Article Text

Download PDFPDF

Original article
Relationship between cumulative exposure to 1,2-dichloropropane and incidence risk of cholangiocarcinoma among offset printing workers
  1. Shinji Kumagai1,
  2. Tomotaka Sobue2,
  3. Takeshi Makiuchi2,
  4. Shoji Kubo3,
  5. Shinichiro Uehara4,
  6. Tomoshige Hayashi4,
  7. Kyoko Kogawa Sato4,
  8. Ginji Endo4,5
  1. 1Department of Occupational and Environmental Management, University of Occupational and Environmental Health, Kitakyusyu, Japan
  2. 2Department of Environmental Medicine and Population Sciences, Graduate School of Medicine, Osaka University, Osaka, Japan
  3. 3Department of Hepato-Biliary-Pancreatic Surgery, Osaka City University Graduate School of Medicine, Osaka, Japan
  4. 4Department of Preventive Medicine and Environmental Health, Osaka City University Graduate School of Medicine, Osaka, Japan
  5. 5Ohara Memorial Institute for Science of Labor, Tokyo, Japan
  1. Correspondence to Professor Ginji Endo, Ohara Memorial Institute for Science of Labor, in Oberlin University, 1-1-12, Sendagaya, Sibuya-ku, Tokyo, Japan; endog{at}med.osaka-cu.ac.jp

Abstract

Objectives This study aimed to evaluate the relationship between cumulative exposure to 1,2-dichloropropane (1,2-DCP) and incidence risk of cholangiocarcinoma among workers in the offset proof-printing section of a small printing company in Osaka, Japan.

Methods We identified 95 workers of a printing company (78 men and 17 women) who had been exposed to 1,2-DCP between 1987 and 2006, and calculated the standardised incidence ratio (SIR) of cholangiocarcinoma from 1987 to 2012. We estimated cumulative exposure to 1,2-DCP and calculated SIRs in four exposure categories. We also calculated incidence rate ratios (RRs) adjusted by sex, age, calendar year and dichloromethane (DCM) exposure for three exposure categories using Poisson regression analysis.

Results Cumulative exposures to 1,2-DCP ranged from 32 to 3433 ppm-years (mean, 851 ppm-years) and the SIR was 1171 (95% CI 682 to 1875). In the analysis of the four exposure categories, SIRs increased significantly in the three highest exposure categories, but not in the lowest category. Adjusted RRs in the middle and high exposure categories were 14.9 (95% CI 4.1 to 54.3) and 17.1 (95% CI 3.8 to 76.2), respectively, in the analysis without lag time, and were 11.4 (95% CI 3.3 to 39.6) and 32.4 (95% CI 6.4 to 163.9), respectively, in the analysis with a 5-year lag. The trend analysis revealed a significant increase in RR in association with increasing cumulative exposure to 1,2-DCP. DCM exposure was not significantly associated with the development of cholangiocarcinoma.

Conclusions The present study demonstrated an exposure–response relationship between exposure to 1,2-DCP and the development of cholangiocarcinoma.

  • 1,2-dichloropropane
  • Bile duct cancer
  • Printing workers

Statistics from Altmetric.com

Request Permissions

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.

What this paper adds

  • It was previously reported that 17 workers had developed cholangiocarcinoma in the offset proof-printing section of a small printing company in Osaka, Japan, and that they had used 1,2-dichloropropane (1,2-DCP) for ink-removal operations.

  • The International Agency for Research on Cancer classified 1,2-DCP as group 1 (carcinogenic to humans) mainly on the basis of the above findings, but the dose–response relationship between 1,2-DCP exposure and cholangiocarcinoma has not been evaluated.

  • This study revealed that the incidence risk of cholangiocarcinoma increased with increasing cumulative exposure to 1,2-DCP among 95 workers of the offset proof-printing section, suggesting that an exposure–response relationship exists.

  • Further studies are necessary to elucidate the mechanism of carcinogenesis resulting from 1,2-DCP exposure, and to determine why the region of tumour development differs between humans and rodents.

