based on legislation on health and safety at work and regular examinations are offered to all construction workers. Although participation is voluntary, over 75% of all invited employees underwent the medical examination during the period of recruitment and were eligible for follow-up. All participants were members of the statutory pension fund and didn't receive a disability pension at baseline examination. They were representative for the underlying population of all construction workers with respect to age, nationality and type of occupation. The study was approved by the ethics committees of the medical faculties of the university clinics of Heidelberg and Ulm and by the Baden-Württemberg State Ministry of Social Affairs.

Data collection

Baseline examination

The health examination at baseline was part of the routine occupational health surveillance examination in which medical and occupational history as well as, among others, information on tobacco and alcohol consumption were obtained through a standardised personal interview.

Follow-up

Information on date and cause of incident occupational disability was obtained from the German pension fund in March 2006, while the cut-off date for follow-up was June 2005. This time lag is needed, because some cases of occupational disability are approved in retrospect. Mean follow-up time per person with respect to risk of occupational disability was 10.8 years in the entire cohort. Cause of occupational disability was coded according to International Classification of Diseases, ninth edition (ICD-9) by trained medical officers from the pension fund. In case of missing data, which occurred mainly due to the move of some workers outside the population covered in the regional pension fund, we also included the information from previous follow-up rounds performed at two points in time (1992–1994 and 1998–2000).18 20 The pension register of the German Pension Fund Baden Württemberg provided information regarding vital status and whether the individual was still working, had retired due to age, was unemployed or under rehabilitation, or whether a disability pension (permanent or temporary) was granted. This information is available as it is directly linked to payments to and from the pension fund. In case of transient or multiple temporary occupational disabilities, first occurrence of disability was taken as endpoint for the analysis.

The criteria for being work disabled and receiving disability pension are under repeated revision. Up to the year 2000, a disability pension was granted in Germany when the earning capacity (ie, the working hours) had been permanently reduced by at least 50% due to injury, illness, or impairment—irrespective of whether the injury was caused by work or not—and whether the worker was able to work in another occupation. In the year 2001, the threshold was set to 3 and 6 h of work ability per day for complete and partial work disability. A patient applies for a disability pension at the local insurance office, which requires a health certificate from the applicant's primary doctor. In addition, a doctor employed by the pension insurance institution examines the applicant and judges whether the patient meets the criteria for a disability pension or whether a rehabilitation measure might be appropriate first.

Statistical analysis

For this analysis, we had to exclude 3065 men with missing data of smoking status (15.8%). Another 1873 men (9.6%) had to be excluded, who had either moved to a different region or changed employment. The very strict confidentiality rules in Germany did not allow us to follow these people further. We used information from earlier follow-up for 1221 persons (6.3%) with missing data in the current follow-up. Hence, the final study population for this analysis comprised 14 483 construction workers who could be successfully linked with the pension register.

Smoking habits at baseline were measured by number of cigarettes smoked a day. Smoking status was categorised as never smoker, former smoker or current smoker of 1–9, 10–19, 20, or more than 20 cigarettes/day or smoker of other tobacco products. Supplementary analyses based on ‘pack-years’ as the measure of smoking habits were also conducted, but are not presented in detail as results were very similar.

Techniques of survival analysis were employed to assess the impact of smoking status on the occurrence of occupational disability. We defined the onset of being work disabled as point of time, from which a disability pension was granted—irrespective of the date of ascertainment of occupational disability. A person was denoted as censored in case of being known not to receive disability pension compensation according to pension fund records, or termination of pension fund insurance due to other reason, such as retirement pension, 65th birthday, death, or change to another insurer, or (in analyses of cause-specific disabilities only) occupational disability due to another cause.

Relative occupational disability hazards according to level of smoking status were calculated using the Cox proportional hazards model. In addition, assessing the independent association of smoking with occupational disability, we also adjusted for age, nationality, type of occupation, body mass index (BMI; kg/m²) and alcohol consumption at baseline as covariates in the multivariable analysis. For age, a linear and a quadratic age term were simultaneously entered into the model while index variables were created for the other, categorical variables.

