**3. Results**

Figure 1 and Table S1 show the percentages stratified by age among persons who suffered from CVDs in the working and general populations. WRCVDs had greater percentage contributions, than in the general population, among persons aged 45–54 (42.16%) and 35–44 years (25.49%) in the working population. In contrast, higher CVD attack rates were found in the older age group ≥65 years (61.69%) in the general population than in the working population.

**Figure 1.** Comparisons of percentage of persons suffering from WRCVDs and CVDs in the working and general populations in Taiwan, 2006–2013, stratified by age. WRCVDs: work related cerebrovascular and cardiovascular diseases. CVDs: cerebrovascular and cardiovascular diseases. y: year.

Table 1 reports the number of event and annual age-specific CVD attack rates (per 100,000) in the working and general populations. In the working population, WRCVD attack rates increased approximately four-fold from 2006 to 2013 (2.10 to 8.60 per 100,000 persons). In the same interval, CVD rates in the general population slowly increased (1113.55 to 1853.32 per 100,000 persons) approximately two-fold. A significant age trend was shown in the general population, but such a trend was not seen in observations of WRCVDs.


**Table 1.** Annual event number and age-specific attack rates (per 100,000) of WRCVDs and CVDs in working and general populations.


**Table 1.** *Cont.*

Event/attack rates (per 100,000) are shown. WRCVDs: work related cerebrovascular and cardiovascular diseases. CVDs: cerebrovascular and cardiovascular diseases.

The trends of annual age-specific WRCVD attack rates and CVD attack rates are displayed in Figure 2a,b. WRCVD attack rates generally increased with age (shown in Figure 2a). Moreover, in the general population, a marked increase occurred in the age group ≥65 years, which had the highest CVD rates (as shown in Figure 2b).

(**a**)

**Figure 2.** *Cont*.

**Figure 2.** Trends of annual age-specific WRCVD rates and CVD rates (**a**) in the working population and (**b**) in the general population. WRCVDs: work related cerebrovascular and cardiovascular diseases. CVDs: cerebrovascular and cardiovascular diseases.

(**b**)

An age–period model was used to estimate the RR and 95% CI for WRCVDs and CVDs, and results are shown in Table 2. The population aged ≥65 years was a very different population in the working population than in the general population. Therefore, persons aged ≥65 years old were excluded from the final analysis.

In the working population, period RRs were strongly significant in 2007 and 2010~2013 as compared to 2006 as the reference. An approximate four-fold increase in the WRCVD risk was indicated from 2006 to 2010–2013. Such significant effects of WRCVDs were also found in workers aged 25–64 years. The RRs for WRCVDs increased with age, but we found that the risk rose only two-fold. The highest risk (RR = 1.84; 95% CI = 1.37–2.46) in the 55–64-year age group was shown in comparison to that aged 35–44 years. In the general population, all RRs for CVD onset in each period and age group were significant.

In order to examine the evolution of workers suffering CVD risk disparities across periods and ages, we presented the RRs of CVDs in Table 2. For each concentration of work effect, adjusted RRs were calculated using RRs from the working and general populations. The adjusted RR slightly increased over time, with a considerable increase in 2012 (adjusted RR = 3.84; 95% CI: 2.38–5.61). Of note, there was an observed increase in the adjusted RR in the age group 25–34 years (adjusted RR = 1.08; 95% CI: 0.66–1.94), which significantly decreased in workers aged ≥45 years. At 55–64 years of age, contributions of the effect of work on the CVD risk were slight (adjusted RR = 0.29) but still strongly significant (95% CI: 0.20–0.37).

PARs contributed by working were also calculated using standardized rates. After age and year standardization, the PAR of working exposure was 13.45% (95% CI = 13.2–13.7%). There was a decreasing PAR trend in the age categories 15–24, 25–34, 35–44, 45–54, and 55–64 years, and the percentages were 17.64% (95% CI = 17.4–17.9%), 16.89% (95% CI = 16.7–17.1%), 16.46% (95% CI = 16.2–16.7%), 10.6% (95% CI = 10.4–10.8%) and 0.65% (95% CI = 0.58–0.69%) when standardized by year.


#### **4. Discussion**

As far as we know, this is the first longitudinal study to present the attributes of WRCVDs considered as occupational diseases by workers' compensation systems for comparison to all CVDs in the general population. By comparing two population-based databases from national labor and health insurance, we found disparities between the working and general populations in CVD attack rates. A PAR of 13.45% was contributed from working exposure in populations aged 15–64. The contribution from occupations to CVD risk significantly decreased with age, from an RRadjusted of 0.63 at 45–54 years to an RRadjusted of 0.29 at 55–64 years, when compared to the reference group (persons aged 35–44 years) The PARs of working exposure also decreased with increasing age. This demonstrates that work and work-related factors are very significant risk factors in younger populations.

Three major findings were shown when we looked at temporal and age differences from the age–period analysis. First, WRCVDs and CVDs attacks increased with aging and period in populations aged 15–75 years from 2006 to 2013. However, their performances were different. Second, the magnitude of the differences in CVDs between the working and general populations substantially varied by age, with the largest disparities observed particularly among young and middle-aged adults in the main labor force. A pattern of adjusted RRs and PAR was shown, with significantly higher impacts focused on WRCVDs in young workers. In contrast, aging prominently led to increased CVD onset in the general population. Finally, disparities in period-specific adjusted RRs from CVDs for 2009 and 2013 were modest. However, they were all >1, indicating a substantial improvement in the reporting and registration of job-related CVDs after the revised guidelines were promulgated.