Introduction

In May 2012, Kumagai and Kurumatani1 reported that at least five workers (including former workers) had suffered from intrahepatic or extrahepatic bile duct cancer (cholangiocarcinoma) in an offset proof-printing company (Company A) in Osaka, Japan; this was the first report of occupational cholangiocarcinoma in the world. Kumagai et al2 also reported that 11 workers (including the above five) had developed cholangiocarcinoma in the plant by the end of 2011. Subsequently, Kubo et al3 identified other workers with cholangiocarcinoma in the plant and the total number reached 17 by the end of 2012. All of the 17 patients had been exposed to high levels of 1,2-dichloropropane (1,2-DCP) for a long term during ink removal operations. Their ages at diagnosis were 25–45 years, and the duration from the first 1,2-DCP exposure to the diagnosis ranged from 7 to 20 years. All were acknowledged to have developed an occupational disease by the Ministry of Health, Labour and Welfare, Japan (hereafter referred to as ‘the Ministry’).4

After this incident became widely known through the mass media, workers with cholangiocarcinoma in other printing companies filed claims for workers' compensation, reaching 75 (excluding the above 17) in total as of January 2015.5 Among the cases, two, two and three patients occurred in three small printing companies and all the patients had also been exposed to 1,2-DCP at high levels for a long term.5–7

These findings strongly suggest that 1,2-DCP can cause cholangiocarcinoma in humans. The International Agency for Research on Cancer (IARC) classified 1,2-DCP as group 1 (human carcinogen) in June 2014.8

Sobue et al9 evaluated the relationship between cumulative duration of exposure to 1,2-DCP and the incidence of cholangiocarcinoma among 106 workers who had worked in the proof-printing section at the Osaka plant of Company A, and found that the incidence risk increased with cumulative duration. However, the study did not evaluate the relationship between cumulative exposure (ppm-years) and incidence of cholangiocarcinoma. In this study, we estimated the exposure concentrations of 1,2-DCP among workers in the proof-printing section at Company A and evaluated the relationship between cumulative exposure to 1,2-DCP and incidence of cholangiocarcinoma.

Subjects and methods

Subjects

Company A had five plants in which 1,2-DCP was used. These include the former Osaka plant (operated from July 1985 to March 1991, ‘Plant O-1’), the current Osaka plant (from April 1991 to 2015, ‘Plant O-2’), the former second Osaka plant (from April 2002 to 2013, ‘Plant O-3’), the former Tokyo branch plant (from February 2002 to September 2008, ‘Plant T-1’) and the former second Tokyo branch plant (from January 2005 to September 2008, ‘Plant T-2’). Since Plant T-2 was in operation only during busy periods, we did not include this plant in this study.

Using employee lists obtained from Company A and the Ministry, we identified 116 workers (94 men and 22 women) who had worked in the proof-printing section of Plant O-1, Plant O-2 and/or Plant O-3 between 1985 and 2012. We excluded eight workers with missing information for date of birth, date of employment or date of retirement and two workers with unknown vital status. Furthermore, we excluded 11 workers who had started working after November 2006 because they had not been exposed to 1,2-DCP. The final number of workers analysed was 95 (78 men and 17 women), of whom 67 (70.5%) worked in Plant O-2, 11 (11.6%) in Plants O-1 and O-2, seven (7.4%) in Plants O-2 and O-3, four (4.2%) in Plants O-2 and T-1, three (3.2%) in Plant O-3, two (2.1%) in Plant O-1 and one (1.1%) in Plants O-2, O-3 and T-1 (we obtained information on workers who transferred from Plant O-2 or Plant O-3 to Plant T-1 through interviews with them and/or their colleagues).

Description of plants

We obtained information on the four plants from Company A and the Ministry. Table 1 summarises the characteristics of the plants.

Table 1

Characteristics of four plants

In Plant O-1, there were four printing machines in the printing room and printing workers printed about 10 sheets for each set of proofs using red, blue, black and yellow inks sequentially (‘proof-printing’). To change colours, they removed ink from the ink roller using kerosene, and from the ink transcription rubber roller (this roller transfers ink from the printing plate to paper, hereafter referred to as ‘blanket’) using chlorinated organic solvents such as 1,2-DCP and dichloromethane (DCM). The workers wore plastic gloves during the ink removal operations, but did not use respiratory protection while working. Two exhaust fans were installed in the printing room. Since the company did not save information regarding the exhaust fans, the ventilation rate was assumed to be 1000 m3/hour per exhaust fan in this study, according to the performance of a normal exhaust fan sold in Japan. Consequently, the total ventilation rate was estimated to be 2000 m3/hour.