Further baseline data available comprised information regarding prevalent diseases at baseline such as respiratory diseases (ICD-9: 460–519) and cardiovascular diseases (ICD-9: 390–459). As these diseases may represent intermediate steps in the causal pathway between smoking and occupational disability, they were not included as confounders in the multivariate modelling. However, additional analyses were carried out in subgroups according to presence or absence of defined types of comorbidity and with respect to cause-specific occupational disability. Within these subgroup specific analyses, we combined the smoking categories to prevent too imprecise effect estimates as follows: never smoker, former smoker, moderate smoker (1–19 cigarettes/day or smoker of other tobacco products) and heavy smoker (20 or more cigarettes/day).

To minimise possible bias due to disease-related reduction or cessation of smoking and due to violation of the proportional hazard assumption, we carried out separate analyses for the first 3 years and all subsequent years of follow-up, and we excluded the first 3 years of follow-up in all subgroup specific analyses. Furthermore, the population attributable risk proportion of occupational disability due to smoking (ie, subsuming all current as well as former smokers to one group) was calculated to quantify the public health relevance of smoking in this population.

All statistical analyses were performed with the SAS statistical software package, release 9.1 (SAS Institute, Cary, North Carolina, USA). The proportional hazards assumption was checked by log (-log) survival plots.

Results

Baseline characteristics

Characteristics of the study population (overall and grouped by smoking status) are shown in table 1. Of all 14 483 men included, 3497 (24%) were never smokers, 2578 (18%) were former smokers, 3093 (21%) were light or moderate smokers (1–19 cigarettes/day or smokers of other tobacco products), while 5315 (37%) were heavy smokers (20 or more cigarettes/day). A majority of 62% of men were classified as being overweight or obese (BMI≥25.0 kg/m²). The mean age of the study population at baseline was 41.7 years, and about 75% of the cohort members were of German nationality. Regular alcohol consumption was very common with a proportion of 41% consuming more than 30 g/day.

Construction workers with higher cigarette consumption were on average younger and consumed larger amounts of alcohol than light or never smokers. The proportion of foreign workers was higher among current smokers of cigarettes than among non-smokers, former smokers or smokers of other tobacco products.

The prevalence of overweight or obesity at baseline was highest among former (76%) and never smokers (68%) and lowest among smokers of 10 or more cigarettes per day (53% to 56%). There were no significant differences in smoking patterns between occupational groups (data not shown). Almost half of the study population (44%) had musculoskeletal disorders at baseline, most of them dorsopathy, followed by cardiovascular disease (31%) as second most frequent comorbidity. Respiratory disease (13%) and disorders of the liver/bile/pancreas (14%) were also common, and their prevalence increased with number of cigarettes smoked a day. The latter was also observed for mental disorders. By contrast, the proportion of persons without any recorded comorbidity (9%) was lowest among heavy smokers.

All cause occupational disability according to smoking

Overall there were 2643 incident cases of occupational disability during a mean follow-up time of 10.8 years. Table 2 presents the association between smoking and all-cause occupational disability with never smokers as reference category. Age adjusted analysis revealed a significantly increased risk for being work disabled in all categories of smoking with by far highest risk (HR 1.95, 95% CI 1.72 to 2.22) among heavy smokers (>20 cigarettes/day). Moderate risk elevations in the magnitude of about 30% to 40% were observed for light (1–9 cigarettes/day), moderate (10–19 cigarettes/day) and former smokers. Further adjustment for BMI, nationality, type of occupation and alcohol consumption did not materially alter the observed associations.

Of the 2643 cases of occupational disability, 452 (17%) occurred during the first 3 years of the follow-up period. Again, the highest risks were observed for heavy smokers.

After exclusion of the first 3 years of follow-up, the results changed only slightly compared to the analysis without left truncation. After adjustment for age, BMI, nationality, type of occupation and alcohol consumption we observed a significantly increasing risk with increasing number of cigarettes smoked a day (p value for test for trend <0.0001), along with significantly elevated risks also among former smokers and smokers of other tobacco products. Within the entire cohort, 27.1% of all occupational disabilities were statistically attributable to smoking.