In our study, the age–period model was used to adjust for age and period influences [22], and RRs denoted the ratios of morbidity rates as a relative indicator for the reference group after adjustment. Taking account of background age and period trends, RRadjusted was used to adjust residual confounding in the general population. In our observations, the annual CVD attack rates from 2006 to 2013 ranged from 1114 to ~1853 per 100,000. These findings are similar to those in developed countries including the United States, Australia, and Britain [23,24]. As far as we know, age is the strongest risk factor for CVDs in the population, and the aging of the population is projected to continue. We adjusted for the age effect in our age–period model analysis. In addition, the NHIRD population is a representative sample of the general population in Taiwan, which was good for excluding residual confounding.

In our analysis, the working population WRCVD attack rate increased with age. However, the age distribution of CVD events differed between the two populations. Among 109,236 CVD events in the general population, 61.69% were among people aged ≥65 years. However, among 408 WRCVDs, 42.16% were among workers 45–54 years old and 25.49% were among workers 35–44 years old. After eliminating background residual confounding, attributable work-related excessive CVD risk mainly occurred in workers aged 45–54 and 35–44 years. The effect was huge, i.e., fourfold (RRadjusted from 0.29 at 55–64 years to 0.6–1.0 at 35–54 years). A Scottish study indicated that premature death from coronary heart disease remains a major contributor to the most affluent groups aged 35–44 and 45–54 years [25,26]. Since middle-aged workers in Japan have been reported to experience prolonged working hours and occupational stress related to CVD risk [27], much evidence has focused on associations between CVDs and occupational factors such as working hours and stress [28].

In a worldwide diagnosis of guidelines for occupational CVDs made in different periods, the governments of Taiwan, Japan, and Korea only recognize CVDs caused by overwork as WRCVDs. The Taiwan Labor Department published "Guidelines for the Diagnosis of Work-Related Cardiovascular Diseases" in 2004 and revised it to be more sensitive in December 2010. Our study showed the higher adjusted RRs of CVDs in 2007 and 2012. The WRCVD attack rate increased from 5.71 in 2010 to 8.60 in 2013 per 100,000 people in the post-2010 revised guideline period. Similarly, the Japanese government recognized WRCVD diagnostic criteria in late 2001, and a gradual increase was found in total compensated CVD occupational diseases [15]. Nevertheless, the Korean government produced occupational health standards in 2003 and found that compensated CVDs accounted for 26% of the total compensated diseases in 2003, which was a dramatic increase, but the proportion dropped to 7% by 2009 [1]. Reducing compensatory CVDs may be attributable to many preventive activities carried out by governments and employers, but occupational health policy advocacy may increase the annual recognition of WRCVDs. The possible directions need further study.

We provided an exploratory descriptive tool to examine the occupation-attributed relative risk by age and period by taking account of the residual confounding of unknown background factors. Most notably, the present study is the first to examine combined national labor and health insurance databases, which was ideal and strengthened the results. The CVD diagnosis was performed according to ICD-9 codes. WRCVD events were identified through the application and review of guidelines from the government, which identified occupational causes of acute circulatory diseases. The population size was representative and was based on national insurance databases of the workforce and general population. Therefore, the findings in this study are reasonable to present as epidemiological evidence.

Three limitations should be considered in this study. First, the application accuracy of occupational diseases is related to a willingness to recognize occupational causes of injuries or health problems. These assessments included personal exposures to environmental risk factors, for example, from evidence of job insecurity, working hours, job intensification, and management [29,30]. This might vary by age, period, social culture, and country. Problems are also related to the historical background of workers' demands for protection and prevention or compensation and their employers seeking to deny or reduce their liability for work-related diseases and injuries. Second, CVD identification in the general population is based on ICD-9 codes. Due to data extraction limitations, all attack cases included recurring events. However, the stratum-specific rate ratio may have a lower impact in estimation because the extraction criteria for each group are the same. Finally, most WRCVD registered events are recorded for males. The estimations are unstable in some strata in separating by gender. This issue should be considered in the future.

#### **5. Conclusions**

CVDs occurred in different periods and age groups in the two insurance system databases. The relative effects attributed to work were more severe in the younger population. Persistent intensive assessment and management of overwork and preventing WRCVDs among young workers are important [31,32]. It should be noted that monitoring regimens across subgroups, as well as the most effective timing and efficacy of primary, secondary, and tertiary preventive interventions for public health policies should be determined in future studies.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/1660-4601/16/6/961/ s1, Table S1: Percentage stratified by age and year among persons who suffered CVDs in the working and general populations.

**Author Contributions:** Data curation, Y.-Y.H. and R.W.; funding acquisition, C.-H.B.; investigation, C.-H.B.; methodology, C.-H.B.; project administration, C.-H.B.; supervision, C.-H.B.; writing—original draft, Y.-Y.H.

**Funding:** This study was supported by the Institute of Labor, Occupational Safety, and Health (awarded to ILOSH) and Ministry of Labor, and Ministry of Science and Technology (MOST 107-2314-B-038-072-MY3) in Taiwan.

**Acknowledgments:** This work was supported by the Council of Labor Affairs, Institute of Labor Occupational Safety and Health in Taiwan.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

International Journal of *Environmental Research and Public Health*