In Plant O-2, there were six or seven printing machines in the printing room and printing workers conducted proof-printing without respiratory protection. Two general ventilation systems were installed in the room and the real ventilation rate was 3300 m3/hour, according to a report of the National Institute of Occupational Safety and Health, Japan (JNIOSH).10 The front room was adjacent to the printing room, and the other workers (front room workers) were in charge of managing proof-printing, making printing plates and/or preparing printing paper. In the front room, workers who delivered printed proofs to their clients (delivery workers) also stood ready to receive printed proofs from printing workers. Although neither 1,2-DCP nor DCM was used in the front room, the contaminated air in the printing room belched into the front room through a gap in the door due to poor air pressure balance, and front room workers and delivery workers were also exposed to these chemicals.

In Plant O-3, there were two or three printing machines (two of the machines were regularly in operation) in the printing room and printing workers were engaged in proof-printing without respiratory protection. An exhaust fan was installed in the room. The ventilation rate was assumed to be 1000 m3/hour. In this plant, ultraviolet-printing (UV-printing) as well as normal proof-printing had been performed since the opening of the plant. The first UV irradiator was installed in the room in 2002, and the second in 2004. A cooling fan was attached to each of the UV irradiators, and air was exhausted to the outside by the fan. The exhaust rate was about 1000 m3/hour per cooling fan. Consequently, the total ventilation rate combining the exhaust fan and cooling fan was 2000 m3/hour in 2002 and 2003, and 3000 m3/hour in 2004 and thereafter.

In Plant T-1, there were two or three printing machines (two of the machines were regularly in operation) in the printing room and printing workers were engaged in proof-printing without respiratory protection. A general ventilation system was installed in the room. Based on a layout drawing of the plant and interviews with printing workers, the type of ventilation system was similar to that of Plant O-2. Given that the volume of the printing room was about half that of Plant O-2, with one-half or one-third of the number of printing machines used, the ventilation rate was assumed to be 1650 m3/hour (ie, half of that in Plant O-2). In this plant, UV-printing as well as normal proof-printing had been performed since the opening of the plant, and a UV irradiator was installed in the room. The exhaust rate of a cooling fan attached to the UV irradiator was about 1000 m3/hour. Consequently, the total ventilation rate combining the general ventilation equipment and the cooling fan was 2650 m3/hour.

Exposure assessment

Chemical components and amounts of blanket cleaners

Data regarding chemical components and monthly purchase volumes of the blanket cleaners in the four plants were obtained from Company A and/or the Ministry. Table 1 summarises the chemical components of the blanket cleaners used during the observation period. The blanket cleaner used from November 1987 to 1992 was a mixture of 1,2-DCP (50−60% by weight), DCM (15−20%) and 1,1,1-trichloroethane (1,1,1-TCE) (15−20%). From 1993 to February 1996, a mixture of 1,2-DCP (40−50%), DCM (40−50%) and petroleum solvent (1−10%) was used. Nearly pure 1,2-DCP (> 95%, BC-C) was used between March 1996 and October 2006 in Plant O-2, between April 2002 and July 2006 in Plant O-3 and between February 2002 and July 2006 in Plant T-1. Thereafter, mixtures of glycol ethers, alcohols and/or cycloaliphatic hydrocarbons were used as blanket cleaners, with no 1,2-DCP or DCM.

Monthly volumes were summed to obtain the annual volume, and the hourly volume was calculated by dividing the annual volume by annual operating hours. However, since monthly volume data from 1991 to 1995 for Plant O-2 were not kept at Company A, the hourly volume was assumed to be equal to the mean of hourly volumes in the following 3 years (1996−1998). As mentioned above, Plant T-2 started operating in January 2005, and we obtained combined monthly volume data for Plants T-1 and T-2 in 2005 and 2006. Since the operation rate ratio at Plants T-1 and T-2 was roughly 8:2 based on interviews with workers, the volume used for Plant T-1 was assumed to be 0.8 times the combined volume. Finally, hourly 1,2-DCP and DCM amounts by weight, G(g/hour), were calculated according to the percentages of these chemicals contained in blanket cleaners.