Further analyses revealed a stronger positive association in relative terms between smoking and occupational disability among younger construction workers than among older workers. For example, the age specific HR for heavy smokers (ie, 20+ cigarettes/day) compared with non-smokers were 2.52 (95% CI 1.85 to 3.43) and 1.77 (95% CI 1.49 to 2.09) for age under 45 years and older ages, respectively.

All-cause occupational disability according to smoking status and comorbidity

Table 3 shows occurrence of occupational disability ≥3 years after recruitment according to smoking and baseline comorbidity. The increase of adjusted HR was more pronounced among men with respiratory diseases at baseline than among men without these disorders. In contrast, the association of smoking with occupational disability was somewhat weaker among construction workers with musculoskeletal disorders, the most frequent comorbidity, compared to those without these disorders, but still remained significant in all smoking classes. The association did not materially change after further stratification for cardiovascular diseases, which was very common among construction workers, or further stratification for mental disorders.

Occupational disability according to smoking and cause of disability

Information on cause of disability could be obtained for 92% of all cases of occupational disability. Among those, with 906 (46%) cases, musculoskeletal disorders represented the most common cause of occupational disability, half of them being dorsopathies. We therefore distinguished in the analysis between dorsopathies and osteoarthritis, the second most frequent musculoskeletal disorder. The second most common cause was cardiovascular diseases (18%), followed by mental disorders (9%), cancer (8%) and respiratory diseases (4%). Frequencies and HR of cause-specific occupational disability after exclusion of the first 3 years of follow-up are shown in table 4. With the exception of osteoarthritis, increased risks of occupational disability due to all of the assessed causes were observed among heavy smokers. With HR of 3.25 and 3.26 the strongest risk elevations were seen for occupational disability due to mental and respiratory diseases. For most of the assessed diseases, namely cancer, cardiovascular diseases, mental as well as respiratory diseases, risk of disability monotonically increased with smoking intensity. For dorsopathy, cardiovascular and mental diseases significant risk elevations were also seen among moderate smokers, and for dorsopathy, risk was significantly elevated even among former smokers.

Discussion

In this large occupational cohort of 14 483 construction workers, age-adjusted analysis revealed that smoking is clearly associated with an increased risk of occupational disability due to all causes. Further adjustment for potential confounding factors such as BMI, alcohol consumption, nationality as well as type of occupation did not materially change this pattern. More than a quarter of all occupational disabilities were statistically attributable to smoking. In addition, a consistent risk increase was found for respiratory, cardiovascular and mental diseases, cancer and dorsopathy, which represent the major causes of disability in this occupational group. The strongest associations were seen between heavy smoking and occupational disability due to respiratory and mental disorders. Risk of occupational disability due to dorsopathy was even elevated at moderate levels of smoking and among former smokers.

The association of smoking with respiratory diseases, including chronic obstructive pulmonary disease, emphysema and pneumonia is well established.3–6 Among construction workers, the risk of respiratory diseases may further be elevated by additional occupational factors, such as exposure to dust.21 The particularly strong association of smoking with occupational disability due to respiratory diseases among workers with pre-existing respiratory disease might point to potential synergistic effects with other factors such as occupational dust exposure,4 22 23 and they underline the particular importance of promotion of smoking cessation in this occupational group.

The association of smoking with cardiovascular diseases, including myocardial infarction, stroke and arteriosclerosis is likewise well established.9–12 In our cohort, 16% of all cases of occupational disability were due to cardiovascular diseases. Previous studies have shown relatively low mortality and disability rates due to cardiovascular diseases in construction workers.24 25 On first view, these findings seem to be in contrast with the high prevalence of smoking (58% vs 37% in men from the general population in Germany)26 and the increased risk of occupational disability due to smoking among construction workers found in our study. Taken together, however, these results may indicate that the risk of cardiovascular diseases could potentially be at very low levels among construction workers, possibly due to the high levels of physical activity in this occupational group, if the very high smoking prevalence could be reduced.