Estimation of exposure concentrations

In 2012, the JNIOSH conducted an experiment to reproduce the working environment of the printing room in Plant O-2.10 The ink removal operation was performed using a mixture of 1,2-DCP (46.4% by volume) and DCM (53.6%) at 1.75 L/hour, which corresponds to 940 g/hour for 1,2-DCP and 1250 g/hour for DCM. Exposure concentrations for printing workers were measured, and time-weighted average values were reported to be 60−210 ppm (mean, 110 ppm) for 1,2-DCP and 130−360 ppm (mean, 240 ppm) for DCM. In this study, we assumed that exposure concentrations were proportional to the amounts of chemicals used, and estimated exposure concentrations Embedded Image for printing workers using the above calculated hourly amount (G(g/hour)).

As mentioned above, although front room workers did not use 1,2-DCP or DCM, they were exposed to both chemicals in Plant O-2. The JNIOSH experiment showed that airborne concentrations in the front room were 40 ppm for 1,2-DCP and 90 ppm for DCM when using 1,2-DCP at 940 g/hour and DCM at 1250 g/hour, respectively, in the printing room. In this study, assuming that airborne concentrations were proportional to the amounts of chemicals used, we estimated the exposure concentrations Embedded Image in the front room workers. As mentioned above, delivery workers stood ready in the front room, received printed proofs from printing workers and went out to deliver the proofs to clients. Therefore, exposure concentrations were assumed to be half the values calculated for front room workers, considering the time spent in the front room.

For Plants O-1, O-3 and T-1, for which no experiment to reproduce the conditions was conducted, we estimated exposure concentrations for printing workers using the following two mathematical models. First, the concentration of direct exposure to chemical vapour generated from the ink removal operation of the worker himself Embedded Image was calculated using a near-field and far-field model.11 Here, Embedded Image is not exposure concentration only during the ink removal operation but the time-weighted average of the exposure concentration during a full shift, including the duration of the ink removal operation. Next, the concentration of indirect exposure to chemical vapour generated from the ink removal operation of other workers Embedded Image was calculated using a well-mixed model.11 Finally, the combined exposure concentration Embedded Image was obtained by adding Embedded Image and Embedded Image. Details on description of the mathematical models are provided in the online supplementary file.

Supplemental material

Calculation of cumulative exposure

The beginning of the 1,2-DCP exposure period was set to November 1987 or the date of employment, whichever came later, for printing workers, and April 1991 or the date of employment, whichever came later, for front room workers and delivery workers. The end of the exposure period was set to October 2006 in Plant O-2 or July 2006 in Plants O-3 and T-1, the date of resignation or the date of diagnosis of cholangiocarcinoma, whichever came first. Exposure concentration in each month during the observation period for each worker was assigned to be the estimated Embedded Image during the corresponding calendar month, and cumulative exposure to 1,2-DCP Embedded Image at a given month was calculated by summing up Embedded Image values from the beginning of the exposure period to the given month and dividing the sum by 12 (months/year). When considering lag time, exposure during the 5 years just prior to the given month was ignored. For workers who had been employed since before February 1996, cumulative exposure to DCM Embedded Image was calculated in a similar manner.

Evaluation of cholangiocarcinoma incidence

Calculation of standardised incidence ratio

We obtained information on vital status of study subjects and development of cholangiocarcinoma from the Ministry, and one of the authors (Kubo) evaluated validity of diagnosis using medical records.3

For each worker, the observation period was from the first 1,2-DCP exposure through December 2012, the date of diagnosis of cholangiocarcinoma or the date of death due to other disease (one worker died of stomach cancer), whichever came earlier. For external comparison, the standardised incidence ratio (SIR) was calculated for cholangiocarcinoma (codes 155.1 and 156.1 in International Classification of Diseases Ninth Edition (ICD-9), C22.1 and C24.0 in 10th ICD). The expected number was calculated using sex, calendar year and age-specific incidence rates of cholangiocarcinoma in the general population in Japan. The 95% CIs of SIR were calculated on the basis of the Poisson distribution.

Next, the workers were classified into four exposure categories (1–499, 500–999, 1000–1900 and 2000–3999 ppm-years) according to cumulative exposure to 1,2-DCP Embedded Image in each month over the observation period, which implies that cumulative exposure was treated as a time-dependent variable. Furthermore, SIRs with 95% CIs in the four categories were calculated.

Trend of incidence risk with increasing cumulative exposure

For internal comparison, cumulative exposure to 1,2-DCP was categorised into three levels (low, middle and high) with a nearly equal number of workers who developed cholangiocarcinoma, and incidence rate ratios (RRs) with 95% CIs of the high and middle exposure categories to the low exposure category were calculated by Poisson regression analysis using STATA V.14 (Stata Corporation, College Station, Texas, USA). Sex, age (≤29, 30–39, 40–49, 50≤), calendar year (1987–1994, 1995–2004, 2005–2012) and DCM exposure (no, yes) were used in the analysis as potential confounding variables. The reason why exposure to DCM was treated as a dichotomous variable but not as a continuous variable is shown in online supplementary file with table S1. The model was given by Embedded Image, where Embedded Image=incidence rate, Embedded Image is explained in table 4 and xi=0 or 1. Incidence rate ratio (RRi) was given as Embedded Image. Akaike's information criterion (AIC) was calculated to evaluate goodness-of-fit of the model. The significance of trend in incidence risk across cumulative exposure level to 1,2-DCP was assessed by Poisson regression analysis while treating cumulative exposure as a continuous variable.12

Results

Worker characteristics

Table 2 summarises the characteristics of workers. The cohort included 65 male and 6 female printing workers, 9 male and 11 female front room workers and 4 male delivery workers. Of these, 48 were born in the 1970s, and 27 in the 1960s. Forty-six male workers and 16 female workers were exposed to 1,2-DCP but not to DCM (‘DCP workers’), and 32 male workers and one female worker were exposed to both 1,2-DCP and DCM (denoted by ‘DCP/DCM workers’).

Table 2

Characteristics of workers and estimated exposure conditions of 1,2-dichloropropane and dichloromethane

Exposure conditions

Figure 1 shows estimated exposure concentrations Embedded Image in the four plants. For printing workers, the exposure concentrations of 1,2-DCP and DCM were 130–210 ppm and 65–270 ppm, respectively, between November 1987 and February 1996, and the exposure concentrations of 1,2-DCP were 84–346 ppm between March 1996 and October 2006. For front room workers, the exposure concentrations of 1,2-DCP and DCM were 51–76 ppm and 45–100 ppm, respectively, between April 1991 and February 1996, and the exposure concentrations of 1,2-DCP were 55–130 ppm between March 1996 and October 2006.

Figure 1

Estimated exposure concentrations of 1,2-dichloropropane (1,2-DCP) and dichloromethane (DCM) in printing workers, front room workers and delivery workers from 1987 to 2006.

Table 2 also shows exposure duration, exposure concentration (average over exposure duration) and cumulative exposure (at the end of the observation period). Among all workers, exposure durations of 1,2-DCP were 0.3–15.1 years (mean, 4.5 years), exposure concentrations were 28–346 ppm (mean, 187 ppm) and cumulative exposures were 32–3433 ppm-years (mean, 851 ppm-years). Cumulative exposure to 1,2-DCP in male workers was more than twice of that in female workers (mean, 948 ppm-years vs 407 ppm-years), and that in DCP/DCM workers was more than twice of that in DCP workers (mean, 1361 ppm-years vs 579 ppm-years). Cumulative exposure to DCM among all workers was less than a fourth of that to 1,2-DCP (mean, 202 ppm-year vs 851 ppm-years).

Cumulative exposure to 1,2-DCP in the 17 workers who developed cholangiocarcinoma was 646–3409 ppm-years (mean, 2044 ppm-years; median, 1873 ppm-years), and the cumulative exposure to DCM in the 11 workers with cholangiocarcinoma exposed to both 1,2-DCP and DCM was 355–1282 ppm-years (mean, 864 ppm-years; median, 890 ppm-years) (not shown in table 2).