Our finding of increased risk of occupational disability due to cancer among heavy smokers is consistent with the well known association of smoking with various malignancies such as lung cancer and malignant tumours of the larynx, pharynx and urogenital bladder.2 7 8

The association of smoking with increased risk of disability due to mental diseases is more difficult to interpret. Despite the longitudinal design of our study, the time sequence and causal relationship between mental disorders and smoking is less unequivocal than for other diseases. In particular, there is clear evidence that smoking habits themselves are strongly influenced by mental disorders27 and that certain mental diseases, including depression, anxiety and schizophrenia are associated with high smoking prevalence.28 29 Furthermore, mental disorders could generate among others psychotropic drug use, which is well known to be associated with a number of diseases, occupational injuries and altered living conditions.30 The most common mental diseases leading to occupational disability in our cohort were alcohol dependencies, which are associated with increased smoking prevalence.

A recent study showed that smoking is a risk factor for occupational injury.31 In our cohort, we could not examine this association in detail, since work-related injuries due to accidents are covered by the workmen's compensation board for construction workers and therefore those injured workers didn't receive a disability pension from the German pension fund. However, chronic forms of injuries other than musculoskeletal disorders as cause of occupational disability (ICD-9: 800–999) were not very common in our cohort and didn't play a role on the association of smoking with occupational disability.

The most important novel finding of our study is the clear association of smoking with increased risk of occupational disability due to dorsopathy. Several epidemiological studies have suggested smoking to be a risk factor for back-related problems32–34 but evidence from longitudinal studies is still limited. The consistent significant associations found in this longitudinal study that persisted after control for potential confounding factors support suggestions of a potential causal role of smoking in the development of dorsopathy. Plausible pathogenetic mechanisms might include—for example, effects of smoking on lumbar disc degeneration and herniation.35 36

The results found in the present study, that occupational disability was related to smoking are in agreement with several investigations demonstrating occupational demands, smoking and alcohol use as strong risk factors for premature mortality.25 37

A number of limitations require careful discussion in the interpretation of our study. First, the data from this cohort were collected during a routinely performed health examination. Information regarding smoking was self-reported and not confirmed by biochemical markers, such as serum cotinine. Second, information on smoking habits was only ascertained at baseline examination. Therefore, changes in smoking behaviour or other risk factors during the follow-up period could not be considered. If workers who quit smoking after the baseline examination had a lower risk of early retirement due to occupational disability than those who continued smoking, this might have led to underestimation of associations with occupational disability.

Although we controlled for major potential confounders including BMI and alcohol consumption there remains a potential for residual confounding as a result of imperfect measurements of covariates. In particular, alcohol consumption was self reported and, like BMI, measured at a single point of time only. Other important factors such as social and labour market-related factors which are known to affect disability risk15 38 were not available in sufficient detail. However, adjustment for type of occupation, which might be an indicator for such factors did not materially change the strong association of smoking with occupational disability. Further research on the impact of job-related factors such as high occupational demands, which could favour smoking, and other occupational hazards (eg, dust, noise, climatic conditions) on occupational disability needs to be performed. Alternative explanations for the association of smoking and occupational disability in this cohort could be lower socioeconomic status, unhealthy lifestyle such as diet factors as well as less participation on preventive measures.

Our study also has particular strengths, including the size of the study population, the length and completeness of follow-up. The large case number of occupational disabilities allowed us to assess the smoking–disability association in great detail, in particular with respect to dose-response-relationships.

In conclusion, smoking was found to be highly prevalent and strongly associated with increased risk of occupational disability even after adjustment for age, BMI and alcohol consumption among construction workers.

Besides the established impact of smoking on respiratory and cardiovascular disease as well as cancer, smoking was found to be a major risk factor of occupational disability due to dorsopathy, the most common cause of occupational disability among construction workers. To reduce these long-term health effects of smoking, health promotion programs in the workplace should be implemented with a major focus on smoking cessation.