Incidence risk of cholangiocarcinoma

Table 3 shows SIRs of cholangiocarcinoma. SIRs in all workers, male workers and female workers were 1171 (95% CI 682 to 1875), 1203 (95% CI 701 to 1927) and 0 (95% CI 0 to 9426), respectively. Among 62 DCP workers and 33 DCP/DCM workers, SIRs were 1019 (95% CI 374 to 2218) and 1275 (95% CI 636 to 2280), respectively.

Table 3

Standardised incidence ratios of cholangiocarcinoma among 1,2-dichloropropane exposed workers

Table 3 also shows SIRs in the four exposure categories of 1,2-DCP among workers. In the analysis without lag time, although none of the workers developed cholangiocarcinoma in the lowest category (1–499 ppm-years), one to nine workers developed cholangiocarcinoma in the higher categories, and SIRs increased with statistical significance. In the two highest categories (1000–1999, 2000–3999 ppm-years), SIRs were more than 10000. Similar results were obtained in the analysis with a 5-year lag.

Table 4 shows the results of Poisson regression analysis. The adjusted RRs in the middle and high 1,2-DCP exposure categories were 14.9 (95% CI 4.1 to 54.3) and 17.1 (95% CI 3.8 to 76.2), respectively, in the analysis without lag time, and 11.4 (95% CI 3.3 to 39.6) and 32.4 (95% CI 6.4 to 163.9), respectively, in the analysis with a 5-year lag. The trend analysis revealed a statistically significant increase in RR in association with increasing cumulative exposure to 1,2-DCP (p<0.001). Adjusted RRs for exposure to DCM were 0.45 (95% CI 0.11 to 1.77) for the analysis without lag time and 0.31 (95% CI 0.07 to 1.34) for the analysis with a 5-year lag. Sex, age and calendar year were not significantly associated with the risk of development of cholangiocarcinoma. AIC values of the models without lag time and with a 5-year lag were similar.

Table 4

Coefficients (βi) of nine variables and incidence rate ratios (RRi) of exposure to 1,2-dichloropropane and dichloromethane estimated by Poisson regression analysis

Discussion

Known risk factors for cholangiocarcinoma include primary sclerosing cholangitis, liver fluke infestation, biliary stones, fibropolycystic liver disease, chemical carcinogen exposure including thorotrast and heavy drinking and smoking.13 Kubo et al3 confirmed that the known risk factors were not found among the 17 workers who had developed cholangiocarcinoma, although 13 patients were smokers, 3 were habitual alcohol consumers and 1 had a history of hepatitis B viral infection. Therefore, we focused on 1,2-DCP, a chemical to which all 17 workers were exposed. We also examined DCM exposure, because 11 of the 17 workers had also been exposed to this chemical and Lane's original study found a significantly increased mortality risk for biliary tract cancer among workers exposed to DCM (SMR=20 (95% CI 5.2 to 56)).14

In this study, assuming that exposure concentrations of 1,2-DCP and DCM were proportional to the amounts of chemical used, we estimated exposure concentrations in Plant O-2 based on data from the JNIOSH experiment. Accordingly, we believe the estimated concentrations to be accurate. With regard to the other three plants (Plants O-1, O-3 and T-1), which no longer exist, we estimated exposure concentrations using mathematical models based on the amount of chemical used, ventilation rate and number of printing machines. We confirmed that if the three parameter values were correct, the concentrations estimated by the model were in good agreement with the corresponding concentrations estimated on the basis of the JNIOSH experiment in Plant O-2, thereby demonstrating the appropriateness of this model (see online supplementary figure S1). Of the three parameters, ventilation rate was assigned as the assumed values in the three plants due to lack of data, which may have led to a larger margin of error in the estimation for these plants compared to Plant O-2. However, even when half or double the assumed values were used in the model, cumulative exposure did not largely change (see the online supplementary file and figure S2). Consequently, our model estimation is considered appropriate.

This study found an extremely high SIR of cholangiocarcinoma among workers in the proof-printing section of the company. SIR values, especially in the two higher categories of cumulative exposure, exceeded 10000. The Poisson regression analysis revealed a significant increase in incidence risk with increasing 1,2-DCP cumulative exposure after adjusting for sex, age, calendar year and DCM exposure, which suggests that an exposure–response relationship exists. These findings suggest that the development of cholangiocarcinoma could be attributed to 1,2-DCP in these workers.

The SIR of cholangiocarcinoma was particularly high among male workers, while no female workers developed cholangiocarcinoma, possibly due to the smaller number of female workers with high cumulative exposure compared to male workers (1000 ppm-years or more, 28 male workers vs 2 female workers). The Poisson regression analysis also did not detect a significant association between sex and incidence risk.

The SIR was higher among DCP/DCM workers, compared to DCP workers (1275 vs 1019). However, since these 95% CIs were wide due to the small number of workers and overlapped each other, the difference in SIR was not conclusive. The Poisson regression analysis also did not show a significant effect of exposure to DCM on incidence risk. Consequently, we could not determine whether DCM contributed to the development of cholangiocarcinoma among our workers.

Yamada et al6 ,7 reported that multiple workers had developed cholangiocarcinoma in their 30s through 50s in each of three other small printing companies in Japan. All of those workers had been exposed to 1,2-DCP at a high level (8-hour TWA, 62–240 ppm) for a long term (7–17 years), with cumulative exposure calculated to be 690–2550 ppm-years (mean, 1690 ppm-years). Our result (range, 646–3409 ppm-years; mean, 2044 ppm-years) is consistent with Yamada's report.

While 1,1,1-TCE had also been used for the ink removal operation in Plant O-1 prior to 1992, nine of the 17 workers who developed cholangiocarcinoma were not exposed to this chemical. Kerosene was used by almost all workers, but exposure concentrations were most likely low given its low vapour pressure. Pigments or synthetic resins contained in inks used at the company might be carcinogenic. However, workers used inks that were typically used for offset printing, and the amount used was less than that used in the general offset printing industry. Consequently, these chemicals are unlikely to be causative chemicals. Kawasaki et al15 found that the combination of 1,2-DCP and four photoinitiators (possibly included in UV ink) significantly induced cytotoxicity and speculated that simultaneous exposure to chlorinated organic solvents and photoinitiators may enhance cytotoxicity in the printing workers. However, photoinitiators could not have contributed to the development of cholangiocarcinoma because at least 11 of the 17 patients had not used UV ink during the 1,2-DCP exposure period.

Two animal experiments had found that inhalation exposure to 1,2-DCP increases the incidence of bronchioloalveolar tumours in mice and that of nasal cavity tumours in rats, although no bile duct tumour development was observed in either species.16 ,17 It is unclear why the region of tumour development differs between humans and rodents. Further studies will be needed to elucidate the mechanism of carcinogenesis resulting from 1,2-DCP exposure.

References

Footnotes

  • Correction notice This article has been corrected since it was published Online First. Table 4 has been reformatted and two p values moved to their correct position under the 'Estimate' columns.

  • Contributors S. Kumagai planned and conducted this investigation, and described this article. T. Sobue and T. Makiuchi conducted the epidemiological data analysis. S. Kubo conducted the data collection and supported this study from the aspect of hepatology. S. Uehara, T. Hayashi and KK. Sato conducted the data collection. G. Endo planned and conducted this investigation, and coordinated the study overall. All authors participated in the interpretation of the data, revised this work critically for important intellectual content and approved the final manuscript. G. Endo and S. Kumagai are responsible to the overall contents of this study.

  • Funding This study was supported by the Ministry of Health, Labour and Welfare Research Grants (Grant number: H25-Labour-designatation-013), by Industrial Disease Clinical Research Grants (Grant number: 14040101-01) and by the Princess Takamatsu Cancer Research Fund (Grant number: 12-24407).

  • Competing interests None declared.

  • Ethics approval This study was approved by the Ethics Committee of Osaka City University (Osaka, Japan) and the Ethics Committee of University of Occupational and Environmental Health (Kitakyushu, Japan).

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement No additional data are available.